Burkholderia cenocepacia isolates and growth conditions

The three clinical B. cenocepacia isolates studied, IST439, IST4113 and IST4134, were recovered from the sputum of a CF patient under surveillance at the CF Centre of Hospital de Santa Maria [7,14,15]. The clonal isolates were obtained between January 1999 and July 2002. Studies involving the use of these isolates were approved by the ethics committee of the Hospital (Comissão de Ética para a Saúde do HSM/FML), and the anonymity of the patients is preserved. No unpublished clinical data is included in this study. Studies regarding retrospective molecular microbiology analyses and functional genomics were authorized by the ethics committee. The specific samples used in this study have also been described in previous publications [7,14,15]. For the proteomic analysis, cell samples of B. cenocepacia isolates were prepared by suspending well isolated colonies from Luria Bertani (LB, Difco, Sparks, MD USA) agar plates in 3 mL LB broth, followed by overnight growth at 37°C with shaking at 250 rpm. These cultures were used to obtain cells in mid-exponential phase (OD_{640 nm} of 0.4±0.05). After dilution to a standardised OD_{640 nm} of 0.2 in NaCl 0.9% (w/v), 100 µL of these cell suspensions were plated onto LB agar plates and incubated for 24 hours at 37°C.

Transepithelial resistance of bronchial epithelial monolayers exposed to B. cenocepacia clonal variants

Polarized epithelial monolayers were established using two independent bronchial epithelial cells, which expressed functional CFTR (16HBE14o-) or were homozygous negative for the ΔF508 mutation (CFBE41o-) (a generous gift from Dr. Dieter Gruenert, University of California, San Francisco). 16HBE14o- cells [19] were seeded onto 0.3 µM permeable filter supports at a density of 1×10⁵ cells per mL with a liquid/liquid interface for 6 days. CFBE41o- cells (CFTR negative) [20,21] were seeded onto 0.3 µM filters at a density of 7×10⁵ cells per mL with a liquid/liquid interface for one day and changed to air/liquid interface for five days. On day five, antibiotic was removed from media and transepithelial resistance (TER) measurements taken using electrodes (World Precision Instruments), to confirm monolayer integrity. Bacterial strains were grown in MM9 media and on day 6, the bacteria were added to the epithelial monolayers at a Multiplicity of infection (MOI) of 50:1. The TERs of both control (either no bacteria added to the culture or Escherichia coli-exposed monolayers) and B. cenocepacia-exposed monolayers were measured at 0, 2, 4, 6, 8, 12 and 24 hours. TER was calculated as follows: (Resistance of cells on filter – resistance of blank filters) × 1.13 cm².

Data represents the average of three separate experiments and error bars represent the standard error of the mean. Statistical analysis was carried out using a one-way ANOVA and Holms-Sidak post-test.

Invasion of B. cenocepacia sequential clonal variants into human lung epithelial cells

CFBE41o- (CFTR negative) and 16HBE14o- (CFTR expressing) cells were seeded into 24 well plates at a density of 4×10⁵ cells per well in antibiotic free media for 24 hours at 37°C in a 5% CO₂ humidified atmosphere. Medium was removed and cells were washed three times with Phosphate Buffered Saline (PBS). Strains were grown overnight in LB media and 4×10⁵ CFU/mL were added to Minimum Essential Medium (MEM) alone. The bacterial cultures were added to the wells at an MOI of 10:1 and the plate was centrifuged at 700g for five minutes to allow bacterial interaction/ attachment to epithelial cells. Cells and bacteria were incubated for two hours to allow invasion. The supernatant was replaced with 1 mL of a highly inhibitory (2 mg/mL) ciprofloxacin solution and incubated for a further two hours to kill the extracellular bacteria. Cells were washed three times vigorously with PBS and 500 µL of cell lysis buffer (PBS, 10 mM EDTA, 0.25% (w/v) Triton X-100) was added for 20 min. at room temperature. The resulting lysate was serially diluted in Ringer’s solution and quantified by viable counts on LB agar after 48 hours.

The effectiveness of the antibiotic treatment in killing the extracellular bacteria was confirmed by plating the cellular washes following antibiotic treatment onto LB agar plates and incubating for 48 hours. No growth was detected on these plates. Data represents the average of three separate experiments and error bars represent the standard error of the mean. Statistical analysis was carried out using a one-way ANOVA and Holms-Sidak post-test.

