Search tips
Search criteria 


Logo of plosonePLoS OneView this ArticleSubmit to PLoSGet E-mail AlertsContact UsPublic Library of Science (PLoS)
PLoS One. 2012; 7(10): e47635.
Published online 2012 October 24. doi:  10.1371/journal.pone.0047635
PMCID: PMC3480411

Identification of a CO2 Responsive Regulon in Bordetella

Nancy E. Freitag, Editor


Sensing the environment allows pathogenic bacteria to coordinately regulate gene expression to maximize survival within or outside of a host. Here we show that Bordetella species regulate virulence factor expression in response to carbon dioxide levels that mimic in vivo conditions within the respiratory tract. We found strains of Bordetella bronchiseptica that did not produce adenylate cyclase toxin (ACT) when grown in liquid or solid media with ambient air aeration, but produced ACT and additional antigens when grown in air supplemented to 5% CO2. Transcriptome analysis and quantitative real time-PCR analysis revealed that strain 761, as well as strain RB50, increased transcription of genes encoding ACT, filamentous hemagglutinin (FHA), pertactin, fimbriae and the type III secretion system in 5% CO2 conditions, relative to ambient air. Furthermore, transcription of cyaA and fhaB in response to 5% CO2 was increased even in the absence of BvgS. In vitro analysis also revealed increases in cytotoxicity and adherence when strains were grown in 5% CO2. The human pathogens B. pertussis and B. parapertussis also increased transcription of several virulence factors when grown in 5% CO2, indicating that this response is conserved among the classical bordetellae. Together, our data indicate that Bordetella species can sense and respond to physiologically relevant changes in CO2 concentrations by regulating virulence factors important for colonization, persistence and evasion of the host immune response.


Many cues, such as temperature, oxygen (O2), iron, pH, osmolarity and bicarbonate, allow bacteria to distinguish between environments within a host and outside of a host, as well as various microenvironments within a host [1]. In sensing multiple cues, bacteria are able to synchronize gene expression to adapt and ultimately thrive [2]. One cue, carbon dioxide (CO2), has been shown to affect regulation of virulence factor expression in many bacterial pathogens. Bacillus anthracis responds to elevated levels of CO2 by increasing expression of the genes encoding edema toxin, lethal factor and protective antigen [3][5]. In response to 10% CO2, Streptococcus pyogenes increases transcription of M protein, an important virulence factor that prevents the deposition of complement onto the bacterial surface [6]. In increased CO2, M protein has been shown to be regulated by a trans-acting positive regulatory protein that binds to the promoter of the emm gene [6], [7]. CO2 regulation in B. anthracis appears to be more complicated since the transcriptional regulator of the toxins is not increased transcriptionally in response to growth in CO2 [3]. Additionally, Staphylococcus aureus, Salmonella enterocolitica and Borrelia burgdorferri are responsive to increased CO2 concentrations, suggesting this ability is useful to a variety of pathogens [8][11].

Bordetella bronchiseptica is a Gram-negative bacterium that infects a wide range of hosts causing respiratory disease varying from asymptomatic persistence in the nasal cavity for the life of the host to lethal pneumonia [12][14]. B. bronchiseptica is very closely related to the other two classical bordetellae, Bordetella pertussis and Bordetella parapertussis, the causative agents of whooping cough in humans [13], [15], [16]. Several virulence factors are produced by B. bronchiseptica such as, pertactin (PRN), filamentious hemaglutinin (FHA), two serotypes of fimbriae, and the two cytotoxic mechanisms, adenylate cyclase toxin (ACT) and the Type III Secretion System (TTSS) [13], [16]. ACT, a member of the repeats-in-toxin (RTX) family, is a bi-functional adenylate cyclase/hemolysin that converts ATP to cAMP, disrupting oxidative burst, phagocytosis, chemotaxis and eventually leads to apoptosis in macrophages and neutrophils [17][19]. ACT has also been shown to contribute to pathology, efficient colonization and persistence of B. bronchiseptica and B. pertussis species [20][22].

Regulation of virulence factors in bordetellae occurs via the BvgAS two-component system [23]. BvgS, the sensor in the cytoplasmic membrane, is thought to directly sense changes in the environment and, through a phosphorylation-transfer mechanism, activates BvgA, the response regulator [24][26]. Once BvgA is activated (Bvg+ phase), it binds to high and low affinity motifs in the genome, resulting in increased expression of the genes encoding toxins and adhesins, while expression of Bvg phase genes involved in motility and uptake of certain nutrients are repressed; the opposite occurs in the Bvg phase [27][32]. An intermediate phase has been described in which a subset of virulence factors are expressed, along with a unique set of factors [33][35]; however the Bvg+ phase has been shown to be necessary and sufficient for host colonization [36]. Although BvgAS appears to be sufficient for regulation of virulence factors the ability to respond to multiple signal inputs to differentially regulate transcriptional networks likely allows for adaptation to different microenvironments within the host. Bordetella species have multiple putative transcription factors within their genomes, indicating that gene regulation is likely to be a more complex regulatory system than is currently appreciated [16].

Here we identify, through screening of a collection of B. bronchiseptica isolates, strains that only produce ACT in response to growth in elevated CO2 conditions. Both strain 761 and the sequenced laboratory reference strain RB50 increased transcription of cyaA and production of ACT when grown in 5% CO2 conditions, although only strain 761 was dependent on 5% CO2 for efficient expression. Several other virulence factor genes were increased in transcription in response to growth in elevated CO2. BvgAS was required for ACT production, but cyaA and fhaB were transcriptionally increased in response to 5% CO2 conditions in the absence of BvgS. Together this indicates that an additional regulatory system increases production of ACT and other virulence factors in various Bordetella species.

Materials and Methods

Ethics Statement

This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Institutional Animal Care and Use Committee at The Pennsylvania State University at University Park, PA (#31297 Bordetella-Host Interactions). All animals were anesthetized using isoflourane or euthanized using carbon dioxide inhalation to minimize animal suffering.

