Advances in Infectious Diseases, 2012, 2, 53-61
http://dx.doi.org/10.4236/aid.2012.23008 Published Online September 2012 (http://www.SciRP.org/journal/aid)
53
A wzt Mutant Burkholderia mallei Is Attenuated and
Partially Protects CD1 Mice against Glanders
Aloka B. Bandara
Center for Molecular Medicine and Infectious Diseases, Virginia-Maryland Regional College of Veterinary Medicine, Virginia Poly-
technic Institute and State University, Blacksburg, USA.
Received April 23rd, 2012; revised May 25th, 2012; accepted June 27th, 2012
ABSTRACT
Burkholderia mallei is the etiologic agent of glanders in solipeds and humans. Lipopolysaccharide (LPS) is a major
component of cell envelop of this pathogen. O-antigen, the most external component of LPS, is a virulence factor and a
protective antigen in many pathogenic bacteria. Two putative proteins named Wzm (integral membrane protein) and Wzt
(hydrophilic ATP-binding protein) are believed to make up an ABC-2 transporter of B. mallei that facilitates transport of
components of O-antigen from cytosol to outer-membrane. We studied the importance of wzt (encoding Wzt) to growth,
LPS O-antigen profile, and pathogenicity of B. mallei. A wzt mutant strain was generated by deleting a portion of the
wzt in B. mallei wild type strain ATCC 23344 by gene replacement. Compared to the wild type strain, the wzt mutant
displayed slower growth in v it r o and less lethality in CD1 mice when inoculated intraperitoneally. The 50% lethal doses
(LD50) of the wild type and the wzt mutant strains were 5.9 × 105 and 9.1 × 105 cfu, respectively. CD1 mice inoculated
with a non-lethal dose of the wzt mutant produced specific serum immunoglobulins IgG1 and IgG2a and were partially
protected against challenge with 11.2 times LD50 of the wild type strain. These findings suggest that the wzt is required
for optimal in vitro growth and pathogenesis of B. mallei, and a wzt mutant protects CD1 mice against glanders.
Keywords: Burkholderia mallei; ABC-2 Transporter; wzm Integral Membrane Protein; wzm Hydrophilic ATP-Binding
Protein; Glanders; CD1 Mice; Pathogenicity; Protection
1. Introduction
Burkholderia mallei, the causative agent of glanders, is a
Gram-negative, aerobic bacillus. This bacterium is pri-
marily responsible for disease in horses, mules, donkeys
and occasionally humans [1-3]. Relatively little is known
about the mechanisms of B. mallei pathogenesis [4,5]. In
gram-negative bacteria, lipopolysaccharides (LPS), com-
monly referred to as endotoxins, are a major component
of cell envelopes [6,7]. Bacterial outermembranes pro-
vide the “barrier function” largely due to the presence of
LPS [8]. Bacterial strains expressing a smooth phenotype
synthesize LPS molecules that are composed of three
covalently linked domains: an O-polysaccharide antigen
(O-antigen), a core region, and a lipid A moiety [9]. The
O-antigen is the most external component of LPS, and it
consists of a polymer of oligosaccharide repeating units.
Chemical composition of O-antigens varies among dif-
ferent bacterial species, as a result of the genetic varia-
tion in the genes involved in O-antigen biosynthesis,
designated the wb cluster. The genetics of O-antigen
biosynthesis have been intensively studied in the En-
terobacteriaceae, and it has been shown that the wb
clusters usually contain genes involved in biosynthesis of
activated sugars, glycosyl transferases, O-antigen poly-
merases, and O-antigen export [10,11]. Burtnick et al. [12]
identified the gene cluster responsible for O-antigen bio-
synthesis in B. mallei ATCC 23344, and determined the
physical structure of the B. mallei ATCC 23344 O-anti-
gen.
In bacteria, LPS O-antigen is exported to the cell sur-
face using three distinctive pathways, as reviewed by
Samuel and Reeves [13]. One of these pathways called
ATP-binding cassette (ABC) transporter-dependent path-
way, which has been found in E. coli O8 and O9 and
Klebsiella pneumoniae O1 and O12 [14-17] is comprised
of an integral membrane protein, Wzm, and a hydrophilic
protein containing an ATP-binding motif, Wzt. The me-
chanism for the biosynthesis of LPS O-antigen in B. mal-
lei is largely unknown. The genome of B. mallei strain
ATCC 23344 carries a wzm gene encoding a putative
Wzm integral membrane protein of an ABC-2 transporter
complex, and a wzt gene encoding a putative Wzt hydro-
philic protein of this complex [12]. In this communica-
tion, we report the influence of wzt on in vitro growth
and in vivo pathogenicity of B. mallei, and the protective

A wzt Mutant Burkholderia mallei Is Attenuated and Partially Protects CD1 Mice against Glanders
54
efficacy of a wzt mutant as a vaccine candidate against
glanders.
2. Materials and Methods
2.1. DNA and Protein Sequence Analyses
The nucleotide sequences of the B. mallei wzm and wzt
genes encoding respectively the putative Wzm and the
putative Wzt proteins were analyzed with DNASTAR
software (DNASTAR, Inc., Madison, WI). The presence
of any signal sequence of Wzm and Wzt proteins was
predicted by using the SignalP 3.0 server (http://www.
cbs.dtu.dk/) [18]. The destination of the Wzm and Wzt
proteins upon translation and processing was predicted
using the Subloc v1.0 server (http://www.bioinfo.tsinghua.
edu.cn/). The identity of B. mallei Wzm and Wzt to pro-
teins of the EMBL/GenBank/DDBJ databases was ana-
lyzed using the BLAST software [19].