Quantitative proteomic analysis based on 2-D DIGE

A 2-Dimensional Difference Gel Electrophoresis (2-D DIGE) quantitative proteomic approach was used to compare the cytoplasmic and the membrane-associated protein enriched fractions of B. cenocepacia clinical isolates IST439 and IST4134. The analysis was performed as previously described [13]. As before, bacterial cells were harvested from solid LB agar plates instead of planktonic cultures, because growth on the solid agar surface is believed to be a more accurate reflection of the physiological conditions experienced by B. cenocepacia in the airways of CF patients. Briefly, cells were washed from 8 LB-agar plates for each B. cenocepacia isolate, independently prepared, with 0.9% NaCl and resuspended in lysis buffer (10 mM Tris base, 100 mM sucrose Halt^TM protease inhibitors cocktail (Pierce, Rockford, USA)). Cell disruption was carried out by sonication and the supernatant (crude extract) was fractionated into cytoplasmic and membrane-associated protein fractions by ultracentrifugation at 100,000g, during 90 min., at 4°C. The supernatant obtained was enriched in the cytoplasmic protein fraction while the pellet corresponded to the membrane-associated protein enriched fraction. The membrane-associated protein fraction was washed twice with lysis buffer followed by a second ultracentrifugation (100,000g, 60 min., 4°C) to reduce cytoplasmic protein contaminants. The membrane-associated protein fraction was resuspended in lysis buffer containing 2% dodecylmaltoside. Protein concentration in the extracts was quantified using a Bicinchoninic Acid quantification kit (Pierce), and 50 µg aliquots of the protein samples were subjected to a clean-up process with the 2-D Clean-up kit (GE Healthcare, Uppsala, Sweden). Fluorescent labelling of proteins with Cy-dye™ (GE Healthcare), isoelectric focusing (IEF), protein separation by molecular weight, gel scanning and analysis, and preparation of gels for posterior protein identification, were also performed as previously described [13]. The software package Progenesis Samespots (Nonlinear Dynamics) was used to automatically detect protein spots and normalise individual spot volumes against total spot volumes for a given gel. Average normalised levels for each isolate were then compared using one-way ANOVA between-group test. Only statistically significant spots (P < 0.05) were selected for analysis. Differential expression between isolates IST439 and IST4134 was determined and a threshold of at least a 1.5-fold increase or 0.67-fold decrease between averaged gels was considered. Spots that showed evidence of saturation were not further analysed. The results represent two independent growth experiments and three technical replicates, in a total of six replicates per sample.

Identification of protein spots of interest was performed by mass spectrometry (MS) in the proteomic unit at Centro Nacional de Investigaciones Cardiovasculares Carlos III, Madrid, Spain (CNIC Foundation), as detailed before [13]. The classification into functional categories was performed using the Burkholderia Genome Database (http://www.burkholderia.com) [22], TIGR database (http://www.tigr.org/), the Role Category Lists, KEGG database (http://www.genome.jp/kegg/) and NCBI (http://www.ncbi.nlm.nih.gov/). The mRNA levels of a randomly selected set of genes, encoding proteins with altered expression among the different isolates, were also determined to reinforce the proteomic results (data not shown).

Siderophore production

Siderophore production by the three isolates under various iron concentrations was measured using the Chrome Azurol S (CAS) assay [23]. Briefly, bacterial supernatants were filtered using 0.22 µM filters to remove cells. 500 μL of cell-free supernatant was mixed with 500 μL CAS solution (150 μM CAS, 15 μM FeCl₃, 0.5 M Piperazine, 1.2 mM HDTMA) (Sigma) and 10 µL shuttle solution (0.2 M 5-Sulfosalicylic acid) (Sigma) and incubated for 5 hours at room temperature. Reference samples of bacterial growth medium containing the relevant concentration of iron were also assayed. Absorbance was measured at 630 nm using MM9 alone as a blank. The percentage siderophore units were calculated as [(Ar - As)/Ar] × 100 where Ar is the absorbance of the reference and As is the absorbance of the sample.

