OUP user menu

Genome-wide expression analysis of iron regulation in Burkholderia pseudomallei and Burkholderia mallei using DNA microarrays

Apichai Tuanyok , H. Stanley Kim , William C. Nierman , Yan Yu , John Dunbar , Richard A. Moore , Patricia Baker , Marina Tom , Jessmi M.L. Ling , Donald E. Woods
DOI: http://dx.doi.org/10.1016/j.femsle.2005.09.043 327-335 First published online: 1 November 2005

Abstract

Burkholderia pseudomallei and B. mallei are the causative agents of melioidosis and glanders, respectively. As iron regulation of gene expression is common in bacteria, in the present studies, we have used microarray analysis to examine the effects of growth in different iron concentrations on the regulation of gene expression in B. pseudomallei and B. mallei. Gene expression profiles for these two bacterial species were similar under high and low iron growth conditions irrespective of growth phase. Growth in low iron led to reduced expression of genes encoding most respiratory metabolic systems and proteins of putative function, such as NADH-dehydrogenases, cytochrome oxidases, and ATP-synthases. In contrast, genes encoding siderophore-mediated iron transport, heme-hemin receptors, and a variety of metabolic enzymes for alternative metabolism were induced under low iron conditions. The overall gene expression profiles suggest that B. pseudomallei and B. mallei are able to adapt to the iron-restricted conditions in the host environment by up-regulating an iron-acquisition system and by using alternative metabolic pathways for energy production. The observations relative to the induction of specific metabolic enzymes during bacterial growth under low iron conditions warrants further experimentation.

Keywords
  • Iron
  • Burkholderia pseudomallei
  • B. mallei

1 Introduction

Burkholderia pseudomallei and B. mallei are bacterial pathogens that cause the diseases melioidosis and glanders, respectively [1]. Melioidosis is a tropical disease, which is particularly prevalent in Southeast Asia and Northern Australia [2,3]. Glanders is a zoonotic disease affecting horses and other animals [4,5] that rarely occurs in humans, but infections in laboratory workers and individuals handling infected animals have been documented [6]. The natural habitat of B. pseudomallei is in the soil [7], in contrast, B. mallei is a host-adapted pathogen which survives poorly in the environment outside the host [5]. At present, there are no effective vaccines available against either of these microorganisms. Since they are of significance as agents of bioterrorism (Category B, Centers for Disease Control, US) and biological warfare, the development of effective vaccines and treatments are of particular concern.

Adaptation to host environments is based on pathogenic microbes having evolved with complex gene circuits that allow bacteria to perceive and respond to different growth conditions. In the environment of the host, free iron is limited. However, iron is critical for bacterial growth as it is required in many biological processes including aerobic metabolism which requires that iron be available as a co-factor for respiratory enzymes. Iron-acquisition systems in most bacteria are subject to iron regulation, and this usually occurs through the direct or indirect action of the global regulator Fur. Little is known about iron regulation of gene expression in B. pseudomallei and B. mallei. Studies have demonstrated that B. pseudomallei produces a siderophore known as malleobactin under iron-limiting conditions [8] and that malleobactin is able to mobilize iron from transferrin and lactoferrin [9]. Siderophore biosynthesis and transport in B. pseudomallei and B. mallei would be expected to be iron regulated.

In the present studies, we have used DNA microarray analysis to examine relative gene expression levels in B. pseudomallei and B. mallei grown under low iron and high iron conditions.

2 Materials and methods

2.1 DNA microarrays

We used a combination of two previously described DNA microarrays, a low density oligo-microarray and a B. mallei whole genome microarray [10,11] and a newly developed B. pseudomallei whole genome microarray in these studies. Details of the low-density and the whole genome microarrays are in the Supplementary Text 1.

2.2 Bacterial strains and growth conditions for microarray analysis

Three bacterial strains including B. pseudomallei strains K96243, 1026b and B. mallei ATCC23344 were used in this study. Low iron medium (TSBDC) consisted of tryptic soy broth treated with Chelex-100 resin (Biorad), dialyzed, and supplemented with 50 mM glutamate and 1% (v/v) glycerol [12]. High iron medium was TSBDC plus 200 μM of ferric chloride (FeCl3) [12]. Thirty hour growth curves for all three bacterial strains grown in low and high iron media were performed at 37 °C with shaking at 250 rpm to determine if low and high iron conditions caused major differences in growth rates.