2.2. Bacterial Strains, Plasmids, and Reagents
B. mallei strains ATCC 23344 and 23344∆sacB [20].
were obtained from our culture collection. B. mallei strains
were grown in trypticase soy broth or trypticase soy agar
(Difco Laboratories, Sparks, MD) supplemented with 4%
glycerol (TSB-G and TSA-G, respectively) at 37˚C in the
presence of 5% CO2 as previously described [21]. Es-
cherichia coli XL1Blue was used for general cloning,
and E. coli S17-1 [22] was used as a mobilizing strain for
constructing mutants. The suicide vector pGRV2 [23]
that carries the counter-selectable marker sacB was em-
ployed in generating the mutant B. mallei strains. Bacte-
ria containing plasmids were grown in the presence of
polymyxin at 15 g/ml. Genomic DNA from B. mallei
strain ATCC 23344 and plasmid DNA from recombinant
E. coli strains were harvested by use of kits obtained
from Qiagen (Qiagen Inc., Valencia, CA). Restriction di-
gests, Klenow reactions, and ligations of DNA were per-
formed using standard procedures [24]. All experiments
with live B. mallei were performed in a Biosafety Level 3
facility in the Infectious Disease Unit of the Virginia-
Maryland Regional College of Veterinary Medicine per
CDC-approved standard operating procedures.
2.3. Construction of a ∆wzt Mutant Strain of
B. mallei
PCR primers were designed to amplify the 5’ and 3’ ends
of the wzt gene (BMA1985) on chromosome I of strain
ATCC 23344 (NC_006348). The primer sequences (5’ to
3’) were as follows: wzt-1, CGGCCATGGATGTCCT
CTGAATTGCCG; wzt-2, CGGGGATCCCCGCACG-
TACATGCCGCTCG; wzt-3 CGGGGATCCGCTGCG
CAACGACGTGGAGT; and wzt-4 CGGCCATGGTT
CAGATTTCATTGCGCC. The 5’ end of the primers
wzt-1 and wzt-4 carried NcoI sites (in bold case), whereas
the 5’ ends of primers wzt-2 and wzt-3 carried BamHI (in
bold case). The PCR amplifications were performed us-
ing nearly 100 ng of ATCC 23344 genomic DNA and 20
pmoles oligodeoxyribonucleotide primers.
The fragment-1 was amplified using primers wzt-1 and
wzt-2, and the fragment-2 was amplified using wzt-3 and
wzt-4. Each fragment was restricted digested with NcoI
and BamHI. The digested two fragments were then cloned
into the NcoI digested plasmid pGRV2 [23] to produce
pABwzt.
The plasmid pABwzt was harvested from XL1Blue,
and introduced into competent E. coli S17-1 cells by
electroporation to produce S17-1 (pABwzt). The plasmid
was then delivered to B. mallei 23344∆sacB [20] via
conjugation with S17-1 (pABwzt) by using a membrane
filter mating technique, as described elsewhere [20].
Strain 23344∆sacB was chosen as the platform strain, as
its resistance to sucrose was useful as a non-antibiotic
marker in selecting the recombinant stains [20]. Poly-
myxin was used to counterselect E. coli. One of the
23344∆sacB: pABwzt colonies was used to inoculate
TSB-G. Ten-fold dilutions of the overnight culture were
spread onto TSA-G + 5% sucrose. Six sucrose-resistant
colonies were screened by PCR for the deletion in the
wzt gene (data not shown). One of the colonies carrying
the deletion was chosen for further work and designated
23344∆sacB∆wzt.
2.4. RNA Isolation and Reverse
Transcription-PCR (RT-PCR)
Extraction of RNA, treatment with DNase, and RT-PCR
were perfomed as described elsewhere [20]. The primers
wbiA-Forward (5’ TAGATTCCATACGAGTAGTC 3’)
and wbiA-Reverse (5’ ATGTGGCGCCTGACGCTCAA
3’) were used for PCR amplifycation of wbiA, whereas,
primers wzt-1 and wzt-4 (see section 2.3) were used for
amplification of wzt.
2.5. Extraction and Analysis of LPS and Other
Cellular Components
Single colonies of the wild type, the sacB mutant, and the
wzt mutant were patched on TSA-G, and incubated at
37˚C for 96 h in 5% CO2. The cells were harvested and
treated with 0.5% phenol for 72 h at 4˚C. The lysate was
used for extraction of LPS using a modified hot aque-
ous-phenol procedure [25,26]. Following extraction, the
resulting phenol and aqueous phases were combined and
dialyzed in distilled water to remove phenol. The dialys-
ates were then clarified by centrifugation and concen-
trated by lyophilization. The crude preparations were
solubilized in 10 mM Tris-HCl (pH 7.5), 1 mM MgCl2, 1
mM CaCl2, 50 μg·ml–1 RNase A and 50 μg·ml–1 DNase I,

A wzt Mutant Burkholderia mallei Is Attenuated and Partially Protects CD1 Mice against Glanders 55
and incubated for 3 h with shaking at 37˚C. Proteinase K
was then added to a final concentration of 50 μg·ml–1 and
the digests were incubated for an additional 3 h at 60˚C.