Bacterial Strains and Growth

B. bronchiseptica strain RB50 is an isolate from a rabbit [36]. RB54 and RB50ΔbscNΔcyaA are previously described derivatives of strain RB50 [33], [37]. B. bronchiseptica strain 761, 448, and 308 were obtained from the CDC in Atlanta, Georgia and have been previously described [15], [38], [39]. B. bronchiseptica strain JC100 has been previously described [15]. B. parapertussis strain 12822 was isolated from German clinical trials and has been previously described [40], [41]. B. pertussis strain 536 is a streptomycin resistant derivative of Tohama I [42] and strain 18323 has been previously described [43]. B. pertussis strain CHOC 0012 was isolated on Regan-Lowe media by the Eunice Kennedy National Insitute of Child Health and Human Development (NICHD) Collaborative Pediatric Critical Care Research Network from a child displaying severe Pertussis. Bacteria were maintained on Bordet-Gengou agar (Difco, Sparks, MD) containing 10% sheep blood (Hema Resources, Aurora OR) and 20 µg/mL streptomycin (Sigma Aldrich, St. Louis, MO). Liquid cultures were grown at 37°C overnight in a shaker to mid-log phase (O.D. 0.7–1.0) in Stainer-Scholte (SS) broth. Bacteria were grown overnight with constant shaking (250 rpm) in standard glass test tubes in either atmospheric concentrations of oxygen and carbon dioxide (atmospheric conditions) or in atmospheric levels of oxygen with the constant controlled addition of 5% carbon dioxide into a sealed incubator 37°C (5% CO2 conditions).

cAMP Assay

Murine macrophage-like cell line, J774, was cultured in Dulbecco modified Eagle medium (DMEM) with 10% fetal bovine serum (FBS) (Hyclone Laboratories, Inc., Logan, UT). Cells were grown to approximately 80% confluency, and bacteria were added at a multiplicity of infection (MOI) of 1. After a 5 minute centrifugation at 250×g, the mixture was incubated for 30 minutes at 37°C. cAMP was measured with a cyclic AMP ELISA system (Tropix, Bedford, MA) according to the manufacturer’s instructions. Results were analyzed using analysis of variance with a Tukey simultaneous test, and a P value of <0.05 was considered significant.

Animal Experiments

C57BL/6 mice were obtained from Jackson Laboratories (Bar Harbor, ME). Mice were bred in our Bordetella-free, specific pathogen-free breeding rooms at The Pennsylvania State University. All animal experiments were performed in accordance with institutional animal care and use committee (IACUC) guidelines. 4 to 6 week old mice were lightly sedated with 5% isoflurane (IsoFlo, Abbott Laboratories) in oxygen and 5×105 CFU were pipetted in 50 ul of phosphate-buffered saline (PBS) (Omnipur, Gibbstown, NJ) onto the external nares. This method reliably distributes the bacteria throughout the respiratory tract [43]. To obtain serum, blood from inoculated or vaccinated mice was obtained 28 days post-inoculation and serum was separated from the blood by centrifugation at 500×g for 5 minutes.

Western Immunoblots

Western blots were performed on whole cell extracts of B. bronchiseptica, B. pertussis and B. parapertussis grown to mid-log phase in SS broth as described previously [38], [39]. Lysates were prepared by resuspending 1×109 CFU in 100 µl of Laemmli sample buffer; total cellular protein content were quantitated using the BCA assay to equalize protein content between samples. 1×108 CFU (10 µl) were run on an 8% sodium dodecyl sulfate-polyacrylamide electrophoresis gels in denaturing conditions and transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA). Membranes were probed with pooled serum from mice inoculated with B. bronchiseptica, B. pertussis, B. parapertussis, or a monoclonal antibody against ACT (anti-ACT) at the following dilutions, 1[ratio]1000, 1[ratio]500, 1[ratio]1000 and 1[ratio]1000, respectively. A 1[ratio]10,000 dilution of goat anti-mouse Ig HRP conjugated antibody (Southern Biotech, Birmingham, AL) was used as the detector antibody. Membranes were visualized with ECL Western blotting detection reagents (Amersham Biosciences, Piscataway, NJ) and quantified using Image J software [44].

RNA Isolation

RNA was isolated from three independent biological replicates of B. bronchiseptica strains RB50, 761, RB54, 536 and 12822 grown in SS broth overnight. Bacteria were subcultured at a starting OD600 of 0.1 into 5 ml of SS broth and grown at 37°C while shaking in either atmospheric or 5% CO2 conditions until the OD600 reached 0.75. Bacteria were harvested and total RNA was extracted using a RNAeasy Kit (Qiagen, Valencia, CA) and treated with RNase-free DNase I (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions.

Preparation of Labeled cDNA and Microarray Analysis

RNA isolated from strains RB50 and 761 were used in microarray experiments. A 2-color hybridization format was used for the microarray analysis. For each biological replicate, RNA extracted from cells grown in 5% CO2 conditions was used to generate Cy5-labeled cDNA and RNA extracted from cells grown in atmospheric conditions was used to generate Cy3-labeled cDNA. Additionally, dye-swap experiments were performed analogously, in which the fluorescent labels were exchanged to ensure that uneven incorporation did not confound our results. Fluorescently-labeled cDNA copies of the total RNA pool were prepared by direct incorporation of fluorescent nucleotide analogs during a first-strand reverse transcription (RT) reaction [39], [45][47]. The two differentially labeled reactions were then combined and directly hybridized to a B. bronchiseptica strain RB50-specific long-oligonucleotide microarray [46]. Slides were then scanned using a GenePix 4000B microarray scanner and analyzed with GenePix Pro software (Axon Instruments, Union City, CA). Spots were assessed visually to identify those of low quality and arrays were normalized so that the median of ratio across each array was equal to 1.0. Spots of low quality were identified and were filtered out prior to analysis. Ratio data from the two biological replicates were compiled and normalized based on the total Cy3% intensity and Cy5% intensity to eliminate slide to slide variation. Gene expression data were then normalized to 16S rRNA. The statistical significance of the gene expression changes observed was assessed by using the significant analysis of microarrays (SAM) program [48]. A one-class unpaired SAM analysis using a false discovery rate of 0.001% was performed. Hierarchical clustering of microarray data using Euclidean Distance metrics and Average Linkage clustering was performed using MeV software from TIGR [49]. All microarray data are available in Tables S1 and S2 and have been deposited in ArrayExpress or ArrayExpress Archive under accession number E-MEXP-2875.

