ABSTRACT
Burkholderia cenocepacia is an opportunistic pathogen that primarily infects cystic fibrosis (CF) patients. Previously, we reported that ShvR, a LysR regulator, influences colony morphology, virulence, and biofilm formation and regulates the expression of an adjacent 24-kb genomic region encoding 24 genes. In this study, we report the functional characterization of selected genes in this region. A Tn5 mutant with shiny colony morphology was identified with a polar mutation in BCAS0208, predicted to encode an acyl-coenzyme A dehydrogenase. Mutagenesis of BCAS0208 and complementation analyses revealed that BCAS0208 is required for rough colony morphology, biofilm formation, and virulence on alfalfa seedlings. It was not possible to complement with BCAS0208 containing a mutation in the catalytic site. BCAS0201, encoding a putative flavin adenine dinucleotide (FAD)-dependent oxidoreductase, and BCAS0207, encoding a putative citrate synthase, do not influence colony morphology but are required for optimum levels of biofilm formation and virulence. Both BCAS0208 and BCAS0201 contribute to pellicle formation, although individual mutations in each of these genes had no appreciable effect on pellicle formation. A mutant with a polar insertion in BCAS0208 was significantly less virulent in a rat model of chronic lung infection as well as in the alfalfa model. Genes in this region were shown to influence utilization of branched-chain fatty acids, tricarboxylic acid cycle substrates, l-arabinose, and branched-chain amino acids. Together, our data show that the ShvR-regulated genes BCAS0208 to BCAS0201 are required for the rough colony morphotype, biofilm and pellicle formation, and virulence in B. cenocepacia.
INTRODUCTION
Burkholderia cenocepacia is an opportunistic pathogen and one of the most common species of the B. cepacia complex (Bcc) isolated from the sputum of cystic fibrosis (CF) patients (37, 51). The Bcc consists of 17 closely related but genotypically distinct species which cause respiratory infections in individuals with CF or chronic granulomatous disease (38, 40, 56, 57). Bcc infections may be transient or chronic with a slow pulmonary decline or in some cases lead to bacteremic infections with a rapid decline in lung function, referred to as “cepacia syndrome,” often leading to death of affected individuals (38). Due to the intrinsic multiple antibiotic resistance of Bcc strains, treatment of these infections is problematic (47).
Previously, we have shown that B. cenocepacia K56-2 spontaneously undergoes conversion from a rough to a shiny morphotype (SHV) on agar medium after shaken or static incubation in liquid culture medium (4). Most SHV were avirulent in an alfalfa seedling infection model and demonstrated approximately 50% reduction in biofilm production. Selected SHV were analyzed for virulence in a chronic lung infection model and although able to establish a chronic infection, these infections resulted in significantly reduced lung histopathology compared to that of infections with the wild-type (WT) strain (4). SHV also varied in expression of other phenotypes, including N-acyl homoserine lactone (AHL) production, protease activity, motility, and production of an extracellular matrix (4). Transposon mutagenesis led to the identification of shvR (BCAS0225), which encodes an LysR-type transcriptional regulator (LTTR). Mutations in shvR resulted in colonies with a shiny morphotype, reduced biofilm production, and avirulence in alfalfa seedling infections. Several of the spontaneous SHV were found to contain point mutations in shvR (4).
Our laboratory has recently reported that ShvR, like other LTTRs, negatively autoregulates its own expression. It also negatively regulates expression of both the cepIR and cciIR quorum sensing (QS) systems, resulting in earlier detection of AHLs in culture medium and altered expression of many QS-regulated phenotypes (44). The shvR mutants have increased zinc metalloprotease B (zmpB) and type II secretion gene expression, which correlates with the increased protease activity. Transcriptome analysis revealed that ShvR influences the expression of over 1,000 genes in B. cenocepacia (44). Approximately 40% of these genes are coregulated by either the CepR or CciR QS transcriptional regulators; however, ShvR is required for the rough colony morphotype and independently regulates genes involved in biofilm formation. The expression of genes in an adjacent 24-kb region that contains the afc antifungal genes is ShvR dependent and was reduced up to 100-fold in the shvR mutant. This genomic region is organized in two putative divergently transcribed transcriptional units consisting of BCAS0224/BCAS0223 (afcCD) and BCAS0222 (afcA) to BCAS0201. ShvR was shown to positively regulate the expression of afcA and afcC promoter::lux fusions in a heterologous Escherichia coli host, demonstrating that ShvR directly regulates the expression of genes in this locus (44).
The function of most genes in the afc region is unknown, although mutations in afcA, afcB, and afcC in Burkholderia pyrrocinia BC11 (formerly Burkholderia cepacia BC11) drastically reduced production of an antifungal metabolite (31). In the present study, we report on the phenotypic characterization of the ShvR-regulated genes BCAS0208 to BCAS0201, located downstream of afcA, and their involvement in colony morphology, biofilm formation, and virulence. We show for the first time that BCAS0208, encoding an acyl-coenzyme A (acyl-CoA) dehydrogenase, is required for rough colony morphology in B. cenocepacia. BCAS0208 and BCAS0201, encoding a putative flavin adenine dinucleotide (FAD)-dependent oxidoreductase, contribute to biofilm formation and are required for virulence.
MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions.The bacterial strains and plasmids used in this study are listed in Table 1. Strains were routinely grown in LB broth or on 1.5% Lennox L agar (LB; Invitrogen, Burlington, Ontario, Canada), supplemented with antibiotics when appropriate, and incubated at 37°C unless specified otherwise. M9 medium (43) was routinely prepared by diluting 5× M9 minimal salts (BD Difco, Mississauga, Ontario, Canada) supplemented with 2 mM MgSO4, 1 mM CaCl2, and either 0.4% glucose or an alternative carbon source. M9 agar plates were prepared by the addition of 30 g per liter Bacto agar (BD Difco, Mississauga, Ontario, Canada). Antibiotics were used at the following concentrations: 100 μg/ml trimethoprim (Tp) and 300 μg/ml tetracycline (Tc) for B. cenocepacia and 100 μg/ml Tp, 50 μg/ml gentamicin (Gm), and 50 μg/ml kanamycin (Km) for E. coli. For biofilm assays, overnight cultures were grown in tryptic soy broth (BD Difco, Mississauga, Ontario, Canada). For animal experiments, cultures were grown for 16 h in a 10-ml volume of LB with shaking. For carbon source utilization, experiment cultures were starved by growing in M9 minimal medium containing 0.05% glucose with shaking for 20 h. All chemicals were purchased from Sigma-Aldrich Canada, Ltd. (Oakville, Ontario, Canada).
Bacterial strains, plasmids, and oligonucleotide primers used in this study
DNA manipulations.DNA fragments used for cloning were amplified by PCR and purified by using the QIAquick gel extraction kit (Qiagen, Mississauga, Ontario, Canada). Oligonucleotide primer synthesis and nucleotide sequencing were performed by the University of Calgary Core DNA Services. Sequence analysis and alignments were performed with the aid of NCBI (BLAST) (1), MultAlin (http://multalin.toulouse.inra.fr/multalin/) (17), and http://www.burkholderia.com/ (62).
