Small colony variants of Burkholderia pseudomallei: alteration of the virulence factors

Abstract

Burkholderia pseudomallei produce small colony variants (SCVs) that can persist in harsh conditions. This study aimed to compare SCV with B. pseudomallei wild-type (WT) in the ability to adhere, invade, survive and produce biofilm. Additionally, proteins responsible for pathogenicity were determined through two-dimensional gel electrophoresis (2-DE) and liquid-chromatography mass-spectrometry (LC-MS). Whole bacterial proteins analysed using 2-DE demonstrated 252 and 323 protein spots in WT and SCV, respectively, with 37 distinctive proteins identified in the SCV. Several isoform proteins (dnaK, groEL1, gapA and tuf), upregulated in SCV, may function as moonlighting proteins. Quantitative proteomic analysis using LC-MS revealed 57 and 99 distinctive proteins in the WT and SCV, respectively. Many distinct proteins communicated by B. pseudomallei SCV were involved in the metabolic pathways and may play a role in the alteration of virulence factors. Thus, the alteration of proteins in SCVs may be an approach of B. pseudomallei to enhance pathogenesis.

KEYWORDS:

• Burkholderia pseudomallei
• small colony variant
• proteomics
• virulence
• persistence

1. Introduction

Melioidosis is a disease caused by Burkholderia pseudomallei and is responsible for at least a 21% death rate. The disease is prevalent in tropical countries including South-East Asia and northern Australia [1]. B. pseudomallei is a facultative intracellular bacterium classified as a Category B pathogen by the Centers for Disease Control and Prevention (CDC). This pathogen has developed a wide range of antimicrobial resistance and is predominantly found in contaminated water and soil [2]. Transmission of B. pseudomallei is highly likely to transpire via inhalation through aerosol transmission or subsequent contact of cuts or scratches with contaminated water or soil [3]. In general, there is a wide-ranging inconsistency in terms of severity as well as the time interval of the ailment among infected patients, ranging from acute or chronic lung infection to blood infection [4]. In addition, relapse infection among the hosts may occur after several years of the first infection with B. pseudomallei [5].

The evidence for the development of subpopulation in the primary culture of B. pseudomallei is noted as small colony variants (SCVs) that have much slower growth than the primary population of bacteria [6]. This colony morphology switching may result in phenotypic variations (i.e. elevated antibiotic resistance, higher biofilm production and greater survivability in phagocytic or monocytic cells) that have been linked with alteration of bacterial protein expression and aid to enhance survival under hostile conditions [7–9]. The SCVs are frequently secluded from relapsed or chronic melioidosis cases and they exhibit transformed phenotypic profiling [9,10]. The establishment of SCVs has been detected in numerous bacterial species such as Staphylococcus aureus [11], Pseudomonas aeruginosa [12] and Pseudomonas fluorescens [13].

From our previous study, B. pseudomallei SCVs have expressed much lesser invasiveness and virulence compared to the wild-type (WT) in Caenorhabditis elegans and human lung epithelial cells (A549) cells [14]. Nevertheless, SCVs have also been described to display increased persistence associated with their WTs in in vitro cell culture models [15–17]. Correspondingly, Chantratita et al. (2007) [18] also established the persistence in an experimental mouse model of B. pseudomallei SCV.

Additionally, Ramli et al. (2012) also observed differences in terms of the virulence and persistence of the colony variants, indicating that strain-dependent variation may exist among colony variants [14]. Thus, here we accomplished a study to show the comparison between B. pseudomallei WT and SCV in terms of the ability to adhere, invade, survive, and produce biofilm in the host as well as profiling of proteins responsible for pathogenicity using two-dimensional gel electrophoresis (2-DE) and liquid-chromatography mass-spectrometry (LC-MS). We believe that this study would complement our previous proteomic analysis of the colony variants [17,19] and provide a comprehensive identification of potential factors that alter the expression of virulence of these morphotypes.

2. Materials and methods

2.1. Bacterial strains and growth curve

The isogenic morphotypes of a B. pseudomallei clinical strain, UMC049L (WT) as well as its SCV, UMC049S, were differentiated based on the size of the colony, identified, and confirmed using Ashdown agar, API 20NE kit (Biomerieux, France), and an in-house PCR technique [20]. The B. pseudomallei culture was prepared in 10 ml Luria Bertani (LB) broth (pH 7.0) according to Vellasamy et al. (2011) [21]. The bacterial pellet was separated from the culture by spinning for 10 min at 10,000 × g and it was later resuspended in fresh LB to obtain a final OD_600nm of 0.1. Then after, 10 µl of the B. pseudomallei culture was further diluted into 10 ml of fresh LB for each strain, respectively, and incubated at 37°C with vigorous agitation (180 rpm). The turbidity of the culture was read at OD_600nm at every 2 h interval from 0 to 48 h and the viable bacterial counts were determined by serial dilution [22]. This experiment was conducted in biological triplicate and repeated three times.