Real-time qPCR (qPCR)

qPCR was performed using a modified protocol previously described [46], [47]. RNA was extracted as described, and 1 µg of RNA from each biological replicate was reverse transcribed using ImProm-II Reverse transcriptase and 0.5 µg of random oligonucleotide hexamers (Promega, Madison, WI). cDNA was diluted 1[ratio]1,000 and 1 µl was used in RT-qPCRs containing 300 nM primers designed with Primer Express software (Applied Biosystems, Foster City, CA, and Integrated DNA Technologies software, (Primer sequences are listed in Table S3) and SYBR Green PCR master mix (Invitrogen, Carlsbad, CA). Samples without reverse transcriptase were included to confirm lack of DNA contamination and dissociation curve analysis was performed to determine cycle threshold (CT) for each reaction. Amplification of the recA RNA amplicon was used as an internal control and for data normalization. Change in transcript level was determined using the relative quantitative method (ΔΔCT) [50]. Results were analyzed using analysis of variance with a Tukey simultaneous test, and a P value of <0.05 was considered significant.

Cytotoxicity Assay

Cytotoxicity assays were carried out as previously described [45]. J774A.1 cells, a murine macrophage cell line, obtained from the ATCC were cultured in DMEM with 10% FBS. Cells were grown to approximately 80% confluency, and bacteria were added at a MOI of 100, 10 or 1. After a 5 minute centrifugation at 250×g, the mixture was incubated at 37°C for the indicated times. Cytotoxicity was determined by measuring lactate dehydrogenase (LDH) release using the Cytotox96 (Promega, Madison, WI) kit according to the manufacturer’s protocol. Results were analyzed using analysis of variance with a Tukey simultaneous test, and a P value of <0.05 was considered significant.

Adherence Assay

Adherence assays were modified from a previously described protocol [47]. Rat epithelial cell line L2, obtained from the ATCC, was cultured in DMEM/Ham’s F12 50–50 mixture with 10% FBS. Cells were grown to approximately 80% confluency, and bacteria were added at an MOI of 100. After a 5 minute centrifugation at 250×g, the mixture was incubated for 30 minutes. Cell culture supernatant was removed and cells were washed 4 times with PBS to remove unbound bacteria. Epithelial cells were then trypsinized and resuspended in 1 mL of tissue culture media. The mixture of cells and bacteria were diluted in PBS and plated on BG agar to determine CFU. Results were analyzed using analysis of variance with a Tukey simultaneous test, and a P value of <0.05 was considered significant.


A B. bronchiseptica Isolate Regulates ACT Expression in Response to 5% CO2 Conditions

B. bronchiseptica isolates are generally β-hemolytic when grown in Bvg+ conditions due to the production of ACT, which causes lysis of red blood cells. It was recently discovered that some B. bronchiseptica isolates do not have the genes required to produce a functional ACT and therefore are not hemolytic on blood agar plates [39]. However, through screening of 73 isolates based on hemolysis on blood agar plates and PCR amplification of the genes encoding ACT, 4 B. bronchiseptica isolates were found to be non-hemolytic, but still retained the genes for production of ACT (Fig. 1, data not shown). When one isolate displaying this phenotype, B. bronchiseptica strain 761, was grown in a tissue culture incubator where the CO2 concentration is increased to 5%, it was hemolytic (Fig. 1, compare panels B and D). RB50 was hemolytic even in ambient air (~0.03% CO2) (Fig. 1, compare panels A and C), but there appeared to be more hemolysis when it was grown in 5% CO2 conditions, suggesting both strains produce more ACT in response to growth in 5% CO2 conditions. Growth rate and pH were not significantly affected by additiona of 5% CO2 (Figure S1.), although other indirect effects are possible.

Figure 1
Differential production of ACT in B. bronchiseptica strains grown in 5% CO2.

To more directly assess the production of ACT, lysates of strains RB50 and 761 grown in liquid cultures (mid-log phase, O.D. 0.7–0.9) in normal atmospheric conditions or 5% CO2 conditions were probed with a monoclonal antibody to the cyaA protein product, ACT. RB50 produced more ACT when grown in 5% CO2 conditions than in normal atmospheric conditions (Fig. 1E). Strain761 grown in normal atmospheric conditions produced no detectable ACT, while 761 in 5% CO2 conditions did produce ACT (Fig. 1E). To determine if strain 761 produces a functional ACT, cyclic-AMP (cAMP) was measured in murine macrophages stimulated with bacteria grown in either normal or 5% CO2 conditions. Cells were stimulated for 30 minutes with strains RB50 or 761 to assess their effects on cAMP levels. Both strains grown in 5% CO2 induced significantly more cAMP than the same strains grown in normal atmospheric conditions (Fig. 1F, G). Together, these data demonstrate by different measures that the prototypical B. bronchiseptica strain, RB50, and strain 761 increase production of functional ACT when grown in 5% CO2 conditions, but only strain 761 appears to be dependent on 5% CO2 conditions for production of ACT.

Since ACT production was increased in 5% CO2, we hypothesized that other antigens might also be differentially regulated in response to these conditions. To test this, Western blot analysis was performed by probing RB50 and 761 lysates, grown in ambient air or in 5% CO2 conditions, with serum antibodies from animals convalescent from RB50 infection. No antigens appeared to be produced in greater amounts in RB50 grown in ambient air, while strain 761 produced antigens of between 72 and 95 kDa in greater amounts in 5% CO2 growth conditions (Fig. 2, arrowheads). Antigens greater than 130 kDa were produced in greater amounts when RB50 was grown in 5% CO2 conditions compared to ambient air (Fig. 2, arrows). Additionally, both 761 and RB50 produced antigens between 36 and 55 kDa in greater amounts when grown in elevated CO2 concentrations than in ambient air (Fig. 2). These data indicate that additional antigens besides ACT are differentially regulated in response to 5% CO2.

