| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 ,3 Departments of Pediatrics1 and Microbiology/Immunology3 , University of Oklahoma Health Sciences Center, Oklahoma City, OK 73104, USA
2 Department of Pediatrics and Communicable Diseases, University of Michigan Medical School, Ann Arbor, MI 48109, USA
Correspondence
Terrence L. Stull
tstull{at}ouhsc.edu
Received 26 April 2005
Accepted 31 August 2005
Abbreviations: Bcc, Burkholderia cepacia complex; CF, cystic fibrosis; CRD, conserved repeat domain; HHRP, haemagglutinin/haemolysin-related protein; LRP, large repetitive protein; nr, non-redundant; Q-PCR, real-time quantitative PCR; RTX, repeat in toxin.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Microbial factors that interact with host targets in the pathogenic process can be classified into five broad categories: adhesins, invasins, evasins, cytotoxins and pabulins (Moxon & Tang, 2000). Although their current description is limited, adhesins and pabulins are the better-characterized factors contributing to Bcc virulence. The binding of cable pilin to respiratory epithelia and mucin has suggested its role in mediating colonization of the lung by B. cenocepacia (Sajjan et al., 1992, 2000; Tomich & Mohr, 2004). Cable pilus-independent mechanisms also contribute to B. cenocepacia adherence to cellular and acellular surfaces (Tomich & Mohr, 2003). In some micro-organisms, adhesins may also function as invasins, promote biofilm formation and stimulate inflammatory responses in epithelial cells (Oelschlaeger et al., 2002). Further studies of adhesins in the Bcc complex are needed to clarify these and other functions.
Systematic investigation of the remaining classes of virulence factors – invasins, evasins and cytotoxins – is just beginning in Bcc. It has been shown that B. cenocepacia can enter, survive and replicate in cultured macrophages and human pulmonary epithelial cells (Cieri et al., 2002; Keig et al., 2002; Martin & Mohr, 2000; Saini et al., 1999).
A type III secretion system utilized for the delivery of virulence-associated proteins into host cells has been demonstrated in an epidemic strain of B. cenocepacia (Tomich & Mohr, 2003) and virulence in a murine model was attenuated in a mutant deficient in type III secretion. Whilst a broad range of cytotoxins, including haemolysins, pneumolysins, cytotoxic necrotizing factors, DNA-damaging cytolethal distending toxins, neurotoxins and repeat in toxin (RTX) cytotoxins, occurs in other bacterial genera, few cytotoxins have been identified at present among Bcc members. As cytotoxins of other bacterial pathogens are associated with significant pathology, such virulence factors are attractive candidates for the cause of the destructive pathology associated with life-shortening epidemic strains of B. cenocepacia.
The large size of the B. cenocepacia genome (approx. 7·6 Mb), its apparent plasticity (Lessie et al., 1996) and the ability to infect a wide range of hosts, including humans, rodents (Bernier et al., 2003; Speert et al., 1999), nematodes (Köthe et al., 2003), unicellular organisms (Fehlner-Gardiner et al., 2002) and plants (Bernier et al., 2003), suggest that B. cenocepacia is likely to encode multiple, potentially novel virulence factors. We employed subtractive hybridization to initiate a systematic approach to the isolation of virulence genes from this pathogen. This is a molecular method by which genes whose presence differs between two strains can be isolated (Agron et al., 2002; Sambrook & Russell, 2001; Winstanley, 2002). As this method of gene isolation is not biased by prior assumptions about function, it provides a global approach to identification of virulence genes and has been utilized successfully in other pathogenic bacteria for this purpose (Adderson et al., 2003; Choi et al., 2002; Li et al., 2003; Winstanley, 2002). In the present study, we used suppressive subtractive hybridization between genomic DNA from an environmental isolate of B. cepacia (strain ATCC 25416T; Bcep25416) and the epidemic B. cenocepacia clinical isolate J2315 (BcenJ2315) to identify nucleotide sequences present only in the clinical isolate. We hypothesized that these sequences would be enriched for virulence genes. Recently completed sequencing of the BcenJ2315 genome by the Sanger Institute (Miller & Mahenthiralingam, 2003) provided a new and powerful tool for the identification, mapping and characterization of the sequences that we isolated by subtractive hybridization. Here, we report the identification of a gene encoding a putative RTX toxin in BcenJ2315, its placement within a putative operon containing three additional genes and initial evidence implicating the function of this gene in haemagglutination.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
for species and numbers of isolates) and Escherichia coli strain TOP10 (Invitrogen) were maintained routinely on Luria–Bertani (LB) agar plates at 37 °C or in LB broth at 37 °C with shaking. The strain panel selected for species distribution included 58 separate isolates distributed over nine species of the complex. Isolates of B. cenocepacia were divided into members of the clonal lineages ET12, PHDC and MW, as well as B. cenocepacia isolates categorized as ‘unique’. These latter isolates were B. cenocepacia that did not belong to any of the major clonal lineages and were different from each other. Isolates of the other species were determined to be different from each other by genotype analysis (data not shown). Antibiotic selection for recombinant E. coli containing pCR2.1-TOPO and its derivatives was performed on 50 µg kanamycin ml–1. LB agar was supplemented with 40 µg X-Gal ml–1 when appropriate.
