OUP user menu

Type III secretion system cluster 3 is required for maximal virulence of Burkholderia pseudomallei in a hamster infection model

Jonathan Warawa, Donald E. Woods
DOI: http://dx.doi.org/10.1016/j.femsle.2004.10.045 101-108 First published online: 1 January 2005


Burkholderia pseudomallei, the etiological agent of melioidosis, is an animal pathogen capable of inducing a highly fatal septicemia. B. pseudomallei possesses three type III secretion system (TTSS) clusters, two of which (TTSS1 and TTSS2) are homologous to the TTSS of the plant pathogen Ralstonia solanacearum, and one (TTSS3) is homologous to the Salmonella SPI-1 mammalian pathogenicity island. We have demonstrated that TTSS3 is required for the full virulence of B. pseudomallei in a hamster model of infection. We have also examined the virulence of B. pseudomallei mutants deficient in several putative TTSS3 effector molecules, and found no significant attenuation of B. pseudomallei virulence in the hamster model.

  • Melioidosis
  • Burkholderia pseudomallei
  • Type III secretion
  • Animal studies
  • Attenuation

1 Introduction

Burkholderia pseudomallei is a Gram-negative saprophytic bacterial pathogen responsible for the animal disease melioidosis, which is endemic in the tropics of Southeast Asia and northern Australia. Melioidosis commonly presents as either pneumonia or skin abscesses and is acquired typically through inhalation, skin penetration or possibly through ingestion [1]. Melioidosis can progress into a highly fatal bacteremia with a mortality rate of ∼66%, and current treatment is limited to a small number of antibiotics that are only effective in halving the mortality of bacteremia to ∼33% [1].

Several B. pseudomallei pathogenicity determinants have been elucidated including both secreted and cell-associated products. Lipopolysaccharide (LPS) and capsule polysaccharide (CPS) are cell-associated determinants that have both been shown to be required for the full virulence of B. pseudomallei in animal models [2,3]. Secreted products have been identified that may play a role in B. pseudomallei virulence including several proteases, phospholipase C, hemolysin, lecithinase and lipase [48].

Three type III secretion systems (TTSSs) have been identified in the sequenced genome of B. pseudomallei, including two TTSS clusters (TTSS1 and TTSS2) homologous to the Ralstonia solanacearum plant pathogen TTSS [9,10] and a third TTSS cluster (TTSS3) homologous to the SPI-1 animal pathogenicity island of Salmonella typhimurium [11]. Several Salmonella TTSS-delivered effector homologues have been identified in the TTSS3 cluster, including homologues of SipB (translocation and apoptosis), SipC (translocation and invasion), SopA (PMN transmigration) and SopE (invasion) [12,13]. The Sip (Salmonella invasion protein) and Sop (Salmonella outer protein) homologues have been named Bip and Bop in B. pseudomallei [14]. In addition, three Bap (Bsa associated protein) proteins are located in the TTSS3 cluster, which have been hypothesized to be effector molecules delivered by TTSS3 [14]. BapA has no significant homologue, while BapB and BapC are homologous to IacP and IagB, respectively, of Salmonella [14], both of which do not have a clearly defined role in virulence.

A recent report has suggested a role for the TTSS3 cluster for the survival and persistence of bacteria within a murine macrophage-like cell line and for the escape of B. pseudomallei from endocytic vacuoles [14]. Furthermore TTSS3 was demonstrated to play a role in the invasion of non-phagocytic cells, where the effector molecule BopE contributed to TTSS-mediated invasion [15]. In this study, we have investigated the role of type III secretion in B. pseudomallei virulence in the hamster infection model.

2 Materials and methods

2.1 Bacterial strains and plasmids

All bacterial strains and plasmids used in this study are described in Table 1. Bacteria were routinely grown with LB media at 37 °C supplemented with antibiotics as appropriate, and liquid cultures were shaken at 250 rpm. Antibiotics were used at the following concentrations: 100 μg/ml carbenicillin (Cb), 25 μg/ml kanamycin (Km), 100 μg/ml streptomycin (Sm) and 50 μg/ml polymyxin B (Pm).