2.2. Adherence, invasion and intracellular survival assays

The adherence assay was performed using A549 cells according to Mariappan et al. (2011; 2013) [23,24] with slight variations. The A549 cells were cultured in Roswell Park Memorial Institute (RPMI) growth medium at 5 × 10⁵ cells per well into a tissue culture grade 24-well plate overnight at 37°C with 5% CO₂. The confluent monolayer of A549 cells was infected with the mid-logarithmic phase B. pseudomallei (WT or SCV), at a multiplicity of infection (MOI) of 1: 10 for 2, 4, 6 and 8 h. Cells were rinsed with PBS thrice and later lysed using 1 ml of PBS comprising 0.5% (v/v) tergitol (Fluca, USA) for 5 min at room temperature. Adherence of viable bacteria to cells was quantified by plating serial dilutions [22]. A similar method was used for invasion assay whereby the cells were treated with RPMI medium containing kanamycin (1000 ug/ml) after 2 h of infection. This was necessary to eradicate the extracellular bacteria. Later, the cells were lysed, and the intracellular bacteria were quantified. The intracellular survival assay was carried out following 2 h of invasion assay and 2 h of treatment with RPMI medium containing kanamycin (1000 ug/ml). The A549 cells were then replaced with an antibiotic-free RPMI with reduced fetal bovine serum (FBS) medium and incubated at various time points such as 2, 4, 6, and 8 h at 37°C. The respective cells were detached and lysed at each time point followed by the quantification of intracellular bacteria by serial dilution. Escherichia coli ATCC 15597 was utilized as a negative control. These assays were repeated three times in biological triplicates.

2.3. Biofilm production assay

The crystal violet staining was used to study the ability of bacteria to produce biofilm as described by Mariappan et al. (2017) [25]. The mid-log bacterial culture (OD_600nm 0.5–0.8) was diluted 100× using LB broth and a small portion of the diluted culture (100 µl) was introduced into all wells of a sterile 96-well plate and cultured at 37°C for 18 h. Later, 1 µl of the bacterial suspension was placed into another 96-well plate encompassing 100 µl of sterile LB media. The plate was further incubated at 37°C for 18 h, and then the supernatant was withdrawn slowly without disturbing the formed biofilm. About 1% crystal violet (100 µl) was introduced into each well for 30 min and discarded. The wells were further washed with PBS (150 µl) to remove extra residues of crystal violet. Lastly, the absorbance was read at OD_570nm after the addition of DMSO to each well. Burkholderia thailandensis E264 (non-biofilm producer) strain was used as a control. This experiment was conducted in biological triplicate and repeated three times.

2.4. Antibiotic susceptibility testing

Both the isolates were tested for in vitro antimicrobial susceptibility to six different antibiotics, namely, meropenem (10 µg), imipenem (10 µg), chloramphenicol (30 µg), ceftazidime (10 µg), gentamicin (10 µg) and trimethoprim/sulfamethoxazole (25 µg) using the disks diffusion assay according to the British Society of Antimicrobial Chemotherapy and Khosravi et al. (2014) [26]. A total of 100 µL of test isolates [106 colony-forming units per millilitre (CFU/mL)], which was adjusted to 0.5 McFarland spectrophotometrically at OD_600nm was lawned on Mueller Hinton Agar and respective disks were positioned on the plates and further incubated at 35°C for 18 h.

2.5. Two-dimensional gel electrophoresis (2D-GE)

Both the WT and SCV of B. pseudomallei were cultured until mid-logarithmic phase and then the cultures were collected to isolate bacterial pellets by centrifugation at 10,000 g for 10 min at 4°C. Bacteria were lysed in lysis buffer (8M urea, CHAPS, pharmalyte pH 3–10) using Omni Sonic Ruptor 4000 Ultrasonic Homogenizer (Omni International, USA) for 10 min on the ice (20% amplitude with 5s pulse intervals). The lysed bacterial mixture was centrifuged (10,000 g for 10 min at 4°C) to isolate the supernatant. The proteins in the supernatant were subjected to 2D-GE, following methods previously described by Mariappan et al. (2011) [23] with minor modifications. Briefly, 550 µg of whole bacterial protein [27] were resuspended in rehydration solution (8M urea, 2% CHAPS, 0.002% bromophenol blue) and the IPG strips (13 cm, pH 3-10) (GE Healthcare, Uppsala, Sweden) were reswelled for 18 h. The proteins were then isoelectrically focused using an IPGphor system (GE Healthcare). Then after, the IPG strips were transferred on a 12% SDS-PAGE [28] and electrophoresed according to Al-Maleki et al. (2014) [17]. The parted proteins were visualized using the Hot Coomassie brilliant blue (GE Healthcare) staining method. Imaging of the gels was done using an Image Scanner (GE Healthcare) and analyzed through the Progenesis Samespots software v3.0 (Nonlinear Dynamics, USA). This software converts peptide-based data to protein expression data using multivariate statistics and pathway analysis. The samples were run concurrently in biological triplicates to deter variation.

