Identification of Burkholderia cepacia strains that express a Burkholderia pseudomallei-like capsular polysaccharide

ABSTRACT Burkholderia pseudomallei and Burkholderia cepacia are Gram-negative, soil-dwelling bacteria that are found in a wide variety of environmental niches. While B. pseudomallei is the causative agent of melioidosis in humans and animals, members of the B. cepacia complex typically only cause disease in immunocompromised hosts. In this study, we report the identification of B. cepacia strains isolated from either patients or soil in Laos and Thailand that express a B. pseudomallei-like 6-deoxyheptan capsular polysaccharide (CPS). These B. cepacia strains were initially identified based on their positive reactivity in a latex agglutination assay that uses the CPS-specific monoclonal antibody (mAb) 4B11. Mass spectrometry and recA sequencing confirmed the identity of these isolates as B. cepacia (formerly genomovar I). Total carbohydrates extracted from B. cepacia cell pellets reacted with B. pseudomallei CPS-specific mAbs MCA147, 3C5, and 4C4, but did not react with the B. pseudomallei lipopolysaccharide-specific mAb Pp-PS-W. Whole genome sequencing of the B. cepacia isolates revealed the presence of genes demonstrating significant homology to those comprising the B. pseudomallei CPS biosynthetic gene cluster. Collectively, our results provide compelling evidence that B. cepacia strains expressing the same CPS as B. pseudomallei co-exist in the environment alongside B. pseudomallei. Since CPS is a target that is often used for presumptive identification of B. pseudomallei, it is possible that the occurrence of these unique B. cepacia strains may complicate the diagnosis of melioidosis. IMPORTANCE Burkholderia pseudomallei, the etiologic agent of melioidosis, is an important cause of morbidity and mortality in tropical and subtropical regions worldwide. The 6-deoxyheptan capsular polysaccharide (CPS) expressed by this bacterial pathogen is a promising target antigen that is useful for rapidly diagnosing melioidosis. Using assays incorporating CPS-specific monoclonal antibodies, we identified both clinical and environmental isolates of Burkholderia cepacia that express the same CPS antigen as B. pseudomallei. Because of this, it is important that staff working in melioidosis-endemic areas are aware that these strains co-exist in the same niches as B. pseudomallei and do not solely rely on CPS-based assays such as latex-agglutination, AMD Plus Rapid Tests, or immunofluorescence tests for the definitive identification of B. pseudomallei isolates.