Figure 2
Differential recognition of antigens in B. bronchiseptica strains grown in atmospheric or 5% CO2 conditions.

Defining a CO2 Responsive Regulon in B. bronchiseptica

To determine which genes are differentially regulated in response to different CO2 concentrations, microarray analyses were performed comparing RB50 grown in atmospheric concentrations of CO2 to growth in elevated CO2 conditions. Transcript abundance of 35 genes increased in RB50 in response to 5% CO2, based on SAM analysis (Table S1, Fig. 3), including genes encoding ACT, as well as genes encoding other known virulence factors such as members of the TTSS locus (bscE, bscF, bopD), FHA (fhaC, fhaB, fhaD, fhaA), fimbriae (fimA, fim3), and Prn, (Fig. 3). Expression of 452 genes were decreased when RB50 was grown in 5% CO2, many of which are known to be expressed in the Bvg phase including, flaA, cheW, cheB, wbmD, flgH, fliS, and cheD (Fig. 3). A similar trend was observed for strain 761; genes encoding known virulence factors were increased in 5% CO2 growth conditions while genes for flagellar assembly and chemotaxis were decreased in these conditions (Table S2, Fig. 3). Expression of 6 genes, including cyaA, was increased in both strains and expression of 41 genes decreased in both strains when grown in 5% CO2 conditions (Table S4). qPCR of 12 genes confirmed the microarray results (Table S3). Overall, these data indicate there is a CO2 responsive regulon in B. bronchiseptica that includes several virulence factors, suggesting a role during infection.

Figure 3
Defining the CO2 responsive regulon in B. bronchiseptica.

Among the 35 genes in strain RB50 increased in expression in 5% CO2 conditions, 19 genes were reported to be positively regulated under Bvg+ conditions, 4 genes negatively regulated by BvgAS, and 13 genes not previously known to be regulated by BvgAS, based on previous analysis [46], [51]. Similarly, of the 452 genes negatively regulated by 5% CO2, 252 were known to be negatively regulated under Bvg+ conditions, 19 were positively regulated and 181 were not previously known to be regulated by BvgAS. The CO2-responsive regulon appears to contain genes that are Bvg-regulated, as well as genes that are not, suggesting a regulatory mechanism that functions independently or cooperatively with BvgAS rather than subordinate to it.

CO2 Responsiveness is Not Conferred by Differences in the bvgAS Loci between Strains

Three additional B. bronchiseptica strains, JC100, 308 and 448, were also observed to be hemolytic only when grown in 5% CO2, but not in normal atmospheric conditions. Changes in virulence factor expression have previously been attributed to variation in bvgAS and since some CO2-responsive genes are Bvg-regulated, the bvg locus of these strains was analyzed revealing that strain JC100 carries a 29 amino acid duplication in the region of the bvgS gene encoding the periplasmic domain (Fig. 4). To determine if this duplication is involved in the CO2/ACT dependent phenotype in JC100, the bvgAS locus from JC100 was expressed in a RB50 knockout of bvgAS (RB55::pBvgASJC100). This strain was hemolytic in the absence of 5% CO2, indicating that transfer of the bvgAS locus does not confer the CO2-dependence for ACT production (Table S5). The reverse was also true; when a plasmid carrying the bvgAS locus from RB50 was introduced into JC100 (MLJC114::pEG100), ACT production remained dependent on growth in 5% CO2 (Table S5). Furthermore, the bvgS gene of 761 did not have this duplication (Fig. 4). These data indicate that the duplication in bvgS in JC100 is neither necessary nor sufficient for the CO2 requirement for hemolysis.

Figure 4
Duplication in the bvgS gene in strain JC100.

CO2 Responsiveness in the B. bronchiseptica Bvg State

Since some virulence genes known to be Bvg-regulated were responsive to 5% CO2 conditions, we sought to determine if they are differentially regulated in response to 5% CO2 in the absence of BvgAS. Transcription of six genes responsive to CO2, cyaA, fhaB, bopD, bopB, cheZ and flgB, were analyzed in B. bronchiseptica RB50 and RB54, a Bvgphase locked derivative of strain RB50, grown in normal atmospheric or 5% CO2 conditions (Fig. 5). For RB50, addition of 5% CO2 increased transcription of cyaA, fhaB, bopD and bopB (Fig. 5A–D), but decreased transcription of cheZ and flgB (Fig. 5 E,F). In RB54, transcription of bopD and bopB was not increased in response to addition of 5% CO2 (Fig. 5C,D), and transcription of cheZ and flgB was not decreased (Fig. 5 E,F). Therefore, the differential transcription of bopD, bopB, cheZ and flgB in response to 5% CO2 is dependent on BvgS. However, in the bvgS mutant RB54, the transcription of genes cyaA and fhaB was increased in response to addition of 5% CO2 (Fig. 5 A,B), indicating that some gene regulation in response to 5% CO2 is independent of BvgS.

Figure 5
Differential transcription of 5% CO2 responsive genes independent of bvgS expression.

To determine whether differential transcription results in differential accumulation of antigens in the absence of BvgS, Western blots were performed. Lysates from RB54 grown in normal atmospheric or 5% CO2 conditions were probed with serum antibodies from mice convalescent from RB50 infection (Fig. 6). Strain RB54 grown in 5% CO2 also produced antigens >250 kDa and ~60 kDa in greater amounts (Fig. 6). RB54 grown in ambient air produced an antigen between 95 and 130 kDa in greater amounts (Fig. 6, arrowhead). These data demonstrate that antigen production is differentially regulated in response to 5% CO2 even when a functional BvgS is absent.