The enzymatic digests were clarified by centrifugation,
and LPS was isolated from the supernatants as precipi-
tated gels following three rounds of ultracentrifugation at
100,000 g and 4˚C. After the final spin, the gel-like pel-
lets were resuspended in pyrogen-free water and lyophi-
lized. The purified LPS were electrophorased using 16%
Tricine gels (Invitrogen) as described elsewhere [27], and
the products were stained with silver using the procedure
of Tsai and Frasch [28].
Western blotting was performed using standard pro-
cedures [24]. Briefly, proteins separated by SDS-PAGE
were transferred to a nitrocellulose membrane by using a
Trans-blot semidry system (Bio-Rad Laboratories, Her-
cules, CA). The membranes were blocked with a solution
of 1.5% non-fat milk powder plus 1.5% bovine serum
albumin. For analysis of O-antigen profile, the mem-
branes were incubated with mouse monoclonal antibody
3D11 that is specific to B. pseudomallei LPS O-antigen
(Research Diagnostics, Inc.) overnight and subsequently
developed with goat anti-mouse IgG (whole molecule)
conjugated with horseradish peroxidase (Sigma Chemical
Co).
2.6. Thin-Section Immune-Electron Microscopy
Bacteria were grown for 3 days on TSA-G and gently
suspended in PBS to 109 colony forming units (CFU)
ml–1. Mouse monoclonal antibody 3D11 was diluted 1:40
in PBS and incubated with the cells for 1 h at 37˚C. The
bacteria were washed with PBS, suspended in 0.5 ml of a
1:20 dilution of protein A-20 nm gold particles (Poly-
sciences, Warrington, PA), incubated at 37˚C for 1 hour,
and washed in PBS at 2000 × g for 15 min. The final
pellet was suspended in 0.1 M phosphate buffer (pH 7.3),
mixed with molten agar, solidified, and cut into small
blocks. The cubes were fixed in 2.5% gluteraldehyde/0.1
M L-lysine for 25 min, then 2.5% gluteraldehyde for 90
minutes at room temperature, and stored in 0.1 M phos-
phate buffer at 4˚C. Samples were dehydrated with a
series of ethanol washes at 30, 50, 70, and 80% ethanol
in 0.1 M phosphate buffer (pH 7.4) for 15 minutes each
at room temperature. Samples were dehydrated once
more with 2 parts LR White, 1 part 80% ethanol for 15
minutes at RT, infiltrated with LR white for 39 hours,
and polymerized at 60˚C for 20 hours. Thin sections on
copper grids were stained with 1.7% lead citrate and 2%
uranyl acetate, and viewed with a JEOL 100 CX-II trans-
mission electron microscope (Zeiss 10C; Carl Zeiss Inc.,
New York, NY) with ×25,000 magnification.
2.7. Serum Bactericidal Assay
The bactericidal activity of 20% guinea pig serum (PCS;
which contains no antibody) for B. mallei was determined
as previously described [29] Control tubes contained heat-
inactivated serum.
2.8. Pathogenicity of B. mallei Strains in Mice
The cultures of the wild type, the sa c B mutant, and the
wzt mutant strains were grown in TSB-G for 24 h at 37˚C
with shaking (200 rpm). The cells were harvested by
centrifugation at 2000 × g for 20 min, washed with PBS,
and resuspended in 10 ml of PBS. The dilutions of cul-
tures were plated on TSA-G plates to determine the
cfu/ml.
Seven-week-old female CD1 mice (Charles River Lab-
oratories, Wilmington, MA) were allowed 1 week of ac-
climatization. Groups of five mice each were intraperi-
toneally injected with saline or three different doses (4.4
× 105, 6.6 × 105, or 8.8 × 105 cfu/mouse) of each strain:
wild type, sacB mutant, and wzt mutant. Survival of ani-
mals for 36 days post-inoculation was monitored, and ab-
normal animal behaviors of surviving animals (any hud-
dling or fur ruffling) were recorded. The 50% lethal dose
(LD50) of the treatment groups was calculated using the
Probit.exe program of STAT 2050 server of the University
of Guelph, Canada (http://www.uoguelph.ca/~jhubert/stat
2050/software/software_2050.html). The number of ani-
mals dying up to 6 d post-inoculation was used in LD50
calculations. The animals that received wild type or sacB
strains and survived were sacrificed by exposing to CO2
on day 36 post-inoculation. Their spleens and livers were
homogenized and cultured to determine the presence of
the inoculated strain [21].
2.9. Immune and Protective Responses in CD1
Mice Inoculated with the B. mallei wzt
Mutant
Blood samples were collected by retro-orbital bleeding
from mice injected/inoculated with saline or the wzt mu-
tant, on day 30 post-inoculation/injection. Sera were col-
lected by centrifugation at 3,000 g for 5 min, and serum
IgG1 and IgG2a levels were determined by enzyme-linked
immunosorbent assay as described elsewhere [20]. The
heat-killed wild type strain ATCC 23344 suspended in
0.06 M sodium carbonate buffer (pH 9.6) was used as the
antigen to coat polystyrene plates.