|
Transformation with plasmids.
Plasmid constructs were transferred into electrocompetent or chemically competent E. coli. Electrocompetent E. coli DH5
was produced by using the method of Sharma & Schimke (1996). Electroporation was performed in an Eppendorf 2510 electroporator set to 17 kV cm–1. Transformation using chemically competent cells was performed by using commercially prepared E. coli TOP10 cells, following the manufacturer's directions.
Subtractive hybridization. Genomic subtractive hybridization between BcenJ2315 and Bcep25416 was performed by using a Clontech PCR Select Bacterial Genome Subtraction kit. The tester DNA (BcenJ2315) and the driver DNA (Bcep25416) were digested for several hours with RsaI prior to performing the suppression PCR-based subtractive hybridization, as detailed in the manufacturer's instructions. Resulting subtracted tester DNA fragments were ligated into pCR2.1-TOPO (Invitrogen) and transformed into E. coli TOP10. Recombinants were selected after plating on LB agar containing kanamycin and X-Gal. Plasmids from selected white colonies were purified by using a Wizard Miniprep kit (Promega) and the nucleotide sequence of the inserts was determined by automated DNA sequencing (ABI Prism model 3700 DNA analyser) at the Recombinant DNA/Protein Resource Facility at the Oklahoma State University Molecular Biology Core Facility, Stillwater, OK, USA.
In silico analysis. Nucleotide sequences from the subtractive hybridizations were analysed to determine boundaries of the insertion. The insertions were also analysed by BLASTX (Altschul et al., 1990) using the non-redundant (nr) database at the NCBI. In addition, each sequence was used in BLASTN analysis against the BcenJ2315 genome sequence at the Sanger Institute (www.sanger.ac.uk). Results were tabulated and analysed. In-house DNA analysis was performed by using the Vector NTI Suite 9.0 (Informax). Annotation was performed by using ARTEMIS 5.0 (Berriman & Rutherford, 2003; Rutherford et al., 2000). Searches using sequences of gene or protein fragments of the bcr operon were performed against all sequences in GenBank/EMBL.
Real-time quantitative PCR (Q-PCR).
The kinetics of bcrA transcription were analysed by using Q-PCR. A 100 ml culture of BcenJ2315 was grown in LB broth for 24 h. A time-zero control was obtained by removing a 1 ml sample and mixing with 2 ml RNA Protect (Qiagen) to stabilize the RNA. Samples (1 ml) for analysis of RNA profiles were taken at 1 h intervals and mixed with 2 ml RNA Protect. The samples were incubated for 10 min at room temperature and then centrifuged at 14 000 g to pellet the cells. The supernatant was aspirated and the pellets were frozen at –20 °C until processing. RNA from each sample was isolated by using an RNeasy Mini kit (Qiagen) as directed by the manufacturer and resuspended in 30 µl RNase-free water. Residual chromosomal DNA was removed by digestion with amplification-grade DNase I (Invitrogen). The RNA samples were used to prepare cDNA in a 20 µl reaction containing 7 µl template RNA, 5·5 mM MgCl2, 500 µM each dNTP, 1x RT buffer, 80 mU RNase inhibitor and 25 U MultiScribe reverse transcriptase (Applied Biosystems). The synthesis reaction was incubated at 25 °C for 10 min followed by 48 °C for 30 min. The reaction was terminated by heating at 95 °C for 5 min. Prior to analysis, the cDNA was diluted by addition of 180 µl RNase-free water. Q-PCR was utilized to examine transcription of 16S rRNA (normalizer), bcrA and cblA. Gene-specific oligonucleotide primers were designed by using Primer Express 2.0 (Applied Biosystems) (Table 2). Primers were tested to determine amplification specificity, efficiency and linearity of the amplification to RNA concentration as described by the manufacturer. A typical 25 µl reaction contained 12·5 µl SYBR Green Master Mix, 250 nM each primer and 5 µl cDNA sample. Quantification reactions for each gene at each time point were performed in triplicate and normalized to concurrently run 16S rRNA levels from the same sample. Relative quantification of gene expression was determined by using the 
method of Livak & Schmittgen (2001), where 
Ct=(Ct,Target–Ct,16 s)Time x–(Ct,Target–Ct,16 s)Control.