View this table:
Table 1

Bacterial strains, plasmids, and oligonucleotides

DD503B. pseudomallei 1026b derivative, PmR, SmR, AGS, TcS[24]
ΔsctUBp1TTSS1 mutantThis study
ΔsctUBp2TTSS2 mutantThis study
ΔsctUBp3TTSS3 mutantThis study
ΔsctUBp1,2TTSS1 and 2 mutantThis study
ΔsctUBp2,3TTSS2 and 3 mutantThis study
ΔsctUBp1,3TTSS1 and 3 mutantThis study
ΔsctUBp1,2,3TTSS1, 2 and 3 mutantThis study
ΔsctUBp1,2,3/pBHR1-sctUBp3This study
DH10bE. coli cloning strain[29]
SY327(λpir)E. coli cloning strain[30]
SM10(λpir)E. coli cloning strain[31]
S17-1E. coli cloning strain[31]
pCR2.1PCR cloning vector, ApRInvitrogen
pSKcloning vector, ApRStratagene
pBHR1Mobilizable cloning vector, CmR, KmRMoBiTec [32]
pKAS46suicide vector, ApR, KmR[33]
pKAS46-ΔsctUBp1This study
pKAS46-ΔsctUBp2This study
pKAS46-ΔsctUBp3This study
pKAS46-ΔbopAThis study
pKAS46-ΔbopEThis study
pKAS46-ΔbapAThis study
pKAS46-ΔbapCThis study
pBHR1-sctUBp3This study
OligonucleotidesSequence (5′→ 3′)b
  • aR, resistant; S, sensitive; Pm, polymyxin B (50 μg/ml); Sm, streptomycin (100 μg/ml); AG, aminoglycoside; Tc, tetracycline; Cm, chloramphenicol; Km, kanamycin (25 μg/ml); Ap, ampicillin (100 μg/ml.

  • bNovel restriction sites used for cloning were incorporated by PCR and are indicated by underlined portions of each oligonucleotide.

2.2 Mutagenesis of TTSS genes

Deletion mutants were generated for each copy of sctU by removing an internal portion of the respective gene and ligating together the 5′ and 3′ portions of sctU using a unique restriction site. This was accomplished through PCR amplification (HotStar Taq polymerase, Qiagen) of the upstream and downstream sequence flanking each sctU gene using the oligonucleotides listed in Table 1. PCR fragments were cloned directly into pCR2.1, and then respective upstream and downstream fragments were assembled in pSK using oligonucleotide incorporated restriction sites. The assembled gene deletions were cloned into pKAS46, which were then delivered into B. pseudomallei strains for mutagenesis as previously described [16]. As a result, a 1026-bp fragment was excised from the 1059 bp sctUBp1 gene, 1050 bp from the 1083 bp sctUBp2 gene and 1063 bp from the 1236 bp sctUBp3 gene, where all gene mutations used the native stop codon except for sctUBp3 which used a novel oligonucleotide-encoded stop codon. Confirmation of chromosomal gene deletions was confirmed by identification of truncated genes by PCR, as described below. Likewise, all combinations of double and triple mutants of sctU genes were generated by stepwise mutagenesis. For example, the ΔsctUBp1,2,3 triple mutant was generated by mutating sctUBp2 in a ΔsctUBp1,3 background. Mutants of several TTSS3 effector molecules were generated as described above, using oligonucleotides listed in Table 1. A 1515 bp fragment was excised from the 1539 bp bopA gene, 711 bp from the 786 bp bopE gene, 2607 bp from the 2643 bp bapA gene and 528 bp from the 564 bp bapC gene (see Fig. 1).

Figure 1

Schematic representation of the three B. pseudomallei TTSS clusters. The putative open reading frames of the three TTSS clusters are depicted to scale, with the sct (secretion and cellular translocation) apparatus genes colored blue, the rts (regulation of type three secretion) regulator genes colored red, the cts (chaperone of type three secretion) chaperone genes colored green, the putative effector molecules colored magenta, the TTSS2-associated pilus pil genes colored orange and unknown putative orf's in cyan. The red asterisk indicates the sctU gene from each TTSS cluster mutated in this study.