At day 36 post-inoculation, those mice that were in-
jected with saline (5 mice) or the wzt mutant (8 mice) and
survived were challenged intraperitoneally with 6.6 × 106
cfu/mouse (11.2 times the LD50) of wild type strain
ATCC 23344. Survival of the mice for 15 days post-
challenge was monitored, and abnormal animal behave-
iors were recorded. On day 15 post-challenge, the sur-
viving animals were killed by CO2 asphyxiation. Their
spleens and livers were homogenized and cultured to
determine the presence of B. mallei.

A wzt Mutant Burkholderia mallei Is Attenuated and Partially Protects CD1 Mice against Glanders
56
3. Results
3.1. Organization of LPS O-Antigen Biosynthetic
Gene Cluster, and Nucleotide and Protein
Sequences of wzm and wzt
The goals of this study were to elucidate the influence of
the wzt gene encoding the putative hydrophilic ATP-
binding protein of the polysaccharide ABC transporter
system (also called Wzt) on LPS O-antigen biosynthesis
and pathogenicity of B. mallei, and evaluate the protect-
tive efficacy of a wzt mutant against glanders infection in
mice. The gene cluster believed to encode O-antigen
biosynthesis is comprised of at least 15 individual genes,
and is located in Chromosome I of B. mallei strain ATCC
23344 (GenBank accession AY028370 and NC006348)
[12].
The wzm gene (BMA1986) encoding the putative per-
mease protein of the polysaccharide ABC transporter
system (also called Wzm) is 833-bp long. The DNA se-
quence analyses predicted that Wzm is a non-secretory
protein without a clear N-terminal signal sequence (Sig-
nal peptide probability: 0.043). The predicted subcellular
localization of Wzm is cytoplasmic (Reliability Index: RI
= 1; Expected Accuracy = 63%). The wzt gene (BMA1985)
encoding the putative ATP-binding protein of the poly-
saccharide ABC transporter system (also called Wzt) is
1397-bp long. The ATG starting codon of wzt is located
just 2-bp downstream the stop codon of wzm. The DNA
sequence analyses predicted that Wzt is also a non-sec-
retory protein without a clear N-terminal signal sequence
(Signal peptide probability: 0.00). The predicted subcel-
lular localization of Wzt is cytoplasmic (Reliability Index:
RI = 6; Expected Accuracy = 98%).
3.2. Genomic and Transcriptomic
Characterization of the wzt Mutant
A wzt mutant strain of B. mallei was constructed by dis-
rupting the wzt gene of the sacB mutant 23344∆sacB [20],
and designated as 23344∆sac B ∆wzt. The sucrose resis-
tance of strain 23344∆sacB was used as the non-antibiotic
marker in selecting the wzt mutant. A PCR assay with the
primer pair wzt-1/wzt-4 (see Materials and Methods)
produced a predicted 1.4-kb amplicon from wild type B.
mallei, and an approximately 0.9-kb amplicon from the
wzt mutant strain 23344∆sacB∆wzt (data not shown), in-
dicating that due to homologous recombination event, a
484-bp region was deleted from the wzt gene.
Reverse-transcription (RT)-PCR with the primer pair
wzt-1/wzt-4 produced a 1.4-kb product from the wild type
and the sacB mutant strains (lanes 1 and 2 of Figure 1),
but no product from the wzt mutant (lane 3 of Figure 1).
These results suggest that both the wild type and the
sacB mutant expressed a full-length wzt mRNA, whereas,
the wzt mutant failed to express a wzt mRNA as a result
of the deletion event in the wzt gene. In order to charac-
terize any polar effect induced by this deletion event, the
expression of mRNA of wbiA was analyzed by RT-PCR
using the primer pair wbiA-Forward/wbiA-Reverse (see
Materials and Methods). The wbiA gene was chosen for
this assay since it is located immediately downstream of
wzt in LPS O-antigen cluster. Just like the wild type and
the sacB mutant strains, the wzt mutant produced an ap-
proximately 0.8-kb amplicon (lanes 4, 5, and 6 of Figure
1), suggesting that expression of wbiA mRNA was not
affected due to the deletion in wzt.
3.3. Growth Characteristics of B. mallei Strains
The growth rates of B. mallei strains in trypticase soy
broth supplemented with 4% glycerol (TSB-G) was meas-
ured. During log phase, the growth of the wzt mutant was
slower (approximately 4.5 h doubling time) than that of
the wild type or the sacB mutant (approximately 2 h
doubling time).
3.4. LPS Profiles of B. Mallei Strains
LPS extracts were electrophoresed in 16% Tricine gels,
and stained with silver to analyze any visible differences
among strains with regard to expression of cellular com-
ponents at translational level (Figure 2). No major dif-
ferences were visible between strains with regard to
products in the range of 20 to 200-kDa. When the LPS
extracts were analyzed by western immunoblot proce-
dure using the mouse monoclonal antibody 3D11 that is
specific to B. pseudomallei LPS O-antigen, no differ-
ences were seen between the wild type and the wzt mu-
tant (data not shown). The results suggest that the mu-
tation in wzt did not affect O-antigen biosynthesis.