IST4113 and IST4134 exhibit a higher capacity for epithelial cell invasion and disruption of epithelial monolayer integrity compared with isolate IST439

To compare the virulence potential of the three B. cenocepacia clonal isolates retrieved during chronic infection examined in this study, polarized epithelial monolayers were separately exposed to isolates IST439, IST4113 and IST4134, and epithelial integrity was analysed. B. cenocepacia IST439 was the first isolate recovered from the patient and is believed to have initiated the infection, while IST4134 was recovered 42 months after the isolation date of IST439 and 9 months after IST4113, immediately before the patient’s death with cepacia syndrome and following progressive deterioration of pulmonary function [8,13,24]. Isolate IST439 significantly reduced the transepithelial resistance (TER) of 16HBE14o- (CFTR expressing) monolayers after 8 hours (30%), whereas IST4113 and IST4134 both had a similar impact on TER values as early as 4 hours (Figure 1A). In addition, both IST4113 and IST4134 continued to decrease TER values in 16HBE14o- cells to 10% of controls (cells with no bacteria and with Escherichia coli added) over 12 hours (P ≤ 0.001) (Figure 1A). In the independent CFBE41o- epithelial monolayers that do not express functional CFTR, all three isolates rapidly disrupted tight junction integrity to 70% of controls in the first 2 hours with no significant difference between isolates (Figure 1A). Furthermore, there was a more rapid drop in TER in these epithelial monolayers compared to the independently derived 16HBE14o- cells which express functional CFTR upon initial exposure to the bacterial strains (Figure 1A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Measure of transepithelial resistance and epithelial cell invasion for B.

cenocepacia clonal isolates.

(A) Effect of B. cenocepacia clonal variants IST439, IST4113 and IST4134 on tight junction integrity of 16HBE14o- (CFTR positive) and CFBE41o- (CFTR negative) cells. (B) Invasion potential of the three B. cenocepacia clonal variants into 16HBE14o- and CFBE41o- epithelial cells. Data represents the mean of 3 individual experiments and error bars represent the standard error of the mean. * P<0.01.

https://doi.org/10.1371/journal.pone.0083065.g001

Remarkably, the results on epithelial invasion also follow the same trend: the later isolates are significantly more invasive than IST439 using either CFTR-positive or CFTR-negative cells, but their virulence potential varies with the type of epithelial cells studied (Figure 1B). While IST4134 achieves a percentage of cellular invasion that is twice that of IST4113 in 16HBE14o- (CFTR expressing) cells, IST4113 is more invasive of CFBE41o- epithelial cells than IST4134 (Figure 1B). Interestingly, it should be noted that CFBE41o- cells, which carry the CF mutation, are significantly more resistant to invasion by IST4134 (uptake of this isolate is reduced by more than half), while the levels of invasion by the first two isolates are comparable for both CFTR-negative and CFTR-positive cells (Figure 1B).

Differential protein expression profile of IST4134 compared to IST439

Using a 2-D DIGE-based quantitative proteomic approach, it was possible to identify major functional groups and cellular functions that are altered in isolate IST4134 compared with IST439. This analysis allowed the separation of around 740 and 850 protein spots in the gels prepared from cytoplasmic and membrane-associated protein fractions, respectively, within a pI range spanning from pH 3 to approximately pH 9, of which 281 protein spots were positively identified by mass spectrometry (Table S1). GRAVY values (grand average of hydropathicity index) were obtained using the software program ProtParam tool [25,26]. Positive GRAVY values are recognised as valid indicators of hydrophobicity and of membrane involvement [27–29]. In total, 16 (20%) of the 80 proteins identified in the membrane-associated protein fraction had positive GRAVY values ranging from +0.273 to +0.006. Together, the GRAVY index, the transmembrane mapping and the subcellular localisation predictions [13] revealed an enrichment of proteins with a hydrophobic nature in the membrane-associated fraction.

Overall, 72 protein spots were found to have a different content (fold-change cut-off: 1.5-fold increase or 0.67-fold decrease; P ≤ 0.05, ANOVA) between isolates IST4134 and IST439. Sixty-one of these protein spots, corresponding to 52 different proteins, overlapped the 89-protein dataset previously reported for the comparison between IST4113 and IST439 [13]. All 52 proteins were altered in the same direction in both isolates (Table 1 and Table S1). This subset of proteins will be given a particular emphasis in light of its potential involvement in the enhanced virulence potential and persistence of the last two isolates.