Generation of pBBRS0208, pBBRS0207-S0205, pBBRS0204-S0202, and pBBRS0201 constructs.A 1,900-bp fragment containing the BCAS0208 gene, a 3,100-bp fragment containing BCAS0207 to BCAS0205, a 2,901-bp fragment containing BCAS0204 to BCAS0202, and a 1,600-bp fragment containing BCAS0201 were separately amplified from K56-2 genomic DNA. Primer pairs used for these amplification products were KpnIS0208F with XbaIS0208R, S0208ApaIF2 with S0204XbaIR2, S0205KpnIF2 with S0201XbaIR2, and KpnIS0201F with XbaIS0201R, respectively (primer sequences are available on request). PCR products were initially cloned into pCR2.1-TOPO and subsequently directionally cloned into pBBR1MCS-3 (32) using the appropriate restriction enzymes (KpnI and XbaI for BCAS0208, BCAS0204 to BCAS0202, and BCAS0201; ApaI and XbaI for BCAS0207 to BCAS0205) to generate pBBRS0208, pBBRS0207-S0205, pBBRS0204-S0202, and pBBRS0201 constructs. The expression of genes in each plasmid construct carried by the BCAS0208::Tn (transposon) strain was verified by reverse transcriptase PCR (RT-PCR).
Mobilization of plasmids into B. cenocepacia strains.All pBBR1MCS-3 plasmid constructs were mobilized from E. coli DH5α to B. cenocepacia by triparental mating using pRK2013 as the helper plasmid (22). Transconjugants were selected on half-strength Pseudomonas isolation agar (BD Difco, Mississauga, Ontario, Canada) plus LB agar (Invitrogen, Burlington, Ontario, Canada) containing 300 μg/ml Tc. Plasmid presence was verified by PCR using primers for pBBR1MCS-3 and a gene-specific primer.
Construction of BCAS0207, BCAS0204, and BCAS0201 insertion mutants.The genes BCAS0207, BCAS0204, and BCAS0201 in K56-2 were mutated by insertional inactivation using pGSVTp-lux as previously described (4). Briefly, a 529-bp, a 509-bp, and a 568-bp internal region of BCAS0207, BCAS0201, and BCAS0204, respectively, were amplified using the primers EcoRIBCAS0207SF with EcoRIBCAS0207SR, BCAS0201F with BCAS0201R, and BCAS0204F with BCAS0204R (primer sequences are available on request). The amplification products were cloned in pCR2.1-TOPO and subsequently cloned into pGSVTp-lux to generate pGSVTplux-BCAS0207, pGSVTplux-BCAS0201, and pGSVTplux-BCAS0204, respectively. The pGSVTp-lux constructs were mobilized into K56-2, and transconjugants were selected on LB agar plates containing 100 μg/ml Tp and 50 μg/ml Gm. Insertions at the correct loci and orientation were verified by PCR using luxCR and gene-specific primers.
Construction of Δ1BCAS0208 and Δ2BCAS0208 mutants.Two independent unmarked BCAS0208 mutants containing internal deletions of 1.6 kb and 1.2 kb were constructed as described by Flannagan et al. (24), with slight modifications. Briefly, pTopoBCAS0208 was used as the template for deletion PCR. Mutagenic primers, overdS0208F with overdS0208R and overdS0208F1 with overdS0208R1, which have overlapping partially complementary sequences flanking the region of BCAS0208 to be deleted, were used to amplify pTopoBCAS0208, followed by digestion with DpnI to remove the template plasmid. The amplified products transformed into DH5α, giving rise to pTopoBCAS0208Δ1 and pTopoBCAS0208Δ2 constructs. pTopoBCAS0208Δ1 and pTopoBCAS0208Δ2 were digested with KpnI and XbaI. Fragments of approximately 200 bp and 600 bp obtained by digestion were cloned into pGPI-SceI, giving rise to mutagenesis plasmids pGPIS0208Δ1 and pGPIS0208Δ2, which were separately introduced into K56-2 by conjugation. The plasmid pDAI-SceI was introduced by conjugation to obtain the double-crossover event. The pDAI-SceI plasmid was resolved by curing the exconjugants by passage in LB broth. All constructs and mutants were confirmed by PCR analysis.
Generation of the BCAS0208 catalytic mutant.A catalytic mutant of BCAS0208 in pBBR1MCS-3 was constructed as described by Edelheit et al. (21). Complementary primers S0208g1336cF and S0208g1336cR were designed using QuickChange primer design software (Agilent Technologies). Briefly, pTopoBCAS0208 was used for two separate PCRs with the two mutagenic primers, and the products were then mixed, denatured, and subjected to slow annealing before the methylated nonmutated parental strands were digested with DpnI. The mutated plasmid was transformed into DH5α to generate pTopoBCAS0208E446Q. This plasmid was digested with KpnI and XbaI, and the excised fragment was inserted into pBBR1MCS-3, giving rise to pBBRS0208E446Q. The expression of BCAS0208 from the construct pBBRS0208E446Q in the Δ1BCAS0208 mutant was verified by RT-PCR.
RNA isolation, RT-PCR, and qRT-PCRs.RNA was prepared using the RiboPure bacterial RNA isolation kit (Ambion, Streetsville, Ontario, Canada) and treated with DNase I as previously described (41). Quantitative RT-PCR (qRT-PCR) was performed in triplicate using the gene coding for NADH dehydrogenase, ndh (BCAM0166), as an internal control for expression levels, and quantification was carried out as described previously (52).The expression of ndh has previously been shown not to be influenced by shvR or cepR mutations (44, 52). For RT-PCR, cDNA was synthesized by using the iScript Select cDNA synthesis kit (Bio-Rad) and subjected to PCR using Platinum Taq DNA polymerase and the primers listed in Table 1. The conditions for the amplification were 1 cycle of 3 min at 95°C, followed by 35 cycles of 30 s at 95°C, 30 s at 55°C, and 30 s at 72°C and a final extension of 10 min at 72°C.
Congo red binding assays.Cultures grown in 96-well microtiter plates were inoculated manually (2 μl) or with a 48-pin replicator onto LB agar plates containing Congo red dye at 0.01% (wt/vol). Plates were incubated at 37°C for 48 h, and the binding properties were determined by the color of the resulting colony. Red was indicative of high affinity, and pink and/or white was indicative of a decrease in binding affinity.
Biofilms on abiotic surfaces.Biofilm assays on abiotic surfaces were performed as previously described (6) by growing bacteria in 96-well microtiter plates (Nunc, Roskilde, Denmark) covered with a 96-peg lid (Nunc) and quantitating the biomass formed on the polystyrene pegs. Tc was included in the medium for complementation experiments. Biofilm formation was measured as A600 of cellular biomass stained with crystal violet normalized to culture optical density at 600 nm (OD600). Each strain was inoculated into at least 3 wells, and the values shown in Fig. 2 and 3 represent the means ± standard deviations of the replicates from a single assay. Each experiment was repeated at least 3 times with similar results.
Pellicle formation.Strains were inoculated into borosilicate glass tubes (16 by 150 mm) containing 5 ml LB to an initial OD600 of 0.05 and incubated statically at 37°C for 5 days. The pellicle at the air-liquid interface was qualitatively evaluated. Each strain was inoculated in triplicate, and the experiment was repeated twice with similar results.
