Burkholderia pseudomallei transcriptional adaptation in macrophages

Infection model and bacterial RNA isolation

Human macrophage-like U937 cells were chosen as the infection model as it has previously been used extensively to study B. pseudomallei interaction with host cells [7, 8]. In this study, bacterial replication within macrophages was observed with a maximum CFU obtained at 6 h post-infection. The calculated doubling time for intracellular B. pseudomallei is about 6.4 h, indicating a slower growth rate compared to the control bacteria in RPMI (doubling time of ~ 1.9 h). The number of viable bacteria decreased consistently after 6 h post-infection and a 2-log reduction in viable cells was noted with prolonged incubation of up to 24 h. A similar decline in bacterial numbers following prolonged incubation was previously reported in RAW264.7 cells [7, 21]. Viability of infected U937 cells decreased over time and was drastically reduced by 6 h post-infection, suggesting a cytotoxic effect of B. pseudomallei on the host cells (Figure 1).

An ultrastructural study of infected U937 cells showed that the nuclei, mitochondria and vacuoles were swollen whilst the nuclear material appeared diluted and the cell membrane totally disrupted (Figure 2A). All these changes indicate that infected U937 cells displayed a phenotype similar to oncotic cells as previously observed in other cells infected with B. pseudomallei[21]. Loss of phagosome membrane in B. pseudomallei infected macrophage cells was established as early as 15 min post-incubation similar to that previously reported [22]. At 2 h post-infection, loss of phagosome membrane is evident and bacilli were seen in the cytoplasm (Figure 2C). Moreover, replication of intracellular bacteria was seen in the cytoplasm of host cells (Figure 2D).

An infection period of 1 to 6 h was selected to study changes in the expression profile as more than 85 % of the infected U937 cells were still viable during this period (Figure 1). This would ensure maximum recovery of intracellular B. pseudomallei from viable cells to obtain sufficient intact bacterial RNA. We adopted the differential lysis method with saponin to prevent bacterial cell lysis and easily recover bacterial RNA. This strategy enriches the bacterial mRNA several thousand-fold [23]. Treatment with saponin specifically lyses eukaryotic cells without affecting bacterial viability [24, 25] and in our study, the use of 0.1 % to 4 % saponin was not detrimental to B. pseudomallei (data not shown). Differential centrifugation was combined with this method to efficiently remove host cellular debris. Based on the RNA integrity number (RIN) and the 16 S/23 S rRNA subunit ratio, the electrophoretic profiles of the total RNA extracted from intracellular bacteria showed no evidence of degradation and only minor host RNA contamination was detectable (Figure 3).

The global gene expression profile

Temporal gene expression profiles obtained from intracellular B. pseudomallei were compared with the transcriptome of control bacteria grown in cell culture medium. Statistical analysis (p < 0.01) combined with a 2-fold variation cut off indicated that 2,797 genes were differentially expressed at 1 h, 2,755 at 2 h, 2,776 at 4 h and 1,918 genes at 6 h. Hierarchical clustering of gene expression levels revealed similar patterns of gene regulation over the period of infection (Figure 4A). The majority of the genes with altered expression were down-regulated throughout the infection period compared to bacteria grown in control medium. Expression levels of these genes continued to decrease up to 4 h before increasing to levels lower or similar to control bacteria at 6 h. It appears that the adaptation of B. pseudomallei within U937 cells is rapid and most of the changes occurred as early as 1 h post-infection (Figure 5). With time, the number of significantly deregulated genes gradually decreased, suggesting that B. pseudomallei had become adapted to the intracellular environment. Functional classification of intracellularly modulated bacterial genes at each time point showed that most of these genes encoded core functions such as metabolism, cell envelope, regulatory functions, transport and binding. Many genes encoding proteins with unknown function or hypothetical proteins were also modulated during infection (Figure 6).

As an independent measure of differential gene expression, we examined the relative expression of six up- or down-regulated genes selected from different functional categories by real-time qPCR (Table 1 and Additional file 1) on the same samples as those used for microarray analysis. Changes in expression were verified by real-time qPCR with a correlation of >0.95 to the microarray data (Figure 7). The strong correlation observed verified the efficiency and robustness of the designed microarray for high throughput screening of the B. pseudomallei transcriptome.

Due to the large number of significantly differentiated genes modulated during the infection, only data related to genes that have some functional information are shown and discussed below. The identified genes are discussed according to functional categories.