3.5. Transport of O-Antigen to the Cell Surface
of B. mallei
In order to find the effect of mutation in wzt on O-antigen
export, the presence of this carbohydrate on cell surface
of strains was determined by immune-electron micros-
copy. The gold particles were found attached to the cell
surface of both the wild type and the wzt mutant (data not
shown). The results suggest that the O-antigen was
transported to the cell surface of the wzt mutant similar to
that of the wild type.
3.6. Resistance of B. mallei Strains to Serum
Bactericidal Killing
The wild type and the wzt mutant strains were completely
resistant to the bactericidal action of guinea pig sera at
20% (v/v) (data not shown). The serum sensitive Fran-
cisella tularensis mutant strain WptIG191V [30] that was
used as the control was completely killed by 20% guinea
pig sera. The results suggest that disruption of wzt did not

A wzt Mutant Burkholderia mallei Is Attenuated and Partially Protects CD1 Mice against Glanders
57
1.0-kb
1.6-kb
0.8-kb
Figure 1. Expression of wzt and wbiA mRNA by B. mallei strains as determined by RT-PCR. For PCR assays, cDNA from the
wild type strain ATCC 23344 (lanes-1 and -4), sacB mutant 23344sacB (lanes-2 and -5), and wzt mutant 23344sacBwzt
(lanes-3 and -6) was used. PCR with forward and reverse primers for wzt produced a 1.4-kb ampicon from the wild type
(lane-1) and sacB mutant (lane-2), but not from wzt mutant (lane-3). PCR with forward and reverse primers for wbiA pro-
duced 0.8-kb ampicons from all three strains; wild type (lane-4), sacB mutant (lane-5), and wzt mutant (lane-6). Lane-M
represents 1-kb molecular size markers.
lated with the wild type strain died, but no mice inocu-
lated with the sacB or wzt mutants died (Figure 3(a)).
When a dose of 6.6 × 105 cfu was used in inoculations,
60% of animals inoculated with the wild type strain and
40% inoculated with the sacB mutant died within 6 days
post-inoculation, but only 20% of animals inoculated
with the same dose of wzt mutant died (Figure 3(b)).
When the mice were inoculated with a dose of 8.8 × 105
cfu, all the mice injected with the wild type strain or the
sacB mutant died within 4 days post-inoculation, but
only 40% of animals inoculated with the wzt mu t a n t died
(Figure 3(c)). The LD50 dose of the wild type strain, the
sacB mutant, and the wzt mutant were 5.9 × 105, 6.6 ×
105, and 9.1 v 105 cfu, respectively. None of the inocu-
lated mice died after day 6 post-inoculation. Those mice
that survived the B. mallei inoculations exhibited clinical
manifestations including huddling during the first 4 days
following inoculations, but remained clinically normal
throughout the rest of the 36-day observation period.
1 2
3.8. Immune and Protective Responses of Mice
Inoculated wzt Mutant
Figure 2. LPS profiles of B. mallei as observed by SDS-PAGE
and silver staining. LPS extracts harvested from the wild
type ATCC 23344 (lane 1), and the wzt mutant (lanes 2)
were electrophoresed using 16% Tricin gel, and stained with
silver.
Specific serum immunoglobulins IgG1 and IgG2a of CD1
mice were measured by enzyme-linked immunosorbent
assay. Serum IgG titers of mice inoculated with the wzt
mutant were 34 to 43-fold higher than naïve controls at
30 days post infection (data not shown). The two IgG
isotypes were present in sera in almost equal amounts.
influence resistance of B. mallei to bactericidal killing.
3.7. Pathogenicity of B. mallei Strains in Mice The protective efficacy of the wzt mutant in CD1 mice
against challenge with the wild type B. mallei (ATCC
23344) was determined. At day 36 post-inoculation, those
mice that were injected with saline (n = 5) or the wzt
mutant (n = 8) and survived were challenged intraperito-
neally with 6.6 × 106 cfu (11.2 times the LD50) of the
wild type strain, and behaviors and survival of mice were
The pathogenicity of the B. mallei strains in CD1 mice
was evaluated by inoculating groups of five mice intrap-
eritoneally with doses of B. mallei strains, and recording
mortality of animals over a course of eight days (Figure
3). When nearly 4.4 × 105 colony forming units (cfu) of
strains were used for inoculations, 20% of mice inocu-

A wzt Mutant Burkholderia mallei Is Attenuated and Partially Protects CD1 Mice against Glanders
58
(a)
(b)
(c)
Figure 3. Survival of CD1 mice inoculated B. mallei strains wild (♦), sacB mutant (▪), and wzt mutant (▲). Groups of five mice
were injected intraperitoneally with 4.4 × 105 (a), 6.6 × 105 (b), or 8.8 × 105 (c) cfu/mouse (n = 5), and the number of survivors
was recorded during a course of eight days post-inoculation.

A wzt Mutant Burkholderia mallei Is Attenuated and Partially Protects CD1 Mice against Glanders 59
monitored. Eighty percent of mice injected with saline
and subsequently challenged with the wild type strain
died within 24 h post-challenge (Figure 4). Among the
mice inoculated intraperitoneally with the wzt mutant and
subsequently challenged intraperitoneally with the wild
type strain, 87.5% survived longer than 15 days post
challenge (Figure 4). The surviving mice did not exhibit
any clinical symptoms during the 15-day post-challenge
period.