2.3 Analysis of gene deletion mutants

PCR analysis was used to identify clones that possessed a truncated gene copy of the gene being mutagenized. Chromosomal DNA was isolated from each mutant clone using Wizard Genomic DNA Purification Kit (Promega), which was subsequently analyzed using gene specific oligonucleotide sets: sctUBp1 (GCG ATG GAA TTC TCC GAT AGC CGA GCC GAC GCA TGA GCG, GTG AAC AGC AGG AGC GAC GGA AAC GCG GAT AGC GC), sctUBp2 (CGG CAT GAA TTC CTC CCG AAG GAA CCG CGG CGA TGG C, GAT GTT CAG CGA CAG CCG GTA CAG CGT CGT GAA CAG C) and sctUBp3 (TGC ATC CCA TGG GCA GCC TGG TGA AGT GAG CGA TGG CC, GCC ATT TCC TGC GCC TCG ATG TCG TCT ATG TTC ACG TCG). PCR reactions were resolved on a 0.8% TAE agarose gel (Fig. 2).

Figure 2

Confirmation of sctU mutants. Chromosomal DNA from each of the TTSS mutant constructs was PCR amplified with gene specific oligonucleotide sets as described in Section 2. The top panel (sctUBp1), middle panel (sctUBp2) and bottom panel (sctUBp3) all correspond to the three separate oligonucleotide sets used to amplify each of the three sctU genes. Lanes are labeled with the short form of the mutant name: Wt = DD503, U12 =ΔsctUBp1,2, U23 =ΔsctUBp2,3, U13 =ΔsctUBp1,3, U123 =ΔsctUBp1,2,3, U1 =ΔsctUBp1, U2 =ΔsctUBp2, U3 =ΔsctUBp3. The arrows indicate the theoretical bp length of the PCR amplicon for either the wildtype or deletion-mutant gene copies of each respective gene, which correlated in size with a DNA ladder. The PCR was resolved on a 0.8% TAE agarose gel.

2.4 Complementation of sctUBp3

To trans complement the ΔsctUBp3 mutation, the sctUBp3 gene was PCR amplified using the oligonucleotides U3FULL(+) AND U3FULL(−) (Table 1), the resulting DNA fragment was cloned into the pBHR1 mobilizable vector using restriction sites NcoI and PinAI, and the resulting plasmid, pBHR1-sctUBp3 was conjugated into ΔsctUBp1,2,3 using Escherichia coli strain S17-1.

2.5 Virulence studies

Syrian golden hamsters were infected with B. pseudomallei strains in accordance with the Canadian Council on Animal Care Guidelines, as described previously [17]. Briefly, groups of six hamsters were infected intraperitoneally with 100 μl of an appropriate dose of B. pseudomallei from a serially diluted logarithmic phase shaking culture. At 48 h post infection, endpoint data were collected, and LD50 values were calculated according to Reed and Muench [18]. The LD50 values obtained were transformed into logarithmic form and used to obtain the geometric mean of LD50 values for each strain tested and for statistical analysis. The data were analyzed for significance by the Wilcoxon signed rank test, and 95% confidence intervals were determined.

3 Results

3.1 Nomenclature of TTSS clusters of B. pseudomallei

Several groups have reported varied naming systems for the genes present in the three TTSS clusters of B. pseudomallei (Fig. 1). In the interest of clarity, we propose the use of the nomenclature system proposed by Hueck [19] for the naming of type III secretion assembly (TTSA) genes throughout all prokaryotes. A summary of this standardized naming system as it applies to B. pseudomallei TTSS genes is provided in Table 2, where previous gene names are indicated in parallel with the new standardized gene names.