2.3 Total RNA isolation and microarray analysis

Total bacterial RNA was isolated from broth cultures using RNAwiz™ reagent (Ambion) with some modifications as previously described [10]. Total RNA was isolated from four different growth phases: exponential phase, OD600∼ 0.5–0.8; late exponential phase, OD600∼ 1.3–1.4 in B. mallei and 1.5–1.6 in B. pseudomallei; stationary phase, OD600∼ 1.5–1.6 in B. mallei and 1.7–1.8 in B. pseudomallei; late stationary phase, OD600∼ 1.7 in B. mallei and 2.0–2.2 in B. pseudomallei. Microarray analyses were performed as previously described [10] and details are in Supplementary Text 2. Data obtained from 8 hybridizations of 4 independent RNA preparations including flip-dye replications were used in each analysis.

2.4 Validation of microarray data by Northern blot hybridization and reverse transcription-PCR

Data from the low-density and the whole genome arrays were confirmed by Northern blot hybridization as previously described [10] and One-Step RT-PCR (QIAGEN). Details of the procedures are in the Supplementary Text 3. At least 9 iron-regulated genes were chosen for the confirmation including BPSL1775/BMA1178 (orbA homolog), BPSL1776/BMA1179 (pvdA homolog), BPSS0244/BMAA1826 (putative heme receptor protein), BPSS0495/BMAA1783 (nitroreductase family protein), BPSS0369/BMAA1800 (putative bacterioferritin-associated ferredoxin protein), BPSS0362/BMAA1511 (hypothetical protein), BPSS0357 (putative exported protein), sodB (BPSL0880), and BPSL0247/BMA3298 (putative iron–sulfur protein). We also used groEL gene (chaperonin, BPSL2697/BMA2001) as a control for a non-iron-regulated gene in this study. Both techniques were performed at least twice from both sources of RNA from B. pseudomallei K96243 and B. mallei ATCC23344.

2.5 Construction of deletion mutants

Deletion mutants B. pseudomallei PB247 and PB362 were constructed in two iron-regulated genes, BPSL0247 and BPSS0362, respectively, using pKAS46, an allelic exchange vector based on rpsL for counter selection [13,14]. Bacterial strains and plasmids used in this study are shown in Table 1, and details of procedures are in the Supplementary Text 4.

View this table:
1

Bacterial strains and plasmid used in this study

2.6 Determination of growth rate of mutant strains

The deletion mutants, PB362 and PB247, and the parent strain DD503 were grown in duplicate cultures of 75 ml of TSBDC and TSBDC supplemented with 200 μM FeCl3 at 37 °C, shaking 250 rpm, for 24 h. Flasks were inoculated with 0.75 ml of an overnight culture. The absorbance (OD600) of each culture was measured every 1–2 h.

2.7 Animal studies

The animal model of acute B. pseudomallei infection has been previously described [15]. Syrian hamsters (females, 6–8 weeks) were intraperitoneally injected with 100 μl of a diluted mid-exponential phase culture of B. pseudomallei DD503, PB362, or PB247, containing 10 CFU ml−1 of bacteria. The number of dead animals in each group (5 per group) was determined after 48 h. All animals used in these studies were cared for and used humanely according to the Guide to the Care and Use of Experimental Animals published by the Canadian Council on Animal Care.

3 Results

3.1 Bacterial growth in low and high iron conditions

Overall, bacterial growth was enhanced in high iron medium versus low iron medium. It was noticed that the B. mallei strain grew much slower compared to the two B. pseudomallei strains. Doubling times of B. pseudomallei and B. mallei strains at the exponential phase were 2 and 2.5 h, respectively. All bacterial cultures had reached stationary phase by the 20 h time-point, as demonstrated by an OD600 of approximately 1.8 and 1.7 for the B. pseudomallei strains K96243 and 1026b, and 1.5 for B. mallei strain ATCC23344. Investigation of stationary phase gene expression was conducted at this time-point.