Figure 6
Differential recognition of antigens of a Bvg

Growth in 5% CO2 Affects Cytotoxicity and Adherence of B. bronchiseptica Strains

Since genes, cyaA and the TTSS genes, associated with the cytotoxicity of B. bronchiseptica were increased when strains RB50 and 761 were grown in 5% CO2 conditions (Fig. 3), we assessed the relative cytotoxicity to J774 murine macrophages of strains grown in atmospheric or 5% CO2 conditions. Similar to previous findings [52], RB50 killed >90% of cells at an MOI of 10 or 100; however, RB50 only killed ~65% at an MOI of 1 (Fig. 7A). RB50 grown in 5% CO2 killed >90% at all MOIs, indicating that RB50 grown in 5% CO2 killed more macrophages at a lower MOI than RB50 grown in atmospheric conditions (Fig. 7A). Strain 761 had detectable (~30%) killing only at high MOIs (10 and 100), while 761 grown in 5% CO2 was cytotoxic at an MOI of 1 (~45%) and comparable to RB50 at higher MOIs (Fig. 7A). These data show that growth in 5% CO2 increased killing of murine macrophages by both strains.

Figure 7
Cytotoxicity and adherence of strains grown in 5% CO2 conditions.

ACT and TTSS have been previously shown to account for all cytotoxicity of macrophages when RB50 is grown in ambient air [37], [52]. Since 5% CO2 increased expression of several genes, we examined whether the increased cytotoxicity is due to increases in these known factors or a new cytotoxic mechanism. Cells were exposed for 4 hours at MOIs of 1, 10 or 100 with wild-type RB50, RB50ΔcyaA, RB50ΔbscN (encoding the ATPase of the TTSS) or a mutant lacking bscN and cyaA (RB50ΔcyaAΔbscN). Growth in 5% CO2 increased cytotoxicity of RB50Δcya to macrophages at MOIs of 1, 10 and 100, while growth 5% CO2 caused increased cytotoxicity of RB50ΔbscN only at an MOI of 100, likely indicating the differential roles of the TTSS and ACT in cytotoxicity (Figure 7B). RB50ΔcyaAΔbscN caused very low levels of cytotoxicity as observed previously [37], and growth in 5% CO2 did not increase cytotoxicity(Fig. 7B). These data suggest that there are no other cytotoxic mechanisms and that increased ACT and TTSS function accounts for the increased cytotoxicity when strains are grown in 5% CO2 conditions.

Since many genes encoding adhesins were increased in transcription in response to growth in 5% CO2 conditions, we hypothesized that strains grown under these conditions, in comparison to growth in ambient air, would be more adherent to epithelial cells. L2 cells were incubated with RB50 or 761 pre-grown in either atmospheric or 5% CO2 conditions. Both strains pre-grown in 5% CO2 conditions adhered to lung epithelial cells more efficiently than bacteria grown in normal atmospheric conditions (Fig. 7C).

B. pertussis and B. parapertussis Modulate Virulence Factor Expression in Response to 5% CO2

Since up-regulation of virulence factors in response to growth in 5% CO2 conditions is common to multiple B. bronchiseptica strains, we hypothesized that B. pertussis and B. parapertussis may also regulate virulence factor expression in response to growth in 5% CO2 conditions. Genes shown to be CO2 responsive (cyaA, fhaB, fimA) or non-responsive to CO2 (bvgS) in B. bronchiseptica (Fig. 3), were chosen to be analyzed by qPCR in B. pertussis and B. parapertussis. B. pertussis and B. parapertussis had increased expression of fhaB and cyaA, but not bvgS in response to CO2 (Fig. 8A, B). B. pertussis, unlike B. parapertussis, also had increased expression of fimA, indicating that the 5% CO2 responsive regulon may be different among the three classical Bordetella species. To further investigate this effect, Western blots with lysates of B. pertussis and B. parapertussis grown in ambient air or 5% CO2 were probed with either B. pertussis-induced sera or B. parapertussis-induced sera (Fig. 8C, D). B. pertussis strains 536, 18323 and a recent clinical isolate CHOC 0012 grown in ambient air showed a different antigenic profile from the lysates prepared from strains grown in 5% CO2, with bands from roughly 55 to 250 kDa which were more numerous and intense in 5% CO2 for strains 536 and CHOC 0012 (Fig. 8C). Notably, B. pertussis strains 536 and CHOC 0012 appeared to increase similar antigens in response to growth in 5% CO2 conditions, while strain 18323 grown in 5% CO2 decreased production of several antigens (Figure 8C). B. parapertussis grown in ambient air produced a more intense band between 130 and 250 kDa, while growth in 5% CO2 produced a more intense band between 36 and 55 kDa (Fig. 8D). Overall, B. parapertussis did not appear to differentially regulate many antigens in response to growth in 5% CO2 conditions (Figure 8D). These data indicate that several antigens in B. pertussis, but few in B. parapertussis isolate 12822, are differentially regulated in response to growth in 5% CO2 compared to growth in ambient air and that there is strain variation in CO2 responsiveness in the bordetellae.

Figure 8
Differential expression of virulence factor genes in B. pertussis and B. parapertussis in response to 5% CO2 conditions.


BvgAS was originally considered an ON/OFF switch, modulating Bordetella species between two distinct states, avirulent (Bvg) and virulent (Bvg+). The discovery of an intermediate phase has led to a view of BvgAS gene regulation as a rheostat visualized as varying along a one dimensional gradient [23], [32], [34]. In this view of the two-component system few signals, temperature and some chemical cues, are known to affect virulence factor regulation through the BvgAS system. However, the respiratory tract contains many microenvironments, and within each environment there is likely to be great variation. For example, CO2 levels are thought to vary between air and epithelial cells of the respiratory tract, although these are separated by a fraction of a millimeter of mucous. These sites also change dramatically in the course of the various stages of an infection, and there is likely to be a selective advantage to any strain that can sense these differences and modulate virulence factor expression in response.

