Content uploaded by Ben Crossett
Author content
All content in this area was uploaded by Ben Crossett
Content may be subject to copyright.
Genomic plasticity of the causative agent of
melioidosis,
Burkholderia pseudomallei
Matthew T. G. Holden
a
, Richard W. Titball
b,c
, Sharon J. Peacock
d,e
, Ana M. Cerden
˜
o-Ta
´
rraga
a
, Timothy Atkins
b
,
Lisa C. Crossman
a
, Tyrone Pitt
f
, Carol Churcher
a
, Karen Mungall
a
, Stephen D. Bentley
a
, Mohammed Sebaihia
a
,
Nicholas R. Thomson
a
, Nathalie Bason
a
, Ifor R. Beacham
g
, Karen Brooks
a
, Katherine A. Brown
h
, Nat F. Brown
g
,
Greg L. Challis
i
, Inna Cherevach
a
, Tracy Chillingworth
a
, Ann Cronin
a
, Ben Crossett
h
, Paul Davis
a
, David DeShazer
j
,
Theresa Feltwell
a
, Audrey Fraser
a
, Zahra Hance
a
, Heidi Hauser
a
, Simon Holroyd
a
, Kay Jagels
a
, Karen E. Keith
h
,
Mark Maddison
a
, Sharon Moule
a
, Claire Price
a
, Michael A. Quail
a
, Ester Rabbinowitsch
a
, Kim Rutherford
a
,
Mandy Sanders
a
, Mark Simmonds
a
, Sirirurg Songsivilai
k
, Kim Stevens
a
, Sarinna Tumapa
e
, Monkgol Vesaratchavest
e
,
Sally Whitehead
a
, Corin Yeats
a
, Bart G. Barrell
a
, Petra C. F. Oyston
b
, and Julian Parkhill
a,l
a
Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SA, United Kingdom;
b
Defence Science and Technology
Laboratory, Porton Down, Salisbury SP4 0JQ, United Kingdom;
c
Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical
Medicine, London WC1E 7HT, United Kingdom;
d
Nuffield Department of Clinical Medicine, John Radcliffe Hospital, University of Oxford, Oxford OX3 9DU,
United Kingdom;
e
Faculty of Tropical Medicine, Mahidol University, Bangkok 10400, Thailand;
f
Laboratory of Hospital Infection, Division of Nosocomial
Infection Prevention and Control, Central Public Health Laboratory, London NW9 5HT, United Kingdom;
g
School of Health Science, Griffith University, Gold
Coast, Queensland 9726, Australia;
h
Department of Biological Sciences, Centre for Molecular Microbiology and Infection, Flowers Building, Imperial College,
London SW7 2AZ, United Kingdom;
i
Department of Chemistry, University of Warwick, Coventry CV4 7AL, United Kingdom;
j
U.S. Army Medical Research
Institute for Infectious Diseases, 1425 Porter Street, Fort Detrick, MD 21702-5011; and
k
Department of Immunology, Faculty of Medicine, Siriraj Hospital,
Mahidol University, Bangkok 10700, Thailand
Edited by E. Peter Greenberg, University of Iowa, Iowa City, IA, and approved August 13, 2004 (received for review May 10, 2004)
Burkholderia pseudomallei is a recognized biothreat agent and the
causative agent of melioidosis. This Gram-negative bacterium
exists as a soil saprophyte in melioidosis-endemic areas of the
world and accounts for 20% of community-acquired septicaemias
in northeastern Thailand where half of those affected die. Here we
report the complete genome of B. pseudomallei, which is com-
posed of two chromosomes of 4.07 megabase pairs and 3.17
megabase pairs, showing significant functional partitioning of
genes between them. The large chromosome encodes many of the
core functions associated with central metabolism and cell growth,
whereas the small chromosome carries more accessory functions
associated with adaptation and survival in different niches.
Genomic comparisons with closely and more distantly related
bacteria revealed a greater level of gene order conservation and a
greater number of orthologous genes on the large chromosome,
suggesting that the two replicons have distinct evolutionary ori-
gins. A striking feature of the genome was the presence of 16
genomic islands (GIs) that together made up 6.1% of the genome.
Further analysis revealed these islands to be variably present in a
collection of invasive and soil isolates but entirely absent from the
clonally related organism B. mallei. We propose that variable
horizontal gene acquisition by B. pseudomallei is an important
feature of recent genetic evolution and that this has resulted in a
genetically diverse pathogenic species.