|
Haemolysis assay. The method used was described previously by Rowe & Welch (1994). Essentially, Bcc cells grown in LB broth were pelleted by centrifugation and the supernatant was aspirated to a fresh tube. The cells were washed three times in 150 mM NaCl and resuspended in the same solution. Supernatant or cell suspension (200 µl) was added to 770 µl 150 mM NaCl solution containing 20 mM CaCl2 and prewarmed to 37 °C. Human red blood cells (30 µl), previously washed three times in 150 mM NaCl solution, were added to start the lytic reaction. To terminate the reaction, the mixture was centrifuged at 4 °C to pellet red blood cells and bacterial cells, and the haemoglobin released was determined by measuring A540. As many Bcc strains produce pigments that might absorb at 540 nm, a negative control containing no red blood cells was included for each strain evaluated. Positive controls were prepared by lysing 30 µl red blood cell suspension in 970 µl H2O (A540, 1·20).
Time course of haemagglutinin expression. BcenJ2315 was grown in LB broth at 37 °C with shaking and sampled at 2 h intervals for 24 h. Bacteria were pelleted and assayed as described above. In addition, viable counts were determined at each time point. Several other isolates were examined similarly to ensure that the BcenJ2315 expression profile was representative of the Bcc.
Species distribution of the bcr operon.
Gene-specific oligonucleotide pairs were designed to target bcrA–D based on the BcenJ2315 sequence data (Table 2
). Following optimization with BcenJ2315 genomic DNA, these primers were utilized to examine the distribution of these genes across the Bcc strain panel. As several primer pairs targeting bcrB resulted in the amplification of multiple bands, the examination of the distribution of this gene was excluded from this study. The conditions used for bcrA, C and D PCR assays were as follows: a colony-lysis method was used to prepare template DNA for PCR. Each 25 µl bcrC and bcrD PCR contained 1 unit Taq polymerase, 2 mM MgCl2, 0·8 mM dNTPs, 0·4 µM each primer, 1x PCR buffer (Invitrogen) and 2 µl bacterial lysate. For the bcrA PCR, the concentrations were the same as above except that the MgCl2 concentration was reduced to 1 mM. A PTC-100 programmable thermal cycler (MJ Research) was used for amplification. Amplification parameters included an initial step of 95 °C for 5 min for each primer pair and 30 cycles of annealing for 30 s, extension at 72 °C for 30 s and denaturing at 95 °C for 30 s, followed by a final extension of 2 min. Individual annealing temperatures were 57 °C for bcrA, 64 °C for bcrB and 58 °C for bcrC.
|
RESULTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
). BLASTX analysis indicated that C09 had significant sequence identity to 33 distinct regions within a putative haemagglutinin/haemolysin-related protein (HHRP) from Ralstonia solanacearum GMI1000 (51–79 % identity) (GenBank accession no. NP_522741).