Here we show that the classical bordetellae share the ability to sense and respond to physiological changes in CO2 concentrations likely to be encountered in the host. In mammalian tissues and blood, CO2 concentrations are higher than inhaled ambient air concentrations of CO2, which are approximately 0.03%. The observed changes in expression of various virulence factors (Fig. 3), and altered phenotypes (adherence and cytotoxicity, Fig. 7) provide additional evidence that the ability to respond to changes in CO2 concentrations allow Bordetella species to adjust to different microenvironments within the host respiratory tract.

Recently, it has been shown that there is a zone of oxygenation between the anaerobic luminal environment and the host epithelium in the gastrointestinal tract, which can be sensed by Shigella flexneri [2]. The presence of oxygen alters the expression of TTSS effectors that are important for invasion of host cells, and this ‘aerobic zone’ may enhance secretion of these effectors thereby increasing invasion of host cells [2]. Similarly, the respiratory tract of mammals contains multiple sites where gradients of CO2 or oxygen likely influence virulence factor expression and how respiratory pathogens interact with host cells. B. bronchiseptica has been isolated from multiple sites within the respiratory tract (e.g. nasopharnyx, trachea, lungs) and the ability to detect these differences could allow this pathogen to respond by expressing the array of factors optimal for success under each condition [2]. As bacteria disseminate from the nasal cavity to the trachea, lung and potentially even invade tissues, CO2 concentrations may increase, serving as a signal for increased transcription of factors such as adhesinsand toxins that subvert the immune response, which is more robust in these regions 17,5254.

B. bronchiseptica strains sense and differentially regulate virulence factor gene expression in response to 5% CO2 (Fig. 3), and differential regulation was observed in multiple strains and species of Bordetella demonstrating that sensing and responding to carbon dioxide levels is an ability shared among the classical bordetellae. Additionally, the transcription of several Bvg+-phase genes increased in response to 5% CO2, suggestive of BvgAS involvement in the response. Interestingly, regulation of some virulence factor genes (bopD, bopB) by BvgAS was epistatic to 5% CO2 regulation. However, not all virulence gene expression (cyaA, fhaB) was dependent on bvgS (Fig. 5A, B), demonstrating an independent mechanism for virulence factor gene regulation in response to 5% CO2.

Standard Bvg+ conditions, without additional CO2, are sufficient for production of ACT (Fig. 1) in RB50, suggesting that additional mechanisms may contribute to increases in production, but are not required. Of 73 B. bronchiseptica strains screened, 4 strains were identified here, 761, 308, 448 and JC100, in which Bvg+ phase conditions are not sufficient for measurable production of ACT. In these strains both Bvg+ phase conditions and 5% CO2 are required for detectable production of ACT. The requirement for 5% CO2 for the production of virulence factors may reflect evolutionary adaption of B. bronchiseptica strains, and suggests that the mechanism of CO2 sensing may confer a selective advantage. Intriguingly, there also appears to be variation in responsiveness of both B. bronchiseptica and B. pertussis strains suggesting that although the ability to respond appears to be conserved the regulon may vary between species and strains.

Collectively these data demonstrate that a CO2 response mechanism contributes to regulation of virulence factors in the classical bordetellae (Fig. 5A, B). This is the first description of a CO2 sensing mechanism that regulates virulence factor expression cooperatively with, or independently of, BvgAS. Our data support the idea that virulence factor gene expression can be fine-tuned in response to signals specific to different microenvironments within the respiratory tract or deeper tissues within the host.

Supporting Information

Figure S1

Growth and pH of B. bronchiseptica strains RB50 and 761 in ambient air or 5% CO2 conditions. The growth of strains RB50 (diamonds) and 761(squares) grown in normal atmospheric oxygen conditions (black) or in elevated 5% CO2 conditions (white) was measured. pH was assessed at the indicated timepoints expressed as the mean ± standard deviation.


Table S1

B. bronchiseptica strain RB50 Expression Arrays 1 & 2.


Table S2

B. bronchiseptica strain 761 Expression Arrays 1 & 2.


Table S3

qRT-PCR Data and Primers.


Table S4

Genes Increased and Decreased in Transcription in response to 5% CO2 conditions in B. bronchiseptica strains RB50 and 761.


Table S5

Colony characteristics of RB50 and JC100 derivatives under different conditions.



We acknowledge Jeff F. Miller, Laura Weyrich and Xuqing Zhang for critical review of this manuscript and all members of the Harvill lab for support and helpful discussion. Additionally, we thank Alexia Karanikas and the Eunice Kennedy National Insitute of Child Health and Human Development (NICHD) Collaborative Pediatric Critical Care Research Network (CPCCRN) for B. pertussis strain CHOC 0012. The USDA is an equal opportunity provider and employer.

Funding Statement

This work was supported by National Institutes of Health grant GM083113 (to ETH) and by the Agriculture and Food Research Initiative Competitive Grants Program Grant no. 2010-65110-20488 from the USDA National Institute of Food and Agriculture. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.


