|Home | About | Journals | Submit | Contact Us | Français|
Burkholderia cenocepacia (formerly Burkholderia cepacia complex genomovar III) causes chronic lung infections in patients with cystic fibrosis. In this work, we used a modified signature-tagged mutagenesis (STM) strategy for the isolation of B. cenocepacia mutants that cannot survive in vivo. Thirty-seven specialized plasposons, each carrying a unique oligonucleotide tag signature, were constructed and used to examine the survival of 2,627 B. cenocepacia transposon mutants, arranged in pools of 37 unique mutants, after a 10-day lung infection in rats by using the agar bead model. The recovered mutants were screened by real-time PCR, resulting in the identification of 260 mutants which presumably did not survive within the lungs. These mutants were repooled into smaller pools, and the infections were repeated. After a second screen, we isolated 102 mutants unable to survive in the rat model. The location of the transposon in each of these mutants was mapped within the B. cenocepacia chromosomes. We identified mutations in genes involved in cellular metabolism, global regulation, DNA replication and repair, and those encoding bacterial surface structures, including transmembrane proteins and cell surface polysaccharides. Also, we found 18 genes of unknown function, which are conserved in other bacteria. A subset of 12 representative mutants that were individually examined using the rat model in competition with the wild-type strain displayed reduced survival, confirming the predictive value of our STM screen. This study provides a blueprint to investigate at the molecular level the basis for survival and persistence of B. cenocepacia within the airways.
Isolates of the Burkholderia cepacia complex are gram-negative opportunistic pathogens. Usually harmless while surviving in the soil, these bacteria can cause devastating infections in patients with cystic fibrosis (CF) and chronic granulomatous disease (3, 23, 24, 30, 31). Pseudomonas aeruginosa is the predominant respiratory pathogen in CF patients (23), but the disease risk for infection with B. cepacia complex in patients with CF is substantially higher than with P. aeruginosa alone or with other bacteria (13). Although the clinical outcome among CF patients infected with B. cepacia complex varies greatly, infections in some patients result in a rapidly progressive and fatal bacteremic disease (30). Furthermore, isolates of the B. cepacia complex have the potential for patient-to-patient spread, both within and outside the hospital (22, 32, 41, 51, 52). Recent taxonomic studies demonstrated that strains identified as B. cepacia comprise a complex of closely related species or genomovars, collectively called the B. cepacia complex (43, 67). This complex consists of at least nine species sharing a high degree of 16S rDNA and recA sequence similarity and moderate levels of DNA-DNA hybridization (for a recent review, see reference 10). Strains of each of the species within the B. cepacia complex have been isolated from CF patients; however, some are more common than others. Burkholderia multivorans (formerly genomovar II) and Burkholderia cenocepacia (formerly genomovar III) isolates comprise about 10 and 83%, respectively, of all B. cepacia complex isolated from CF patients in Canada (62).
Isolates of the B. cepacia complex are inherently resistant to many antimicrobial agents (49, 54, 55). The widespread antibiotic resistance of these microorganisms has proven extremely problematic for the treatment of infections. Furthermore, the lack of sensitivity to antibiotics commonly used for genetic selection complicates genetic studies in this pathogen by drastically limiting the choice of antibiotic resistance gene markers for mutagenesis and complementation experiments (38).
Several bacterial factors may play a role in the infections caused by isolates of the B. cepacia complex. Some isolates have the ability to survive intracellularly within eukaryotic cells such as macrophages, respiratory epithelial cells, and amoebae (6, 34, 44, 46, 56). Other potential virulence factors that have been described include cable pili (57), flagella (65), a type III secretion system (50, 64), surface exopolysaccharide (7, 9), production of melanin (71), catalase (37), up to four types of iron-chelating siderophores (16), proteases and other secreted enzymes (12, 29, 47, 68), quorum-sensing systems (40, 42), and the ability to form biofilms (66). Not all strains produce each of the proposed virulence factors and, to date, none of these individual factors has been clearly demonstrated to be a major contributor to human disease.