3.2 Identification of iron-regulated genes in B. pseudomallei and B. mallei

Gene expression levels were compared between cells grown in low iron and high iron media at exponential phase and at stationary phase from at least two different time-points of each growth phase. This was initially studied using the low-density microarray. At the beginning of the exponential phase, gene expression patterns of most of the genes were similar regardless of the difference in the iron content in the media, but more differences were noted at the late-exponential phase. During stationary phase, the expression patterns of a number of specific genes were significantly different between low and high iron grown cells. Scatter plots of the relative gene expression levels during stationary phase for the three bacterial strains are shown in Fig. 1. In both B. pseudomallei strains, comparison of overall gene expression levels showed that during stationary phase, 2 genes responsible for iron uptake mechanisms, BPSL1776 (pvdA homolog) and BPSL1775 (orbA homolog) were highly induced. In addition, the sodB gene (BPSL0880), known to be involved in high-iron stress response, was down regulated. The results seen with known iron-regulated genes demonstrated that the maximum regulatory response to iron restriction occurs during the stationary phase of growth. In B. mallei, only the increased expression of BMA1179 (pvdA hololog) was seen among those iron-regulated genes using the low-density microarray. The hierarchical cluster analysis (Fig. 2) enables visualization of gene expression in B. pseudomallei K96243 grown under low iron conditions at four selected time-points including exponential phase, late-exponential phase, stationary phase and late-stationary phase. It clearly shows a highly differential expression of low-iron responsive genes during stationary phase.

1

Scatter plots of comparative gene expression levels seen in low iron (LEX.E) versus high iron (LEX.R) growth media during stationary phase based on 200 gene microarray analysis. Gene BPSL1775 is a homolog of orbA gene in B. cenocepacia; genes BPSL1776 and BMA1779 are homologs of pvdA gene in P. aeruginosa. Red data points are low-iron induced genes; green data points are low-iron repressed genes; and yellow data points are equally expressed genes.

2

Hierarchical cluster analysis of gene expression in B. pseudomallei K96243 grown in low iron medium from four different growth phases using 200-gene microarrays. Most known iron-regulated genes were differentially expressed when the bacteria were grown at stationary phase. Gene BPSL1775 is a homolog of orbA gene in B. cenocepacia; genes BPSL1776 and BMA1779 are homologs of pvdA gene in P. aeruginosa. Red bars and green bars indicate up-regulation and down-regulation in low iron condition, respectively.

Furthermore, whole genome microarray analysis of B. pseudomallei strain K96243 revealed that at least 198 (3.4%) genes were differentially expressed, while analysis of B. mallei strain ATCC23344 showed 278 (5.3%) differentially expressed genes based on the 95% confidence level. A summary of the responsive genes in B. pseudomallei K96243 and B. mallei ATCC23344 during growth in low iron concentration is provided in Table 2, and details of all differentially expressed genes are listed in the Supplementary Tables S1 and S2, respectively. Overall, gene expression profiles among B. pseudomallei K96243 and B. mallei ATCC23344 responded to low-iron growth condition were similar, except some low-iron repressed genes seen in B. pseudomallei such as sodB and genes for formate and succinate dehydrogenases were not differentially expressed in B. mallei. Microarray data obtained from these studies were reproducible among technical replicates (flip-dye replication) and biological replicates (independent sets of RNA preparation).

View this table:
2

Genes of B. pseudomallei K96243 and B. mallei strain ATCC23344 responsive to growth under low iron conditions

3.3 Validation of microarray data by Northern blot hybridization and RT-PCR

Iron regulation, as identified by microarray analysis, was confirmed for 9 representative genes using Northern blot hybridization and RT-PCR. Results showed that both techniques confirmed the microarray data obtained for the 9 tested genes. Figs. 3 and 4 show the results of selected experiments obtained from the Northern blot hybridization and the RT-PCR, respectively.

3

Northern blot hybridization: lanes 1 and 2, total RNA from B. mallei ATCC23344 grown in low iron and high iron media, respectively; lanes 3 and 4, total RNA from B. pseudomallei K96243 grown in low iron and high iron media, respectively; and lane R, RNA ladder (Millenium Markers, Ambion). Genes BMA1179/BPSL1776 (pvdA homolog), BPSL1775/BMA1178 (orbA homolog), and BMAA1800/BPSS0369 (bacterioferritin-associated ferredoxin gene) are identified as the low-iron induced genes, while the BMA2271/BPSL0880 (sodB) and BMA3298/BPSL0247 (iron–sulfur protein gene) are the high-iron induced genes. The groEL gene is a non-iron regulated gene and used as the control in this study. Ribosomal RNA bands (23S, 16S and 5S rRNA) were also used as the molecular markers throughout the study.

4

RT-PCR of selected iron-regulated genes in B. pseudomallei. RT-PCR products were obtained from RNA isolated from each strain grown under low (L) and high (H) iron conditions. Lane M, 1-Kb plus DNA ladder (Invitrogen); lanes 1 and 2, BPSS0362; lanes 3 and 4, BPSS0357; lanes 5 and 6, BPSL0247; lanes 7 and 8, pvdA; lanes 9 and 10, BPSS0244; lanes 11 and 12, BPSS0495; lanes 13 and 14, groEL.