Intracellular metabolism and ion transport

In host-pathogen interactions, sufficient nutrition levels are necessary for the successful survival of the pathogen. Once intracellular, the bacteria must make metabolic adjustments to adapt to changes in nutrient availability. We noted a robust shutdown in the expression of B. pseudomallei genes involved in metabolism. Genes involved in glycolysis and oxidative phosphorylation were consistently down-regulated in intracellular B. pseudomallei. In addition, most genes involved in energy metabolism such as ATP synthase and NADH dehydrogenase were down-regulated (Figure 4B). Collectively, these data suggest that intracellular B. pseudomallei have lower energy requirements and limit their energy production during the initial stage of infection. Several anaerobic metabolism genes were induced in intracellular B. pseudomallei. Anaerobic metabolism pathway genes such as BPSS1279 (threonine dehydratase), BPSL1771 (cobalamin biosynthesis protein CbiG) and BPSS0842 (benzoylformate decarboxylase) were up-regulated throughout the infection period. Nevertheless, none of the components of the anaerobic respiratory chain showed significant changes in expression except for BPSL2311 (putative respiratory nitrate reductase delta chain) and BPSL2312 (putative respiratory nitrate reductase gamma chain) that were induced at the early stage of infection.

Other induced genes were catAC genes, which are involved in benzoate degradation, indicating that intracellular B. pseudomallei utilized aromatic compounds as a source of carbon. Increased expression of phenylacetic acid (PA) pathway genes at the later stage of infection (4 h and 6 h post-infection) was also observed (Figure 4B). The major nitrogen source in the intracellular compartment is most likely methylamine and purine as suggested by the increased expression of methylamine utilization protein (BPSS0404) and allantoicase (BPSL2945). Allantoicase is involved in purine metabolism and provides a secondary nitrogen source under nitrogen limiting conditions [26].

Expression of virulence and virulence-associated factors

We observed the repression of genes encoding proteins that are well characterised as B. pseudomallei virulence factors. These include the main capsular polysaccharide biosynthesis (BPSL2787-BPSL2810) genes, two potential surface polysaccharide biosynthesis gene clusters (BPSS0417-BPSS0429 and BPSS1825-BPSS1834), majority of genes in the lipopolysaccharide (LPS) biosynthesis cluster and genes encoding for flagella assembly and chemotaxis. We also noted the repression of bspR, a regulator recently shown to control the expression of TT3SS-3 genes [27]. The inhibition of this regulator leads to the reduced expression of T3SS-3 controlled genes in intracellular B. pseudomallei.

One of the six clusters of the type VI secretion system, the tss-5 cluster (BPSS1493-BPSS1511), was up-regulated up to 182-fold during intracellular infection (Figure 8). This observation is consistent with previous reports on the induction of three genes in this cluster, tssH-5 tssI-5 and tssM-5, upon invasion of macrophages [18]. We also observed the induction of genes flanking the tss-5 cluster, bimA (B urkholderia i ntracellular m otility A)(BPSS1492) and BPSS1512 at 2 to 6 h post-infection. Moreover, the hemolysin activator-like protein precursor, fhaC (BPSS1728) gene was significantly up-regulated during intracellular infection. Consistently, the large filamentous hemagglutinin precursor, fhaB (BPSS1727) gene, a potential virulence factor of B. pseudomallei[20], was induced between 2 to 6 h post-infection.

Stress responses genes

The general stress-responsive alternative sigma factor rpoS transcribes genes involved in bacterial survival under conditions of environmental stress. In B. pseudomallei rpoS is a regulator of carbon starvation and oxidative stress [28].We observed down-regulation of rpoS throughout intracellular growth. Furthermore, the expression of most of the genes encoding sigma factors was repressed in the intracellular bacteria. Under oxidative stress, rpoS regulates oxyR and katG dpsA operons [29]. As expected, we observed down regulation of oxyR and katG dpsA expression at 1 and 2 h post-infection. The expression of other genes in the oxy R regulon was either repressed or did not change significantly relative to control cells. Moreover, class I stress response genes including dnaJ dnaK hrcA and groEL were not significantly regulated, except for groES, which was induced at 6 h post-infection.

DNA topology and growth arrest within macrophages

Modification of DNA topology plays a major role in assisting DNA replication and protein synthesis. Prokaryotic DNA is usually maintained in a negatively supercoiled form by topoisomerases and the level of supercoiling can be affected by several environmental parameters [30]. The ability of DNA gyrase to generate negative supercoils in DNA is inhibited when ATP levels are reduced [31]. In this study, DNA topoisomerase IV subunit A (parC), DNA topoisomerase IV subunit B (parB) and DNA gyrase subunit A (gray) genes were down-regulated. The repression of gyrase suggests a relaxation in bacterial DNA supercoiling during B. pseudomallei intracellular growth in U937. Additionally, the nucleotide-associated regulator protein Fis was also negatively regulated. Fis expression depends on superhelical density whereby maximum density is observed at high levels of negative supercoiling [32].