|
). Analysis of the first ORF indicated a predicted product of 2888 aa that had a high degree of similarity to a putative HHRP in R. solanacearum GMI1000, identified above. Initial examination of the predicted BcenJ2315 protein sequence indicated that the central region of the protein contained multiple copies of a dimeric repeat sequence. Each copy contained a region (referred to as repeat A) of approximately 104 aa and a second region (referred to as repeat B) of approximately 85 aa. The putative protein contains 11 AB repeats and a single A repeat with no associated B region. Thus, the arrangement of the core section was denoted as being ‘(AB)11A’. The regions corresponding to the 12 C09 homologies, identified in the subtractive hybridization, span the 12 A repeats (Fig. 1). Sequence alignments between the A and B repeats indicated that they were similar to each other and that they each contained three highly conserved regions, denoted as conserved repeat domains (CRDs). These regions were also shared by the R. solanacearum protein (Fig. 2). Using CRD1, CRD2 and CRD3 in BLAST searches, we identified a number of other proteins from a diverse range of human and plant pathogens that contained these sequences. These were large, repetitive proteins (LRPs) with multiple repeats of the CRD1, CRD2 and CRD3 motifs. Fig. 2 shows the alignment of the individual CRD regions with the putative HHRP of R. solanacearum, a putative adhesin from Pseudomonas aeruginosa PA01 (GenBank accession no. NP_250565) and two other LRPs from Salmonella enterica subsp. enterica serovar Typhi TY2 (GenBank accession no. NP_806354) and Burkholderia pseudomallei strain KG6243 (GenBank accession no. YP_108272). Immediately adjacent to the (AB)11A repeat region are two regions sharing partial similarity with the repeat sequences. As these were not as conserved as each A and B repeat, these were referred to as repeat-like regions.
|
|
). In addition, BLAST analysis of the region from the first residue to the start of the first repeat-like region (amino-terminal 208 aa) identified significant similarity to the R. solanacearum HHRP (52 % identity) and a haemolysin-type, RTX-related protein from Nitrosomonas europaea ATCC 19718 (42 % identity). BLAST analysis of the carboxy terminus of our protein, distal to the RTX region, identified significant similarity to the carboxy terminus of the previously identified R. solanacearum, N. europaea and S. enterica proteins (73, 45 and 44 % identity, respectively). Additionally, it demonstrated 45 % identity to the carboxy terminus of a putative RTX family exoprotein from E. coli O157 : H7 EDL933.
Three other potentially co-transcribed ORFs were identified downstream of the putative BcenJ2315 RTX-related gene. Homologues of the predicted products of these genes were located contiguous to all of the related proteins discussed above. Homology suggests that the ORF immediately downstream of the RTX-related gene encodes a putative outer-membrane efflux protein of the TolC family. The predicted product of the next ORF is a putative composite ATP-binding transmembrane ABC transporter containing fused permease and ATPase domains. The predicted product of the last ORF is a putative haemolysin-type secretion transmembrane protein. Together, these predicted products constitute a type I secretion system, previously demonstrated to be required for secretion of other RTX proteins (Welch, 2001). We have assigned the designation bcr (Burkholderia cenocepacia RTX) to this newly discovered operon, containing gene bcrA (the RTX protein) and three genes distal to this RTX determinant, designated bcrB, C and D, respectively (Fig. 1).
Temporal expression of bcrA
Q-PCR was used to investigate the expression of the BcenJ2315 RTX–haemagglutinin gene (bcrA) in broth cultures. The expression profile of bcrA expression is shown in Fig. 3. Early exponential-phase cultures (
108 c.f.u. ml–1) contained low levels of bcrA transcripts. A 3·4-fold increase (range, 3·1- to 3·8-fold) in expression of this gene occurred as the culture density increased to 5x108 c.f.u. ml–1. When the culture entered early stationary phase, the expression of this gene had increased 35-fold (range, 30- to 40-fold). Subsequently, transcription of bcrA decreased markedly, but remained above levels expressed in early-exponential cultures (2·4- to 9-fold) and returned to basal levels in late stationary-phase cells.
|
). The observed pattern of bcrA expression differed from that of cblA. Unlike bcrA, maximal transcription of cblA was observed in exponentially growing cells. Expression of cblA declined coincident with the culture entering late-exponential phase and expression of this gene at 24 h had fallen by 128-fold (range, 112- to 148-fold) compared with the level observed in early exponentially growing cells.