Best-hit homologs to B. cenocepacia J2315 genes were found for all proteins identified in the comparison between IST4134 and IST439, with the exception of the ABC transporter-related protein A9AH88, with 97% of identity to the L-glutamate ABC transporter ATP-binding protein encoded by gene Bcen_0189 from B. cenocepacia AU1054, a strain that was recovered from the blood of a CF patient [30] (Table 1). Proteins that were suggested to be differently expressed in both IST4113 and IST4134, compared with IST439, were clustered into functional categories (Table 1 and Figure 2). Considering only the analysis of IST4134 vs. IST439, the most prominent category is “Energy metabolism” (n = 19 proteins; Table 1 and Figure 2), similarly to what was previously described for IST4113 [13]. The second main category is “Translation”, including a total of 10 proteins (9 of which are over-expressed) with altered content in IST4134. “Cell envelope biogenesis”, “Amino acid metabolism” and “Nucleotide metabolism” follow, each comprising 5 proteins with altered content in IST4134. The categories “Protein folding” and “Iron transport” are noteworthy since all proteins that were considered altered in IST4134 are over-expressed compared to IST439. A more detailed description of the main results of this quantitative proteomic analysis can be found below.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Clustering of the proteins differently expressed in B.

cenocepacia clonal variants, based on biological function.

Number of proteins differently expressed in isolates IST4113 [13] or IST4134 (this work), compared to IST439, grouped by functional categories (as detailed in Table 1), based on the information available in the Burkholderia Genome database and in the KEGG Pathways Database. All categories that include less than 2 proteins were grouped together as “Others”; this category includes proteins involved in “Coenzyme metabolism”, “Intracellular trafficking and secretion”, “Secondary metabolites biosynthesis”, “Transcription” Transcriptional regulation”, “Replication”, “Adaptation to atypical conditions” and “Adhesion”.

https://doi.org/10.1371/journal.pone.0083065.g002

Proteomic differences between IST4113 and IST4134 detected indirectly

Based on the fold-change ratios determined for each protein spot in the comparison of IST4134 vs. IST439 [13] and IST4134 vs. IST439 (Table 1), we were able to determine indirect proteomic differences between IST4134 and IST4113, with possible relevance to understand disease progression. Considering a cut-off of 1.5-fold increase or 0.67-fold decrease in IST4134 (P < 0.05) 40 proteins are altered significantly from IST4113 to IST4134 (Table 1). Considering a more significant fold-change cut-off (2-fold increase or 0.5-fold decrease), 9 proteins were singled-out, including the heat shock chaperone ClpB, the C. elegans virulence factor GatA, the nucleotide metabolism protein PurH, and 2 proteins involved in cell envelope biogenesis, BCAL2783 (lipid synthesis) and the assembly factor YaeT – all with reduced content in IST4134 vs. IST4113.

The content of proteins involved in purine and pyrimidine synthesis is higher in isolate IST4134 compared to IST439

In general, proteins involved in the de novo purine synthesis pathway, namely PurM, PurH and GuaB, exhibit a higher content in the later isolates than in IST439 [13] (Table 1). The exception is PurA, an adenylosuccinate synthetase that catalyses the first step of the conversion of inosinate into adenosine monophosphate (AMP) in the aforementioned pathway. This protein is less abundant in IST4134 than in IST439 (Table 1), as previously described for IST4113 [13]. The higher content of proteins involved in the alternative step that converts inosinate into guanosine monophosphate (GMP), in particular GuaB, in both IST4113 and IST4134, reinforces the previously suggested preference for the synthesis of GTP versus AMP in both isolates, compared to IST439 [13]. The content of proteins involved in pyrimidine synthesis is also increased in both IST4113 and IST4134 (Table 1), in particular CarB, one of the two subunits of the first enzyme involved in this anabolic pathway [31].