2.6. Two-dimensional gel electrophoresis (2-DE) protein identification and in silico analysis

Following image analysis, protein spots that were differentially and greatly expressed with a fold-change of ≥2.5 (p-value ≤ 0.05) in both isolates were picked from the gels and pooled together. MALDI-TOF/TOF from Proteomics International Pty Ltd (Nedlands, Western Australia) was used to further analyse those gel plugs. The data were later discovered using MASCOT (Matrix Science, London, UK) search counter to the B. pseudomallei K96243 genome (reference) catalogue. The search was carried out using the Basic Local Alignment Search Tool (BLAST) and non-redundant NCBI library database (http://www.ncbi.nlm.nih.gov/) including the annotated proteins of the reference database. In silico analyses were performed using PSORTb v.3.0.2 (http://www.psort.org/psortb2/index.html) (in order to predict molecular mass and cellular localization) and SignalP v.4.1 (http://www.cbs.dtu.dk/services/SignalP/) (to predict the signal peptides of the proteins). To consent to a comprehensive conception of the expected active pathways (in both up-and down-regulated proteins), the proteins were analysed using the KEGG Pathway Database (http://www.genome.jp/kegg/pathway.html).

2.7. B. pseudomallei (WT and SCV) liquid chromatography-mass spectrometry (LC-MS) analysis

In addition to the 2-DE analysis, LC-MS analysis was also carried out on the B. pseudomallei (WT and SCV) to obtain a complete proteome profile of these isolates. Briefly, the bacterial pellets were extracted from the mid-log culture (1 × 10⁸ cfu/ml) of B. pseudomallei (WT and SCV). The pellets were lysed in 500 µL PBS using a sonicator Omni Sonic Ruptor 4000 Ultrasonic Homogenizer (Omni International, USA) for 10 min on the ice (20% amplitude with 5s pulse intervals), only supernatant was used for the next step after centrifugation at 10,000 g for 10 min at 4°C. The protein content of the supernatant was measured using the Bradford assay (1976) [27]. The content of protein in the supernatant (200 µL) was protected using protease inhibitor (100X) of 10% (v/v) solution and isolated using a detergent-free total protein isolation kit (Norgen Biotek Corp., Canada) based on the manufacturer’s directions. The total content of proteins was reduced with 100 µL 10 mM dithiothreitol (DTT) for 10 min and alkylated with 100 µL of 20 mM iodoacetamide solution in the dark for 30 min.

The protein digestion was performed overnight using 6.0 µL trypsin stock solution (1 µg/µL). The protein evaluation was presented using Agilent 1260 Infinity HPLC-Chip/MS System with an Agilent 6540 Ultra High Definition (UHD) Accurate-Mass Quadrople Time-of-Flight (Q-TOF) LC/MS (Agilent, USA). The Proteomic Chip: ProtID-Chip-150 (II), 150 mm 300A C18 chip w/40 nL trap column was used with 0.1% formic acid (mobile phase A) and 0.1% formic acid in acetonitrile (mobile phase B). The samples (5 µL of 1 µg/µL protein) were injected into the column with a flow rate of 0.3 µl/minute and the temperature of the column was maintained at 50°C with a column pressure of 200 bars (biological triplicate with three technical replicates). Reference mass correction was carried out using standard reference mass, 229.294457 and 1221.990637 g/mol. Data acquisition and analysis were achieved using the PEAKS 7 software (Bioinformatics Solutions Inc., Canada). The proteins’ functional categorization and sub-cellular distribution were performed based on Swiss-Prot/TrEMBL database search.

2.8. Statistical analysis

In order to analyse the results, GraphPad Prism 7 (GraphPad Software, USA) was used. One-way and two-way ANOVA and Tukey’s multiple comparison tests were applied for evaluation and the significance of data was established at p < 0.05.

3. Results

3.1. Growth curve of B. pseudomallei wild type and small colony variant

Both B. pseudomallei WT and SCV have demonstrated noticeably diverse growth profiles (Figure 1). Both the strains showed slow growth (lag phase) up to 4 h post-inoculation. B. pseudomallei WT started increasing rapidly from 4 h (OD_600nm = 0.258; viable count = 0.97 × 10⁵ cfu/ml) to 40 h (OD_600nm = 7.942; viable count = 5.97 × 10⁹ cfu/ml). Whereas the SCV was relatively slow-growing with the log phase starting from 4 h (OD600nm = 0.219; viable count = 0.81 × 10⁵ cfu/ml) to 44 h (OD_600nm = 6.382; viable count = 9.81 × 10⁸ cfu/ml). Finally, both WT and SCV strains reached the stationary phase at 40 and 44 h, respectively. In relation to WT, the SCV has a relatively slower growth rate, and the optical density of B. pseudomallei WT and SCV were observed to correlate with the bacterial viable count. Additionally, we also determined the doubling time of the bacterial population based on their generation times. It was obvious that WT has a shorter generation time (172 min) as compared to SVC (212 mins) (p < 0.05).