pseudomallei, the causative agent of melioidosis, and members of the Burkholderia cepacia complex that can colonize and infect immunocompromised hosts, especially cystic fibrosis and chronic granulomatous disease patients (1).Melioidosis is prevalent in tropical countries with the highest number of cases reported in Southeast Asia and northern Australia.B. pseudomallei infections are typically acquired through contact with contaminated soils or water and predominantly occur in individuals with underlying risk factors who reside in endemic areas (2).The clinical presentations of melioidosis are diverse and range from localized skin abscesses to acute pneumonia and sepsis, the latter of which can be rapidly fatal.Like most Burkholderia species, B. pseudomallei is naturally resistant to many commonly used antibiotics and no licensed melioidosis vaccines currently exist (3).Because of these issues, the rapid and accurate diagnosis of melioidosis is critical.
The current gold standard method for diagnosing melioidosis is the culture and identification of B. pseudomallei from clinical samples.This is time-and labor-intensive and lacks sensitivity (4).In order to reduce the time to obtain a presumptive iden tification, various approaches have been used, including latex agglutination with a monoclonal antibody (mAb) 4B11 against the 6-deoxyheptan capsular polysaccharide (CPS) (5).This approach has been used in clinical and research laboratories in Thailand and Laos for many years, and although cross-reactions with other Burkholderia species from environmental samples have been reported, this has not previously been reported among clinical isolates (6).Such presumptive misidentifications are inevitably misleading and might result in patients being treated inappropriately for melioidosis unless the error is recognized.
The 6-deoxyheptan CPS expressed by B. pseudomallei is a key virulence determinant encoded by a 34.5-kb gene cluster located on chromosome I (7,8).Experimental evidence has shown that CPS reduces phagocytosis of B. pseudomallei by host cells, preventing complement factor C3b deposition on the bacterial surface (9).Several studies indicate that CPS is a promising vaccine candidate and that antibodies against CPS provide protection in animal models of melioidosis (10)(11)(12)(13).CPS is also considered an attractive antigen for the development of rapid point-of-care diagnostics since it is highly conserved among B. pseudomallei isolates and is known to be shed and circulate throughout host tissues during active infections (14,15).In addition to the latex agglutination assay mentioned above, lateral flow immunoassays (LFIs) such as the Active Melioidosis Detect (AMD, InBios International, Inc.) assay that employs the CPS-specific mAb 4C4 have been developed and tested using a variety of Burkholderia isolates and clinical samples and have been shown to be highly specific for B. pseudomal lei (16)(17)(18).
Burkholderia thailandensis is a closely related, non-pathogenic, near-neighbor species that co-exists in the same environmental niches as B. pseudomallei (19,20).While lack of 6-deoxyheptan CPS production was once considered to be a key differentiating feature between B. thailandensis and B. pseudomallei, several recent studies have identified B. thailandensis variants that express a B. pseudomallei-like CPS (21)(22)(23)(24)(25)(26).B. thailandensis E555, a soil isolate from Cambodia, is the most well-characterized CPS-expressing variant strain to date.E555 has been shown to harbor a highly similar CPS biosynthetic gene cluster to that of B. pseudomallei and expresses a structurally identical CPS; however, this isolate is avirulent in mice (21,23).Although B. thailandensis infections have been reported in humans, these appear to be exceedingly rare (24,(27)(28)(29).
In this study, we report the identification and characterization of three B. cepacia isolates that produce a B. pseudomallei-like CPS along with two previously undescribed CPS-expressing B. thailandensis strains.Importantly, two of the CPS-expressing B. cepacia strains are clinical isolates that were identified in areas where B. pseudomallei is endemic.

Strains
The B. cepacia and B. thailandensis strains used in this study are described in Table 1.All strains were obtained from the Lao-Oxford-Mahosot Hospital-Wellcome Trust Research Unit (LOMWRU) Microbiology Laboratory strain collection in Vientiane, Laos, or the Mahidol-Oxford Tropical Medicine Research Unit (MORU) strain collection at Mahidol University, Bangkok, Thailand.Work with B. pseudomallei strain K96243 (7) was conduc ted in CDC-registered biosafety level 3 (BSL-3) facilities in compliance with the rules and regulations of the U.S. Federal Select Agent Program.

Clinical isolates
B. cepacia strain 39628 was collected in Laos as a routine diagnostic specimen in 2014 with oral consent.B. cepacia strain 10223 was obtained from a sputum sample in 2010 and was stored anonymously as per routine research lab practice at the time.B. cepacia strain U668 was collected from a patient in NE Thailand in 1990 and was stored anonymously as part of routine lab practice at that time.The researchers did not have access to information that could identify individual participants after strain collection.

Latex agglutination assays
The latex agglutination assay is based on a specific reaction between mAb 4B11 and CPS of B. pseudomallei.The latex agglutination reagent was prepared by Mahidol University and used for testing as previously described (30).Briefly, the assay was performed by mixing single bacterial colonies with 10 µL of latex reagent on a glass slide.Agglutination was observed by eye within 1-2 min after mixing.

Biochemical tests and antibiograms
Isolates in Laos were presumptively identified as species using the API 20NE (bioMérieux) kit according to the manufacturer's instructions, and antimicrobial susceptibility tests were conducted by disk diffusion according to the methods of the Clinical Labora tory Standards Institute that were current at the time, interpreted according to local guidelines (31).