Phenotypic characterization of BcenJ2315
The presence of an RTX-like gene with significant similarity to a putative haemagglutinin in the genome of BcenJ2315 led us to compare the abilities of BcenJ2315 and Bcep25416 to cause haemagglutination. Stationary-phase cells of BcenJ2315, washed and resuspended in PBS to a density of 109 cells ml–1, caused reproducible agglutination of human red blood cells after 30 min incubation at room temperature. Bcep25416 cells failed to cause haemagglutination. A representative example comparing the agglutination of red blood cells by the two strains is shown in Fig. 4(a). Addition of PBS to the suspension of red blood cells served as a negative control. As shown by the negative control (lane 1), non-agglutinated red blood cells pelleted in the centre of the conical wells within 30 min of PBS addition. In contrast, erythrocytes that agglutinated after interaction with BcenJ2315 remained dispersed throughout the wells (lane 2). Clumping of erythrocytes was evident visually in the erythrocyte–J2315 cell mixtures. Suspensions of BcenJ2315 cells in either PBS or 150 mM NaCl containing 20 mM CaCl2 did not cause spectrophotometrically detectable lysis of human red blood cells after incubation at room temperature or at 37 °C. The appearance of the human erythrocyte suspension 30 min after mixing with Bcep25416 cells from companion cultures (lane 3) did not differ from that of the negative control. Neither cell-free medium from BcenJ2315 cultures nor membrane-free cell lysates brought about haemagglutination (data not shown). Trypsin treatment of BcenJ2315 cells prior to their addition to a red blood cell suspension blocked their haemagglutinating ability (data not shown). Thus, the haemagglutinating activity of BcenJ2315 appeared to be both cell-bound and protein-based, and not an artefact associated with haemolysis. This agglutination phenotype was not observed in the environmental isolate used in the subtractive hybridization.
|
. Concentrated BcenJ2315 cells from exponentially growing cultures did not agglutinate human erythrocytes. Onset of haemagglutinating activity of BcenJ2315 cells began in late-exponential phase (
109 c.f.u. ml–1) and continued for an extended time during stationary phase. At no time did strain Bcep25416 exhibit haemagglutinating activity. The data in Fig. 4(b) also show that the inability of Bcep25416 to cause haemagglutination cannot be explained as an artefact of low cell titre related to a reduced growth rate or that the culture was in a different growth phase from BcenJ2315. The phenotypic change at 109 c.f.u. ml–1 was reproducible in multiple experiments.
Autoagglutinating ability of BcenJ2315
Another phenotypic difference that was observed between the BcepJ2315 and Bcep25416 strains was in their abilities to autoagglutinate (Fig. 5). Suspensions of 109 BcenJ2315 cells in PBS autoagglutinated, whereas cell suspensions of strain Bcep25416 did not. BcenJ2315 cell suspensions agglutinated visibly in 2–12 h, but suspensions of Bcep25416 remained dispersed for 24 h upon standing at room temperature (Fig. 5a). Turbidimetric measurements (OD600) confirmed the more rapid settling out of BcenJ2315 cell suspensions compared with suspensions of Bcep25416 cells (Fig. 5b). The rapidly settling BcenJ2315 cultures could be identified reliably by turbidimetric measurements, typically within 60 min of vortexing of the cell suspensions. Expression of the autoagglutination phenotype by BcenJ2315 cells also was growth phase-dependent. Exponential-phase BcenJ2315 cells did not autoagglutinate (Fig. 5b). Onset of the ability to autoagglutinate was concomitant with the onset of haemagglutinating activity (data not shown) and remained present in stationary-phase cultures for at least 36 h.
|
Distribution of the bcr operon and haemagglutination activity among Bcc isolates
The species distribution of the bcr operon was determined by using 58 repesentative Bcc isolates. The bcrA gene was present in only the ET12 lineage of B. cenocepacia (Table 1). bcrC was detected in B. multivorans (three of five isolates), whilst bcrD was observed in both B. multivorans (four of five isolates) and Burkholderia stabilis (one of five isolates). PCR analysis indicated that, in all of the ET12 isolates, bcrA was always associated with bcrC and bcrD. Examination of haemagglutination indicated that this phenotype was specific to the ET12 lineage, with 100 % of isolates tested being positive.
|
DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Subtractive hybridization, a differential genomic technique, has been employed to identify virulence factors, including a heat-resistant haemagglutinin involved in E. coli urinary-tract infections (Srinivasan et al., 2003), a novel fimbrial operon distinct to Salmonella typhimurium (Morrow et al., 1999) and numerous genetic elements distinct to certain strains of Helicobacter pylori (Akopyants et al., 1998). This technique is ideally suited to the identification of novel genetic elements within the Bcc, due to the high degree of genetic relatedness between the species.