Figure 1. Growth curve of Burkholderia pseudomallei WT and SCV colony morphotypes.

3.2. Adherence, invasion assay and intracellular survival assays of B. pseudomallei wild type and small colony variant

The capacity of both isolates to adhere to the A549 cells was established using a MOI of 1:10 and monitored at 2 h post-infection. It was found that the WT was more (p < 0.01) adherent compared to SCV (Table 1). At 2 h post-infection the percentage of WT and SCV adherence was 1.6 ± 0.24% and 1.1 ± 0.21%, respectively. Similarly, the SCV (0.21 ± 0.28% and 1.1 ± 0.22%) showed a significant (p < 0.001) reduction in invasive characteristics relative to its WT (0.54 ± 0.26% and 1.42 ± 0.19%) at 2 h post-infection and consecutively up to 8 h post-infection (Figure 2). It was also found that the invasion is interrelated with the proportion of adherence.

Figure 2. Invasion assay of Burkholderia pseudomallei WT and SCV colony morphotypes. The experiment was carried out in biological triplicates (***p < 0.001).

Table 1. Adherence assay of Burkholderia pseudomallei WT and SCV in A549 cells.

Adherence rate (%)
B. pseudomallei	MOI 1:10	Significant level
WT	1.6 ± 0.24	(p < 0.01)
SCV	1.1 ± 0.21

An intracellular survival assay was performed to identify the aptitude of WT and SCV to survive in the host cells. About 2.33 ± 0.34 × 10² cfu/ ml of WT and 2.12 ± 0.26 × 10² cfu/ ml of SCV was seen at 2 h meanwhile a significant difference was observed only at 6 h, with the count of CFU at 2.64 ± 0.1 × 10⁴ cfu/ ml for SCV and 3.02 ± 0.18 × 10⁴ cfu/ ml for WT (p < 0.05). However, replication of SCV was found to be highly increased with 3.1 ± 0.17 × 10⁵ and 3.3 ± 0.17 × 10⁵ at the 10th and 12th h (Figure 3). Overall, clear conclusions on the intracellular replication potential of UMC049L (WT) and UMC049S (SCV) are not possible as statistical significance between these strains was only observed at 6 h.

Figure 3. Intracellular survival assay of Burkholderia pseudomallei WT and SCV colony morphotypes. The experiment was carried out in biological triplicates (*p < 0.05).

3.3. Biofilm formation

The colourimetric method was used to investigate the ability of both the WT and the SCV strains to produce biofilm. Both the strains were able to produce biofilm, however, the biofilm production in SCV was significantly higher (p-value ≤ 0.05) (OD_570nm 0.384 ± 0.054) than the WT (OD_570nm 0.231 ± 0.082) (Figure 4).

Figure 4. Biofilm formation of WT and SCV of Burkholderia pseudomallei. The experiment was carried out in biological triplicates (*p < 0.05).

3.4. Antibiotic susceptibility testing

Both the strains (WT and SCV) were found to be resistant to gentamicin and sensitive to the remaining antibiotics, except that the SCV was categorized as susceptible, increased exposure to chloramphenicol.

3.5. Two-dimensional gel electrophoresis of B. pseudomallei isogenic morphotype wild type and small colony variant

Both WT and SCV were found to be significantly expressed (p-value ≤ 0.05) with more than 2-fold change whereby 252 and 323 proteins were spotted, respectively (Figure 5). However, only 67 protein spots were found to be differentially expressed with a cut-off of ≥2.5 (p-value ≤ 0.05), this will help to remove uncertain/low quantity proteins and at the same time increase the credibility of differential expression analysis. The majority of proteins were not highly expressed in SCV (40/67) meanwhile 27 of those protein spots were highly up-regulated (Figure 6(A,B)), however one protein spot was not discovered due to a low assurance score (spot 13; down-regulated in SCV, Figure 5). Of those up-regulated proteins in SCV, four proteins were found to have more than one spot identified which may be due to isoforms of the proteins. These included three protein spots for chaperone protein DnaK (dnaK), 10 protein spots for 60 kDa chaperonin 1 (groEL1), two protein spots for glyceraldehyde-3-phosphate dehydrogenase (gapA) and elongation factor Tu (tuf1), respectively (Table 2). Meanwhile, the majority of the down-regulated protein seemed not to have isoforms. A total of 31 and six proteins [antioxidant AhpC/T family (ahpC), ilvE-1 branched-amino acid aminotransferase (ilvE), arginine deiminase (arcA), dihydrolipoyl dehydrogenase (IpdV), putative extracellular ligand-binding protein, glutamate/aspartate periplasmic binding protein (gltl)] were found to be distinctively down- and up-regulated in SVC, respectively. In addition, gapA, groEL1, tuf1, dnaK, 30S ribosomal protein S1 (rpsA), superoxide dismutase (sodB), ATP synthase subunit beta 1 (atpD1), and glutamate-binding periplasmic protein (glnH) were the eight proteins that have high and low expression in SCV.