DNA sequencing and polymerase chain reaction 16s RNA sequencing
For the 16s RNA assay, primers 16SU17F and 16s 1541R (PH) were used and adapted from Edwards et al. (32).Purified (Qiagen, Germany) amplicons of ~1.5 kbp were sequenced (Macrogen, Korea) and analyzed using NCBI databases to confirm the species.

recA sequencing
For the recA polymerase chain reaction (PCR) assay, primers BCR1 and BCR2 were used, as described by Mahenthiralingham et al. (33).For recA sequencing, these primers were combined with two others: Bcc seqF1 and Bcc seqR2, as described by Turton et al. (34).BioNumerics was used to analyze the sequences, which were clustered with type and reference strains using the neighbor-joining method.

Matrix-assisted laser desorption/ionization biotyping
Bacterial isolates were prepared and analyzed by the matrix-assisted laser desorp tion/ionization (MALDI) Biotyper system as previously described (35).Briefly, all isolates were cultured on Columbia agar at 37°C under aerobic conditions for 24 h, extracted with formic acid, and 1 µL of supernatant was spotted onto an MSP-384 polished steel target plate (Bruker Daltonics, Germany) and dried.Following this, 1 µL of a saturated solution of MALDI matrix, α-cyano-4-hydroxycinnamic acid (Bruker Daltonics, Germany) was applied to each sample and dried.Measurements were performed with the Bruker MALDI Biotyper system using FlexControl software (version 3.4.135;Bruker Daltonics, Germany).Spectra ranging from 2,000 to 13,000 m/z were analyzed using the MALDI-Biotyper software (version 3.1; Bruker Daltonics, Germany) and a reference database supplemented with a Burkholderia library (35).An identification score of ≥2.3 indicated reliable species identification, a score of 2.0-2.29 indicated probable species level identification, a score of 1.7-1.9indicated probable genus level identification, and a score of <1.7 indicated no reliable identification (35).

Carbohydrate purification
B. cepacia and B. thailandensis strains were grown overnight in 1 L of Luria Bertani-Lennox (LBL) broth at 37°C with aeration.Bacterial cultures were pelleted by centrifugation (10 min at 8,000 × g) and the resulting cell pellets were extracted using a modified hot aqueous-phenol procedure essentially as previously described (36,37).Following extraction, the phenol and aqueous phases were combined and dialyzed against distilled water to remove the phenol.The dialysates were clarified by centrifugation (10 min at 10,000 × g) and the supernatants were concentrated by lyophilization.The samples were then treated with RNase A, DNase I, and proteinase K (50 µg/mL), and the resulting carbohydrates were isolated as precipitated gels following successive rounds of ultracentrifugation.The gel-like pellets were resuspended in ultrapure water, lyophilized, and weighed to determine the yield of total carbohydrate.

CPS-specific lateral flow immunoassays
Active Melioidosis Detect (AMD) Plus Rapid Tests (InBios International, Inc.) were kindly provided for clinical work at LOMWRU and used following the manufacturer's instruc tions.The AMD Plus Rapid Tests were purchased for research use at the University of Nevada, Reno, and used per the manufacturer's instructions.Purified total carbohydrate samples (1 µg) from five B. cepacia strains and one B. thailandensis strain were individu ally loaded onto rapid tests followed by chase buffer.Results were determined after incubation at room temperature for 15 min.