Signature-tagged mutagenesis (STM) was originally developed to facilitate the detection of Salmonella enterica serovar Typhimurium genes required for in vivo survival (28). STM has since been applied to several different bacteria and fungal pathogens (for a recent review, see reference 48). STM is a comparative hybridization technique that employs a collection of transposons, each modified by the incorporation of a unique DNA sequence tag (28). These transposons are introduced into the pathogen to be studied and, following random transposition, insertional mutants are isolated. Mutants containing a transposon insertion in a gene which codes for a function required for in vivo growth or survival will fail to pass through the in vivo selection. Individual mutants can be distinguished from each other based on unique tags carried by the transposon insertions of each strain. By comparing the tags present in the input pool of mutants initially inoculated into the animal model with those unique tags present in the bacteria recovered after infection, it is possible to identify mutants that fail to grow exclusively in vivo. Subsequently, the DNA sequence surrounding the insertion can be analyzed to identify the gene or genes required for virulence. Traditional STM involves using randomly generated tags that can be detected by conventional hybridization techniques (28), which may result in cross-hybridization signals and high background, increasing the proportion of false positives (53). This was a significant concern in the case of B. cenocepacia due to the high moles percent G+C content of the genome. A PCR-based modification of STM can potentially eliminate pitfalls inherent to hybridization and increase the specificity during the PCR screening step (39).
In this study, we describe a modified STM procedure to isolate and identify candidate genes of B. cenocepacia required for the in vivo survival in 102 transposon mutants that could not be recovered from intratracheal lung infections in rats. The modifications to the STM strategy included the use of a real-time PCR-based screening step for the detection of each transposon mutant within a pool. In addition, we used a plasposon backbone instead of a classical transposon to facilitate the identification of flanking chromosomal regions once the element is integrated onto the chromosome, since the transposable element within the plasposon contains an Escherichia coli plasmid replication origin (18). Inherent problems of STM such as cross-amplifying tags were removed prior to preparing the pools for the in vivo passage, thereby making these modifications useful as a general screening method to identify genes required for survival in vivo.
Unless otherwise indicated, all strains were grown at 37°C. B. cenocepacia strain K56-2 (16) was grown in Luria-Bertani (LB) medium. This strain is a CF isolate that belongs to the same clonal group as the strain J2315, whose genome has recently been sequenced (http://www.sanger.ac.uk/Projects/B_cenocepacia/), and it was previously shown to cause a chronic lung infection in the agar bead rat model (61). E. coli DH5α [F− 80 lacZΔM15 endA1 recA1 hsdR17(rK− mK+) supE44 thi-1 ΔgyrA96 (ΔlacZYA-argF)U169 relA1] was used as a host for the plasposon constructs and for the helper plasmid pRK2013 (see below). Transposon mutants were grown in Typticase soy agar, Vogel-Bonner minimal medium (VBM) (69), or LB medium as required, all of which were supplemented with trimethoprim at a final concentration of 100 μg/ml. E. coli strains were grown in LB medium supplemented with 50 μg of trimethoprim/ml or 40 μg of kanamycin/ml, as appropriate. All chemicals were purchased from Sigma Chemical Co., St. Louis, Mo., unless otherwise indicated.
Forty-one plasposons (designated pTnMod-OTp-tag1 to pTnMod-OTp-tag41) were constructed as depicted in Fig. Fig.1.1. The plasposon backbone was from pTnMod-OTp′ (18). Forty-one different oligonucleotides of 21 bp (Table (Table1)1) were annealed to complementary molecules to yield double-stranded oligonucleotides. The ends of the double-stranded oligonucleotides were phosphorylated and separately ligated into the blunted KpnI site of pTnMod-OTp′ to produce 41 different plasposons pTnMod-OTp-tagn, where n is the oligonucleotide tag number 1 to 41. Ligations were verified by PCR using each of the 21-bp oligonucleotides (Table (Table1)1) as one primer and primer STM-common (5′-TCGATTTCGTTCCACTGAGCG-3′) as the second primer. PCR amplifications were performed in a PTC-0200 DNA engine (MJ Research, Incline Village, Nev.) by using Taq polymerase (Roche Diagnostics, Laval, Quebec, Canada). The 811-bp product was amplified as follows: 7 min at 95°C, two cycles of 95°C for 1 min, 65°C for 2 s (followed by decreases of 0.2°C per second up to 55°C), and 72°C for 1 min; and then 10 cycles of 95°C for 1 min, 55°C for 1 min, and 72°C for 1 min. The product was visualized on a 0.7% (wt/vol) agarose gel.