Figure 5. Representative 2-DE proteome map of Burkholderia pseudomallei (A) wild type and (B) small colony variant. A total of 252 and 323 protein spots were detected on the WT and SCV gels, respectively. The protein spots circled in red indicate the differentially expressed proteins with a fold-change of ≥2.5 (p-value ≤ 0.05), which were identified using MALDI-TOF.

Figure 6. The Venn diagram demonstrates (A) the number of differentially expressed proteins between WT and SCV with a fold-change of ≥2.5 (p-value ≤ 0.05) identified using MALDI-TOF analysis (B) the number of proteins identified using liquid chromatography-mass spectrometry (LC-MS) analysis, that was common between the WT and SCV and proteins that were observed to be present only in UMC049L (WT) or UMC049S (SCV).

Table 2. Summary of proteins with isoforms from SCV and WT of Burkholderia pseudomallei.

Protein name	Spot number
Chaperone protein DnaK (dnaK)	41, 54–55
60 kDa chaperonin 1 (groEL 1)	44, 46–50, 57–58, 60, 64
Glyceraldehyde 3-phosphate dehydrogenase (gapA)	52–53
Elongation factor Tu (EF-tu)	62–63

3.6. B. pseudomallei protein in silico analysis

Based on the computational analysis revealed that 66 proteins were found to be significantly expressed in both B. pseudomallei WT and SCV whereby 27 of them were highly expressed meanwhile remaining expression of 39 proteins was reduced. The functionality of each protein was presented according to “Clusters of Orthologous Groups” (COG) (Supplementary Table 1). Based on the analysis, proteins associated with metabolism were 33, cellular processes were 24, as well as information storage were seven. In addition, the function of the two proteins was poorly revealed. The majority of these proteins associated with metabolism were found to be specifically involved in the metabolism of carbohydrates, lipids, amino acids and secondary metabolites. On the other hand, 36.36% of proteins that fall under cellular processes were linked with the development of basic cell structure including cell envelope, outer membrane, post-translation modification and chaperones as well cell movement. About 10.61% of proteins under the function of information storage and processes were found to be specifically linked with the synthesis of ribosomal, protein translation, and DNA transcription, replication as well as repair (Figure 7(A)).

Figure 7. (A) Functional classification (Swiss-Prot/ TrEMBL) (B) Subcellular localization (PSORT) of differently expressed proteins of Burkholderia pseudomallei SCV and WT identified using two-dimensional gel electrophoresis analysis.

The specific location of the protein in the cells was analysed using PSORTb v.2.0. Only one out of 67 recognized proteins were extracellular meanwhile 66 proteins were predicted to be intracellular (53; cytoplasmic proteins, 1; cytoplasmic membrane-associated protein, 1; outer membrane protein and 8; periplasmic proteins or ABC transporter) and two proteins were not known for a specific location in the cell (Figure 7(B)). Along with this, the categorization of all identified proteins was done through SignalP v3.0, whereby the presence of signal peptides was predicted in 4.55% of the proteins identified. Gene ontology (GO) enrichment analysis of the biological processes was determined to govern these networks significantly (p < 0.001 based on false discovery rate (FDR) correction), which includes energy production, cellular process, metabolic process, glycolytic process, carbohydrate metabolic process, cellular protein metabolic process and small molecule metabolic process (Table 3).

Table 3. Significant (p < 0.001) pathway enrichments observed in the protein-protein interaction of WT and SCV using GO biological processes.

Protein	p-value	n
Purine nucleoside monophosphate metabolic process	1.879e-5	7
Purine ribonucleoside monophosphate metabolic process	1.879e-5	7
Primary metabolic process	1.879e-5	25
Purine ribonucleoside metabolic process	1.879e-5	7
Purine nucleoside metabolic process	1.879e-5	7
Single-organism carbohydrate metabolic process	2.570e-5	8
Cellular metabolic process	2.570e-5	25
Ribonucleoside monophosphate metabolic process	2.570e-5	7
Purine ribonucleoside metabolic process	2.570e-5	7
Purine nucleoside metabolic process	3.139e-5	7
Nucleoside monophosphate metabolic process	3.160e-5	7
Ribonucleoside metabolic process	5.089e-5	7
Glycolic process	5.310e-5	4
Organic substance metabolic process	5.310e-5	25
Ribonucleoside metabolic process	5.310e-5	7
Purine-containing compound metabolic process	5.429e-5	7
Carbohydrate metabolic process	6.050e-5	9
Metabolic process	6.669e-5	32
Nucleoside metabolic process	8.299e-5	7
Glycosyl compound metabolic process	8.229e-5	7
Ribose phosphate metabolic process	1.140e-4	7
Cellular process	1.140e-4	27
Single-organism carbohydrate catabolic process	2.139e-4	4
Carbohydrate catabolic process	2.590e-4	4
Generation of precursor metabolic and energy	2.590e-4	6
Small molecule metabolic process	2.590e-4	14
Pyruvate metabolic process	4.360e-4	4
Cellular protein metabolic process	4.360e-4	9