M
elioidosis is a bacterial infection caused by Burkholderia
pseudomallei, an environmental Gram-negative sapro-
phyte present in wet soil and rice paddies in endemic areas
(1–3). The majority of infections are reported from east Asia
and northern Australia, the highest documented rate being in
northeastern Thailand, where melioidosis accounts for 20% of
all community-acquired septicaemias (4). Disease occurs after
bacterial contamination of breaks in the skin or by inhalation
after contact with water or soil. A pneumonic form of the
disease can also result from the inhalation of contaminated
dusts and was reported in U.S. helicopter pilots during the
Vietnam War. The potential for the bacterium to cause disease
after inhalation has also resulted in the inclusion of this
pathogen on the Centers for Disease Control list of potential
biothreat agents as a Category B agent (5). The most frequent
clinical picture is a septicaemic illness associated with bacterial
dissemination to distant sites, such that metastatic pneumonia
and hepatic and splenic abcesses are common. However,
clinical manifestations are protean and have led to the infec-
tion being termed ‘‘the great mimicker’’ (6). Of the cases in
Thailand, one-fifth occur in children under the age of 14 years,
for whom the overall mortality of infected individuals is 51%
(3). Death usually occurs within the first 48 h as a result of
septic shock and in a setting where optimal antimicrobial
chemotherapy is given. Of equal concern, there is evidence
that the bacterium does not cause overt disease in all individ-
uals exposed to the bacterium but is able to persist at unknown
sites in the body to become reactivated later in life. Possibly the
best documented examples of this are in Vietnam veterans who
developed disease, in one case 26 years later, after returning
to the U.S. (7–9).
Here we describe the 7.25-megabase pair (Mb) genome of B.
pseudomallei strain K96243, isolated from a case of human
melioidosis. Comparative analysis highlights the role that
horizontal gene acquisition has played in the evolution of the
genome and clarifies the genetic relationship of B. pseudoma-
llei with Burkholderia mallei, the genome of which is described
in an accompanying manuscript (10). The work presented here
also provides insights into the molecular basis of environmen-
tal survival, virulence, and antimicrobial resistance.
Materials and Methods
Bacterial Strain, Growth, and DNA Isolation. B. pseudomallei strain
K96243 was isolated in 1996 from a 34-year-old female diabetic
patient in Khon Kaen hospital in Thailand. K96243 is sensitive
to imipenem, ceftazidime, chloramphenicol, ciprofloxacin,
and augmentin and resistant to minocycline, gentamicin, co-
trimoxazole, and streptomycin. The API 20NE profile of the
bacterium was 1156576. The median lethal dose in Porton
strain mice by the intraperitoneal route was 262 colony-
forming units. Bacteria were cultured in L broth at 37°C for
This paper was submitted directly (Track II) to the PNAS office.
Freely available online through the PNAS open access option.
Abbreviations: GI, genomic island; CDS, coding sequence; Mb, megabase pair; IS, insertion
sequence.
Data deposition: The sequences reported in this paper have been deposited in the Euro-
pean Molecular Biology Laboratory database (accession nos. BX571965 and BX571966).
l
To whom correspondence should be addressed. E-mail: parkhill@sanger.ac.uk.
© 2004 by The National Academy of Sciences of the USA
14240–14245
兩
PNAS
兩
September 28, 2004
兩
vol. 101
兩
no. 39 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0403302101
18 h and pelleted at 10,000 ⫻ g. The cells were resuspended in
30 ml of lysis solution (10 mM NaCl兾20 mM Tris䡠HCl, pH 8.0兾1
mM EDTA兾0.5% SDS) and incubated at 50°C overnight.
Three milliliters of 5 M sodium perchlorate was added and
incubated for1hatambient temperature. After phenol
chloroform extraction, the DNA was precipitated with etha-
nol, spooled into deionized water, and stored at ⫺20°C.
Whole-Genome Sequencing. The sequence was assembled, fin-
ished, and annotated as described in ref. 11 by using
ARTEMIS to
collate data and facilitate annotation (12); detailed information
is available in Supporting Text and Tables 3–6, which are
published as supporting information on the PNAS web site.
Comparative Genomics. Comparison of the genome sequences was
facilitated by using the ACT (Artemis Comparison Tool) suite of
programs (K.R., unpublished data; see also www.sanger.ac.uk兾
Software兾ACT), which enabled the visualization of BLASTN and
TBLASTX comparisons (13) between the genomes. Orthologous
proteins were identified as reciprocal best matches by using
FASTA with subsequent manual curation. Pseudogenes had one
or more mutations that would prevent correct translation; each
of the inactivating mutations was subsequently checked against
the original sequencing data.
Multiplex PCR. The presence or absence of 11 GIs was defined for
40 B. pseudomallei isolates by using PCR. Twenty isolates were
recovered from patients with melioidosis admitted to Sappasiti-
prasong Hospital (Ubon Ratchatani, Thailand) during 2001, and
20 isolates were recovered from the soil samples from the
surrounding area. Target genes within the GIs were amplified by
two multiplex reactions according to standard methodology (14,
15) after optimization by using B. pseudomallei K96243 genomic
DNA. Detailed information, including the primer sequences and
PCR conditions, is available in Supporting Text.
Results and Discussion
The complete genome of B. pseudomallei strain K96243 con-
sists of two circular replicons (European Molecular Biology
Laboratory accession nos. BX571965 and BX571966) of 4.07
Mb and 3.17 Mb (Fig. 1) each that have been designated
chromosome 1 and chromosome 2 and encode 3,460 and 2,395
coding sequences (CDSs), respectively (for a summary of the
features of the chromosomes, see Table 7, which is published
as supporting information on the PNAS web site). A skew in
strand-specific G兾C content was seen for both chromosomes,
which enabled the prediction of the origins of replication (Fig.