Microbial adhesion to solvents.This experiment was carried out as described by Chavant et al. (13) with minor modifications. Briefly, bacteria were grown in LB to stationary phase, harvested by centrifugation at 4°C, and resuspended in 0.15 M NaCl. The OD600 was adjusted to 0.80, and a 1-ml sample was taken (sample A0). A total of 2.4 ml of the microbial solution was vortexed for 1 min with 0.4 ml of solvent (either chloroform or hexadecane), and the mixture was allowed to stand for 15 min to completely separate the two phases. Another 1-ml sample was carefully taken from the aqueous phase (sample A). The turbidities of samples were determined at 600 nm. The percentage of bacterial cells present in each solvent was calculated: percent affinity = 100 × [1 − (A/A0)]. Each strain was tested at least three times in triplicate, and results are presented from one representative experiment.
Alfalfa infection assay.Alfalfa seedling infections were carried out as previously described (4). Each strain was inoculated onto at least nine seedlings in at least two assays with similar results, and the results presented are from one assay.
Animal studies.Male Sprague-Dawley rats (150 to 175 g; Charles River Canada, Inc.) were tracheotomized under anesthesia and inoculated with approximately 107 CFU of the appropriate strain embedded in agar beads as previously described (10). At 7 days postinfection (p.i.), the lungs were harvested and lungs from four to five animals per group were used for either quantitative bacteriology or histopathology analysis as previously described (5). Infiltration of hematoxylin-and-eosin-stained lung sections with inflammatory cells and exudates was quantitated using Image Pro Plus (Media Cybernetics, Bethesda, MD).
Phenotype microarray assays using Biolog plates and growth assays in minimal medium.Carbon source utilization was performed using phenotype MicroArray plates 1 and 2A (Biolog, Hayward, CA) containing 190 different carbon sources. Cultures were grown overnight at 37°C on Biolog universal growth (BUG) agar plates, inoculated into the plates according to the manufacturer's instructions, and incubated at 37°C without shaking. The increase in purple color at 600 nm was measured every 24 h using a 96-well plate reader (Wallac Victor multilabel reader) for up to 72 h. Area graphs were plotted for each strain with OD600 values (OD600 of each well − OD600 of negative control) and used to calculate their average metabolic response compared to that of K56-2. Selected Biolog growth phenotypes were confirmed by assessing growth on M9 minimal medium agar and liquid supplemented with 0.4% of the carbon source of interest. The inocula for these assays were prepared by growing cultures in low-glucose (0.05%) M9 medium. Cultures were normalized to an OD600 of 0.4, and 2 μl of 10-fold serial dilutions of the inocula was spotted onto the plates and incubated for 72 h at 37°C. For liquid assays, strains were inoculated at an initial OD600 of 0.05 and growth was monitored for 72 h. Each experiment was carried out in duplicate.
Statistical analysis.Analysis of variance (ANOVA) and t tests were performed with GraphPad Prism software (GraphPad Software, San Diego, CA). A P value of <0.05 was considered statistically significant.
RESULTS
A genomic region adjacent to shvR is involved in colony morphology.A Tn-SHV mutant, which was not rhamnose inducible, was identified while a pool of transposon mutants generated using the pSCrhaBout transposon mutagenesis approach described by Bernier et al. (4) was being screened. The sequence flanking the transposon insertion was determined, and comparison to the B. cenocepacia J2315 genome sequence (http://www.burkholderia.com) revealed that the Tn insertion was located within BCAS0208, one of the ShvR-dependent genes in the adjacent afc genomic region (BCAS0208::Tn strain) (Fig. 1A). BCAS0208 is located within a 22-gene putative operon downstream of afcA that includes genes whose putative products are involved in energy production and conversion, lipid metabolism, and amino acid transport and metabolism (http://www.burkholderia.com; DOOR [Database for Prokaryotic Operons]) (19, 30).
Analysis of the BCAS208 to BCAS201 genomic region. (A) Schematic diagram of genes adjacent to shvR highlighting the open reading frames (ORFs) from BCAS0209 to BCAS0201. Black inverted triangles represent locations of transposon insertions in K56-2 shvR::Tn and BCAS0208::Tn mutants. Horizontal lines indicate the BCAS0207 to BCAS205 and the BCAS0204 to BCAS0202 fragments cloned in pBBR1MCS-3. ORFs downstream of BCAS0209 are indicated as arrows with different patterns to represent the various pBBR1MCS-3 constructs. Mutant strains are listed below, and the expression statuses of genes as determined by RT-PCR are indicated by plus or minus signs. Expression of three upstream genes, BCAS0215, BCAS0220, and BCAS0225 (shvR), was also confirmed in each mutant. (B) Predicted functions assigned to products encoded by genes BCAS0208 to BCAS0201.
The predicted protein products of genes encoding BCAS0208 to BCAS0201 are shown in Fig. 1B. BCAS0208 is predicted to code for an acyl-CoA dehydrogenase and shows 84 to 99% similarity to orthologs in other Burkholderia spp. and 47% similarity to fadE5 of Mycobacterium tuberculosis H37Rv. BCAS0207, BCAS0206, and BCAS0205 are predicted to encode a citrate synthase, a putative methyltransferase, and a taurine catabolism dioxygenase protein, respectively. BCAS0204, BCAS0203, and BCAS0202 encode products that exhibit similarity to an ABC transport system. BCAS0201 encodes a putative FAD-dependent oxidoreductase. Genes BCAS0208 to BCAS0201 are conserved in the sequenced genomes of Burkholderia cenocepacia AU1054, HI2424, and MC0-3 as well as in Burkholderia ambifaria AMMD. However, orthologs of genes in this region were not identified in genome sequences of Burkholderia vietnamiensis G4, Burkholderia multivorans ATCC17616, or the non-Bcc species Burkholderia thailandensis E264, Burkholderia xenovorans LB400, and Burkholderia phymatum STM815. The genes BCAS0208 to BCAS0202 are also present in Burkholderia mallei and Burkholderia pseudomallei; however, there is no BCAS0201ortholog adjacent to these genes.
Since the Tn insertion in BCAS0208 likely had polar effects on downstream genes, we focused our study on genes BCAS0208 to BCAS0201 in order to define their contribution to colony morphology. When BCAS0208 was provided in trans, the shiny colony morphology of the BCAS0208::Tn strain was converted to an intermediate rough morphotype (Table 2). BCAS0208 in trans, however, did not complement the SHV morphotype of an shvR mutant (Table 2).
Effect of BCAS0208-S0201 on colony morphology, Congo red binding, and alfalfa virulence
Since the majority of SHV and shvR mutants have reduced biofilm formation, the biofilm phenotype of the BCAS0208::Tn strain was determined and compared to that of the WT and shvR::Tn strains. Evaluation of normalized biomass revealed that the BCAS0208::Tn (pBBR1MCS-3) strain exhibited a 90% reduction in biofilm formation compared to that of K56-2 (pBBR1MCS-3) (P < 0.001, one-way ANOVA; Fig. 2A) and was not significantly different from that of the shvR::Tn (pBBR1MCS-3) strain. Biofilm formation was partially restored to the BCAS0208::Tn strain carrying pBBRS0208 (P < 0.001, one-way ANOVA; Fig. 2A). The levels of biofilm formation by the BCAS0208::Tn (pBBRS0208) strain were still only 30% of that of K56-2 (pBBR1MCS-3) (P < 0.001, one-way ANOVA) but at least 3-fold higher than the biofilm levels of the shvR::Tn (pBBRS0208) strain (P < 0.001, one-way ANOVA).