Proteins involved in translation and protein folding have a higher content in IST4134 compared to IST439

A total of 9 proteins involved in translation processes were found to have a higher content in isolate IST4134, compared to IST439, including components of the ribosome (RpsA, 3 spots, with fold-changes ranging from 2.0 to 2.5), elongation factors (Tuf and FusA), and tRNA synthetases (LeuS, PheT, AlaS, ArgS, GatA and IleS) (Table 1). The content of all these proteins is also altered in the highly antimicrobial resistant isolate IST4113, compared to IST439 [13] (Table 1). Overall, the results indicate that translation is more active in IST4134 than in the isolate that initiated the infection, as suggested previously for IST4113 based on transcriptomic and proteomic data [12,13].

The content of several proteins belonging to the category “Protein folding” is higher in IST4134 than in IST439, including the chaperone protein DnaK, the heat shock protein ClpB, and the trigger factor Tig (Table 1). DnaK was described in the Gram-negative human pathogen Brucella suis as playing an essential role in the protein repair system that protects the bacteria from the environment in the macrophage phagosome [32]. Of the 72 protein spots whose content is altered in isolate IST4134 compared to IST439, the most significantly altered protein is ClpB, identified in two different spots (Table 1).

The content of proteins involved in cell envelope biogenesis is altered in isolate IST4134 compared to IST439

Proteins involved in lipopolysaccharide (LPS) biosynthesis, namely the phosphomannomutase ManB and the NAD-dependent epimerase, were found to have a lower content in the last isolate compared to IST439, as previously described for IST4113 [13] (Table 1). ManB is involved in lipid A synthesis while the NAD-dependent epimerase protein is mainly involved in the synthesis of core oligosaccharide and O-antigen, which are components of the lipopolysaccharide molecule. Both forms of the outer membrane protein assembly factor YaeT, which is involved in protein extrusion to the outer membrane maintaining a homeostatic LPS to protein ratio [33,34], have a higher content in isolate IST4134 than in IST439.

Late isolates exhibit an increased content of iron uptake proteins and are able to capture iron more efficiently from the environment

Four proteins involved in iron uptake are more abundant in IST4134, compared to IST439 (Table 1). Three of these are also over-expressed in IST4113 compared to IST439: putative siderophore-interacting protein, putative pyochelin receptor protein FptA and TonB-dependent receptor [13]. OrbC, a putative iron transport-related ATP-binding protein whose gene expression is up-regulated by OrbS under iron deprivation conditions [35], was only over-expressed in isolate IST4134, compared to IST439 (Table 1). These results led us to examine the production of siderophores by the three clonal variants, using the CAS assay with different iron concentrations (Figure 3). Significant levels of siderophores were secreted by the three B. cenocepacia isolates within two hours of exposure to an iron-depleted environment, demonstrating the ability of this species to rapidly adjust to low iron availability (Figure 3). The results obtained show that IST4113 and IST4134 are more tolerant to low iron concentrations than IST439. The latter produces significantly higher levels of siderophores below 6 µM of Fe³⁺ (Figure 3A), while isolates IST4113 and IST4134 only up-regulate siderophore production below 5 µM or 4 µM of Fe³⁺, respectively (Figure 3B, C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Siderophore production measured by CAS assay for the three B.

cenocepacia clonal variants.

Siderophore production was measured for isolates IST439 (A), IST4113 (B) and IST4134 (C) under different iron concentrations (1 to 11 μM of Fe³⁺) added to minimal M9 media.

https://doi.org/10.1371/journal.pone.0083065.g003

Alteration of proteins involved in metabolic processes

Energy and central intermediary metabolism functions are strikingly overrepresented among proteins whose content is altered in IST4134, compared to IST439 (Table 1, Figure 2). This is the case of several proteins of the tricarboxylic acid (TCA) cycle (OdhL, AcnA, SdhA and FumC) and other pathways like glycolysis/gluconeogenesis (Putative aldehyde dehydrogenase family protein), pyruvate metabolism (PpsA, AceE, OdhL and PckG), and oxidative phosphorylation (NADH-quinone oxidoreductase, SdhA and AtpA) (Table 1). In total, 19 of the 23 protein spots that are differently abundant in isolate IST4134 compared to IST439 are also altered in IST4113 [13] (Table 1). However, a number of proteins with functions in cell metabolism that were altered in IST4113 did not have their expression altered in IST4134, or exhibited a less significant variation.