1). Chromosome 1 shares components near to the origin that
are common to other prokaryotic genomes, such as dnaA and
dnaN homologues. Chromosome 2 has a weaker G ⫹ C
deviation pattern, and the predicted origin of replication
contains features similar to those associated with plasmid
replication, such as parA and parB homologues. Identification
of CDSs on chromosome 2 that are involved in central
metabolism and essential functions has led us to designate this
component of the genome as a chromosome rather than a
megaplasmid. A direct example of the distribution of essential
features can be found in the tRNA genes on both chromo-
somes: 53 are encoded on chromosome 1, and 8 are encoded
on chromosome 2. One of the eight tRNAs on chromosome 2
is unique to this replicon (Ser
GGA
); however, genes encoding
Ser by using this codon are found in equal proportions on both
chromosomes (12% of Ser codons). In addition to the presence
of tRNA genes, chromosome 2 also contains an rRNA gene
cluster, components of the replication (polA and dnaG) and
transcription machinery (rpoD), and components of central
Fig. 1. Schematic circular diagrams of the large and small chromosomes of the B. pseudomallei genome. Where appropriate, categories are shown as pairs
of concentric circles representing both coding strands. Rings from outside to inside: GIs represented by red segments; scale (in Mb); annotated CDSs
colored according to predicted function; additional CDSs compared to the sequenced B. mallei strain ATCC 2944; the percentage of G ⫹ C content plot;
(G ⫺ C)兾(G ⫹ C) deviation plot (⬎0%, olive; ⬍0%, purple). Color coding for CDSs: dark blue, pathogenicity兾adaptation; black, energy metabolism; red,
information transfer; dark green, surface-associated; cyan, degradation of large molecules; magenta, degradation of small molecules; yellow, central兾
intermediary metabolism; pale green, unknown; pale blue, regulators; orange, conserved hypothetical; brown, pseudogenes; pink, phage plus IS elements; gray,
miscellaneous.
Holden et al. PNAS
兩
September 28, 2004
兩
vol. 101
兩
no. 39
兩
14241
MICROBIOLOGY
metabolism involved in amino acid biosynthesis and energy
metabolism that are absent from chromosome 1.
Analysis of the predicted functions of the CDSs on the two
chromosomes reveals distinct partitioning of core and accessory
functions (Fig. 2). Chromosome 1 contains a higher proportion
of CDSs involved in core functions, such as macromolecule
biosynthesis, amino acid metabolism, cofactor and carrier syn-
thesis, nucleotide and protein biosynthesis, chemotaxis, and
mobility. Chromosome 2, by contrast, contains a greater pro-
portion of CDSs encoding accessory functions: adaptation to
atypical conditions, osmotic protection and iron acquisition,
secondary metabolism, regulation, and laterally acquired DNA.
In addition, chromosome 2 contains a greater proportion of
CDSs with matches to hypothetical proteins or proteins that have
no database matches at all. Comparison of the two chromosomes
reveals that there is very little similarity, except in the regions of
the rRNA clusters. This partitioning of core and accessory
functions is reminiscent of the arrangement in the actinomycete
soil-dwelling bacterium, Streptomyces coelicolor strain A3 (2),
where the 8-Mb single linear chromosome is broadly described
as having a central portion that encodes the majority of essential
and housekeeping functions flanked by two arms encoding
accessory functions (16).
Analysis of B. pseudomallei proteins with orthologue matches
to proteins encoded in other previously sequenced bacterial
genomes revealed the highest number of matches with Ralstonia
solanacearum, another member of the Burkholderiaceae (2,535
of the 5,855 B. pseudomallei proteins had matches in R. so-
lanacearum; see Table 8, which is published as supporting
information on the PNAS web site) (17). Notably, this organism
also has a bipartite genome structure, containing a 3.7-Mb
chromosome and a 2.0-Mb megaplasmid. Orthologous matches
were identified on both chromosomes; 57% of CDSs on chro-
mosome 1 and 25% of CDSs on chromosome 2 have matches.
Comparison to the more distant pseudomonads, Pseudomonas
aeruginosa, Pseudomonas putida, and Pseudomonas syringae, and
Xanthomonas campestris, all of which have a single chromosome,
produced fewer total orthologue matches (Table 8) (18–21).
Chromosome 2 appears to be the more divergent of the B.
pseudomallei replicons, containing a smaller percentage of CDSs
with matches for each comparison.