Transposon mutagenesis of B. cenocepacia K56-2 was performed 41 separate times using each pTnMod-OTp-tagn plasposon to produce 96 individual K56-2 transposon mutants for each tagged transposon. Triparental matings were carried out as previously reported (15) to mobilize each plasposon into K56-2 with E. coli DH5α(pRK2013) as a helper strain (21). After transposition, insertion mutants were plated on minimal VBM media supplemented with trimethoprim to prevent the recovery of auxotrophic mutants, and finally the individual mutants were stored in 96-well plates.
Infections were done using the agar bead model of chronic lung infection in rats as previously described (61). Two Sprague-Dawley male rats (150 to 160 g; Charles River Canada) were used for each infection pool of 37 B. cenocepacia transposon mutants containing 105 viable bacteria embedded in agar beads. The lungs were removed 10 days following intratracheal instillation into the left pulmonary lobe and homogenized using a Polytron homogenizer (Brinkman Instruments, Westbury, N.Y.). Serial dilutions in phosphate-buffered saline were plated on tryptic soy agar and on B. cepacia selective agar (BCSA) (27) containing 100 μg of trimethoprim/ml. The CFU were counted and bacteria collected for further analysis following incubation at 37°C overnight.
To determine the competitive index (CI) of selected STM mutants, mutant and wild-type bacteria were grown overnight and adjusted to the same optical density at 600 nm. Agar beads prepared from a mixture of parent and mutant combined to give approximately equal numbers (5 × 105 CFU of each strain) were used to inoculate groups of five rats. The ratio of mutant to wild-type bacteria in the inoculum was verified by plating agar beads containing bacteria on BCSA medium with and without trimethoprim to determine viable counts. At 10 days postinfection the lungs were removed, homogenized, diluted, and plated in triplicate on BCSA and BCSA containing 100 μg of trimethoprim/ml. The CI was calculated as the mean output ratio of mutant to wild type divided by the input ratio of mutant to wild-type organisms.
Real-time PCR was performed on B. cenocepacia transposon mutants recovered from lung homogenates using the Light Cycler (Roche Diagnostics). Chromosomal DNA was extracted from the bacterial pools and used as a template for real-time PCR using a Faststart DNA Master SYBR Green I kit for DNA amplification and detection (Roche Diagnostics). The forward primers used are listed in Table Table1,1, and the reverse primer was STM-LC (5′-AAGGGAGAAAGGCGGACAGGTA-3′). The conditions used for real-time PCR to amplify the 342-bp product were as follows: activation for 10 min at 95°C, followed by 35 cycles of 95°C for 15 s, 55°C for 5 s, and 72°C for 11 s. A melting curve was generated by decreasing the temperature to 65°C, followed by a 0.2°C per second increase in the temperature up to 95°C. Quantitation curves and melting curves were generated for each 35-sample PCR run and analyzed using the Light Cycler software version 3.5 (Roche Diagnostics). Real-time PCR screening for the second screen was performed exactly the same as for the initial screen.
The chromosomal sequences surrounding the transposon insertions were identified using the self-cloning strategy as described previously (18). Briefly, chromosomal DNA from the B. cenocepacia transposon mutants that passed through the second screen was isolated and subjected to restriction endonuclease digestion by either SalI or NotI (Roche Diagnostics). The digests were ligated under dilute conditions to favor intramolecular ligations with T4 DNA ligase (Roche Diagnostics) and transformed into competent E. coli JM109 (70). Transformants were selected on LB agar supplemented with trimethoprim (50 μg/ml). Plasmid DNA was isolated using a High Pure plasmid isolation kit (Roche Diagnostics) and sequenced at the Core Molecular Biology Facility (York University, Ontario, Canada) using primers OTp-3′ (5′-TGTGGCTGCACTTGAACG-3′) or pTnMod (5′-TTCCTGGTACCGTCGACA-3′). The sequences obtained were compared to the GenBank database by BLAST to identify homologous sequences and to data from the B. cenocepacia sequencing group at the Sanger Institute for B. cenocepacia strain J2315 (http://www.sanger.ac.uk/Projects/B_cenocepacia/).