Biofilm formation by the BCAS0208::Tn, BCAS0201::lux, BCAS0204::lux, and BCAS0207::lux mutants. (A) Complementation of biofilm formation in the BCAS0208::Tn mutant by pBBRS0208. (B) Comparison of biofilm formation in K56-2 and the BCAS0208::Tn and shvR::Tn mutants in the presence of pBBRS0207-S0205 and pBBRS0204-S0202. (C) Comparison of biofilm formation in K56-2 and the BCAS0208::Tn and shvR::Tn mutants in the presence of pBBRS0201. (D) Biofilm formation by the BCAS0201::lux and BCAS0207::lux mutants. The results are the means of normalized biomass ± the standard deviations from three replicate cultures and are representative of at least three independent trials. Asterisks indicate that the values are significantly different between strains compared (***, P < 0.001; **, P = 0.001 to 0.01; *, P = 0.01 to 0.05; one-way ANOVA with Bonferroni multiple comparisons posttest).
As BCAS0208 was predicted to be part of an operon with the downstream genes BCAS0207 to BCAS0201 and BCAS0208 was able to only partially restore biofilm formation and the rough colony morphotype to the BCAS0208::Tn strain, we determined the effect of the Tn insertion on expression of both up- and downstream genes. qRT-PCR was carried out with primers specific for BCAS0208, BCAS0204, BCAS0201, BCAS0220, BCAS0215, and shvR, and RT-PCR was performed with primers specific for BCAS0207, BCAS0206, BCAS0205, BCAS0203, and BCAS0202. The expression of BCAS0208 and all downstream genes was drastically reduced in the BCAS0208::Tn strain compared to that of the WT (data not shown), whereas the expression of the upstream genes BCAS0220, BCAS0215, and shvR was unaffected. Therefore, the transposon insertion in BCAS0208 has a polar effect on downstream gene expression (Fig. 1A; data not shown).
To investigate the possibility that the genes downstream of BCAS0208 may also be involved in determination of rough colony morphology and biofilm formation, complementation experiments with various downstream genes were carried out to assess their ability to restore these phenotypes to the BCAS0208::Tn strain. Since the predicted products encoded by genes BCAS0207 (citrate synthase), BCAS0206 (methyl transferase), and BCAS0205 (taurine catabolism dioxygenase) showed similarity to proteins involved in metabolic pathways, these three genes were cloned together to generate the construct pBBRS0207-S0205. Similarly, products encoded by genes BCAS0204 (ATP-binding protein), BCAS0203 (ABC transporter protein), and BCAS0202 (permease) appeared to be involved in ABC transporter functions. These three genes were cloned as a single fragment to generate the construct pBBRS0204-S0202. BCAS0201 encoding a putative FAD-dependent oxidoreductase was cloned separately. Complementation experiments with the BCAS0208::Tn strain and pBBRS0207-S0205, pBBRS0204-S0202, or pBBRS0201 in trans failed to restore either colony morphology or biofilm formation, although the expression of each gene in these constructs was confirmed in the BCAS0208::Tn strain by RT-PCR. The levels of biofilm formation for the BCAS0208::Tn strain carrying these plasmids were significantly lower than those of K56-2 (one-way ANOVA, P < 0.0001) (Fig. 2B and C and Table 2).
The shvR::Tn (pBBR1MCS-3) strain formed only 10% of the biomass of K56-2 (pBBR1MCS-3) (P < 0.001, one-way ANOVA) (Fig. 2A). Comparison of biofilm formation levels of the shvR::Tn (pBBR1MCS-3) mutant with the shvR::Tn (pBBRS0208), shvR::Tn (pBBRS0207-S0205), shvR::Tn (pBBRS0204-S0202), and shvR::Tn (pBBRS0201) mutants revealed that they were not significantly different, indicating that none of these genes could restore biofilm levels in an shvR mutant (Fig. 2A to C and Table 2). Similarly, the shvR::Tn strain could not be restored to the rough morphotype in the presence of any of these genes (Table 2). This suggests involvement of additional shvR-regulated genes in both biofilm formation and the rough colony morphotype.
Taken together, these experiments indicate a role for BCAS0208 in both colony morphology and biofilm formation. The lack of full restoration of the rough colony morphology and failure to obtain maximum levels of biofilm formation with the BCAS0208::Tn strain in the presence of BCAS0208 alone also suggest contribution by one or more downstream genes.
BCAS0207 and BCAS0201 are required for optimum biofilm formation but do not influence colony morphology.To determine the potential contribution of genes downstream of BCAS0208 to biofilm formation, pGSVTplux insertional mutations were constructed in BCAS0207, BCAS0204, and BCAS0201 to create the BCAS0207::lux, BCAS0204::lux, and BCAS0201::lux mutants. RT-PCR using RNA from cultures of the BCAS0207::lux and BCAS0204::lux mutants was carried out with primers specific for BCAS0206, BCAS0205, BCAS0204, BCAS0203, BCAS0202, and BCAS0201. In both the BCAS0207::lux and BCAS0204::lux mutants, all downstream genes were found to be expressed, indicating that the mutations were not polar (Fig. 1A, data not shown). Insertional mutations in each of these three genes resulted in strains with the WT rough morphotype (Table 2). The BCAS0207::lux and BCAS0201::lux mutants had 25% and 29% biofilm formation compared to that of K56-2 (one- way ANOVA, P < 0.001), which could be restored by pBBSRS0207-S0205 and pBBRS0201, respectively, in trans (Fig. 2D). The biofilm formation of the BCAS0204::lux mutant was generally lower but not significantly different from that of K56-2 (data not shown).
BCAS0208 is required for rough colony morphology and biofilm formation.Since the insertion in the BCAS0208::Tn strain was polar on downstream genes, we attempted to generate nonpolar unmarked deletion mutants of BCAS0208. Two constructs were made; the first one contained a 1.6-kb in-frame deletion (pGPIS0208Δ1), while the second construct contained a 1.2-kb in-frame deletion within BCAS0208 (pGPIS0208Δ2). The BCAS0208 mutants generated were designated the Δ1BCAS0208 and Δ2BCAS0208 mutants. qRT or RT-PCR analysis of RNA isolated from the Δ1BCAS0208 and Δ2BCAS0208 mutants using primers specific for the upstream genes BCAS0220 and BCAS0215, and each gene from BCAS0208 to BCAS0201, was performed to determine if these genes were expressed. The Δ2BCAS0208 mutant expressed all genes at a level similar to that of the WT, whereas the expression of BCAS0208 to BCAS0201 was at least 1,000-fold downregulated in the Δ1BCAS0208 mutant, which was similar to that observed for the BCAS0208::Tn strain. Therefore, the 1.6-kb deletion in the Δ1BCAS0208 mutant resulted in a polar mutation, while the 1.2-kb deletion in the Δ2BCAS0208 mutant was nonpolar (data not shown). Both the Δ1BCAS0208 and Δ2BCAS0208 mutant colonies appeared shiny and had a biofilm phenotype similar to the BCAS0208::Tn strain (Table 2). The Δ1BCAS0208 (pBBRS0208) mutant had an intermediate rough morphotype similar to that of the BCAS0208::Tn (pBBRS0208) strain (Fig. 3B and Table 2). On the other hand, the Δ2BCAS0208 (pBBRS0208) mutant had a rough WT morphotype (Fig. 3B and Table 2). Biofilm formation was not significantly increased in the Δ1BCAS0208 (pBBRS0208) mutant but was significantly increased in the Δ2BCAS0208 (pBBRS0208) mutant (one-way ANOVA, P < 0.0001; Fig. 3C). These results indicate that BCAS0208 is crucial for the shiny-to-rough-morphotype conversion as well as biofilm formation in K56-2.