Analysis of the levels of gene order conservation in compar-
ison with R. solanacearum also highlights the relative genetic
diversity of the two B. pseudomallei replicons. The large repli-
cons contain discrete regions of conserved gene order, but there
has been a large amount of recombination reciprocal about the
origin of replication (see Fig. 4A, which is published as support-
ing information on the PNAS web site), as is common with other
intergenomic comparisons between related species (22). When
the small replicons were compared, no conserved gene order was
detected, suggesting that extensive recombination may have
shuffled the gene order beyond recognition. Equally, because
the number of orthologous matches is relatively low (579 CDSs),
it may be that the two secondary replicons do not share a
common ancestry. Shared genes may have been transferred
independently to secondary replicons since the two organisms
diverged from a common ancestor. This model is also suggested
by a lack of similarity of genes in the regions around the origins
of replication. Similar genome comparisons with other organ-
isms, including the pseudomonads, failed to reveal any signifi-
cant levels of gene order conservation for either B. pseudomallei
chromosome.
At 7.3 Mb, the B. pseudomallei genome is large in compar-
ison with the typical prokaryotic genome. Horizontal acqui-
sition of DNA appears to have been intrinsic to the evolution
of this organism. Many regions within both chromosomes
showed some of the characteristics of GIs acquired through
very recent lateral transfer, such as anomalies in the percent-
age of G ⫹ C content or dinucleotide frequency signature of
the DNA in these regions. Additionally, these regions often
contained CDSs with similarities to genes associated with
mobile genetic elements, such as insertion sequence (IS)
elements, bacteriophages, and plasmids. Twelve putative GIs
have been identified on chromosome 1, and four have been
identified on chromosome 2, each comprising ⬇7.6% and
⬇4.2% of the DNA of these replicons, respectively (Table 1).
In addition, there are also several other regions of the genome
that may have arisen by means of gene acquisition, although
the evidence is less clear.
At least three of the GIs appear to be prophages (GIs 2, 3,
and 15). It is unclear from sequence analysis alone whether all
of these prophages are capable of lysogeny, although we have
shown that the K96243 strain produces at least one lysogenic
phage after UV induction,
K96243 (see Fig. 5, which is
published as supporting information on the PNAS web site).
This phage can be classified as a member of the order
Caudovirales and the family Myoviridae (23) and can also infect
B. mallei. Notably, B. mallei strains that do not produce
lipopolysaccharide O-antigen were not infected with
K96243,
suggesting that this surface molecule is the phage receptor.
Sequencing of DNA fragments from the
K96243 genome
reveals that this bacteriophage corresponds to GI 2. Analysis
of the integration site of this prophage identified a 45-bp
repeat corresponding to the 3⬘ end of the tRNA-Phe as the
likely attachment site.
Several of the other GIs encode large numbers of hypothet-
ical proteins with no database matches plus a few phage-like
proteins. It is unclear whether these are prophage-like entities,
such as defective or satellite phages. In addition to GIs whose
transfer may be bacteriophage mediated, some of the islands
may be transferred by other mechanisms. One of the small GIs
identified, GI 11, contains CDSs with similarities to plasmid
conjugal transfer proteins, suggesting that it may be an inte-
grated conjugative element (24). The other GIs identified do
not share similarity with any characterized mobile elements.
Some of these elements are located next to tRNA genes and
are flanked by small repeats (GIs 4, 5, 7, 9, 10, and 13). Many
of the GI regions contain multiple phage-like integrases (e.g.,
GIs 3 and 4), suggesting that they may be composite in nature,
having arisen from multiple insertion events over a period of
time.
B. mallei is the etiological agent of glanders, a disease of horses
that can be transmitted to humans to cause a life-threatening
Fig. 2. Distribution of the CDSs belonging to different functional classes on
the two chromosomes of B. pseudomallei. Figures for functional classes on
each chromosome are expressed as a percentage of the total number of CDSs
on that replicon.
14242
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0403302101 Holden et al.
illness (10). Compared with B. pseudomallei, the distribution and
host range of B. mallei is more narrow. This organism does not
survive in the environment, and its natural reservoir is thought
to be equines (25). Before the sequencing of both genomes,
DNA–DNA hybridization suggested that these two species were
closely related (26). A recent multilocus sequence typing study
of B. pseudomallei and B. mallei isolates from around the globe
has revealed that B. mallei can be considered to be a clone of B.
pseudomallei (27).
Although B. mallei strains appear to have recently evolved
from a B. pseudomallei ancestor, there are large differences in
the sizes of the genomes; the B. pseudomallei genome is 1.31 Mb
larger than that of B. mallei. Comparative analysis of chromo-
somes reveals that 16% of chromosome 1 and 31% of chromo-
some 2 of B. pseudomallei are unique with respect to the
equivalent chromosomes of B. mallei (Fig. 1). Conversely, ⬎1%
of chromosome 1 and 4% of chromosome 2 of B. mallei are
unique with respect to the equivalent chromosomes of B.
pseudomallei. Notably, the majority of unique regions of the B.
mallei chromosome 2 consist of DNA regions of ⬎2 kb. The
single largest region of difference of B. mallei chromosome 2 is
a 41-kb fragment, which is found on chromosome 1 of B.