SDS-PAGE, Western immunoblotting, and silver staining
Purified carbohydrate samples resuspended in water were mixed 1:1 with 2× SDS-PAGE sample buffer and heated to 100°C for 5 min prior to electrophoresis on 12% Tris-gly cine gels (Invitrogen).For immunoblot analysis, the antigens were electrophoretically transferred to nitrocellulose membranes.The membranes were blocked with 3% skim milk in high-salt Tris-buffered saline (HS-TBS; 20 mM Tris, 500 mM NaCl, pH 7.5) for 60 min at room temperature and then incubated for 60 min at room temperature with a 1/1,000 or 1/2,000 dilution of a CPS-specific mAb (3C5, 4C4, or MCA135) or with a 1/400 dilution of a type A lipopolysaccharide (LPS)-specific mAb (Pp-PS-W).To facilitate detection, the membranes were incubated for 60 min at room temperature with a 1/5,000 dilution of anti-mouse IgG or IgM horseradish peroxidase conjugates (Southern Biotech).The blots were visualized using Pierce ECL Western Blotting Substrate (Thermo Scientific) and a ChemiDoc XRS imaging system (BioRad).CPS purified from B. pseudomallei RR2683 was used as a positive control (11).For analysis of LPS, the purified carbohydrate samples were electrophoresed on 12% Tris-glycine gels as described above.Silver staining was conducted essentially as previously described (38).
Genomic DNA isolation, genome sequencing, and analysis B. cepacia (strains LNT40, 39628, and 10223) and B. thailandensis (strains SBXCC001 and SBXPR001) were grown overnight in LBL broth at 37°C with aeration.DNA was extracted from the strains using a Wizard Genomic DNA Purification Kit (Promega) as per the manufacturer's instructions.DNA preparations were further purified by ethanol precipitation using a standard protocol.Sequencing of the genomic DNA samples was conducted at the Institute for Genome Sciences (IGS) Genomics Resource Center (Baltimore, MD, USA).PacBio single-molecule real-time (SMRT) sequencing was conducted on a PacBio RS II instrument to ~16× coverage (strains 39268 and LNT40) or ~24× coverage (strains 10223, SBXCC001, and SBXPR001) using 20 kb SMRTbell libraries and P6C4 chemistry.PacBio genomic data were assembled using the Hierarchi cal Genome Assembly Process algorithm version 3 (10) implemented in PacBio SMRT Portal version 2.3.0 for (strains 39268 and LNT40) (39) or Celera Assembler version 8.2 (strains 10223, SBXCC001, and SBXPR001) (40).Assemblies were reorganized relative to the B. pseudomallei K96243 genome (41).
The IGS Annotation Engine was used for structural and functional annotation of the sequences (https://ae.igs.umaryland.edu[42]).Manatee was used to view annota tions (http://manatee.sourceforge.net/).Submission of the genomes to GenBank and comparative analysis of the annotated genomes were conducted by the IGS Informatics Resource Center (University of Maryland).Sequence Read Archive (SRA) and GenBank accession numbers for each genome are listed in Table 2.
For comparison of the CPS operons, the three B. cepacia and two B. thailandensis genomes including the reference genome B. pseudomallei K96243 (NC_006350) were run through a comparative analysis pipeline to generate protein ortholog clusters using Jaccard-filtered bi-directional best blast matches.Sybil (http://sybil.sourceforge.net/documentation.html), a web-based graphical user interface, was used to search and view ortholog clusters.The genomic comparative view pictures of the clusters in the CPS operon region were generated by selecting the genomes of interest.

CPS-specific immunofluorescence assay
Immunofluorescence assays (IFAs) using the CPS-specific mAb 4B11 were conducted essentially as previously described (43).Briefly, bacteria were cultured in LB broth at 37°C overnight following which 1 mL of culture was centrifuged at 10,000 rpm for 5 min, washed three times with PBS, and fixed with 500 µL of 2% paraformaldehyde in PBS for 15 min.The fixed bacteria were washed again with PBS and stained with IFA reagents (containing mAb 4B11 and Alexa Fluor 488 conjugated-goat anti-mouse IgG at a dilution of 1:1,000 in PBS) for 20 min at room temperature.Bacteria were observed using a laser scanning confocal microscope (LSM 700; Carl Zeiss) using a 100× objective lens with oil immersion and Zen software (2010 edition, Zeiss, Germany).