The catalytic site of BCAS0208 is required for morphotype, biofilms, and virulence. (A) Schematic representation of putative polypeptide of BCAS0208 showing PFAM-predicted domains (top) and alignment of the C-terminal region of BCAS0208 with ACAD-fadE5 family members (below); 1, acyl-CoA dehydrogenase (ACDH) N-terminal; 2, ACDH middle domain; 3, ACDH C-terminal domain. E-446 (glutamate at 446) is one of the amino acids required for catalytic activity of ACDHs. The ORFs used for alignment with BCAS0208 are Rsc1762 (NP_519883), Rv0244c (NP_214758), SCO3800 (NP_733617), and Mlr5624 (NP_106251). Effect of pBBRS0208 and pBBRS0208E446Q on colony morphology and Congo red binding (B), biofilm formation (C), and alfalfa virulence (D) of K56-2 and the Δ1BCAS0208 and Δ2BCAS0208 mutants. Values shown are the means ± the standard deviations from three replicate cultures and are representative of values obtained from at least three independent trials. Asterisks indicate that the values are significantly different between strains compared (***, P < 0.001, one-way ANOVA with Bonferroni multiple comparisons posttest). Disease symptoms on alfalfa seedlings were noted as yellowing or browning of leaves associated with necrosis 5 days p.i.
A catalytically active BCAS0208 is important for determining colony morphology and biofilm formation.BCAS0208 is similar to putative acyl-CoA dehydrogenases containing the fadE5 conserved domain (E value of 2.73e−70, ACAD_fadE5; cd01153). Analysis of BCAS0208 using PFAM revealed that it contains acyl-CoA dehydrogenase N-terminal, middle, and C-terminal domains (23) (Fig. 3A). The conserved residues for FAD and substrate binding are present in the middle domain, whereas the conserved catalytic residues are present in the C-terminal domain. Multiple sequence alignment of the conserved regions of representative proteins of the ACAD_fadE5 subfamily with BCAS0208 showed the presence of the highly conserved catalytic residues glutamate and glycine at positions 446 and 447 (Glu-Gly class of acyl-CoA dehydrogenase) (Fig. 3A). We constructed a catalytic variant of BCAS0208 with a g1336c mutation which converts glutamate at position 446 to glutamine (designated BCAS0208E446Q). To determine whether the catalytic activity of BCAS0208 is important for its role in colony morphology and biofilm formation, complementation experiments were carried out in the polar Δ1BCAS0208 and nonpolar Δ2BCAS0208 mutants by providing pBBRS0208E446Q in trans. Both the Δ1BCAS0208 and Δ2BCAS0208 mutant colonies remained shiny and defective for biofilm formation in the presence of pBBRS0208E446Q (Fig. 3B and C). The lack of complementation of the nonpolar Δ2BCAS0208 mutant suggests that a catalytically active BCAS0208 is required for the rough colony morphotype and maximum biofilm formation.
Genes other than BCAS0208 may influence colony morphotype.Previously, we reported that mutations in addition to those in shvR may lead to the SHV morphotype. In order to determine whether mutations in BCAS0208 were responsible for the shiny morphotype in spontaneous SHV previously described that did not contain shvR mutations (4), pBBRS0208 was introduced into K56-S15, -S86, and -S92 mutants to determine if the WT rough morphotype was restored. Partial complementation was observed in the S15 mutant but not in the S86 or S92 mutants. BCAS0208 was sequenced from the S15 mutant and determined to be identical to that in K56-2, indicating that mutations in neither shvR nor BCAS0208 could explain the shiny phenotype in these SHV. We did not confirm the expression level of BCAS0208 in these SHV. It is possible that a mutation affecting S0208 expression or mutations in genes in addition to shvR and BCAS0208 may lead to an SHV morphotype.
Both BCAS0208 and BCAS0201 contribute to pellicle formation and cell surface hydrophobicity.Since BCAS0208 (shiny) and BCAS0201 (rough) mutants were deficient for biofilm formation and previously several SHV were observed to have decreased pellicle formation, these mutants were assessed for their ability to form a pellicle at an air-liquid interface. The Δ2BCAS0208 (nonpolar) and BCAS0201::lux mutants were not appreciably affected for pellicle formation. However, the polar BCAS0208::Tn strain and Δ1BCAS0208 mutant were drastically decreased for pellicle formation (Fig. 4 and data not shown). Complementation with either BCAS0208 or BCAS0201 in trans increased pellicle formation by these mutants, suggesting that both genes contribute to this phenotype (Fig. 4). Complementation of the shvR::Tn mutant with either BCAS0208 or BCAS0201 in trans marginally increased pellicle formation but to a lesser extent than observed for the polar BCAS0208 mutants.
Pellicle formation by Δ1BCAS0208, Δ2BCAS0208, BCAS0201::lux, and shvR::Tn mutants. Tubes showing cultures of K56-2 (A) and the Δ2BCAS0208 (B), BCAS0201::lux (C), BCAS0208::Tn (D), and shvR::Tn (E) mutants after 5 days of static incubation. For each strain: a, pBBR1MCS-3; b, pBBRS0208; c, pBBRS0201.
To determine if changes in cell surface hydrophobicity might influence pellicle and biofilm formation, partitioning experiments were performed in polar and apolar solvents. The K56-2, shvR::Tn, BCAS0208::Tn, Δ2BCAS0208, and BCAS0201::lux strains with either pBBR1MCS-3, pBBRS0208, or pBBRS0201 in trans were tested for their affinity to chloroform and hexadecane, and all strains showed a higher affinity to chloroform (55 to 65%; data not shown) than to hexadecane (Fig. 5), indicating that the surfaces of these strains were more hydrophilic. The affinity of the shvR::Tn (pBBR1MCS-3) and BCAS0208::Tn (pBBR1MCS-3) strains to hexadecane was not significantly different from that of the K56-2 (pBBR1MCS-3) strain; however, the Δ2BCAS0208 (pBBR1MCS-3) mutant showed significantly lower affinity for hexadecane (Fig. 5A; P < 0.01) than K56-2. The presence of pBBRS0208 in trans resulted in significant increases in affinity to hexadecane for both the shvR::Tn and Δ2BCAS0208 mutants (Fig. 5A; P < 0.01). Similarly, pBBRS0201 in trans resulted in increased affinity to hexadecane for both the K56-2 and shvR::Tn strains compared to that of the K56-2 (pBBR1MCS-3) and shvR::Tn (pBBR1MCS-3) strains (Fig. 5B; P < 0.01); however, pBBRS0201 in trans did not significantly alter the affinity of the BCAS0201::lux strain to hexadecane (Fig. 5B). These results indicate that the presence of pBBRS0208 or pBBRS0201 in trans may lead to changes in cell surface properties for some strains, and this might contribute to alterations in pellicle formation.