4. Discussion
The O-antigen is a key component of LPS in the outer
membrane of many gram-negative bacteria. It consists of
repeats of an oligosaccharide unit (O unit), which usually
contains two to eight residues of a broad range of both
common and rarely occurring sugars and their derivatives.
Sugar nucleotides are the activated precursors for cell
surface polysaccharides. Tosynthesize O-antigens, mono-
mers are assembled on a lipid carrier (undecaprenol pho-
sphate) by enzymes encoded in the wb gene cluster be-
fore their incorporation into the LPS molecule. The ABC-
2 transporters consist of an integral membrane protein,
Wzm, and a hydrophilic protein containing an ATP-bind-
ing motif, Wzt. Involvement of the transporter with the
translocation of the polymer has not been proven ex-
perimentally and details of the process are not clear at
this stage [9].
The O-antigen biosynthetic cluster of B. mallei is com-
prised of a wzm gene encoding a protein identical to
membrane component of the ABC transporter of other
bacteria, and a wzt gene encoding a protein identical to
ATP-binding component of the ABC transporter. Ac-
cordingly, it can be speculated that O-antigen in B. mallei
is exported by an ABC transporter pathway just like in E.
coli O8 and O9 [15], and K. pneumoniae O1 and O12 [14,
16,17]. The B. mallei wzm and wzt shared substantial
identity with wzm and wzt proteins of a large number of
other bacterial species suggesting that these proteins are
conserved and are likely important for the O-antigen bio-
synthesis, growth and/or persistence of many bacteria.
Both wzm and wzt were predicted to localize in the cy-
toplasm of the cell, an observation consistent with the
predicted functions of these two proteins.
Western immunoblotting assays revealed that disrupt-
tion of wzt did not influence O-antigen biosynthesis.
Burtnick et al., [12] reported that the O-antigen moiety is
required for resistance of B. mallei to the bactericidal
action of serum. Our bactericidal assays revealed that the
wzt mutant and the wild type B. mallei were similarly
resistant to serum bactericidal killing. These observations
suggest that the wzt mutant strain produced O-antigen
uninterruptedly. Electron micrographs suggested that dis-
ruption of wzt did not prevent transport of O-antigen to
the cell surface.
Figure 4. Protective efficacy of the wzt mutant against challenge with virulent wild type B. mallei. Thirty-six days after injec-
tion with saline (n = 5) (♦) or inoculation with the wzt mutant (n = 8) (▪), the mice that survived were challenged intraperito-
neally with 6.6 × 106 cfu/mouse of the wild type strain ATCC 23344. The number of survivors was recorded during a course
of seven days post-challenge.

A wzt Mutant Burkholderia mallei Is Attenuated and Partially Protects CD1 Mice against Glanders
60
When CD1 mice were inoculated intraperitoneally, the
B. mallei wzt mutant caused less murine mortality than
the wild-type strain or the sacB mutant. These observa-
tions suggest that functions of wzt are critical for viru-
lence of B. mallei in vivo. The wzt mutant displayed
slower growth in vitro, suggesting that wzt functions are
important for normal growth. The less pathogenicity of
this mutant can partly be attributed to its slower growth
rate. The slow growing mutant may be less capable of
evading the effects of host immune system.
Since B. mallei is a facultative intracellular pathogen,
a live attenuated vaccine may be the best strategy to in-
duce protective cell-mediated and antibody-mediated im-
mune responses. When the CD1 mice were inoculated
with a non-lethal dose of wzt mutant, both IgG1 and IgG2a
were induced. When those mice were challenged with a
11.2 time LD50 of the wild type, a greater proportion sur-
vived relative to uninoculated controls suggesting that
the mutant induces protection in mice against glanders.
The mice that survived after wzt inoculation and subse-
quent challenge with the wild type strain did not exhibit
prolonged clinical symptoms such as huddling or fur
ruffling. In contrast, the B. mallei mutant strains gener-
ated by disrupting the capsule biosynthesis [21], type III
secretion system [31], branched-chain amino acid biosyn-
thesis [23], and quorum-sensing network [32] failed to
induce reasonable protection against glanders in mice or
Syrian Hamsters. Taking all these observations into ac-
count, we speculate that the wzt mutant has potential as a
vaccine candidate against glanders.
5. Conclusion
The functions of wzt were found important for B. mallei
in its growth in culture medium and pathogenicity in
CD1 mice. The attenuated wzt mutant induced partial pro-
tection in mice against B. mallei challenge, and therefore,
has potential as a vaccine candidate against glanders.
6. Acknowledgements
The author thanks Anna Champion for assistance in LPS
extraction and SDS-PAGE, Kathy Lowe for assistance in
electron microscopy, Lynn Heffron and Dustin Lucas for
expert handling of mice, and Kay Carlson and Nancy
Tenpenny for technical assistance.
REFERENCES
[1] E. Yabuuchi, Y. Kosako, H. Oyaizu, I. Yano, H. Hotta, Y.
Hashimoto, T. Ezaki and M. Arakawa, “Proposal of Burk-
holderia gen. nov. and Transfer of Seven Species of the
Genus Pseudomonas Homology Group II to the New Ge-
nus, with the Type Species Burkholderia cepacia (Pal-
leroni and Holmes 1981) comb. nov.,” Microbiol Immu-
nol, Vol. 36, No. 12, 1992, pp. 1251-1275.