Identification of B. cepacia strains testing positive in the melioidosis latex agglutination test
The first cross-reacting B. cepacia isolate (LNT40) was recognized during the re-examina tion of isolates from an environmental study undertaken in Laos in 2009 (44).Although this isolate agglutinated strongly with the anti-CPS monoclonal antibody-based latex reagent and gave a strong positive reaction with the AMD test (17), the colony mor phology was atypical.Subsequent examination by API 20NE, antibiogram (specifically resistance to co-amoxiclav), 16S rDNA sequencing, and PCR to distinguish between B. pseudomallei, B. thailandensis, and B. cepacia (45) suggested that it was actually a member of the B. cepacia complex (data not shown).
In 2014, an oxidase-positive Gram-negative bacillus that agglutinated strongly with the latex reagent was isolated at Mahosot Hospital from the sputum of an outpatient.Further examination of the isolate (designated strain 39628) confirmed the agglutination reaction but API 20NE and antibiogram suggested that the isolate was B. cepacia (data not shown).Testing of 32 further clinical isolates of B. cepacia complex from the LOMWRU collection revealed that strains LNT40 and 39628 were the only two isolates that gave this cross-reaction.In addition, DNA from these 32 isolates was extracted and sent to the Antimicrobial Resistance and Healthcare Associated Infections Reference Laboratory (AMRHAI), London, UK, for the identification of specific genomovars by recA sequencing.Both LNT40 and 39628 were identified as B. cepacia (genomovar I) as were four of the other non-cross-reacting isolates.Of the other strains tested, 25 were identified as B. cenocepacia IIIA, one each was B. seminalis and Taxon K and one was unassignable (data not shown).

MALDI Biotyper analysis of B. cepacia strains
To further characterize the two cross-reacting B. cepacia isolates (LNT40 and 39628) from the LOMWRU collection, MALDI biotyping experiments were conducted.For comparative purposes, a Thai clinical isolate of B. cepacia (strain 10223) that cross-reacted with the latex reagent obtained from the MORU collection and B. pseudomallei strain K96243 were also included in this analysis.The three B. cepacia isolates (strains LNT40, 39628, and 10223) and B. pseudomallei strain K96243 were subjected to MALDI-TOF-MS and analyzed using MALDI Biotyper system software with a supplemented Burkholderia reference database.Results showed that all of the latex-positive B. cepacia isolates were identified as belonging to the B. cepacia complex (score values ≥ 2.300) as opposed to the B. pseudomallei complex (Table 3).The three isolates of B. cepacia also demonstrated similar protein profile patterns.Peaks were observed at approximately m/z of 2,600, 2,880, 3,130, 3,250, 3,770, 4,410, 4,810, 5,200, 6,250, 6,500, 7,540, 8,100, and 9,610 for all B. cepacia strains and B. pseudomallei (Fig. 1).Importantly, peaks that were unique to the B. cepacia isolates were at approximately m/z of 2,200 and 4,700 and peaks at m/z of 2,050 and 5,800 were only observed in B. pseudomallei.In addition, a peak at m/z of 2,330 was only observed in B. cepacia strain LNT40 while peaks at m/z 2,180 and 5,870 were found only in B. cepacia strain 10223.Taken together, these data are consistent with the results of the latex agglutination assays, API 20NE tests, antibiograms, recA and 16S sequencing,

Reactivity of B. cepacia strains with CPS-specific mAbs
Total carbohydrate was extracted from B. cepacia strains LNT40, 39628, and 10223 along with one clinical and one soil isolate of B. cepacia (strains 2.1B and U668) that tested negative in the latex agglutination assay.In addition, three soil isolates of B. thailandensis (strains E555, SBXCC001, and SBXPR001) had also been found to agglutinate with the latex reagent and/or were known to express the 6-deoxyheptan CPS antigen (21)(22)(23).The yields of total carbohydrate obtained from these strains ranged from 8.8 to 27.1 mg/L and are shown in Table 1.To determine if the purified carbohydrate samples contained the CPS antigen of interest, they were tested with AMD Plus LFIs that use the CPS-specific mAb 4C4 (17).As expected, B. cepacia strains LNT40, 39628, and 10223, as well as B. thailandensis strains E555, SBXCC001, and SBXPR001 exhibited positive results on the AMD Plus tests while B. cepacia strains 2.1B and U668 were negative (Fig. 2).
To further characterize the total carbohydrate samples extracted from the B. cepacia and B. thailandensis strains, three different CPS-specific mAbs (3C5, 4C4, and MCA147) were used in Western immunoblot analyses.As shown in Fig. 3, B. cepacia strains LNT40, 39628, and 10223 and B. thailandensis strains E555, SBXCC001, and SBXPR001 reacted strongly with mAbs 3C5 and 4C4, but B. cepacia strains 2.1B and U668 did not.Similar results were observed with mAb MCA147 (data not shown).These findings were consistent with the results of both the latex agglutination and AMD Plus immunoassays.
To examine the LPS expressed by the B. cepacia and B. thailandensis strains, SDS-PAGE and silver staining were conducted on all of the purified carbohydrate samples.As shown in Fig. 4, B. cepacia strains 2.1B and U668 displayed LPS banding patterns that were