Cytosolic proteins were obtained for two-dimensional gel electrophoresis by two passages in a French press (American Instrument Co., Inc., Silver Spring, Md.) at 20,000 lb/in2, and the protein concentration was estimated by Bradford assays (4) using Bio-Rad protein assay dye reagent concentrate. Protein from cell lysates (300 μg) was precipitated with 10% trichloroacetic acid, and the pellet was washed first with 5% trichloroacetic acid and then with acetone (25). Precipitated proteins were then subjected to rehydration on Immobiline DryStrip (pH 3 to 10, 13 cm; Amersham Biosciences, Uppsala, Sweden) for 10 h in rehydration solution (8 M urea, 2% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate [CHAPS] and IPG buffer [Amersham Biosciences], pH 3 to 10) containing dithiothreitol, followed by isoelectric focusing for 1 h at 500 V, 1 h at 1,000 V, 5 h at 5,000 V, and 6.25 h at 8,000 V (IPGphor; Amersham Biosciences). After isoelectric focusing, IPG strips were equilibrated in equilibration buffer (50 mM Tris-HCl, pH 8.8, 6 M urea, 30% (vol/vol) glycerol, 2% (wt/vol) SDS, trace bromophenol blue) and applied to a 1-mm thick 12.5% polyacrylamide gel to resolve the second dimension. The gel was overlaid with 0.125% (wt/vol) agarose in SDS electrophoresis buffer containing bromophenol blue (3.02 g of Tris base/liter, 14.42 g of glycine/liter, 1% (wt/vol) SDS in distilled water). Gels were run at 10 mA per gel for 15 min followed by 20 mA for 5 h in SDS electrophoresis buffer. Proteins were visualized by SYPRO orange (Bio-Rad Laboratories) according to the manufacturer's instructions and then stained with Coomassie brilliant blue (0.25% Coomassie brilliant blue, 45% methanol, 10% glacial acetic acid; distained with 45% methanol and 10% glacial acetic acid). Gels were dried using a Bio-Rad gel dryer model 583 (Bio-Rad Laboratories). To assess reproducibility and standardize the preparation of extracts and the conditions for running the two-dimensional gel electrophoresis, gels with samples from the wild-type strain K56-2 were run four times, and those with samples from the mutants strains were run twice. Spots were analyzed using ImageMaster 2D Elite software (NonLinear Dynamics, Durham, N.C.).
To distinguish individual transposon mutants within a pool, 41 oligonucleotide tags, each consisting of 21 bp, were incorporated into the transposon backbone of pTnMod-OTp′ near one of the inverted repeats to generate 41 unique plasposons pTnMod-OTp′-tagn (Fig. (Fig.1).1). The 21-bp oligonucleotides were designed such that the 14 residues at the 5′ end were identical, while the 7-bp region near the 3′ end was unique to each tag (Table (Table1)1) (39). The pTnMod-OTp′ plasposon was previously shown to successfully integrate at random into the chromosome of B. cepacia complex isolates following mobilization from an E. coli host strain (18-20). The presence of a dhfr gene within the transposable element confers resistance to trimethoprim, one of the few antibiotics to which most strains of the B. cepacia complex are sensitive (38, 49). Prior to the construction of the transposon mutant library for infection in the animal model, we examined each of the 41 plasposons by PCR to investigate possible cross-amplification among the individual tags. This control was necessary since cross-amplifications would be detrimental to the STM experiment, as potential mutants unable to survive in the in vivo animal model would be misidentified. We tested each of the 41 pTnMod-OTp-tagn plasmids against each other using a primer set consisting of an upstream primer that anneals 811 bp upstream of the 21-bp oligonucleotide tag within the transposon (STM-common) and a downstream primer which was the 21-bp oligonucleotide tag itself (Fig. (Fig.1).1). Figure Figure2A2A shows a representative gel in which the PCR was performed with the tag2-specific and STM-common primers, with the various pTnMod-OTp-tagn plasmids as DNA templates. The absence of PCR amplification in the reactions with all the plasposons not containing tag2 indicated that this tag did not cross-amplify with any of the other tag oligonucleotides. A series of PCRs was performed for each tag primer in a similar manner as with tag2. These experiments allowed us to eliminate four pTnMod-OTp-tagn plasposons carrying the oligonucleotide tags 13, 21, 24, and 30, which cross-amplified with other tags (Table (Table11 and data not shown). The tags in the remaining 37 pTnMod-OTp-tagn plasposons did not cross-amplify during PCR amplifications and were thus used for further steps in the STM procedure.