pseudomallei (CDS coordinates BPSL3040–BPSL3078), result-
ing from an interchromosomal transposition. Comparison of the
two genomes also reveals that there has been a significant
amount of intrachromosomal rearrangements in B. mallei, re-
Table 1. Genomic islands of B. pseudomallei
Island Size, kb CDS Coordinates Integrases GC (%) D K Bp08 Functional note
Chromosome 1
GI 1 12.6 BPSL0081–BPSL0092 1 – (61.5) – ⫹ – Miscellaneous island; contains
lipoprotein and hypothetical
proteins
GI 2 36.2 BPSL0129–BPSL0176 1 – (65.4) –– – Prophage
K96243
GI 3 51.3 BPSL0548–BPSL0589 3 ⫹ (56.6) –– – Putative prophage; contains
dienelactone hydrolase family
protein
GI 4 39.7 BPSL0745–BPSL0772 3 (1) ⫹ (56.8) ⫹ ––Miscellaneous island; contains
putative RNA
2⬘-phosphotransferase and
putative helicases
GI 5 11.7 BPSL0944–BPSL0953 1 ⫹ (58.7) –– – Contains putative type I restriction
system and plasmid replication
protein
GI 6 15.0 BPSL1137–BPSL1157 1 ⫹ (58.8) –– – Prophage-like
GI 7 5.9 BPSL1384–BPSL1393 (1) ⫹ (58.6) –– ⫹ Prophage-like; contains exported
avidin family protein
GI 8 92.3 BPSL1637–BPSL1709 —–(62.0) – ⫹⫹Miscellaneous island; contains
transport protein,
hemolysin-related protein, amine
catabolism, various regulators,
and YadA-like exported protein
GI 9 9.8 BPSL2568–BPSL2586 1 – (64.0) –– – Prophage-like
GI 10 6.6 BPSL3114–BPSL3118 — ⫹ (54.5) –– ⫹ Restriction and modification system
GI 11 15.3 BPSL3257–BPSL3270 0 ⫹ (55.9) –– ⫹ Putative integrated plasmid or
integrated conjugative element;
contains recombinase, conjugal
plasmid transfer, and replication
proteins
GI 12 11.5 BPSL3342–BPSL3353 1 ⫹ (57.3) ⫹ ––Prophage-like
Chromosome 2
GI 13 19.0 BPSS0378–BPSS0391A 1 ⫹ (58.4) – ⫹ – Prophage-like, contains
hypothetical proteins
GI 14 18.6 BPSS0652A–BPSS0666 — ⫹ (55.1) ⫹⫹ ⫹ Miscellaneous island; contains
putative collagenase
GI 15 34.6 BPSS1047–BPSS1089 0 – (65.2) ⫹ ––Putative prophage; contains
hypothetical proteins and ParA
homologue
GI 16 61.8 BPSS2051–BPSS2090 0 ⫹ (59.3) ⫹⫹ ⫹ Metabolic island; contains various
ABC transporters, large cell
surface protein,
L-asparaginase,
and fatty aldehyde
dehydrogenase
The size and CDS coordinates of the putative genomic islands are indicated. The number of putative integrases contained within islands are given; —, not
applicable. Partial genes and pseudogenes are indicated in parentheses. Whether anomalies are present in the properties of DNA within islands is indicated in
the GC, D, and K columns, representing the percentage of G ⫹ C content, G ⫹ C deviation, and Karlin dinucleotide frequency signature, respectively: ⫹, anomolies
are present; –, anomolies are not present. The Bp08 column indicates regions of the K96243 genome that are identical to islands previously identified by using
representative difference analysis of B. pseudomallei strain 08 against B. thailandensis: ⫹, match with B. thailandensis; –, no match with B. thailandensis. GI 7
is the only GI found in the B. mallei ATCC 23344 genome. ABC, ATP-binding cassette.
Holden et al. PNAS
兩
September 28, 2004
兩
vol. 101
兩
no. 39
兩
14243
MICROBIOLOGY
sulting from recombination between the large number of re-
peated IS elements in its genome (see Fig. 4B) (10).
The unique regions in the B. pseudomallei genome with respect
to B. mallei have been summarized in Table 9, which is published
as supporting information on the PNAS web site. One of the main
sources of additional DNA are GIs; all but one of the islands
identified in Table 1 (GI 7) are absent in the B. mallei genome. In
the case of GI 7, fragments of this island were found in the B. mallei
genome, suggesting that a similar island may have been present in
an ancestral strain that subsequently decayed in B. mallei ATCC
23344. Although many of the B. pseudomallei islands contain DNA
with properties that suggest that they have recently been acquired,
it is very unlikely that all of these islands have been acquired by
K96243 in the time since B. mallei diverged from a B. pseudomallei
ancestor. Rather, these islands can be seen as representative of the
diversity within B. pseudomallei and specifically as being variably
present between K96243 and the B. pseudomallei strain that was the
immediate ancestor of B. mallei. This result suggests that the
distribution of these islands is highly variable between different B.
pseudomallei strains and that the islands are therefore actively
transmissible. This interpretation is supported by probing phyloge-
netically diverse clinical and environmental isolates of B. pseudoma-
llei from Thailand with multiplex PCR of target genes within a
selection of the K96243 GIs (Fig. 3). The screen identified consid-
erable variation in the distribution of the GIs: GI 3 was ubiquitous;
GIs 2, 11, 12, and 16 were detected in 10–70% of strains; GIs 5, 6,
and 9 were detectable in 10% or less of strains; and GI 4 was unique
to K96243.