3.7 Liquid chromatography-mass spectrometry (LC-MS) analysis

The specific protein from the digested polypeptide of B. pseudomallei SCV and WT was confirmed using de novo sequencing approach via PEAKS Studio v7.5. Approximately, 435 proteins from WT and 447 proteins from SCV were discovered and selected for further analysis. Followed by the combination of several subcategory databases involving annotated FASTA sequences of B. pseudomallei SCV and WT which resulted in the rapid discovery of low-quality proteins since most of the other labelling approaches tend to ignore those proteins which could be useful. As a result, UMC049L (WT) and UMC049S (SCV) were found to express 57 and 99 proteins, respectively (Figure 7(B)) (Supplementary Table 2). These proteins were further analysed using in silico approaches.

It was found that most (48%) of the proteins present only in UMC049L (WT) or UMC049S (SCV) were involved in various metabolism pathways such as energy production and conversion, nucleotide passage and metabolism, lipid transport and metabolism, inorganic ion passage and metabolism, and secondary metabolites biosynthesis, transport and catabolism. Nearly 23% of proteins were found to be associated with cell cycle arrest, cell proliferation, chromosome partition, cell motility, protein as well as ribosomal translation. Meanwhile, 16% of proteins were associated with information storage and processes and the remaining 13% of the proteins were found to be poorly characterized (Figure 8(A)).

Figure 8. (A) Functional classification (Swiss-Prot/ TrEMBL) (B) Subcellular localization (PSORT) of proteins present only in Burkholderia pseudomallei UMC049L (WT) or UMC049S (SCV) were identified using liquid chromatography-mass spectrometry (LC-MS) analysis.

The cellular locations of the proteins present only in UMC049L (WT) or UMC049S (SCV) were predicted using PSORTb v.2.0. Most of the proteins were expected to be involved as cytoplasmic protein (42%) and the remaining were predicted to be present at the cytoplasmic membrane (30%) and unknown locations (17%). A total of seven proteins were projected as part of the outer membrane and periplasmic regions, respectively. Only a small percentage (2%) of proteins, namely flagellin (fliC), penicillin-binding protein 1A and an uncharacterized protein were present as extracellular proteins (Figure 8(B)).

4. Discussion

SCVs are a sub-population of bacteria characterized by slower growth when equated with the WTs. Interest in the SCVs in general increased exponentially over the years, and rigorous studies have been described on their distinguishing phenotypic and pathogenic characteristics [29]. Additionally, various pathogenic bacteria have been reported to differentiate into SCVs including, Burkholderia cepacia [30], P. aeruginosa [31], and Salmonella [32].

In this study, the parental strain (WT) was observed to differentiate into a single morphotype i.e. the SCV. Although only a pair of WT and SCV were used in this study, a comparison with other studies was performed in order to identify similarities or differences in virulence characteristics and protein expressions. The SCV displayed a slow growth rate in contrast to the WT, as reported in other studies [14,17]. Additionally, this SCV also demonstrated lesser adherence and invasion associated with its WT counterpart. Like our current results, the SCV-infected C. elegans in our previous study have been shown to survive longer compared to the WT-infected one indicating the fewer invasiveness properties of SCV [14]. In contrast, another study conducted by our laboratory found that adherence of an unrelated B. pseudomallei SCV was reported to be higher compared to its WT when assayed at 2-hour time points [17]. Given the differing growth rates of SCV and WT, we performed the adherence and invasion of the two isolates at multiple time points. This was done to facilitate a better understanding of the aptitude of the morphotype variants to adhere and invade over time.

Moreover, according to Al-Maleki et al. (2015) [19], the WT seemed to have higher replication within the A549 cells compared to the SCV between 0 and 12 h post-infection. However, based on this study, SCV has a slow rate of intracellular survival compared to WT, but at the 12th hour, the rate was higher than WT. The SCV was able to withstand a steady rise in intracellular survival and replication, while the WT established a reduction following long hours of survival in the host cells. This result needs to be treated with caution as the reduced number of WT bacteria recovered at 12 h post-infection may also be recognized as the reduction in the cell viability due to the higher replication of WT or a point of saturation reached. Similar to the findings by Ramli et al. (2012) [14], we also found that SCV has higher biofilm production compared to its WT counterpart. In conjunction with that, others described that SCVs may persist slowly and remain extracellularly in a protected biofilm matrix [33,34]. Additionally, we also found that B. pseudomallei SCV and WT had a similar antibiotic susceptibly pattern. This is in agreement with the study by Gracia et al. (2013) [35], which reported that in general, the antibiotic susceptibly is universally comparable for SCVs and their phenotype corresponding strain for most antibiotics.

In order to further understand the need for phenotype switching, global gel-based (2D-GE) and non-gel-based (LC-MS) proteomic analyses were performed on B. pseudomallei SCV and WT to discover possible reasons that modify the manifestation of pathogenesis. We believe that this study may provide an advantage by utilizing both the gel- and non-gel-based proteomics to obtain a high-throughput snapshot of those proteins that are distinctly present in both the SCV and WT (which might have been missed through the utilization of only the gel-based method in our previous study).