[2] A. Srinivasan, C. N. Kraus, D. DeShazer, P. M. Becker, J.
D. Dick, L. Spacek, J. G. Bartlett, W. R. Byrne and D. L.
Thomas, “Glanders in a Military Research Microbiolo-
gist,” The New England Journal of Medicine, Vol. 345,
No. 4, 2001, pp. 256-258.
doi:10.1056/NEJM200107263450404
[3] M. S. Redfearn, N. J. Palleroni and R. Y. Stanier, “A Com-
parative Study of Pseudomonas pseudomallei and Bacil-
lus mallei,” Journal of General Microbiology, Vol. 43,
No. 2, 1966, pp. 293-313.
doi:10.1099/00221287-43-2-293
[4] L. Wilkinson, “Glanders: Medicine and Veterinary Medi-
cine in Common Pursuit of a Contagious Disease,” Medi-
cal History, Vol. 25, No. 4, 1981, pp. 363-384.
[5] G. C. Whitlock, D. M. Estes and A. G. Torres, “Glanders:
Off to the Races with Burkholderia mallei,” FEMS Mi-
crobiology Letters, Vol. 277, No. 2, 2007, pp. 115-122.
doi:10.1111/j.1574-6968.2007.00949.x
[6] M. R. Burns, S. A. Jenkins, N. M. Vermeulen, R. Bala-
krishna, T. B. Nguyen, M. R. Kimbrell and S. A. David,
“Structural Correlation between Lipophilicity and Lipo-
polysaccharide-Sequestering Activity in Spermine-Sulfo-
namide Analogs,” Bioorganic & Medicinal Chemistry Let-
ters, Vol. 16, No. 24, 2006, pp. 6209-6212.
doi:10.1016/j.bmcl.2006.09.026
[7] M. R. Burns, S. A. Jenkins, S. J. Wood, K. Miller and S.
A. David, “Structure-Activity Relationships in Lipopoly-
saccharide Neutralizers: Design, Synthesis, and Biologi-
cal Evaluation of a 540-Membered Amphipathic Bisamide
Library,” Journal of Combinatorial Chemistry, Vol. 8, No.
1, 2006, pp. 32-43. doi:10.1021/cc0500755
[8] H. Nikaido, “Molecular Basis of Bacterial Outer Mem-
brane Permeability Revisited,” Microbiology and Mole-
cular Biology Reviews, Vol. 67, No. 4, 2003, pp. 593-656.
doi:10.1128/MMBR.67.4.593-656.2003
[9] C. R. Raetz and C. Whitfield, “Lipopolysaccharide Endo-
toxins,” Annual Review of Biochemistry, Vol. 71, 2002,
pp. 635-700.
[10] C. A. Schnaitman and J. D. Klena, “Genetics of Lipopoly-
saccharide Biosynthesis in Enteric Bacteria,” Microbiol-
ogy Reviews, Vol. 57, No. 3, 1993, pp. 655-682.
[11] C. Whitfield, “Biosynthesis of Lipopolysaccharide O An-
tigens,” Trends in Microbiology, Vol. 3, No. 5, 1995, pp.
178-185. doi:10.1016/S0966-842X(00)88917-9
[12] M. N. Burtnick, P. J. Brett and D. E. Woods, “Molecular
and Physical Characterization of Burkholderia mallei O
Antigens,” Jo urnal of Bacteriology, Vol. 184, No. 3, 2002,
pp. 849-852. doi:10.1128/JB.184.3.849-852.2002
[13] G. Samuel and P. Reeves, “Biosynthesis of O-Antigens:
Genes and Pathways Involved in Nucleotide Sugar Pre-
cursor Synthesis and O-Antigen Assembly,” Carbohy-
drate Research, Vol. 338, No. 23, 2003, pp. 2503-2519.
doi:10.1016/j.carres.2003.07.009
[14] B. R. Clarke and C. Whitfield, “Molecular Cloning of the
rfb Region of Klebsiella pneumoniae Serotype O1:K20:
the rfb Gene Cluster Is Responsible for Synthesis of the
D-Galactan I O Polysaccharide,” Journal of Bacteriology,
Vol. 174, No. 14, 1992, pp. 4614-4621.

A wzt Mutant Burkholderia mallei Is Attenuated and Partially Protects CD1 Mice against Glanders 61
[15] H. C. Flemming and K. Jann, “Biosynthesis of the O9 An-
tigen of Escherichia coli. Growth of the Polysaccharide
Chain,” European Journal of Biochemistry, Vol. 83, No.
1, 1978, pp. 47-52.
doi:10.1111/j.1432-1033.1978.tb12066.x
[16] L. Izquierdo, N. Coderch, N. Pique, E. Bedini, M. M. Cor-
saro, S. Merino, S. Fresno, J. M. Tomas and M. Regue,
“The Klebsiella pneumoniae wabG Gene: Role in Bio-
synthesis of the Core Lipopolysaccharide and Virulence,”
Journal of Bacteriology, Vol. 185, No. 24, 2003, pp. 7213-
7221. doi:10.1128/JB.185.24.7213-7221.2003
[17] L. Izquierdo, S. Merino, M. Regue, F. Rodriguez and J.