An additional control was done to ensure that the oligonucleotide tags could be detected once the transposon was inserted in the B. cenocepacia chromosome. Because the genome of B. cenocepacia has an approximate G+C content of 67 mol%, PCR using chromosomal DNA as a template can often yield no amplification products or several nonspecific bands, possibly due to mispriming. In addition, high stringency conditions are required for the PCR amplifications due to the high sequence similarity between the oligonucleotide tags. PCR was performed using the STM-common and tag-specific primers to demonstrate, as shown in Fig. Fig.2B,2B, that the appropriate tag could be detected from B. cenocepacia chromosomal DNA. Furthermore, specific tags could also be detected when DNA from mixed cultures was used as a template (Fig. (Fig.2B).2B). These preliminary tests demonstrated that the tags and plasposon constructed could be used for the STM screening in B. cenocepacia.
The 37 plasposons were then separately mobilized into B. cenocepacia K56-2 by triparental conjugation to produce 96 transposon mutants for each of the tags. The 37 banks of B. cenocepacia transposon mutants were next arranged into pools such that any individual pool contained one of each of the tag-specific transposon mutants. Thus, each pool contained 37 distinct transposon mutants which could be individually distinguished from one another by the presence of the unique 21-bp oligonucleotide tag within the integrated transposon. The pools of transposon mutants were separately incorporated into agar beads, and the beads were used for intratracheal lung infection in rats. Each infection was performed in duplicate and allowed to proceed for 10 days, an interval which results in chronic lung infection (C. Kooi and P. A. Sokol, unpublished data). After the 10-day infection, the lungs were removed and the recovered bacteria were collected for DNA preparation and PCR screening to identify the surviving mutants in each original pool.
The 21-bp oligonucleotide tags containing unique sequences within each transposon enabled differentiation of mutants that were present from those mutants which were absent from each of the output pools. DNA from the B. cenocepacia mutants that were recovered from the lung infections served as templates for PCR amplification. Because infections with each pool of mutants were carried out in duplicate, screening of mutants was done with the infection that exhibited the highest bacterial yield, as this would result in the lowest probability of false negatives. We used real-time PCR amplification with primer sets similar to those employed to test cross-reactivity (STM-LC and each specific tag oligonucleotide). Quantitation and melting curves were generated for each of the PCR amplifications. Figure Figure33 illustrates the data analysis of one PCR run. For each run, a positive and a negative control were included together with 33 sample reactions. The product quantitation curve (Fig. (Fig.3A)3A) indicated the amount of double-stranded PCR product generated following each cycle. In the case of DNA prepared from bacteria recovered from rat lungs, we observed quantitation curves ranging from no product to more product than with the positive control. After each round of PCR amplification, a melting curve was generated for each run of 35 samples (Fig. (Fig.3B).3B). This demonstrated whether a PCR product generated was specific, by comparing it with the product from the positive control reaction, which consistently had a peak at about 88°C. The melting curve was more informative than the quantitation curve because it indicated whether or not the observed amplification was specific. Figure Figure3C3C shows a subset of the melting curves in Fig. Fig.3B,3B, including the positive and negative control plus three sample reactions. In this example, the tag7 mutant in pools F4b and H8b did not amplify with the tag7-specific primer, and therefore these mutants were not present in the pool and would pass through this screening procedure. Seventy-one pools of B. cenocepacia transposon mutant chromosomes were screened in this manner with the 37 tag primers. One important limitation of STM for in vivo screening is that the inoculated pool must have a dose such that each mutant is well represented and is able to establish an infection. To determine if this was the case with our pools, the input pools were randomly examined by real-time PCR, and the amplified DNA from the mutants in these pools gave similar quantitation curves, suggesting that the individual mutants were present at similar concentrations (data not shown). Mutants attenuated for survival in the animal model were scored as those whose genomic DNA did not reveal any PCR amplification after 28 cycles on the quantitation curve or, alternatively, had low or nonspecific melting peaks (peaks not at about 88°C) on the melting curve. According to these criteria, 260 of the 2,627 mutants that we assayed were putatively identified as having a survival defect in the animal model.