The 2-DE analysis identified more than 200 protein spots, of which 27 were significantly up-regulated in SCV by >2.5-fold. Among these 27, 17 spots corresponding to four proteins (dnaK, groEL1, gapA and tuf1) were found to be isoforms that are localized specifically in the cytoplasmic region. Isoforms are different forms of the protein coded from the same gene and these proteins are acknowledged as multifunctional known as “moonlighting proteins” [36]. These proteins accomplish various autonomous and frequently distinct functions without segregating the functions into the diverse domain of the protein. Isoforms are known to increase the parameter of numerous critical biological progressions as well as cell death, proliferation, and cell cycle control [37]. Parallel findings have also been reported in other bacteria such as Burkholderia spp [17,38] and Helicobacter pylori [37], whereby most of the isoform proteins present were found to be involved in metabolic pathways in the cytoplasm and are essential for bacterial colonization. SCV is a slow-growing sub-population and thus, up-regulation of the isoforms may help in bacterial cell survival. However, additional studies are required to comprehend the association of isoforms in SCV pathogenesis.

There were six proteins (ahpC, ilvE, arcA, IpdV, putative extracellular ligand-binding protein, gltl) found to be distinctively up-regulated in SVC. Thus, we believe that these proteins may have important roles in the phenotypic variation demonstrated by B. pseudomallei. In addition, based on our findings, there were several proteins (gapA, groEL1, tuf1, dnaK, rpsA, sodB, atpD1, and glnH) identified to be both up-and down-regulated in SCV. This is a common attribute as several other studies have also demonstrated that the same proteins could be both up and down-regulated in SCV [18,19,25].

In our previous study, we performed proteomics analysis of an unrelated pair of B. pseudomallei SCV and WT in pre-exposure [17] and post-exposure [19] to A549 cells, where a large number of proteins were identified via MALDI-TOF analysis. In comparison with our current study, only five proteins [Gro-EL1, enolase (eno), phosphoglycerate kinase (pgk), ATP synthase subunit alpha 1 (atpA1) and tuf1], and seven proteins [GapA, Gro-EL1, tuf1, enolase, fliC, succinyl Co-A ligase (sucD) and malate dehydrogenase (mdh)] were found to be similarly identified in the pre-exposure [17] and post-exposure [19] conditions, respectively (Table 4). Collectively, these findings indicate the vast difference among these strains, which may be attributed to strain-specificity and differences in the inherent genetic factors.

Table 4. B. pseudomallei SCV and WT proteins identified in this study compared with B. pseudomallei SCV and WT proteins from a previous study, Al-Maleki et al., 2014; 2015 [17,19].

		2DE MALDI-TOF (this study)		LC-MS (this study)		2DE MALDI-TOF [17]		2DE MALDI-TOF WT vs SCV (post-exposure to A549) [19]		2DE MALDI-TOF WT (pre-exposure vs post-exposure to A549) [19]		2DE MALDI-TOF SCV (pre-exposure vs post-exposure to A549) [19]
Protein name	Locus name	Down-regulated in SCV	Up-regulated in SCV	Unique in SCV	Unique in WT	Down-regulated in SCV	Up-regulated in SCV	Down-regulated in SCV	Up-regulated in SCV	Down-regulated in WT pre-exposure	Up-regulated in WT pre-exposure	Down-regulated in SCV pre-exposure	Up-regulated in SCV pre-exposure
Flagellin	BPSL3319	√		√								√
Putative polyketide synthase	BPSS1008	√		√
Superoxide dismutase	BPSL0880	√	√
60 kDa chaperonin	BPSL2697	√	√
Elongation factor Tu	BPSL3215	√	√
Chaperonin DnaK	BPSL2827	√	√
30S ribosomal protein S1	BPSL2515	√	√
Glyceraldy-3-phosphate dehydrogenase	BPSL2952	√	√					√		√
ATP synthase subunit beta 1	BPSL3396	√	√
Glutamate-binding periplasmic protein	BPSS2337	√	√
Phosphoglycerate kinase	BPSL0796	√					√
60 kDa chaperonin 1	BPSL2697	√	√			√	√			√			√
Enolase	BPSL2270	√				√			√		√	√
ATP synthase subunit alpha 1	BPSL3398	√				v
Elongation factor Tu	BPSL3215	√	√			√				√
Malate dehydrogenase	BPSS1722							√
Succinyl CoA ligase	BPSL0780							√		√