We found that bacterial genes involved in cell division (ftsABH) were down-regulated indicating reduced cell division activity. The repression of minD and minE genes, which mediate spatial regulation of cytokinesis in bacteria [33], and chromosome partitioning genes, parA and parB was also observed. Furthermore, replicative DNA helicase dnaB, DNA polymerase III subunit alpha dnaE and DNA polymerase III subunit chi genes were also down-regulated. Consistent with the down regulation of catabolic and cell replication genes, the expression of RNA polymerase (rpoABCZ) and ribosomal subunit genes (S1, S10, S21, L11 and L28) was also reduced throughout the infection period. Collectively, these support the view that cell division and replication processes are interrupted during infection, leading to slower growth kinetics of intracellular B. pseudomallei.

Bacterial strain and growth conditions

B. pseudomallei D286, a clinical isolate previously described by Lee [57], was grown on Ashdown agar for 48 h or in Luria-Bertani (LB) broth overnight at 37 °C. Prior to infection, overnight cultures were diluted to 1:50 in LB broth and grown to mid-logarithmic phase (OD₆₀₀ = 0.4-0.6) at 37 °C, 250 rpm for 3 h.

Cell culture and infection model

Human monocyte-like U937 cells (CRL-1593.2) were maintained in RPMI1640 medium (Gibco) supplemented with 10 % fetal bovine serum (Hyclone), 2 mM L-glutamine, 10 mM HEPES and 1 mM sodium pyruvate (Invitrogen). For infection assays, 2 × 10⁵ cells/well were seeded in 12-well cell culture plates, while for RNA extraction, approximately 3 × 10⁷ cells/flask were seeded in T175 flasks. U937 cells were supplemented with 10 ng/ml phorbol myristate acetate (PMA) (Sigma) to induce macrophage differentiation 48 h prior to infection [58]. Induced U937 cells were washed once with Hanks’ Balance Salt Solution (Gibco) to remove traces of PMA before addition of bacteria at a multiplicity of infection of 10. After 2 h at 37 °C, extracellular bacteria were removed by extensive washing with PBS. Fresh media containing 250 μg/ml kanamycin was added to each plate or flask and infected cells were incubated at 37 °C until lysed.

Intracellular survival and eukaryotic cell viability determination

Intracellular survival of B. pseudomallei in U937 was estimated as previously described with some modifications [8]. At different time points following the initial 2 h infection with B. pseudomallei, infected macrophage monolayers were washed three times with PBS and intracellular B. pseudomallei were harvested by adding 500 μl of 1 % saponin in PBS to each well. After 5 min incubation at 37 °C, cell lysates were collected and serially diluted 10-fold in PBS and aliquots were plated onto Ashdown agar to assess viable bacterial counts.

To determine the overall viability of macrophage cells following bacterial infection, trypan blue exclusion was used. At selected time points after infection with bacteria, the monolayers were washed three times with PBS and gently scrapped off from the wells. Trypan blue solution was added to cell suspensions and stained infected cells were visualized under an inverted microscope. Assays were performed in triplicate and repeated at least three times. The number of intact viable cells was expressed as a percentage relative to viable uninfected cells.

Transmission electron microscopy

Transmission electron microscopy for infected U937 cells was performed as previously described [22] with some modifications. Briefly, cells were fixed for 24 h in 4 % glutaraldehyde at 4 °C and washed three times in PBS. After a secondary fix for 2 h in 1 % osmium tetroxide at 4 °C, the specimens were washed three times in PBS. The specimens were then dehydrated in a graded series of acetone/water containing 30 % to 100 % acetone. The specimens were infiltrated with acetone-resin mixture, embedded into resin hard mix and left to polymerize at 60 °C for 24 to 48 h. Ultrathin sections were stained with lead citrate and viewed on a Hitachi H-7100 transmission electron microscope.

RNA extraction

At each time point (1, 2, 4 and 6 h post-infection), macrophage monolayers were washed and infected macrophages from three T175 flasks were combined and lysed in 1 % saponin in PBS (sterile and filtered) for 5 min at 37 °C [25]. Lysates were collected and subjected to differential centrifugation; first at 800 × g for 5 min to sediment eukaryotic cells and cellular debris, and secondly at 8,000 × g for 10 min to pellet bacterial cells. Pellets were immediately snap-frozen in liquid nitrogen and kept at -80 °C until extraction. Bacterial RNA was prepared using the Qiagen’s RNeasy Mini Kit and on-column DNase I digestion was performed. Total RNA obtained was further purified by ethanol/ammonium acetate precipitation. Control RNA from in vitro grown bacteria was obtained by diluting overnight cultures and growing stationary at 37 °C to mid-logarithmic phase in complete RPMI1640 medium under 5 % CO₂. These conditions mimicked those used for the cell infection experiments. Control bacteria were treated similarly and RNA was isolated as described above. Control U937 RNA was isolated using Trizol and purified with the RNeasy Mini Kit. The concentration, quality and integrity of all RNA isolated were analysed using the Nanodrop® ND-1000 and Agilent 2100 Bioanalyser.