View this table:
Table 2

Nomenclature of Burkholderia pseudomallei TTSS genesa

New gene namePrevious gene name (gene annotation)
sctCbpscC (BPSS1390)BpscC2 (a and b) (BPSS1603 and 1592)bsaO (BPSS1545)
sctDbpscD (BPSS1407)BpscD2 (BPSS1614)bsaM (BPSS1547)
sctF√ (BPSS1409)√ (BPSS1612)bsaL (BPSS1548)
sctGOrf4 (BPSS1399)√ (a and b) (BPSS1607 and 1622)NI
sctIOrf5? (BPSS1398)HrpJ homologue? (BPSS1623)bsaK (BPSS1549)
sctJBpscJ (BPSS1397)BpscJ2 (BPSS1624)bsaJ (BPSS1550)
sctKOrf6 (BPSS1396)√ (BPSS1625)√ (BPSS1551)
sctLbpscL (BPSS1395)BpscL2 (BPSS1626)√ (BPSS1552)
sctMNI√ (BPSS1629a)√ (BPSS1553)
sctNbpscN (BPSS1394)BpscN2 (BPSS1627)bsaS (BPSS1541)
sctOOrf7? (BPSS1393)HrpD homologue? (BPSS1628)bsaT (BPSS1540)
sctPNI√ (BPSS1630)bsaU (BPSS1539)
sctQbpscQ (BPSS1403)bpscQ2 (BPSS1618)bsaV (BPSS1538)
sctRbpscR (BPSS1404)bpscR2 (BPSS1617)bsaW (BPSS1537)
sctSBpscS (BPSS1405)bpscS2 (BPSS1616)bsaX (BPSS1536)
sctTbpscT (BPSS1392)bpscT2 (BPSS1629)bsaY (BPSS1535)
sctUbpscU (BPSS1400)bpscU2 (BPSS1621)bsaZ (BPSS1534)
sctVbpscV (BPSS1401)bpscV2 (BPSS1620)bsaQ (BPSS1543)
sctWNINIbsaP (BPSS1544)
  • aThe Hueck nomenclature for TTSS genes is proposed for the three TTSS clusters, which will include the subscriptBp plus the cluster number. Previous gene names are indicated for TTSS1, TTSS2 [10] and TTSS3 [14], as are the gene annotation identifiers. Where no gene homologue has been identified, it is marked ‘NI', and where a gene homologue exists yet has not been previously reported, it is indicated with a ‘√'.

All B. pseudomallei structural TTSS genes are to be named based on the nomenclature previously proposed by Hueck whereby genes are named sct (secretion and cellular translocation) followed by the subunit identifier, followed by the species identifier in subscript [19]. The previous nomenclature for TTSS clusters 1 and 2 originate from Rainbow et al. [10], whereas the cluster 3 nomenclature was proposed by Stevens et al. [14]. Attree and Attree had previously used the sct nomenclature for B. pseudomallei TTSS3 [11], before the proposed nomenclature by Stevens et al. Winstanley and coworkers [10] had also initially used the sct nomenclature for TTSS1 before later altering the nomenclature. This study focuses on the sctU family of genes in the three TTSS clusters; the alternative to the sct nomenclature would require somewhat convoluted discussions of bpscU from TTSS1, bpscU2 from TTSS2 and bsaZ from TTSS3.

TTSS clusters 1 and 2 are homologous to the TTSS cluster from Ralstonia solanacearum. There are a number of TTSS genes that appear to be plant TTSS specific, while there are other TTSS genes that appear to be well conserved in both animal and plant TTSS clusters (i.e. sctCDLNQRSTUV). We propose that some of the plant specific TTSS genes can be also reclassified into the sct nomenclature. For example, HrpY is the major component of the plant TTSS needle as is SctF for the animal TTSS needle [20,21]. Based on functionality as well as similar protein sizes and some sequence alignment, we propose that the hrpY homologues in B. pseudomallei TTSS clusters 1 and 2 be included as members of the sctF family. We have also found that HrpK from R. solanacearum is homologous to YscG from Y. pestis (25% identity, 36% similarity over 107aa), and therefore propose that the hrpK homologues from B. pseudomallei TTSS clusters 1 and 2 be included in the sctG family. HrpH from R. solanacearum is 21.9% identical to YscK from Y. pestis over a global alignment, suggesting that B. pseudomallei hrpH homologues belong in the sctK family. Several plant-like TTSS genes remain to be classified under the sct nomenclature system, although based on localization within the TTSS, we predict that future studies will demonstrate that HrpX, HrpJ and HrpD are functionally related to SctE, SctI and SctO, respectively.