Overall, we noticed that the mass difference of the proteins that were differentially communicated are those involved in the metabolic pathway. Many researchers have conducted experiments to prove that these proteins may also play other roles, especially in bacterial virulence including adherence, invasion and survival of the bacteria in the host and promoting pathogenesis [39,40]. GapA is a key glycolytic enzyme that is engaged in the preliminary step of the sub-pathway that produces pyruvate from D-glyceraldehyde 3-phosphate. According to Egea et al. (2009) [41], GapA could perform as a virulence factor that might contribute to pathogenesis. In support of this, Seidler and Seidler (2013) [42] suggested that GapA develops pathogenesis by encouraging bacterial adhesion and invasion, constraints host lysozyme and causing apoptosis in macrophages. Eno is also a glycolytic enzyme and a constituent of RNA degradosome that is involved in the processing of RNA as well as gene regulation. Weng et al. (2016) [43] reported that eno is vital for the virulence of P. aeruginosa in an acute pneumonia mice model. Transformation of eno gene amplified bacterial predisposition to neutrophil-mediated killing, due to lesser acceptance of oxidative stress. Similarly, the function of mdh, which is part of the tricarboxylic acid cycle, has also been implicated in bacterial virulence. Han et al. (2014) [44] indicated that mdh is not only an immunogenic protein but is also related to bacterial pathogenesis with roles in metabolism, pathogenesis and immunity.

Based on a study by Vanghele et al. (2010) [45], DnaK and GroEL1 (also present as isoforms) were thought to be involved with protein folding, however, the actual role is to help bacteria to survive under environmental stress by incorporating virulence phenotypes in polypeptides during invasion [46]. In addition, the tuf protein was also found to be expressed in both the WT and SCV. This protein is reported to be commonly found in many bacteria including E. coli [47] and has been indicated with multiple roles and attributes as virulence factors in pathogenic bacteria [48].

Similarly, high expression of arcA in SCV compared to WT indicates the role in protecting bacteria from the acidic environment which can cause damage to the bacteria. This has been reported in various other bacteria including streptococci and P. aeruginosa [49], and B. pseudomallei [50]. The ability of SCV to survive in an acidic environment could lead to the increment of ammonia that leads to the up-regulation of pH in the surrounding environment of the host [49]. This is further supported by the regulation of amino acid biosynthetic genes such as ilvE as amino acid biogenesis may result in the regulation of acid pressure of the host resulting preservation of pH related to bacterial survival. In this study, the Tsa family/AhpC were also found to be up-regulated in the SCV. In agreement with this, Master et al. 2002 [51] have also observed an up-regulation of AhpC in Mycobacterium tuberculosis. AhpC is an essential component of the peroxiredoxin family that is widely utilized in bacteria to exhibit resistance towards various reactive oxygen and nitrogen species, production of biofilm, virulence, persistence, and establishment [52–55]. Thus, high expression of AhpC in SCVs could lead to persistence towards oxidative stress. Collectively, regulation of proteins related to intracellular survival and persistence (ArcA and AhpC) suggests longer survival of SCV in A549 cells compared to WT.

Additionally, we found two proteins in SCV, flagellin and putative polyketide synthase, that were down-regulated in the gel-based and present distinctively in non-gel-based techniques. A main concern in the relative quantitation of proteins is the question of the dynamic range of the technique used. A remarkably narrow range causes difficulty in the study of proteins displaying greatly modified quantities. This observation may need to be further confirmed as the non-gel-based techniques are usually known to have a broader variety compared to the colloidal Coomassie-stained 2D gel. Thus, the reason for this discrepancy in the result obtained remains to be elucidated.

Up-regulation of flagellin in SCV may enhance the ability to migrate and communicate with immune cells within the host [18,56]. On the other hand, the synthesis of putative polyketide synthase is associated with the production of natural biofilm in the host which is responsible for at least 80% of bacterial infections [57,58]. Alteration of exopolysaccharides (major components of the extra polymeric substance in biofilms) is considered important in the development of biofilm-associated human infections [59]. These infections could be extremely pathogenic and express strong resistance towards antimicrobials and host immune systems. A study has shown that microbial polyketide synthase of M. tuberculosis and Pseudomonas spp, have exhibited similar pathogenic mechanisms associated with biofilm infections [60]. Similarly, we also found that B. pseudomallei strains have produced natural biofilms indicating B. pseudomallei virulence and pathogenesis [8,14].

5. Conclusions

In conclusion, various differentially expressed proteins were identified in SCV compared to WT using the 2-DE and LMCS. Based on previous studies, these proteins were found to be involved in different pathways and have been implicated in virulence, survival, and persistence. Therefore, we postulate that the occurrence of SCVs type in B. pseudomallei could be an approach to expedite virulence and persistence based on functions and possible dynamic interplay between these proteins. The differential characteristics shown in B. pseudomallei colony variants may be mainly due to strain dependency. However, although strain-dependent variations exist, the metabolic pathway-related proteins produced by B. pseudomallei SCV could be the main factor in determining the phenotypic trait and alteration in virulence that presents potential mechanisms related to disease development in the host.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data will be available to the public upon request

Additional information

Funding

This work was supported by Ministry of Higher Education (MOHE) Malaysia through the Fundamental Research Grant Scheme (FRGS) [grant number FRGS/1/2020/SKK0/UM/02/7 or FP018-2020].