Genome sequencing of the latex agglutination positive B. cepacia strains
To determine whether the CPS-expressing B. cepacia isolates LNT40, 39628, and 10223 harbored the genes necessary for CPS expression in B. pseudomallei, whole genome sequencing was conducted.The genomes of two CPS-expressing B. thailandensis strains SBXCC001 and SBXPR001 were also sequenced for comparative purposes.The genome characteristics of each of the strains sequenced are shown in Table 2.All three B. cepacia strains harbored three chromosomes each, with genome sizes totaling ~8.4 to 8.7 Mb with G+C contents of ~66.4% to 66.8%.Interestingly, strain 10223 also harbored a plasmid of ~212 kb.The two B. thailandensis strains each harbored two chromosomes with overall genome sizes of ~6.8 and ~7 Mb with G+C contents of ~67.7% and ~67.5%, respectively.
To determine if homologs of the B. pseudomallei CPS biosynthesis genes were present in the CPS-expressing B. cepacia and B. thailandensis strains, the K96243 CPS gene cluster (locus tags BPSL2787-BPSL2810) was used as a reference.The 34.5 kb region of B. pseudomallei K96243 containing 24 genes responsible for CPS biosynthesis was compared to the genome sequences of B. cepacia strains LNT40, 39628, and 10223 and B. thailandensis strains SBXCC001 and SBXPR001.The resulting alignments are shown in Fig. 5, and the location of CPS gene clusters (locus tags) for each of the strains is listed in Table 2.While homologs for the majority of CPS biosynthesis genes were identified in the B. cepacia and B. thailandensis strains, some genes were notably absent.For example, wzt2 and wzm2 that encode for a putative ABC transporter involved in CPS export were both absent from the B. cepacia strains, and were truncated in the B. thailandensis strains.
To determine if the CPS antigen was expressed on the surface of the latex-positive B. cepacia strains, IFAs based on the CPS-specific mAb 4B11 were conducted.As shown in Fig. 6, B. cepacia strains 10223, 39628, LNT40, and B. thailandensis strain E555 demonstra ted robust fluorescence, while B. cepacia U668 did not.These findings are consistent with the latex agglutination assay results, the LFI and Western immunoblotting results, and the presence or absence of CPS biosynthetic genes in these strains.