M. Tomas, “Synthesis of a Klebsiella pneumoniae O-An-
tigen Heteropolysaccharide (O12) Requires an ABC 2
Transporter,” Journal of Bacteriology, Vol. 185, No. 5,
2003, pp. 1634-1641.
doi:10.1128/JB.185.5.1634-1641.2003
[18] J. D. Bendtsen, H. Nielsen, G. von Heijne and S. Brunak,
“Improved Prediction of Signal Peptides: SignalP 3.0,”
Journal of Molecular Biology, Vol. 340, No. 4, 2004, pp.
783-795. doi:10.1016/j.jmb.2004.05.028
[19] S. F. Altschul, W. Gish, W. Miller, E. W. Myers and D. J.
Lipman, “Basic Local Alignment Search Tool,” Journal
of Molecular Biology, Vol. 215, No. 3, 1990, pp. 403-410.
[20] A. B. Bandara, D. DeShazer, T. J. Inzana, N. Srirangana-
than, G. G. Schurig and S. M. Boyle, “A Disruption of
ctpA Encoding carboxy-Terminal Protease Attenuates Burk-
holderia mallei and Induces Partial Protection in CD1
Mice,” Microbial Pathogenesis, Vol. 45, No. 3, 2008, pp.
207-216. doi:10.1016/j.micpath.2008.05.005
[21] D. DeShazer, D. M. Waag, D. L. Fritz and D. E. Woods,
“Identification of a Burkholderia mallei Polysaccharide
Gene Cluster by Subtractive Hybridization and Demon-
stration That the Encoded Capsule Is An Essential Viru-
lence Determinant,” Microbial Pathogenesis, Vol. 30, No.
5, 2001, pp. 253-269. doi:10.1006/mpat.2000.0430
[22] R. Simon, U. Priefer and A. Puhler, “A Broad Host Range
Mobilization System for in Vivo Genetic Engineering:
Transposon Mutagenesis in Gram Negative Bacteria,”
Bio/Technology, Vol. 1, No. 9, 1983, pp. 784-791.
[23] R. L. Ulrich, K. Amemiya, D. M. Waag, C. J. Roy and D.
DeShazer, “Aerogenic Vaccination with a Burkholderia
mallei Auxotroph Protects against Aerosol-Initiated Glan-
ders in Mice,” Vaccine, Vol. 23, No. 16, 2005, pp. 1986-
1992. doi:10.1016/j.vaccine.2004.10.017
[24] J. F. Sambrook and E. F. Maniatis, “Molecular Cloning:
A Laboratory Manual,” 2nd Edition, Cold Spring Harbor
Laboratory Press, Cold Spring Harbor, 1989.
[25] P. J. Brett, M. N. Burtnick, D. S. Snyder, J. G. Shannon,
P. Azadi and F. C. Gherardini, “Burkholderia mallei Ex-
presses a Unique Lipopolysaccharide Mixture That Is a
Potent Activator of Human Toll-Like Receptor 4 Com-
plexes,” Molecular Microbiology, Vol. 63, No. 2, 2007,
pp. 379-390. doi:10.1111/j.1365-2958.2006.05519.x
[26] M. B. Perry, L. L. MacLean, T. Schollaardt, L. E. Bryan
and M. Ho, “Structural Characterization of the Lipopoly-
saccharide O Antigens of Burkholderia pseudomallei,”
Infection and Immunity, Vol. 63, No. 9, 1995, pp. 3348-
3352.
[27] U. K. Laemmli, “Cleavage of Structural Proteins during
the Assembly of the Head of Bacteriophage T4,” Nature,
Vol. 227, No. 5259, 1970, pp. 680-685.
doi:10.1038/227680a0
[28] C. M. Tsai, R. Boykins and C. E. Frasch, “Heterogeneity
and Variation among Neisseria meningitidis Lipopolysac-
charides,” Journal of Bacteriology, Vol. 155, No. 2, 1983,
pp. 498-504.
[29] T. J. Inzana and P. Anderson, “Serum Factor-Dependent
Resistance of Haemophilus influenzae Type b to Antibody
to Lipopolysaccharide,” The Journal of Infectious Dis-
eases, Vol. 151, No. 5, 1985, pp. 869-877.
[30] J. Li, C. Ryder, M. Mandal, F. Ahmed, P. Azadi, D. S.
Snyder, R. D. Pechous, T. Zahrt and T. J. Inzana, “Atten-
uation and Protective Efficacy of an O-antigen-Deficient
Mutant of Francisella tularensis LVS,” Microbiology,
Vol. 153, No. Pt 9, 2007, pp. 3141-3153.
doi:10.1099/mic.0.2007/006460-0
[31] R. L. Ulrich and D. DeShazer, “Type III Secretion: A Vir-
ulence Factor Delivery System Essential for the Patho-
genicity of Burkholderia mallei,” Infection and Immunity,
Vol. 72, No. 2, 2004, pp. 1150-1154.
doi:10.1128/IAI.72.2.1150-1154.2004
[32] R. L. Ulrich, D. Deshazer, H. B. Hines and J. A. Jeddeloh,
“Quorum Sensing: A Transcriptional Regulatory System
Involved in the Pathogenicity of Burkholderia mallei,” In-
fection and Immunity, Vol. 72, No. 11, 2004, pp. 6589-
6596.