Furthermore, subtractive hybridisation analysis between B.
pseudomallei strain 08 and Burkholderia thailandensis, a non-
pathogenic relative of B. pseudomallei, indicates that GIs are key
determinants of genome plasticity (28). Of 92 sequence frag-
ments found to be present in B. pseudomallei strain 08 and absent
from B. thailandensis, only 53 (58%) can be mapped to the
K96243 genome. Of these fragments, 27 fell in GIs 7, 8, 10, 11,
14, and 16 (Table 1), confirming these GIs as likely to have been
recently acquired by B. pseudomallei K96243. As a corollary to
this finding, 42% of the sequences found to be unique to strain
08 were not present in the K96243 genome, confirming the large
degree of variability in the carriage of these islands among B.
pseudomallei strains.
The genomic diversity observed in the closely related but
host-restricted B. mallei genome sheds light on differences in
lifestyle and niche adaptation. Comparison of B. mallei and B.
pseudomallei genomes reveals that both have features of
genetic plasticity, but the observed variation points toward
different modes of change. The absence of GIs in the B. mallei
genome suggests that gene acquisition does not play as signif-
icant a role in the genetic variation of this species as it does in
B. pseudomallei. Conversely, gene loss appears to have been an
important source of genetic variation in the recent evolution
of B. mallei. Several of the unique regions in the B. pseudoma-
llei genome contained CDSs at their 5⬘ and 3⬘ ends that are
truncated and adjacent to one another in the B. mallei genome,
suggesting that these regions have been deleted in B. mallei
ATCC 23344 since its divergence from B. pseudomallei (see
Table 9). This observation is consistent with other host-
restricted bacteria that have evolved from versatile ancestors
and have undergone gene loss (29, 30). B. mallei also contains
a much larger number of IS elements and detectable pseudo-
genes in comparison with B. pseudomallei (K96243 contains 42
IS elements and 26 pseudogenes), which points toward recent
host adaptation and a reduction in selective pressure on parts
of the B. mallei genome (10, 30, 31). In B. pseudomallei IS
elements and pseudogenes do not appear to contribute to the
genetic variation of this species to the same extent. Both
species’ genomes contain large numbers of small sequence
repeats. In B. mallei, these small sequence repeats have been
shown to play a role in the mutation of some of the detectable
pseudogenes and a source of proteome variation (10). Com-
parison of small sequence repeats in the two species identified
variation in the size of some of the repeats; however, the extent
to which they contribute to variation in B. pseudomallei strains
has yet to be determined.
The B. pseudomallei genome contains many genes that promote
survival in diverse and challenging environments and a large
selection of genes that modulate pathogenicity and host–cell inter-
action (Table 2). Comparative analysis of B. pseudomallei and B.
mallei has identified many CDSs that may contribute to the
phenotypic differences between the two species (Table 2). These
phenotypes include known virulence determinants, such as flagella
(32) and a type III protein secretion system (33); potential virulence
determinants, such as surface polysaccharides, exoproteins, fim-
briae, pili, and putative adhesins; drug resistance determinants; and
potential environmental survival functions, including various sec-
ondary metabolite pathways, numerous catabolic pathways, trans-
port systems, and stress-response proteins. A detailed description of
these features and the role they play in the biology of B. pseudoma-
llei is presented in Supporting Text.
The dual existence of B. pseudomallei as soil colonizer and
effective human pathogen requires flexibility, a fact reflected in the
genome. At the most rudimentary level, the size of the genome
endows the organism with an expansive inventory of CDSs encod-
ing diverse functions that promote survival and success in different
environments. This capacity is evident from the large metabolic
repertoire encoded and in the redundancy seen for some of the
virulence determinants, such as the type III protein secretion
systems and fimbriae (Table 2). The bipartite structure of the
genome and distinct partitioning of encoded functions also suggest
genetic suppleness. The reduced orthology and lack of conservation
of gene order of chromosome 2 with other genomes suggest that this
replicon may have its origins in the carriage of accessory functions
that promote survival in diverse niches.
The content of the genome is supplemented by horizontally
acquired DNA, an important feature in the recent evolution of
B. pseudomallei. The diversity of the GIs also suggests that
transfer occured by a variety of means. Subsequent analysis of
the distribution of these islands in environmental and clinical
strains of B. pseudomallei has detected variability in the
complement of islands, suggesting a large pool of islands
throughout the population. The presence of such genetic
fluidity has important implications for the future study of
disease pathogenesis and in the development of vaccination
Fig. 3. Prevalence of GIs in environmental and invasive clinical isolates of B.
psuedomallei. Distribution of GIs in environmental and invasive clinical B.
pseudomallei isolates (n ⫽ 40) as determined by multiplex PCR.