DISCUSSION
Accurate diagnosis of B. pseudomallei infections is necessary for the prompt administra tion of effective treatments and for improving the outcomes of melioidosis patients.Due to the lack of highly sensitive, specific, rapid point-of-care tests, and the time-con suming nature of bacterial culture, the diagnosis of melioidosis can be challenging.The 6-deoxyheptan CPS expressed by B. pseudomallei has been pursued as a diagnos tic target, and several tests use this antigen for the identification of bacterial isolates following bacterial culture and detection of the pathogen in patient samples (5,(14)(15)(16)(17)(18)43).In this study, we report the identification of clinical isolates of B. cepacia that express the same CPS antigen as B. pseudomallei as determined by reactivity with CPS-specific mAbs in multiple different assays.Total carbohydrate preparations extracted from these strains were positive for CPS using an LFI as well as Western immunoblots.IFAs confirmed the expression of CPS on the surface of the latex-positive B. cepacia isolates.Furthermore, whole genome sequencing revealed the presence of homologs of the B. pseudomallei CPS biosynthetic genes in the latex-positive B. cepacia strains.
Since there is evidence of frequent horizontal gene transfer within and between Burkholderia species, the occurrence of serological cross-reactivity between isolates within the genus Burkholderia is perhaps not surprising, nor is it unexpected to find that cross-reacting strains from the environment may occasionally be detected in human samples, although this is the first time to our knowledge that this has been reported.This potential cross-reactivity, even when using reagents designed to be specific for B. pseudomallei CPS, is an important pitfall of which anyone working on both environmen tal and clinical samples should be aware.A failure to identify such isolates at the species level could lead to misleading results in environmental surveys, as it did in our previous study, which mistakenly reported the isolation of B. pseudomallei from the environment in Luang Namtha Province (44,46).
In the context of clinical samples, the consequences can be even more significant, with the possibility of inappropriate treatment being given to patients.Fortunately, this phenomenon appears to be restricted to a minority of isolates of B. cepacia and B. thailandensis, although further surveillance is important to monitor this.Since the initial identification of this phenomenon, the Mahosot Microbiology Laboratory has only isolated one further B. cepacia and one B. thailandensis from clinical samples that have shown genuine cross-reactivity in the latex agglutination test, along with a number of other isolates of various species that have given non-specific agglutination (i.e., agglutinate with latex beads that are not coated with anti-CPS monoclonal antibody).Nonetheless, it is important that staff working in melioidosis-endemic areas are aware of this phenomenon and do not rely on CPS-based assays such as latex-agglutination, AMD, or IFA alone for the identification of B. pseudomallei.
Fortunately, B. pseudomallei-specific PCR-based diagnostics are available that are rapid, sensitive, and have high discriminatory power (47)(48)(49)(50)(51).The most widely used single target PCR assay that can differentiate between B. pseudomallei, B. thailandensis, B. mallei, and B. cenocepacia is based on open reading frame 2 (orf2) of the cluster 1 type three secretion system (TTS1) (47,49,51,52).TTS1-orf2 PCR is highly specific for B. pseudomallei and has demonstrated a specificity of 99-100% when used for testing clinical samples (53,54).Recently identified targets including BPS0745, BPSS1187, and BPSS1498 have also been evaluated in comparison to TTS1-orf2 in real-time PCR assays, and all were shown to be highly specific for the detection of B. pseudomallei (50,53,55).Based on this, PCR-based diagnostics should be useful for differentiating B. pseudomallei isolates from CPS-expressing B. cepacia isolates.
Collectively, the results obtained in this study provide compelling evidence that B. cepacia strains expressing the same CPS as B. pseudomallei co-exist in the environ ment alongside B. pseudomallei and B. thailandensis.Future studies will be necessary to investigate the clinical significance of the CPS-expressing B. cepacia isolates and whether or not CPS expression in these strains might enhance their virulence in animal models of infection.In addition, studies aimed at isolating the CPS from these strains for structural determination will be needed to confirm the exact chemical composition of the antigen.

FIG 1 FIG 2
FIG 1 MALDI-TOF spectra of B. cepacia strains LNT40, 39628, 10223 and B. pseudomallei strain K96243.The characteristic spectra of the bacterial isolates were generated by MALDI-TOF MS.The relative intensities of ions are shown on the y axis, and the mass to charge ratio (m/z) is shown on the x axis.

FIG 5 FIG 6
FIG 5 Alignment of CPS biosynthesis gene clusters.Figure showing a 34.5-kb region of B. pseudomallei (Bp) K96243 containing 24 genes (BPSL2787 to BPSL2810, wcbT-manC) responsible for CPS biosynthesis.Similar regions in B. cepacia (Bc) strains 10223, 39628, LNT40, and B. thailandensis (Bt) strains SBXCC001 and SBXPR001 are aligned below the reference genome.Locus tags are shown below each of the open reading frames identified.

TABLE 1 B
. cepacia and B. thailandensis strains used in this

study Species Strain Sample type Country of isolation Latex agglutination result Purified total carbohydrate yield (from 1 L of culture)
(23)eviously described by Sim et al.(23).b ND, not determined in this study.

TABLE 2
Genome characteristics of B. cepacia and B. thailandensis strains sequenced in this study

TABLE 3
Identification of B. cepacia and B. pseudomallei by Bruker MALDI Biotyper system

by Bruker MALDI Biotyper system a Score value b
a Extended reference profile database for Burkholderia species.b Score value > 2.3 indicates species identification.