To eliminate potential false positives due to the large number of mutants in each pool, the 260 mutants obtained from the first screen were rearranged into 30 smaller pools containing 6 to 13 mutants per pool. Each pool was encased into agar beads and used to infect duplicate rats for 10 days as before. The output pools were reanalyzed by real-time PCR as for the first screen, except that this second screen was much more stringent. Only those mutants that did not show any amplification after 35 cycles in the quantitation curve and had no demonstrable peak in the melting curve were considered to pass this second screen. After the second screen, 102 mutants were identified, each containing transposon insertions in candidate genes required for the survival of B. cenocepacia in the rat agar bead model. Our finding that 102 out of 2,627 mutants (3.8%) were unable to survive in an in vivo model correlated well with figures obtained from previous STM experiments using other bacterial pathogens, which have detected a frequency of survival-defective mutants ranging from 0.2 to 13% (48). The survival-defective mutants did not appear to have growth defects in vitro, as all of them grew in nutrient broth at levels comparable with the wild-type K56-2 (data not shown).
The nucleotide sequence of the DNA flanking the site of transposon insertion obtained for each of the 102 attenuated mutants was used to search the GenBank databases for homologous genes. The results of this analysis are shown in Table Table2,2, located within a putative replication region. We speculate that these insertions may result in the instability of the 90-kb plasmid in vivo. It is possible that the instability or even loss of the plasmid in vivo may impair the survival of the plasmid-free bacterial cells by mechanisms involving killing or growth reduction of these cells (for a recent review, see reference 26). Further research is currently under way in our laboratory to explore this possibility. In addition, four survival-attenuated mutants had insertions within two rRNA clusters. While the role of rRNA clusters in virulence is not known at this time, others have found that five of the seven rRNA operons present in E. coli are necessary to support near-optimal growth on complex media, while all seven operons are necessary for rapid adaptation to nutrient and temperature changes (11). Therefore, it is conceivable that a similar situation occurs for B. cenocepacia, and adaptation to the environmental conditions found in vivo may require a full complement of rRNA operons.
We detected several mutants with insertions in putative regulatory genes, including members of the LysR family of transcriptional regulators, an H-NS-like protein, a TetR repressor, a regulator from the cold shock family, and two members of the sensor kinase two-component regulatory systems. It is conceivable that the survival-defective phenotype of these mutants may be due to the dysregulation of effector genes whose expression is controlled or modulated by these proteins rather than the lack of the regulator itself. To determine whether these regulatory genes indeed encode global regulators, we compared the profiles of cytosolic proteins of each mutant and the K56-2 parental isolate by two-dimensional gel electrophoresis (Fig. (Fig.5).5). The profiles of mutants 4H4 (predicted LysR transcriptional regulator homologue), 4G5 (predicted H-NS regulatory gene), 34A1 (σ54 homologue), 38C6 and 39H4 (cold shock family of transcriptional regulator), and 40H2 (Tet-repressor regulator homologue) were compared to that of the wild-type K56-2. The mutant 40H2, containing an insertion in the TetR family repressor homologue, revealed additional protein spots not observed in the parental K56-2 lysate, whereas mutants containing insertions in positive regulators lacked several protein spots that are present in the lysate from the parental K56-2 strain (Fig. (Fig.5).5). These results are consistent with the predicted roles for these regulatory proteins. Ongoing research is underway to confirm that these regulators are indeed involved in the different protein expression profiles observed, to characterize the regulated proteins from the proteome of K56-2, and to determine the components of each regulon.