Microbial adhesion to hexadecane for K56-2 and shvR::Tn, BCAS0208::Tn, Δ2BCAS0208, and BCAS0201::lux mutants. Values are expressed as percent affinity with pBBRS0208 in trans (A) and with pBBRS0201 in trans (B). Values shown are the means ± the standard deviations from three replicate cultures and are representative of values obtained from at least three independent trials. Asterisks indicate that the values are significantly different between strains compared (***, P < 0.001; **, P < 0.01; one-way ANOVA with Bonferroni multiple comparisons posttest).
Effect of mutations in BCAS0208, BCAS0207, BCAS0204, and BCAS0201 on virulence.Previously, SHV were demonstrated to be less virulent in a chronic lung infection model, and most SHV, including shvR mutants, were avirulent in an alfalfa seedling infection model (4). To determine if BCAS0208 and the genes downstream influenced virulence, K56-2 and the polar Δ1BCAS0208 mutant were compared for their ability to cause persistent infections and lung histopathology using the rat agar bead model of infection (10). Although three of five animals infected with the Δ1BCAS0208 mutant had bacterial counts recovered from lungs of infected animals 7 days postinfection that were 4 logs less than those from K56-2-infected animals, there was no significant difference in bacterial counts between animals infected with this mutant and K56-2-infected animals (Fig. 6 A). Quantitative lung histopathology analysis, however, revealed significantly less inflammation in Δ1BCAS0208 mutant-infected lungs compared to that in K56-2-infected lungs (Student's t test, P < 0.05; Fig. 6B). The level of inflammation was similar to that observed in animals inoculated with sterile beads. These results indicate that a mutation affecting BCAS0208 and/or genes downstream of it results in reduced pathological changes in infected lungs compared to K56-2-infected lungs, although bacterial persistence is not significantly reduced.
Virulence of the Δ1BCAS0208 mutant in rats. (A) Scatter plot of the K56-2 and Δ1BCAS0208 mutant CFU recovered 7 days p.i. from lungs of three or five animals per group. (B) Scatter plot of the percentage of the lung with inflammatory exudates at 7 days p.i. detected in hematoxylin-and-eosin-stained lung sections from four animals per bacterial strain. The K56-2 Δ1BCAS0208 mutant was significantly different from K56-2 (P < 0.05; Student's t test).
To determine which genes in this region contributed to virulence, we used the alfalfa seedling model of infection (5) in order to test more mutants and perform complementation analyses. Both the BCAS0208::Tn and the polar Δ1BCAS0208 mutants were avirulent on alfalfa seedlings and failed to regain virulence when BCAS0208 was provided in trans (Table 2). The introduction of pBBRS0207-S0205, pBBRS0204-S0202, or pBBRS0201 in trans to the BCAS0208::Tn and Δ1BCAS0208 strains also did not restore virulence to these mutants. These observations indicated that more than one gene or combinations of genes in the region may be involved. To investigate this possibility, virulence assays were carried out with the BCAS0201, BCAS0204, BCAS0207, and BCAS0208 (nonpolar) mutants. Both the BCAS0201::lux and Δ2BCAS0208 (nonpolar) mutants were avirulent in the alfalfa infection model. The BCAS0207::lux strain was partially virulence deficient, while the BCAS0204::lux mutant was as virulent as the WT. Complementation of the BCAS0201::lux and Δ2BCAS0208 strains with pBBRS0201 and pBBRS0208, respectively, restored full virulence (Fig. 3D, Table 2). The partially virulence-deficient BCAS0207::lux strain was restored to WT levels by providing pBBRS0207-S0205 in trans (Table 2). To determine whether the catalytic activity of BCAS0208 was important for virulence, alfalfa infections were carried out with the K56-2, Δ1BCAS0208, and Δ2BCAS0208 strains containing BCAS0208E446Q in trans. The avirulent phenotype exhibited by the nonpolar Δ2BCAS0208 mutant could not be restored with pBBRS0208E446Q, indicating the requirement for a catalytically functional BCAS0208 in virulence (Fig. 3D).
Metabolic phenotypes of BCAS0208, BCAS0207, BCAS0204, and BCAS0201 mutants.Based on annotation, it was predicted that the product of BCAS0208 might function in fatty acid metabolism. Fatty acids are important constituents of bacterial membranes, and perturbation in their metabolism might explain the colony morphotype alteration of BCAS0208 mutants. While it was difficult to predict the pathway for the BCAS0201-encoded protein, it is likely that the product of BCAS0207 might be involved in intermediary metabolism. The substrate transported by the predicted product of BCAS0204 was unknown, although it is annotated as a cobalt transporter. Therefore, metabolic responses of BCAS0208, BCAS0204, and BCAS0201 mutants were investigated in the presence of various carbon sources, including amino acids, fatty acids, sugars, and glycoconjugates, using Biolog PM1 and PM2A plates. The substrates for which growth differences were observed for the Δ2BCAS0208, BCAS0204::lux, and BCAS0201::lux strains in the Biolog assay compared to the K56-2 strain are shown in Table 3. The Δ2BCAS0208 mutant exhibited lower levels of utilization of α-ketobutyric acid and hydroxybutyric acid, although it did not show any difference on butyric acid, capric acid, or caproic acid. Both the BCAS0201::lux and BCAS0204::lux mutants were less able to utilize all three branched-chain amino acids (BCAAs): valine, leucine, and isoleucine. The BCAS0204::lux mutant showed a drastic defect in utilization of l-arabinose as the sole carbon source, although it grew in d-arabinose. All three mutants also showed a lower level of utilization of α-ketoglutarate, one of the intermediates of the tricarboxylic acid (TCA) cycle.