We used Bcep25416 and BcenJ2315 in this study to isolate putative virulence factors distinct to the latter organism. B. cepacia ATCC 25416T was originally characterized as a phytopathogen, isolated from an onion (Allium cepa). In contrast, B. cenocepacia J2315 is a member of the transmissible ET12 lineage and is responsible for numerous fatalities among CF patients (Govan et al., 1993; Pitt et al., 1996). In addition, the genome of this isolate has been sequenced and assembled into three chromosomes and a plasmid element. A fragment isolated by this procedure led to the identification of an operon predicted to encode proteins with similarity to a putative HHRP and associated type I secretory system. The first of these predicted proteins, BcrA, has a size of 278 kDa, contains multiple conserved dimeric repeats and has four Ca2+-binding domains characteristic of RTX proteins.
While RTX proteins are classically associated with cytotoxic effects, some RTX-containing proteins lack these activities and are associated with other functions (Welch, 2001). In all organisms containing an RTX, secretion of the toxin is mediated by an associated type I secretion system. The most studied of the RTX toxins is HlyA from E. coli. HlyA, like most other RTX toxins, is over 100 kDa and contains multiple Ca2+-binding domains and a conserved carboxy terminus required for correct secretion. This conserved carboxy terminus was not present in BcrA; however, BcrA shared a highly conserved carboxy-terminal region with most of its repeat-containing homologues. This suggests a unique secretion signal specific to this family of proteins. Following synthesis of HlyA, the protein is activated in the cytoplasm by acylation via HlyC. The acylated HlyA is then secreted across the inner membrane and periplasm via a complex of HlyB and HlyD. This complex interacts with TolC, which is located in the outer membrane. Together, the HlyB–HlyD–TolC complex acts as a translocase (Koronakis, 2003). Examination of the ORFs located downstream of bcrA identified three putative proteins with homology to the TolC family, HlyB-like ATPase–permease family and HlyD-like transmembrane secretion protein. The BcenJ2315 RTX operon that we have described lacks a homologue of HlyC. A BLAST search failed to identify a homologue of the E. coli hlyC activator gene in the BcenJ2315 genome, but functional analogues of hlyC may exist. We have assigned the designation bcr (Burkholderia cenocepacia RTX) to this newly discovered operon containing gene bcrA (the RTX protein) and three genes distal to this RTX determinant, designated bcrB, C and D, respectively.
Our original subtractive hybridization was designed to identify genes specific to the human pathogen BcenJ2315 that were absent in the environmental isolate Bcep25416. As we identified a single, small fragment of the bcr locus, it was important to determine whether this result indicated a complete absence of the specific region of bcrA, the entire gene or the bcr operon in Bcep25416. Our results indicate that this isolate lacks the entire bcr locus. Expansion of this study across the Bcc suggests that the locus appears only in B. cenocepacia. Within this species, the locus is restricted to the highly transmissible ET12 lineage. Within other pathogenic species, the presence of RTX-like proteins in certain isolates is associated with elevated transmissibility (Lin et al., 1999). The presence of homologues of the type I secretion-accessory proteins BcrC and BcrD in other species may indicate the presence of other toxins or secreted proteins at these loci, distinct from BcrA.
It has been shown in B. cenocepacia and other organisms that virulence factors exhibit growth phase-dependent expression regulated by the quorum-sensing regulon (Köthe et al., 2003; Lewenza & Sokol, 2001). In addition, other RTX-like proteins show similar profiles of expression (Daborn et al., 2001). Examination of the time course of bcrA transcription indicated dramatically increased expression of the gene in late exponential-phase culture (109 c.f.u. ml–1). The expression profile of bcrA is suggestive of cell density-dependent regulation, possibly by the quorum-sensing regulon.