3.2 TTSS3 is a virulence determinant in hamsters

Numerous gene products contribute to the formation of a TTSS apparatus. We chose to perform mutagenesis on the sctU gene since previous studies of the Yersinia homologue (yscU) indicated that this gene would encode one of the major inner membrane subunits [22], and this gene is conserved in all TTSSs investigated [23]. Indeed, a homologue of sctU was found in each of the three TTSS clusters of B. pseudomallei. Each of the three copies of sctU was subjected to mutagenesis as described in Section 2.

To investigate the role of TTSSs in B. pseudomallei virulence, DD503 and the TTSS mutants were investigated in the previously described Syrian golden hamster model of infection [17]. We determined the log LD50 values of the parent strain and each of the mutants. Isolates were injected intraperitoneally into hamsters, and lethality was determined by observation for 2 days (Table 3). The 95% confidence intervals were determined to be 1.340–3.610.DD503 was previously reported to have a 48 h log LD50 value of less than 1 [24], and we further estimated the log LD50 to be 0.6021 (Table 3), which is consistent with the published log LD50 of B. pseudomallei in hamsters of 0.7782 [25]. Both the ΔsctUBp1 and ΔsctUBp2 mutants had LD50 values similar to the parental strain; however, the ΔsctUBp3 mutant was significantly attenuated with a log LD50 of 3.3802 cfu.

View this table:
Table 3

Virulence of Burkholderia pseudomallei strains in the hamster model

StrainLog LD50
  • aSignificantly elevated over DD503 (p < 0.05, 95% CI = 1.340–3.610).

  • bSignificantly elevated over all other strains (p < 0.05).

To determine if there were synergistic effects of the three TTSS clusters, we also generated all variations of double and triple mutants and examined these strains in the hamster infection model. The presence of a ΔsctUBp1 or ΔsctUBp2 mutant did not contribute to any significant attenuation of B. pseudomallei virulence, as indicated by the ΔsctUBp1,2 double mutant (Table 3). However, both of the double mutants that shared a ΔsctUBp3 mutation (ΔsctUBp1,3 and ΔsctUBp2,3) had statistically significant attenuated log LD50 values of 3.0414 and 3.1461, respectively (p < 0.05, Table 3). These LD50 values are not significantly different from one another or from the LD50 of the ΔsctUBp3 single mutant. The log LD50 value of the ΔsctUBp1,2,3 triple mutant was, however, significantly elevated (p < 0.05) over all of the strains examined, indicating that there may be some contribution to virulence from TTSS1 and TTSS2.

To confirm the importance of the sctUBp3 gene in B. pseudomallei virulence, we complemented the ΔsctUBp3 mutation in the ΔsctUBp1,2,3 triple mutant, generating the strain ΔsctUBp1,2,3/pBHR1-sctUBp3 which was tested in the hamster model. We observed that virulence was enhanced to Wt levels with a log LD50 of 1.3424 (Table 3), where the difference in the log LD50 value between that of the complemented strain (1.3424) and that of the Wt (0.6021) was not statistically significant (p > 0.05). This complementation confirms that the ΔsctUBp3 mutant was non-polar and could be complemented by a single gene copy.

3.3 Role of TTSS3 effector molecules in hamster infection

Given the importance of TTSS3 in mediating B. pseudomallei virulence in hamsters, we decided to investigate the contribution to virulence of individual TTSS3 effector molecules. We therefore generated deletion mutants of the bopA, bopE, bapA and bapC putative effector molecule genes and examined the mutants in the hamster infection model. The LD50 values were calculated and it was found that no effector molecule mutant contributed to a significant attenuation of B. pseudomallei virulence in the acute hamster model (Table 3), suggesting that B. pseudomallei virulence is not critically dependent upon the activity of any single one of the putative effector molecules tested.