Growth phenotypes on Biolog plates
To confirm these phenotypes, growth assays were performed on M9 agar plates and liquid containing selected substrates as the sole carbon source. Since in the Biolog assay the Δ2BCAS0208 mutant was less efficient in the utilization of α-ketobutyric acid and hydroxybutyric acid, growth assays were carried out with these and other branched-chain fatty acids, isobutyric acid and isovaleric acid. It was not possible to demonstrate a growth defect in M9 agar or in liquid with α-ketobutyric acid, isobutyric acid, or hydroxybutyric acid as carbon sources with the Δ2BCAS0208 mutant compared to growth of K56-2; however, the Δ2BCAS0208 mutant was severely affected for growth in isovaleric acid both on M9 agar and liquid medium. This growth defect, however, could not be corrected by providing BCAS0208 in trans (Table 4). The BCAS0204::lux strain was tested on l-arabinose-supplemented minimal medium since it showed a defect on this substrate in the Biolog plates. The BCAS0204::lux strain grew as well as K56-2 on M9 agar plates containing glucose but showed no growth in the presence of l-arabinose. The growth defect of the BCAS0204::lux (pBBR1MCS-3) mutant was partially restored in the presence of pBBRS0204-S0202 (Table 4). In M9 liquid medium, the BCAS0204::lux (pBBR1MCS-3) mutant did not grow in the presence of l-arabinose, and this defect was not corrected in the BCAS0204::lux (pBBRS0204-S0202) strain. The growth of the BCAS0207::lux strain was assessed separately on M9 plates containing various citric acid cycle intermediates as substrates. The BCAS0207::lux strain had a growth defect in the presence of glucose, acetate, citrate, succinate, and fumarate both in M9 agar plates and in liquid, which could be partially complemented in the BCAS0207::lux (pBBRS0207-S0205) mutant (Table 4). These results suggest a role for BCAS0207 in the TCA cycle and for BCAS0204 in l-arabinose uptake. Our observations also suggest a role for BCAS0208 in branched-chain fatty acid metabolism and BCAS0201 in BCAA metabolism.
Complementation of growth defect in M9 liquid and M9 agar
DISCUSSION
B. cenocepacia SHV mutants were previously reported to arise due to mutations in an LysR regulator, ShvR (4). In this study, we demonstrated that the shvR-regulated BCAS0208 to BCAS0201 genes are major determinants influencing colony morphology, biofilm formation, and virulence in this species. We previously showed that ShvR was important for virulence in chronic lung and alfalfa seedling models of infection (4), although an shvR mutation did not affect virulence in invertebrate infection models (54). This study indicates that the BCAS0208 to BCAS0201 genes are some of the primary genes responsible for the shvR mutant virulence phenotype.
Other Bcc colony morphotype variants have been described and reported to be important for persistence in animal models of infections (14, 28) or in adaptation to different host environments, such as lungs of CF patients or the rhizosphere (58). In these studies, the genetic basis for the colony morphology variants was not determined, although the shiny persistent variant of Bcc strain C1394 produced increased amounts of extracellular polysaccharide (14). Neither K56-2 shvR (4) nor BCAS0208 to BCAS0201 were required for persistence but were important for causing lung damage in infected animals and necrosis in infected plants. A few of the spontaneous K56-2 SHV could not be restored to the rough morphotype by complementation with shvR (4) or BCAS0208 in trans, indicating that additional genes may be involved in determining this phenotype.
Based on available genome sequences, orthologs BCAS0208 to BCAS0201 appear to be unique to Burkholderia spp. and conserved in B. cenocepacia, indicating that their role may be species specific. Orthologs of BCAS0208 to BCAS0201 are present in Bcc species B. cenocepacia, B. ambifaria, and Burkholderia lata but not B. multivorans or B. vietnamiensis. B. cenocepacia infections in CF patients generally have a poorer prognosis and a higher mortality rate, whereas B. multivorans, although currently the most frequently isolated Bcc species from CF patients, is typically associated with more mild and often transient infections (38). B. cenocepacia and B. ambifaria have also been shown to be more virulent in a variety of infection models than B. multivorans and B. vietnamiensis, which lack BCAS0208 to BCAS0201 homologs (5, 9, 11, 49, 54). BCAS0208 is also found in the highly pathogenic B. mallei and B. pseudomallei but not present in the closely related B. thailandensis, which is not a human pathogen (26, 61). Interestingly, although these non-Bcc species have BCAS0208 to BCAS0202, they do not have either BCAS0201 or shvR orthologs. Therefore, BCAS0208 to BCAS0201 orthologs are for the most part present in genomes of more virulent Burkholderia spp. and absent in the less virulent species, suggesting that they may contribute to the pathogenic potential of these species.
In addition to being regulated by ShvR, the afc genes are also regulated by QS systems (44, 45). Using global transcriptional profiles, the expression of most of the genes in the afc locus was found to be decreased in a cepR mutant but increased in a cciR mutant (45). It was also shown that CepR indirectly regulates the expression of afc genes (44). The expression of BCAS208 to BCAS0201 is also modulated by their immediate environment (63). Genes in this locus were found to be induced in a B. cenocepacia soil isolate (HI2424) in a soil medium (63). These studies show that proper modulation of expression of genes in the shvR-regulated 24-kb region of B. cenocepacia is important.
In our attempts to generate BCAS0208 mutants, deletion of a larger internal fragment of BCAS0208 resulted in a polar mutant (the Δ1BCAS0208 mutant), whereas a small internal deletion resulted in a nonpolar mutant (the Δ2BCAS0208 mutant). Sequencing of the truncated genomic region in the polar BCAS0208 mutant did not reveal any frameshift mutations in BCAS0208. The reasons for the polar nature of the Δ1BCAS0208 mutant are not clear, although it could be speculated that there may be an effect on RNA stability or transcription initiation as a result of this particular deletion.
Biofilm formation is an important trait of members of Bcc that are more virulent and considered to have a key role in persistent infections (15, 16). Components of both the bacterial cell surface and the biofilm matrix are important for bacterial adhesion and biofilm formation (25). In this study, mutations in BCAS0208 (shiny), BCAS0207 (rough), and BCAS0201 (rough) affected biofilm formation to different degrees. Although both polar BCAS0208 mutants (the BCAS0208::Tn and Δ1BCAS0208 mutants) were severely reduced for biofilm formation, partial complementation was obtained only for the BCAS0208::Tn (pBBRS0208) mutant; however, BCAS0208 in trans was found to be expressed in both strains. The lack of a significant increase in biofilm formation of the Δ1BCAS0208 (pBBRS0208) mutant could be due to interference by truncated versions of a BCAS0208-encoded protein. Similarly, in the Δ2BCAS0208 (pBBRS0208) mutant, biofilm formation is not restored to wild-type levels, although there is a significant increase compared to that in the Δ2BCAS0208 (pBBR1MCS-3) mutant. It is possible that the levels of BCAS0208 expressed in trans in the Δ2BCAS0208 mutant are not sufficient to result in wild-type levels of biofilm formation. Both polar BCAS0208 mutants (the BCAS0208::Tn and Δ1BCAS0208 mutants) and the shvR::Tn mutant to a lesser extent could be complemented for pellicle formation in the presence of pBBRS0208 and pBBRS0201. However, the nonpolar BCAS0208 and BCAS0201 mutants were not appreciably affected for pellicle formation at the air-liquid interface. This suggests that more than one gene influences pellicle formation and that overexpression of either of these genes likely alters cell surface properties to favor pellicle formation, while the absence of either BCAS0208 or BCAS0201 causes a much lesser defect. Partitioning experiments in the presence of hexadecane revealed that the presence of pBBRS0208 or pBBRS0201 in trans increased hydrophobicity for the shvR::Tn strain but not for all strains tested. Therefore, changes in cell surface hydrophobicity are not the only factor involved, and other changes in cell surface properties responsible for formation of both biofilm and pellicle might account for the phenotypes observed with these mutants.