A preliminary control experiment demonstrated that incubation in 1 % saponin for 5 min at 37 °C did not affect the viability of B. pseudomallei (data not shown). Additionally, when comparing the transcriptional profile of B. pseudomallei before and after treatment with saponin and differential centrifugation, we found that these treatments caused no significant changes in bacterial gene expression (data not shown), consistent with studies on other bacterial pathogens [24, 59].

Construction of B. pseudomallei DNA microarrays

Microarray containing probes to the annotated ORFs of the B. pseudomallei reference strain K96243 was designed using Agilent’s eArray 5.0 web-based tool and synthesized using Agilent’s 60-mer Sure Print technology. Probes were filtered using a perfect match filter to eliminate probes with identical sequences and a similarity score filter to discard probes with significant similarity to other parts of the target genome. Quality control of the probes was also done based on their base composition (BC) score. BC score is a numerical value that defines the quality of the probe based upon its base composition and distribution, with BC_1 being the best and BC_Poor the worst. Filtered probes with higher BC scores, namely BC_1 and BC_2, were included in the array, while probes with lower BC scores, such as BC_3 and BC_4, were excluded. A total of 5,721 probe sequences passed the filtering and quality control criteria and were replicated and randomly distributed in a microarray of 15,000 probes to fit Agilent 8 × 15 K microarray format. In summary, the B. pseudomallei 8 × 15 K microarrays (GEO reference GPL13233) allowed for the analysis of 5,721 non-cross-hybridizing ORFs of B. pseudomallei K96243.

Sample labelling and hybridization

Prior to labelling, bacterial RNA was polyadenylated and reverse transcribed. Polyadenylation of bacterial RNA was based on the PAP method using A-plus^TMPoly(A) Polymerase Tailing kit (Epicentre). Poly-a tailing was carried out to provide a priming site for the synthesis of first strand cDNA. Polyadenylation was terminated by ethanol/ammonium acetate precipitation. cDNA synthesis, labelling and hybridization were done according to the Agilent one-colour microarray protocols (available at http://www.chem.agilent.com/Library/usermanuals/Public/G4140-90040_GeneExpression_One-color_v6.5.pdf). Washes were also conducted according to standard Agilent protocols. An additional wash with acetonitrile was conducted to completely dry the arrays. Arrays were scanned with the Agilent Technologies Scanner model G2505B. Spot intensities and other quality control features were extracted with Agilent’s Feature Extraction Software version 9.5.3.1. In this study, 3 independent hybridizations using RNA samples isolated from 3 separate assays (biological replicates) were performed for each incubation time point and control bacteria.

Data analysis

Microarray quality was assessed through the use of Agilent’s control features and only arrays that passed the recommended criteria were included in the analysis. Processed signals obtained from Feature Extraction were used as signal intensities for analysis. The data was filtered using a signal to noise ratio criteria and only features for which the background-subtracted signal was 2.6 times above the background standard deviation for that feature in at least 13 of the 15 arrays, were retained. A total of 5,391 genes passed the filtering process. Arrays were median normalized with BRB-ArrayTools (http://linus.nci.nih.gov) to adjust the scale of intensities across samples and arrays. Filtered and normalized data were subjected to Significance Analysis of Microarray (SAM) analysis in which a two class unpaired analysis was performed and genes with a false discovery rate (FDR) < 0.01 and fold change ≥ 2 were defined as significantly differentially expressed. Hierarchical clustering analysis with Euclidean correlation was performed using TIGR-MeV software version 4.3.2 (http://www.tigr.org). Functional classifications were carried out based on Comprehensive Microbial Resources (CMR) annotations (http://www.cmr.jcvi.org). Gene function enrichment analysis was performed on the DAVID 6.7 database (http://david.abcc.ncifcrf.gov/home.jsp) [60] using a Fisher Exact test with Benjamini and Hochberg multiple testing correction (p < 0.05). All microarray results have been deposited in the GEO database (http://www.ncbi.nlm.nih.gov/geo/) with the GEO series accession number GSE27558.

Real-time quantitative PCR