14244
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0403302101 Holden et al.
strategies. Adaptability in the human host may underlie the
protean disease manifestations and the ability of this organism
to cause chronic disease and to recrudesce many years after
apparently effective treatment. There is a growing recognition
that the global burden of melioidosis is much greater than
current estimates, a result of limitations in diagnostic facilities
(3). Coupled with the status of B. pseudomallei as a biothreat
agent, there is therefore a pressing need to gain a better
understand of the role that horizontal gene transfer plays in
the genomic plasticity of this potent pathogen.
We thank William Nierman and the staff at The Institute for Genomic
Research for sharing data before publication. We also acknowledge the
support of the Wellcome Trust Sanger Institute core sequencing and
informatics groups. This work was supported by the Wellcome Trust
through its Beowulf Genomics initiative. S.J.P. holds a Wellcome Trust
Career Development Award in Clinical Tropical Medicine.
1. Dance, D. A. B. (1991) Clin. Microbiol. Rev. 4, 52–60.
2. Dance, D. A. B. (2002) Curr. Opin. Infect. Dis. 15, 127–132.
3. White, N. J. (2003) Lancet 361, 1715–1722.
4. Chaowagul, W., White, N. J., Dance, D. A. B., Wattanagoon, Y., Naigowit, P.,
Davis, T. M. E., Looareesuwan, S. & Pitakwatchara, N. (1989) J. Infect. Dis. 159,
890–899.
5. Rotz, L. D., Khan, A. S., Lillibridge, S. R., Ostroff, S. M. & Hughes, J. M. (2002)
Emerg. Infect. Dis. 8, 225–230.
6. Yee, K. C., Lee, M. K., Chua, C. T. & Puthucheary, S. D. (1988) J. Tropic. Med.
Hyg. 91, 249–254.
7. Mays, E. E. & Ricketts, E. A. (1975) Chest 68, 261–263.
8. Koponen, M. A., Zlock, D., Palmer, D. L. & Merlin, T. L. (1991) Arch. Intern.
Med. (Moscow) 151, 605–608.
9. Sanford, J. P. & Moore, W. L. (1971) Am. Rev. Respir. Dis. 104, 452–453.
10. Nierman, W. C., DeShazer, D., Kim, H., Tettelin, H., Nelson, K. E., Feldblyum,
T. V., Ulrich, R. L., Ronning, C. M., Brinkac, L. M., Daugherty, S. C., et al.
(2004) Proc. Natl. Acad. Sci. USA 101, 14246 –14251.
11. Parkhill, J., Achtman, M., James, K. D., Bentley, S. D., Churcher, C., Klee,
S. R., Morelli, G., Basham, D., Brown, D., Chillingworth, T., et al. (2000) Nature
404, 502–506.
12. Rutherford, K., Parkhill, J., Crook, J., Horsnell, T., Rice, P., Rajandream, M. A.
& Barrell, B. (2000) Bioinformatics 16, 944–945.
13. Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. (1990) J.
Mol. Biol. 215, 403–410.
14. Elnifro, E. M., Ashshi, A. M., Cooper, R. J. & Klapper, P. E. (2000) Clin.
Microbiol. Rev. 13, 559–570.
15. Henegariu, O., Heerema, N. A., Dlouhy, S. R., Vance, G. H. & Vogt, P. H.
(1997) BioTechniques 23, 504–511.
16. Bentley, S. D., Chater, K. F., Cerdeno-Tarraga, A. M., Challis, G. L., Thomson,
N. R., James, K. D., Harris, D. E., Quail, M. A., Kieser, H., Harper, D., et al.
(2002) Nature 417, 141–147.
17. Salanoubat, M., Genin, S., Artiguenave, F., Gouzy, J., Mangenot, S., Arlat, M.,
Billault, A., Brottier, P., Camus, J. C., Cattolico, L., et al. (2002) Nature 415,
497–502.
18. Stover, C. K., Pham, X. Q., Erwin, A. L., Mizoguchi, S. D., Warrener, P.,
Hickey, M. J., Brinkman, F. S. L., Hufnagle, W. O., Kowalik, D. J., Lagrou, M.,
et al. (2000) Nature 406, 959–964.
19. Nelson, K. E., Weinel, C., Paulsen, I. T., Dodson, R. J., Hilbert, H., dos Santos,
V., Fouts, D. E., Gill, S. R., Pop, M., Holmes, M., et al. (2002) Environ.
Microbiol. 4, 799– 808.
20. Buell, C. R., Joardar, V., Lindeberg, M., Selengut, J., Paulsen, I. T., Gwinn,
M. L., Dodson, R. J., Deboy, R. T., Durkin, A. S., Kolonay, J. F., et al. (2003)
Proc. Natl. Acad. Sci. USA 100, 10181–10186.