Some of the genes between BCAS0208 and BCAS0201 code for functions involved in lipid metabolism, suggesting that they might participate in membrane or cell surface component biosynthesis and thereby influence colony morphotype and biofilm formation. Membrane fatty acid production is energetically expensive and needs to be highly coordinated with other components of intermediary metabolism (65). The coregulation of genes involved in lipid metabolism, like BCAS0208 and BCAS0201 with BCAS0207, involved in the TCA cycle might facilitate fine control of their expression. Bacteria have evolved tightly controlled mechanisms to regulate membrane lipid biogenesis (64, 65). The lipid transcriptional regulator FadR, a GntR family regulator, was first identified in E. coli (18, 29). Other lipid transcriptional regulators recently reported include FasR of Streptomyces coelicolor (an activator of fatty acid biosynthesis) and MabR of M. tuberculosis (a repressor of genes in the cell wall mycolic acid biosynthesis pathway) (2, 48). In both these bacteria, fatty acids produced in excess of the membrane phospholipid requirement may be diverted to storage granules (3, 46, 60). Mutations in FasR of Streptomyces coelicolor lead to reduced total cellular lipid content and rate of lipid synthesis from acetate. We do not know if lipid homeostasis is altered in shvR mutants or BCAS0208 to BCAS0201 mutants. However, the nonpolar BCAS0208 mutant had a slower growth rate with acetate as the sole carbon source (data not shown). In future studies, the lipid content of shvR and BCAS0208 to BCAS0201 mutants will be investigated.
Other LysR regulators are involved in the regulation of cell surface properties or biofilm formation, including PycR of Pseudomonas aeruginosa (35), PhcA of Ralstonia solanacearum (8), AlsR of Vibrio cholerae (33), and CrgA of Neisseria meningitides (20); however, very few studies have addressed the roles of target genes of these LysR-type regulators in determining changes in cell surface or biofilm formation. Several genes under PycR regulation were implicated in biofilm formation, although it is not clear if the regulation is direct (55, 59). Overexpression of the AlsR target acgA, an EAL domain-containing protein, reduced biofilm formation in V. cholerae, likely due to alteration of cyclic di-GMP levels (33). BCAS0208 to BCAS0201 represent additional novel genes involved in biofilm formation and suggest the importance of lipid metabolism in this process.
Acyl-CoA dehydrogenases are enzymes that participate in alpha and beta dehydrogenation of acyl-CoA esters in both fatty acid and amino acid metabolism (53). Examination of the closely related B. cenocepacia J2315 genome sequence revealed that there are several paralogs of acyl-CoA dehydrogenases, and these enzymes may act on structurally different fatty acids or fatty acids of different lengths (27, 30). These paralogs are apparently not functionally redundant, as they do not compensate for the defects due to mutations in BCAS0208. BCAS0208 is similar to the fadE5 family of acyl-CoA dehydrogenases. In M. tuberculosis, fadE paralogs are suggested to have diverse roles in pathogenicity, as they were shown to have differential patterns of expression in macrophages and are induced in animal infection models (34, 36). An fadE mutant of M. tuberculosis also had a reduced ability to kill eukaryotic cells (12). Similarly, in Salmonella enterica serovar Typhimurium, a fatty acyl dehydrogenase is important for survival under conditions of carbon starvation (50). In both these examples, fadE genes appear to be important for nutrition in vivo or in the absence of other carbon sources. We do not know if BCAS0208 is induced in vivo, but the lack of growth defect in fatty acids as sole carbon sources and the alteration of colony morphotype due to the BCAS0208 mutation suggest that the role of this gene has more influence on cell surface properties than growth or nutrition. The BCAS0208 polar mutant was able to persist as well as the WT in the chronic lung infection model, which also suggests that it is not required for nutrition in vivo, although the genes in this region were clearly required for maximum virulence in two infection models, adding to the data from other studies which suggest that fadE genes have important roles in pathogenicity.
To identify potential metabolic pathways influenced by genes BCAS0208 to BCAS0201, substrate utilization assays were carried out with Biolog plates (7). Biolog plates are more sensitive than standard growth measurements, and even slight differences in metabolic responses may be detected. The lack of growth defect in fatty acid substrates other than isovaleric acid may be explained by the functional redundancy of paralogs of BCAS0208, which may allow growth, although they are not redundant as far as biofilm and virulence properties. The severe growth defect of the Δ2BCAS0208 mutant on isovaleric acid, an intermediate of BCAA catabolism, suggests a role for BCAS0208 in this pathway. However, the polar BCAS0208 mutants did not show growth defects on isovaleric acid for reasons that are unclear. The BCAA pathway in bacteria consists of steps common to all three BCAAs catalyzing their conversion to acyl-CoA derivatives followed by reactions specific for each branched-chain acyl-CoA (42). The common pathway includes reactions catalyzed by a keto-acid dehydrogenase, and the first reaction after the common pathway is catalyzed by an acyl-CoA dehydrogenase (42). Interestingly, poorer utilization of BCAAs leucine, isoleucine, and valine by the BCAS0201 mutant suggested that this pathway was affected in this mutant. BCAS0201 codes for an FAD-dependent oxidoreductase/dehydrogenase, and PFAM analysis of the BCAS0201-encoded protein shows the presence of an FAD binding domain (PF01565) and FAD-linked oxidase C-terminal domain (PF02913). Our data and the sequence features of enzymes coded by BCAS0208 and BCAS0201 suggest that they might participate in different steps of the same metabolic pathway related to generation of cell surface lipids.
BCAS0204 codes for an ATP binding component of a transporter complex and is annotated as a cobalt-ion transporter (http://www.burkholderia.com), although growth experiments suggest that BCAS0204 might function in l-arabinose uptake. Examination of the J2315 genome sequence revealed the presence of two ABC transporters for l-arabinose: AraF (BCAL3405, BCAM2798), AraH (BCAL3403, BCAM2796), and AraG (BCAL3404, BCAM2797) (30). The BCAS0204 protein sequence shows 45% similarity to BCAL3404- and 49% similarity to BCAM2797-encoded proteins. It is possible that the other two l-arabinose transporters may not be expressed under conditions used for growth in this study.
In this study, we have shown that BCAS0208 to BCAS0201, an shvR-regulated genomic region, is required for determining colony morphology, virulence, and biofilm formation. While alteration of colony morphotypes has been reported previously, this is the first comprehensive phenotypic analysis of an shvR-regulated genomic cluster with a key role in determining the rough colony morphotype of B. cenocepacia. This study delineates the extent of contribution of genes in this region to these phenotypes and provides a basis for future studies aimed at identifying the alterations in cell surface that lead to changes in colony morphology and biofilm formation and the role of the BCAS0208 to BCAS0201 genes in this process.
ACKNOWLEDGMENTS
These studies were supported by the Canadian Institutes of Health Research (operating grant MOP-42510 to P.A.S.).
We thank D.F. Viteri and S.A. McKeon for performing the animal infection experiments and D.E. Woods for the histopathology analysis. We thank E.P. O'Grady for helpful discussions and critical reading of the manuscript.
FOOTNOTES
- Received 19 February 2011.
- Returned for modification 31 March 2011.
- Accepted 1 June 2011.
- Accepted manuscript posted online 20 June 2011.
- Copyright © 2011, American Society for Microbiology. All Rights Reserved.