1. Mekalanos JJ (1992) Environmental signals controlling expression of virulence determinants in bacteria. Bacteriol. 174: 1–7. [PMC free article] [PubMed]
2. Marteyn B, West NP, Browning DF, Cole JA, Shaw JG, et al. (2010) Modulation of Shigella virulence in response to available oxygen in vivo. Nature 465: 355–358. [PubMed]
3. Fouet A, Mock M (1996) Differential influence of the two Bacillus anthracis plasmids on regulation of virulence gene expression. Infect. Immun.. 64: 4928–4932. [PMC free article] [PubMed]
4. Sirard JC, Mock M, Fouet A (1994) The three Bacillus anthracis toxin genes are coordinately regulated by bicarbonate and temperature. J. Bacteriol. 176: 5188–5192. [PMC free article] [PubMed]
5. Sterne M (1937) Variation in Bacillus anthracis. Onderstepoort J. Vet. Sci. Anim. Ind. 8: 271–349.
6. Caparon MG, Geist RT, Perez-Casal J, Scott JR (1992) Environmental regulation of virulence in group A streptococci: transcription of the gene encoding M protein is stimulated by carbon dioxide. J. Bacteriol. 174: 5693–5701. [PMC free article] [PubMed]
7. Okada N, Geist RT, Caparon MG (1993) Positive transcriptional control of mry regulates virulence in the group A streptococcus. Mol. Microbiol. 7: 893–903. [PubMed]
8. Hyde JA, Trzeciakowski JP, Skare JT (2007) Borrelia burgdorferi alters its gene expression and antigenic profile in response to CO2 levels. J Bacteriol. 189: 437–445. [PMC free article] [PubMed]
9. Miller SI, Mekalanos JJ (1990) Constitutive expression of the phoP regulon attenuates Salmonella virulence and survival within macrophages. J. Bacteriol. 172: 2485–2490. [PMC free article] [PubMed]
10. Skaar EP, Humayun M, Bae T, Debord KL, Schneewind O (2004) Iron-Source Preference of Staphylococcus aureus Infections. Science 305: 1626–1628. [PubMed]
11. Torres VJ, Stauff DL, Pishchany G, Bezbradica JS, Gordy LE, et al. (2007) A Staphylococcus aureus Regulatory System that Responds to Host Heme and Modulates Virulence. Cell Host & Microbe 1: 109–119. [PMC free article] [PubMed]
12. Goodnow RA (1980) Biology of Bordetella bronchiseptica. Microbiol. Rev. 44: 722–738. [PMC free article] [PubMed]
13. Mattoo S, Cherry JD (2005) Molecular Pathogenesis, Epidemiology, and Clinical Manifestations of Respiratory Infections Due to Bordetella pertussis and Other Bordetella Subspecies. Clin. Microbiol. Rev. 18: 326–382. [PMC free article] [PubMed]
14. Stavely CM, Register KB, Miller MA, Brockmeier SL, Jessup DA, et al. (2003) Molecular and antigenic characterization of Bordetella bronchiseptica isolated from a wild southern sea otter (Enhydra lutris nereis) with severe suppurative bronchopneumonia. J. Vet. Diagn. Invest. 15: 570–574. [PubMed]
15. Diavatopoulos DA, Cummings CA, Schouls LM, Brinig MM, Relman DA, et al. (2005) Bordetella pertussis, the Causative Agent of Whooping Cough, Evolved from a Distinct, Human-Associated Lineage of B. bronchiseptica. PLoS Pathog 1(4): e45. [PMC free article] [PubMed]
16. Parkhill J, Sebaihia M, Preston A, Murph LD, Thomson N, et al. (2003) Comparative analysis of the genome sequences of Bordetella pertussis, Bordetella parapertussis and Bordetella bronchiseptica. Nat. Genet. 35: 32–40. [PubMed]
17. Hewlett EL, Donato GM, Gray MC (2006) Macrophage cytotoxicity produced by adenylate cyclase toxin from Bordetella pertussis: more than just making cyclic AMP! Mol. Microbiol. 59: 447–459. [PubMed]
18. Khelef N, Zychlinsky A, Guiso N (1993) Bordetella pertussis induces apoptosis in macrophages: role of adenylate cyclase-hemolysin. Infect. Immun. 61: 4064–4071. [PMC free article] [PubMed]
19. Weingart CL, Weiss AA (2000) Bordetella pertussis virulence factors affect phagocytosis by human neutrophils. Infect. Immun. 68: 1735–1739. [PMC free article] [PubMed]
20. Harvill ET, Cotter PA, Yuk MH, Miller JF (1999) Probing the function of Bordetella bronchiseptica adenylate cyclase toxin by manipulating host immunity. Infect. Immun. 67: 1493–1500. [PMC free article] [PubMed]
21. Khelef N, Sakamoto H, Guiso N (1992) Both adenylate cyclase and hemolytic activities are required by Bordetella pertussis to initiate infection. Microb. Pathog. 12: 227–235. [PubMed]
22. Weiss AA, Goodwin MS (1989) Lethal infection by Bordetella pertussis mutants in the infant mouse model. Infect. Immun. 57: 3757–3764. [PMC free article] [PubMed]
23. Cotter PA, Jones AM (2003) Phosphorelay control of virulence gene expression in Bordetella. Trends Microbiol. 11: 367–373. [PubMed]
24. Steffen P, Goyard S, Ullmann A (1996) Phosphorylated BvgA is sufficient for transcriptional activation of virulence-regulated genes in Bordetella pertussis. EMBO J 15(1): 102–109. [PubMed]
25. Stibitz S, Yang MS (1991) Subcellular localization and immunological detection of proteins encoded by the vir locus of Bordetella pertussis. J. Bacteriol. 173: 4288–4296. [PMC free article] [PubMed]
26. Uhl MA, Miller JF (1994) Autophosphorylation and phosphotransfer in the Bordetella pertussis BvgAS signal transduction cascade. Proc. Natl. Acad. Sci. 91: 1163–1167. [PubMed]
27. Boucher PE, Maris AE, Yang MS, Stibitz S (2003) The response regulator BvgA and RNA polymerase a subunit C-terminal domain bind simultaneously to different faces of the same segment of promoter DNA. Mol. Cell 11: 163–173. [PubMed]
28. Boucher PE, Stibitz S (1995) Synergistic binding of RNA polymerase and BvgA phosphate to the pertussis toxin promoter of Bordetella pertussis. J. Bacteriol. 177(22): 6486–6491. [PMC free article] [PubMed]