4 Discussion

These data constitute the first report that the principal TTSS cluster required for maximal virulence of B. pseudomallei in the hamster model is cluster 3. This role for TTSS3 in B. pseudomallei animal virulence is supported by the recent report that TTSS3 is required for the full virulence of B. pseudomallei in a murine model of infection [26]. Given that these two separate animal models generate similar conclusions, it appears more likely that TTSS3 plays an important role in B. pseudomallei virulence in all of its animal hosts. The related pathogen Burkholderia mallei has also been recently studied in both hamster and murine models, with respect to the role of the TTSS3 cluster homologue, where mutations in homologues of sctUBp3 and sctVBp3bsaZ and bsaQ) resulted in the attenuation of B. mallei in both animal hosts [27]. The TTSS mutants were attenuated in hamsters to LD50 values of >106 from the Wt LD50 of <13 [27]. This report suggests that B. mallei is much more dependent upon TTSS for its virulence in animals than B. pseudomallei, but also suggests that these related pathogens may share some TTSS-mediated mechanisms of host subversion.

Our results demonstrate that TTSS1 and TTSS2 are not critically required for B. pseudomallei virulence in hamsters, although there may be some minimal but statistically significant contribution to virulence from these clusters. This result was anticipated in part due to the in silico analysis indicating that both TTSS1 and TTSS2 are homologous to the Ralstonia solanacearum TTSS cluster required for plant interaction [9,10]. However, it remains possible that these plant-like TTSS clusters may play a supportive role during animal infection that remains to be identified. In support of this possibility, we have recently reported on a whole-genome microarray study which indicates that a growth condition that represses expression of TTSS3 also represses expression of TTSS2 [28], indicating that B. pseudomallei TTSS clusters are co-regulated under some conditions. It is also telling that the three related bacteria B. thailandensis, B. mallei and B. pseudomallei all share homologues of both TTSS2 and TTSS3 (data not shown) yet a plant host has never been identified for any of these bacterial species, raising the question as to why these bacteria have not lost the plant-like TTSS2 cluster if it is only required for plant interaction.

We have also demonstrated that mutations in four different TTSS3 effector genes (ΔbopAΔbopE, ΔbapA and ΔbapC) do not significantly attenuate B. pseudomallei virulence in hamsters, suggesting that these putative effector molecules are not individually, critically important for B. pseudomallei pathogenesis in hamsters. However, we have demonstrated that abolition of secretion from the TTSS3 apparatus (i.e. ΔsctUBp3) does lead to significant attenuation, suggesting that there are unidentified critical TTSS3 effector molecules, or that the TTSS3 effector molecules work in concert to enhance virulence.

Of the four TTSS3 effector molecule mutants that we generated and examined, two were also examined in a recent study by Stevens et al. using a murine model of infection (ΔbopA and ΔbopE) but were similarly found to not significantly attenuate the virulence of B. pseudomallei [26]. In addition, Stevens et al. had generated a non-attenuated mutant of the gene BPSS1514, and had designated this gene as bopB based on the premise that this gene is a homologue of sopB from Salmonella. However, the 210 aa BPSS1514 protein is approximately half the size of SopB, ranging from 416 to 563 aa in size. In addition, BPSS1514 protein alignment with SopB from Salmonella dublin (Accession No. AF060858.1) was insignificant (data not shown), and the assertion by Stevens et al. that BPSS1514 possesses the CX5R motif, typical of SopB and some phosphatases, is incorrect. We propose that the label of BPSS1514 as BopB by Stevens et al. is in error, and future studies will be required to investigate whether BPSS1514 is a TTSS3 effector molecule.

In conclusion, we have demonstrated that TTSS3 is required for the full virulence of B. pseudomallei in the hamster model, which supports other studies using murine models to study TTSS3 from B. pseudomallei and the TTSS3 cluster homologue in B. mallei. Likewise, our hamster virulence data support the murine model demonstrating that bopA and bopE mutants are alone incapable of significantly attenuating B. pseudomallei virulence in animals. We also report that the two additional B. pseudomallei TTSS clusters (TTSS1 and 2) may play some role in B. pseudomallei virulence in hamsters, singly or synergistically; however, inactivation of the putative effector molecules BapA or BapC does not lead to attenuation in the hamster infection model.


This project was funded by the Canadian Institutes of Health Research MOP-36343. Oligonucleotide generation and sequencing was performed by UCDNA Services at the University of Calgary. D.E.W. is a Canada Research Chair in Microbiology.


  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].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
View Abstract