21. da Silva, A. C. R., Ferro, J. A., Reinach, F. C., Farah, C. S., Furlan, L. R.,
Quaggio, R. B., Monteiro-Vitorello, C. B., Van Sluys, M. A., Almeida, N. F.,
Alves, L. M. C., et al. (2002) Nature 417, 459–463.
22. Eisen, J. A., Heidelberg, J. F., White, O. & Salzberg, S. L. (2000) Genome Biol.
1, research0011.1–0011.9.
23. Ackermann, H. W. (2003) Res. Microbiol. 154, 245–251.
24. Burrus, V., Pavlovic, G., Decaris, B. & Guedon, G. (2002) Plasmid 48, 77–97.
25. Howe, C. (1950) in The Oxford Medicine, ed. Christian, H. A. (Oxford Univ.
Press, New York), pp. 185–202.
26. Rogul, M., Brendle, J. J., Haapala, D. K. & Alexander, A. D. (1970) J. Bacteriol.
101, 827– 835.
27. Godoy, D., Randle, G., Simpson, A. J., Aanensen, D. M., Pitt, T. L., Kinoshita,
R. & Spratt, B. G. (2003) J. Clin. Microbiol. 41, 2068–2079.
28. Brown, N. F. & Beacham, I. R. (2000) J. Med. Microbiol. 49, 993–1001.
29. Cole, S. T., Eiglmeier, K., Parkhill, J., James, K. D., Thomson, N. R., Wheeler,
P. R., Honore, N., Garnier, T., Churcher, C., Harris, D., et al. (2001) Nature 409,
1007–1011.
30. Parkhill, J., Sebaihia, M., Preston, A., Murphy, L. D., Thomson, N., Harris,
D. E., Holden, M. T. G., Churcher, C. M., Bentley, S. D., Mungall, K. L., et al.
(2003) Nat. Genet. 35, 32–40.
31. Parkhill, J., Wren, B. W., Thomson, N. R., Titball, R. W., Holden, M. T. G.,
Prentice, M. B., Sebaihia, M., James, K. D., Churcher, C., Mungall, K. L., et al.
(2001) Nature 413, 523–527.
32. Chua, K. L., Chan, Y. Y. & Gan, Y. H. (2003) Infect. Immun. 71, 1622–1629.
33. Stevens, M., Wood, M. W., Taylor, L. A., Monaghan, P., Hawes, P., Jones,
P. W., Wallis, T. S. & Galyov, E. E. (2002) Mol. Microb. 46, 649–659.
Table 2. Survival and virulence functions encoded in the B. pseudomallei genome
Function Notes and examples
Survival
Secondary metabolism Fourteen clusters encoding possible antibiotic, surfactant, and siderophore biosynthesis pathways, including
putative hydroxamate (BPSL1774–BPSL1779) and pyochelin (BPSS0581–BPSS0588*) siderphores
Drug resistance Seven Ambler class A, B, and D

-lactamases, including cephalosporinase (BPSS0946) and oxacillinase
(BPSS1997*); six multidrug efflux systems, including aminoglycoside-*, macrolide-*, and polymyxin
B-specific systems; aminoglycoside acetyltransferase (BPSS0262)
Intracellular stress Superoxide detoxification, superoxide dismutases (BPSL0880 and BPSL1001), and catalases (BPSS0993 and
BPSS2214*); nitric oxide detoxification; flavohaemoglobin (BPSL2840)
Motility and chemotaxis Five gene clusters encode the components of a single flagella system; 38 chemotaxis-associated proteins*
Virulence
Secretion Type I,* type II, type III*, and type V protein secretion systems, including three type III systems (TTS1*,
BPSS1390–BPSS1408; TTS2, BPSS1613–BPSS1629; and TTS3, BPSS1543–BPSS1552) direct the secretion of
proteins outside the bacterium
Lipopolysaccharide and capsule Capsular polysaccharide synthesis and export cluster (BPSL2787–BPSL2810*); lipopolysaccharides
biosynthetic cluster (BPSL2672–BPSL2688); two other potential surface polysaccharides biosynthetic
clusters (BPSS1825–BPSS1834* and BPSS0417–BPSS0429*).
Exoproteins The genome contains many secreted exoenzymes that can break down host tissues, including
phospholipases C (BPSL0338*, BPSL2403, and BPSS0067), metalloprotease A (BPSS1993*), and a
homologue of the P. aeruginosa MucD Ser protease (BPSL0808) and a putative collogenase (BPSS0666*)
Adhesins Several surface proteins potentially modulate host–cell interactions, these including seven Hep_Hag repeat
family proteins (BPSL1631, BPSL1705*, BPSL2063, BPSS0796, BPSS0908, BPSS1434*, and BPSS1439).
Fimbriae and pili Thirteen clusters encoding components of type I fimbriae*, type IV pili, and tad-type pili*
*Gene or genes associated with this function are absent in the B. mallei ATCC 23344 genome (see Supporting Text).
Holden et al. PNAS
兩
September 28, 2004
兩
vol. 101
兩
no. 39
兩
14245
MICROBIOLOGY