3.4 Determination of growth rate in mutant strains

We constructed 2 deletion mutants, PB362 and PB247, of 2 iron-regulated genes in B. pseudomallei, BPSS0362 and BPSL0247 genes, respectively. The BPSS0362 has been annotated as a hypothetical protein, and this gene was highly induced during bacterial growth in low iron medium (see Supplementary Table S1) suggesting that the BPSS0362 gene might have a role function in iron-uptake mechanism. Another gene, BPSL0247, is a putative iron–sulfur protein gene [16]. This gene was repressed during bacterial growth in low iron conditions, suggesting that the expression of this gene was dependent upon the availability of iron. We suspected that if these 2 genes had an important physiological role when cells were grown in low or high iron conditions. The mutants PB362 and PB247 were grown in low and high iron media, and growth rates were compared to the parent strain DD503 [14]. All 3 strains grew at similar rate in both media. We noticed that all 3 strains grew slightly faster in high iron medium compared to low iron medium. However, these studies demonstrated that loss of the genes BPSS0362 and BPSL0247 had no significant effect on their growths in low and high iron conditions.

3.5 Animal studies

B. pseudomallei PB362 and PB247 were tested for virulence in the Syrian hamster model of melioidosis [15]. No difference in virulence was found between the mutants PB362 and PB247 and the parent strain DD503 (data not shown).

4 Discussion

These studies were designed to examine those biological processes associated with the growth of B. pseudomallei and B. mallei under iron-restricted conditions. Microarray analyses suggest that when B. pseudomallei and B. mallei are grown in low iron medium (TSBDC) most primary metabolic systems are repressed, and these bacteria respond to this situation by using alternative metabolic pathways, especially during bacterial growth at stationary phase. This suggests that the iron concentration in TSBDC medium is highly restricted during stationary phase. During stationary phase growth in low iron medium in B. pseudomallei K96243 and B. mallei ATCC23344, genes encoding NADH/formate/succinate-dehydrogenases, cytochrome oxidases and ATP synthase enzymes are repressed. These enzymes catalyze the oxidative and reductive reactions for generating energy in aerobic bacteria. We observed that genes for formate and succinate dehydrogenases were not differentially expressed in B. mallei; this lack of apparent regulation by iron may reflect the loss of gene regulatory mechanisms as a result of B. mallei evolving in an intracellular environment with relatively constant iron levels. In addition, genes encoding iron–sulfur proteins (BPSL0247, BMA3298), a component of many electron carriers, are also repressed in both bacterial strains. This indicates that ferric iron regulates the expression of respiratory enzymes involved in metabolic processes in B. pseudomallei andB. mallei.

Under iron-restricted conditions, alternative metabolic processes available to B. pseudomallei and B. mallei appear to play a major role in bacterial survival. The recently published genome sequence of B. pseudomallei strain K96243 [16] revealed 14 clusters of genes encoding the secondary metabolism that may be needed for bacterial survival under specific conditions. These clusters also included siderophore biosynthesis pathways of putative hydroxamate (BPSL1779–BPSL1774) and pyochelin (BPSS0581–BPSS0588) siderophores. Pyochelin genes are absent in the B. mallei genome [11]. Our present studies revealed that genes from both types of siderophore were induced in response to iron limitation growth of B. pseudomallei. In B. mallei, at least 5 hydroxamate siderophore synthesis genes (BMA1181–BMA1177) were induced. These findings suggest that the hydroxamate siderophore plays an important role in iron uptake in B. pseudomallei and B. mallei. A study by Ong et al. [17] demonstrated that both B. pseudomallei K96243 and B. mallei ATCC23344 produced siderophores in low-iron growth condition and we also confirmed the siderophore production using CAS assay [18] in all B. pseudomallei and B. mallei strains grown in TSBDC medium (data not shown). The roles of hydroxamate siderophore synthesis and iron-acquisition system have been studied thoroughly in P. aeruginosa and B. cenocepacia [19,20]. Pyoverdine and ornibactin are hydroxamate siderophores encoded by pvdA gene in P. aeruginosa and B. cenocepacia, respectively [20]. The transport of ferric-ornibactin complex in B. cenocepacia is associated with an outer membrane receptor encoded by gene orbA [21]. Leoni et al. [22] showed that pvdA transcripts were monocistronic and iron regulated at the transcriptional level. In our present study, the Northern blot analysis revealed that the pvdA gene (BPSL1176/BMA1179) were possibly monocistronic and/or polycistronic transcribed as seen in 2 positive bands of different mRNA sizes (ca. 1.5 and 3.5 Kb, see Fig. 3). A similar observation was also seen in a downstream gene (BPSL1775/BMA1178, orbA homolog). This may suggest that the pvdA and orbA genes can be monocistronic transcribed or polycistronic transcribed with other genes in the same operon. In addition, Leoni et al. [22] also demonstrated that the transcription of the pvdA gene was indirectly controlled by Fur. The fur gene was previously identified and characterized in B. pseudomallei; the expression analysis showed no increased fur mRNA levels in response to various stresses and iron conditions and Fur was found to play roles as a positive regulator of FeSOD (ferric-superoxide dismutase) and peroxidase [23]. In the present studies, we found that the fur gene was not differentially expressed in different iron concentrations; however, the sodB gene (FeSOD, BPSL0880), one of the Fur-regulated genes, was down regulated in low iron conditions, as shown previously [24]. Currently, details on the global regulation of iron-regulated genes by Fur in B. pseudomallei and B. mallei are unknown.