We found 14 attenuated mutants with insertions compromising genes involved in transport functions, including a cation efflux protein (mutants 3A3 and 26H6), several permeases, and ABC transporters (Table (Table2).2). One other insertion was found just upstream of a gene encoding a predicted 96-kDa outer membrane heme-binding receptor protein (mutant 18F4). This is the first gene of a putative operon that may be involved in heme uptake. Heme-binding proteins of molecular masses ranging from 96 to 100 kDa have been recently described in many strains of the B. cepacia complex (60). Expression of heme-binding proteins leading to heme binding and accumulation is associated with virulence in several gram-negative pathogens (36). For example, in Porphyromonas gingivalis the accumulation of heme appears to protect bacterial cells against damage from hydrogen peroxide (59). Thus, an outer membrane heme-binding protein may contribute to in vivo survival by enabling B. cenocepacia to withstand oxidative stresses in inflammatory exudates in the lung.
Mutant 6E3 has an insertion interrupting a gene that encodes a homologue of the MgtC protein. This protein is required in S. enterica and Mycobacterium tuberculosis for growth under Mg2+-limiting environments, as well as for survival within macrophages and virulence in vivo (2, 5). Preliminary experiments indicate that, unlike the wild-type strain K56-2, the growth of this mutant is drastically reduced at Mg2+ concentrations of 25 μM or lower (K. Maloney and M. A. Valvano, unpublished data). Since B. cepacia complex strains can survive intracellularly within amoebae and macrophages (44, 56), we speculate that the MgtC protein homologue may be required for survival in vivo during lung infection in the rat model.
Additional mutants within this category had insertions in genes encoding enzymes required for the synthesis of O-antigen LPS (mutants 32D2, 33H3, 34D8, and 38C2). The isolation of genes involved in LPS synthesis for survival in vivo is not unexpected, as nearly all STM studies of gram-negative pathogens have reported similar findings (8, 17, 28, 33, 63). The detailed characterization of the O-antigen LPS biosynthesis cluster in K56-2 and J2315 strains will be described in detail elsewhere (X. Ortega, T. A. Hunt, and M. A. Valvano, manuscript in preparation).
The majority of the genes identified in our STM screen do not correspond to classical virulence factors proposed by others for B. cenocepacia. Given the large size of the B. cenocepacia genome (8.056 Mbp), it is possible that our screen has not uncovered all the genes required for in vivo survival. Also, it is important to point out that our screen was specifically designed to detect factors involved with in vivo survival and persistence in the lungs from our rat model. Early colonization steps are bypassed in our lung infection model, as bacteria instilled intratracheally are previously encased in agar beads. Thus, we did not expect to find mutants in genes involved with mucosal colonization. Also, STM screens do not usually detect extracellular factors that may be required for infection, since these factors may be provided by other mutants in the pool. The real-time PCR screening described in this study greatly reduced the time required for the screening step while increasing the sensitivity and specificity by allowing quantification of the PCR products produced. In addition, the plasposons and the bank of transposon mutants we generated will permit us to screen for bacterial genes required for survival in other in vivo models. An important limitation of STM screens is that the insertions may cause polar effects on the expression of downstream genes, which is especially true for those insertions in genes that are part of operons and those located upstream of coding regions, which may affect promoter elements or other regulatory elements. Therefore, our results should not be overinterpreted, as they provide a list of genes that are prime candidates as survival genes but that requires a gene-by-gene detailed analysis. We are currently investigating the survival of B. cenocepacia mutants in various infection models including alfalfa (1), amoebae (44), and the nematode Caenorhabditis elegans (35), which represent different habitats that can all be potentially encountered by B. cepacia complex isolates. Elucidation of the genes required for B. cepacia survival in these models will provide us with a more complete picture of requirements for infection of this opportunistic pathogen.
We thank J. Vicarioli, M. Visser, and K. Maloney for useful comments and technical assistance.
This work was supported by the Thompson Family Fund (to M.A.V.) from the Canadian Cystic Fibrosis Foundation and by the Special Program Grant Initiative “In Memory of Michael O'Reilly” funded by the Canadian Cystic Fibrosis Foundation and the Cardiovascular and Respiratory Health Institute of the Canadian Institutes of Health Research (to M.A.V. and P.A.S.). T.A.H. was supported by a fellowship from the Canadian Cystic Fibrosis Foundation. M.A.V. holds a Canada Research Chair in Infectious Diseases and Microbial Pathogenesis.
Editor: J. N. Weiser
University at Buffalo, South Campus
Buffalo, NY 14215