29. Karimova G, Bellalou J, Ullmann A (1996) Phosphorylation-dependent binding of BvgA to the upstream region of the cyaA gene of Bordetella pertussis. Mol. Microbiol. 20: 489–496. [PubMed]
30. Merkel TJ, Barros C, Stibitz S (1998) Characterization of the bvgR locus of Bordetella pertussis. J Bacteriol. 180: 1682–1690. [PMC free article] [PubMed]
31. Merkel TJ, Boucher PE, Stibitz S, Grippe VK (2003) Analysis of bvgR expression in Bordetella pertussis. J Bacteriol. 185: 6902–6912. [PMC free article] [PubMed]
32. Williams CL, Boucher PE, Stibitz S, Cotter PA (2005) BvgA functions as both an activator and a repressor to control Bvg phase expression of bipA in Bordetella pertussis. Mol Microbiol. 56: 175–188. [PubMed]
33. Deora R, Bootsma HJ, Miller JF, Cotter PA (2001) Diversity in the Bordetella virulence regulon: transcriptional control of a Bvg-intermediate phase gene. Mol. Microbiol. 40: 669–683. [PubMed]
34. Stockbauer KE, Fuchslocher B, Miller JF, Cotter PA (2001) Identification and characterization of BipA, a Bordetella Bvg-intermediate phase protein. Mol. Microbiol. 39: 65–78. [PubMed]
35. Vergara-Irigaray N, Chávarri-Martínez A, Rodríguez-Cuesta J, Miller JF, Cotter PA, et al. (2005) Evaluation of the role of the Bvg intermediate phase in Bordetella pertussis during experimental respiratory infection. Infect. Immun. 73: 748–760. [PMC free article] [PubMed]
36. Cotter PA, Miller JF (1994) BvgAS-mediated signal transduction: analysis of phase-locked regulatory mutants of Bordetella bronchiseptica in a rabbit model. Infect. Immun. 62: 3381–3390. [PMC free article] [PubMed]
37. Mann P, Goebel E, Barbarich J, Pilione M, Kennett M, et al. (2007) Use of a genetically defined double mutant strain of Bordetella bronchiseptica lacking adenylate cyclase and type III secretion as a live vaccine. Infect. Immun. 75(7): 3665–3672. [PMC free article] [PubMed]
38. Buboltz AM, Nicholson TL, Karanikas AT, Preston A, et al. (2009) Evidence for horizontal gene transfer of two antigenically distinct O antigens in Bordetella bronchiseptica. Infect. Immun. 77(8): 3249–3257. [PMC free article] [PubMed]
39. Buboltz AM, Nicholson TL, Parette MR, Hester SE, Parkhill J, et al. (2008) Replacement of adenylate cyclase toxin in a lineage of Bordetella bronchiseptica. J Bacteriol. 190: 5502–5511. [PMC free article] [PubMed]
40. Heininger U, Cotter PA, Fescemyer HW, Martinez de Tejada G, Yuk MH, et al. (2002) Comparative phenotypic analysis of the Bordetella parapertussis isolate chosen for genomic sequencing. Infect. Immun. 70: 3777–3784. [PMC free article] [PubMed]
41. Heinger U, Stehr K, Schmitt-Grobe S, Lorenz C, Rost R, et al. (1994) Clinical characteristics of illness caused by Bordetella parapertussis compared with illness caused by Bordetella pertussis. Pediatr. Infect. Dis. J. 13: 306–309. [PubMed]
42. Preston A, Allen AG, Cadisch J, Thomas R, Stevens K, et al. (1999) Genetics basis for lipopolysaccharide O-antigen biosynthesis in bordetellae. Infect. Immun. 67: 3763–3767. [PMC free article] [PubMed]
43. Harvill ET, Cotter PA, Miller JF (1999) Pregenomic comparative analysis between Bordetella bronchiseptica RB50 and Bordetella pertussis Tohama I in murine models of respiratory tract infection. Infect. Immun. 67: 6109–6118. [PMC free article] [PubMed]
44. Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image analysis. Nature Methods 9: 671–675. [PubMed]
45. Buboltz AM, Nicholson TL, Weyrich LS, Harvill ET (2009) Role of the type III secretion system in a hypervirulent lineage of Bordetella bronchiseptica. Infect. Immun. 77(9): 3969–3977. [PMC free article] [PubMed]
46. Nicholson TL (2007) Construction and validation of a first-generation Bordetella bronchiseptica long-oligonucleotide microarray by transcriptional profiling the Bvg regulon. BMC Genomics 8: 220. [PMC free article] [PubMed]
47. Nicholson TL, Buboltz AM, Harvill ET, Brockmeier SL (2009) Microarray and functional analysis of growth phase-dependent gene regulation in Bordetella bronchiseptica. Infect. Immun. 77(10): 4221–4231. [PMC free article] [PubMed]
48. Tusher V G, Tibshirani R Chu G (2001) Significance analysis of microarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. 98: 5116–5121. [PubMed]
49. Saeed AI, Sharov V, White J, Li, Liang W, et al. (2003) TM4: a free, open-source system for microarray data management and analysis. Biotechniques 34: 374–378. [PubMed]
50. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2-ΔΔCT method. Methods 25: 402–408. [PubMed]
51. Cummings CA, Bootsma HJ, Relman DA, Miller JF (2006) Species- and Strain-Specific Control of a Complex, Flexible Regulon by Bordetella BvgAS. J. Bacteriol. 188: 1775–1785. [PMC free article] [PubMed]
52. Yuk MH, Harvill ET, Cotter PA, Miller JF (2000) Modulation of host immune responses, induction of apoptosis and inhibition of NF-kappaB activation by the Bordetella type III secretion system. Mol. Microbiol. 35: 991–1004. [PubMed]
53. Cotter PA, Yuk MH, Mattoo S, Akerley BJ, Boschwitz J, et al. (1998) Filamentous hemagglutinin of Bordetella bronchiseptica is required for efficient establishment of tracheal colonization. Infect. Immun. 66(12): 5921–5929. [PMC free article] [PubMed]
54. Mattoo S, Miller JF, Cotter PA (2000) Role of Bordetella bronchiseptica fimbriae in tracheal colonization and development of a humoral immune response. Infect. Immun. 68(4): 2024–2033. [PMC free article] [PubMed]

Articles from PLoS ONE are provided here courtesy of Public Library of Science