Another 2 genes, BPSS0357 (putative exported protein) and BPSS0362 (conserved hypothetical protein) were highly induced in low iron conditions. We were successful in inactivating the BPSS0362 gene in B. pseudomallei, and the growth of the mutant strain (PB362) showed no difference in growth rate in low and high iron medium compared to the parent strain. This suggests that the induction of the BPSS0362 gene has no direct role in the acquisition of the ferric iron (Fe3+), and the inactivation of this gene has no effect on bacterial growth under ferric iron-limiting conditions. However, the functions of these two genes are unknown.

To survive under low iron conditions, both bacterial species may use alternative metabolic pathways and available electron donors/receptors to generate energy. One potential pathway of choice is nitrogen metabolism. The BPSS0495 gene in B. pseudomallei and BMAA1783 in B. mallei encoding a nitroreductase family protein was highly induced during bacterial growth in iron-restricted conditions. This would suggest that B. pseudomallei and B. mallei reduce nitrogen compounds which can be nitrate and/or nitrite during growth in the limitation of iron. Another interesting observation was the induction of putative bacterioferritin-associated ferredoxin (bfd) genes, BPSS0369 in B. pseudomallei and BMAA1800 in B. mallei. This shows that the bacteria may use ferredoxin as electron donor, while most general electron donors such as cytochrome and/or ubiquinol-associated enzymes are unavailable. It was surprising to note that the putative bfd gene was induced during bacterial growth in low iron, while other genes encoding iron-containing proteins were repressed. In E. coli, the bacterioferritin-associated ferredoxin is encoded by the bfd gene located upstream of the bacterioferritin gene (bfr) encoding a 64 amino acid-residue protein identical to a region of NifU, a [2Fe–2S] protein found in nitrogen-fixing bacteria [25,26]. Location and putative protein structure of this gene in B. pseudomallei and B. mallei are similar to those in E. coli.

In summary, microarray analysis has been proven to be a useful tool for the evaluation of gene expression in B. pseudomallei and B. mallei, and the results from the current studies provide a better understanding of gene expression in B. pseudomallei and B. mallei in response to iron-limiting conditions as seen in the host environment.

Appendix A Supplementary data

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.femsle.2005.09.043.

List of supplementary information:

Supplementary Text 1: Details of the low-density and the whole genome microarrays.

Supplementary Text 2: Details ofmicroarrays analysis procedures.

Supplementary Text 3: Details ofthe Northern blot hybridization and One-Step RT-PCR procedures.

Supplementary Text 4: Details of mutagenesis procedures.

Supplementary Table Sl: Comparative gene expression analysis of B. pseudomallei strain K96243 in response to growth in low iron versus high iron medium.

Supplementary Table S2: Comparative gene expression analysis of B. mallei strain ATCC23344 in response to growth in low iron versus high iron medium.

Supplementary Table S3: List ofPCR primers used in northern blot hybridization and RT -PCR.

Acknowledgements

This work was funded by the Canadian Institutes of Health Research Grant MOP-36343 to D.E.W. and National Institute of Allergy and Infectious Disease grants R01AI50565 and R01AI056006 to W.C.N. D.E.W. is a Canada Research Chair in Microbiology. We wish to thank the staff at the Southern Alberta Microarray Facilities for technical support, and we also thank Paige Pardington and Dr. Robert Cary at Los Alamos National Laboratory for printing the B. pseudomallei whole genome microarray (supported by a Los Alamos Research and Development grant to J.D.).

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
View Abstract