Whilst we identified and characterized the bcr locus, the function of the bcrA gene product remains to be established. In order to determine a possible phenotype, we sought a surface-associated or secreted activity limited to bcrA+ isolates that was protein-mediated, had onset at approximately 109 c.f.u. ml–1 and was consistent with a function in virulence. As the original BLAST analysis identified similarity to a putative HHRP and as RTX-containing proteins often have haemolytic activity, we hypothesized that the predicted product of bcrA is a secreted protein with haemolytic and/or haemagglutinating activity. Under the conditions of our studies, we were unable to demonstrate haemolytic activity associated either with culture supernatants or PBS-washed cells in either the absence or presence of Ca2+. We demonstrated a haemagglutinating activity of BcenJ2315 that was absent from Bcep25416. This activity was shown to be protein-mediated and surface-associated. Examination of this phenotype across the species indicated that it was limited to isolates previously determined to be bcrA+. This activity was observed beginning in late exponential- or early stationary-phase cultures, consistent with the bcrA expression profile. During the haemagglutination studies, it was also noticed that BcenJ2315 exhibited autoagglutination activity. This activity was only observed in B. cenocepacia isolates that were bcrA+ and haemagglutinating. Onset of these two activities co-occurred.
Our findings suggest that transition from exponential growth to stationary phase is accompanied by changes in the adhesive properties of the cell. Experiments examining autoagglutination in mixed suspensions of Bcep25416 and BcenJ2315 indicate that the factor(s) mediating this activity is species-specific. In addition, experiments in which exponential- and stationary-phase BcenJ2315 cells were mixed indicate that early-phase cells lack this factor(s) that is expressed in stationary-phase cells. An alternative explanation is that changes in stationary-phase cell-surface architecture expose an existing mediator of autoagglutination. Whilst a previous study suggested that expression of the cable pilus masks autoagglutination (Tomich & Mohr, 2003), our study indicates that autoagglutination (and haemagglutination) occurs prior to a significant decrease in cblA mRNA levels. Our observations are consistent with de novo synthesis of a surface-associated adhesive factor.
Whilst this study demonstrates an association between an RTX-like gene locus and haemagglutination/autoagglutination, there are several caveats that should be noted. Our original hypothesis of BcrA function was based on similarity to a putative HHRP from R. solanacearum GMI1000. The actual biological function of the HHRP and BcrA may be very different. Although the limited number of isolates employed in this study suggests an association between bcrA and the two phenotypes, issues related to clonality and species distribution require the examination of a larger and more comprehensive strain panel. Mutational analysis is required to definitively establish the relationship between bcrA and haemagglutination.
Irrespective of whether bcrA encodes a haemagglutinin, genomic presence and expression of this RTX-like gene have several clinical implications. This gene is a good candidate for a virulence factor contributing to lung pathology. Accumulating evidence indicates the involvement of RTX determinants in pathological changes associated with pulmonary infection by other bacterial pathogens. For example, not only do the RTX toxins cause degranulation of alveolar macrophages and neutrophils (Czuprynski et al., 1991; Maheswaran et al., 1992), but also, some are potent inducers of inflammatory cytokines (Fullner et al., 2002; Yoo et al., 1995), cause loss of paracellular tight junction of epithelial cells (Fullner et al., 2001) and have a concentration-dependent effect on the severity of lung lesions (Ames et al., 1985). Whilst we have shown an association between BcrA and haemagglutination/autoagglutination, these activities may be simply a facet of this protein's activity in vivo. In other organisms, mutational alteration of the RTX gene improves clinical signs of lung disease, decreases serum levels of proinflammatory mediators and increases bacterial load necessary to cause lethality in infection models (Chidambaram et al., 1995; Fullner et al., 2002; Highlander et al., 2000; Tatum et al., 1998). Finally, RTX toxins from other pulmonary pathogens are highly immunogenic. Antibodies against selective domains provide protective immunity in murine infection models (Betsou et al., 1995; Seah et al., 2002). Therefore, the BcrA protein is a potential candidate for the development of a vaccine to prevent high-risk patients from infection with an epidemic, highly pathogenic strain of B. cenocepacia.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
CT method. Methods 25, 402–408.[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  P. W. Whitby, T. M. VanWagoner, J. M. Springer, D. J. Morton, T. W. Seale, and T. L. Stull Burkholderia cenocepacia utilizes ferritin as an iron source J. Med. Microbiol., June 1, 2006; 55(6): 661 - 668. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
method (
