ABSTRACT
Many of the pathogenic species of the genus Bordetella have an absolute requirement for nicotinic acid (NA) for laboratory growth. These Gram-negative bacteria also harbor a gene cluster homologous to the nic cluster of Pseudomonas putida which is involved in the aerobic degradation of NA and its transcriptional control. We report here that BpsR, a negative regulator of biofilm formation and Bps polysaccharide production, controls the growth of Bordetella bronchiseptica by repressing the expression of nic genes. The severe growth defect of the ΔbpsR strain in Stainer-Scholte medium was restored by supplementation with NA, which also functioned as an inducer of nic genes at low micromolar concentrations that are usually present in animals and humans. Purified BpsR protein bound to the nic promoter region, and its DNA binding activity was inhibited by 6-hydroxynicotinic acid (6-HNA), the first metabolite of the NA degradative pathway. Reporter assays with the isogenic mutant derivative of the wild-type (WT) strain harboring deletion in nicA, which encodes a putative nicotinic acid hydroxylase responsible for conversion of NA to 6-HNA, showed that 6-HNA is the actual inducer of the nic genes in the bacterial cell. Gene expression profiling further showed that BpsR dually activated and repressed the expression of genes associated with pathogenesis, transcriptional regulation, metabolism, and other cellular processes. We discuss the implications of these findings with respect to the selection of pyridines such as NA and quinolinic acid for optimum bacterial growth depending on the ecological niche.
IMPORTANCE BpsR, the previously described regulator of biofilm formation and Bps polysaccharide production, controls Bordetella bronchiseptica growth by regulating the expression of genes involved in the degradation of nicotinic acid (NA). 6-Hydroxynicotinic acid (6-HNA), the first metabolite of the NA degradation pathway prevented BpsR from binding to DNA and was the actual in vivo inducer. We hypothesize that BpsR enables Bordetella bacteria to efficiently and selectively utilize NA for their survival depending on the environment in which they reside. The results reported herein lay the foundation for future investigations of how BpsR and the alteration of its activity by NA orchestrate the control of Bordetella growth, metabolism, biofilm formation, and pathogenesis.
INTRODUCTION
Members of the genus Bordetella are Gram-negative bacteria that include human, animal, and avian pathogens and many strains isolated from soil. Of the nine species known to date, the three known as the “classical species” are closely related genetically. Bordetella bronchiseptica infects humans, animals, and marine mammals. It causes a dramatically diverse range of respiratory diseases and chronic asymptomatic colonization. Bordetella pertussis, the agent of whooping cough, is strictly a human pathogen, whereas Bordetella parapertussis can infect both humans and sheep (1–3).
The classical species of Bordetella have an absolute requirement for either nicotinic acid (NA) or nicotinamide for laboratory growth and are routinely grown in medium containing micromolar concentrations of either of these compounds (4, 5). NA, along with nicotinamide is a precursor for the coenzymes NAD and NADP in vivo, which participate in many critical cellular pathways (6). While many bacteria produce NAD de novo from aspartic acid, Bordetella and other bacteria that lack this pathway utilize the salvage or the recycling pathway. This pathway utilizes NAD precursors (NA or nicotinamide) or breakdown products. NA is converted in three steps to NAD by the conserved “Preiss-Handler pathway” by the conversion to nicotinic acid mononucleotide and nicotinic acid adenine dinucleotide. For the production of NAD from nicotinamide, nicotinamidases convert it to NA for the Preiss-Handler pathway (6, 7). The presence of NA in millimolar concentrations during Bordetella growth results in a phenomenon known as phenotypic modulation. This requires the BvgAS signal transduction system and is characterized by large-scale alterations in the expression of genes, including pathogenesis-associated or virulence genes (8–11).
In addition to serving as a source of NAD, NA is also a carbon, nitrogen, and energy source for many bacteria which catabolize it by both anaerobic and aerobic routes (12). In Pseudomonas putida strain KT2440, aerobic degradation of NA via the maleamate pathway involves the genes of the nic cluster (Fig. 1A and B). The expression of the nic cluster is tightly controlled in P. putida and involves the activity of two transcriptional repressors, NicR and NicS, and induction by nicotinic acid (Fig. 1). The orthologues of nic genes are also present in many soil bacteria that colonize the plant rhizosphere (13, 14).
(A) Comparison of the nic cluster of P. putida strain KT2440 with homologous clusters from the classical Bordetella species. Each arrow indicates an ORF. For the Bordetella species, the gene organization is extended to include the adjacent bpsA-D locus that is regulated by bpsR (15, 39, 40). Homologous genes are indicated by the same color, and the protein designation of the Bordetella nic genes is based on their amino acid homology with the corresponding Nic proteins of P. putida KT2440 (13). For the other Bordetella genes, the functions are assigned on the basis of the annotation in the genomic databases. Diagonal stripes indicate annotations as pseudogenes. Cross marks represent insertions of a transposase and thus inactive genes. (B) Predicted aerobic NA degradation pathway in Bordetella bronchiseptica based on its homology to nic cluster in P. putida KT2440 (13, 14).
The conservation of nic genes in bacteria associated with the plant rhizosphere is consistent with plants being the major source of NA in the environment. The degradation of NA by the activity of nic genes will enable these bacteria to survive in the soil environment. Surprisingly, a search for homologs of the nic genes of P. putida KT2440 revealed the conservation of the nic genes, including bpsR, the homolog of nicR, in B. bronchiseptica (Fig. 1) (13). In B. bronchiseptica, BpsR functions as a transcriptional repressor of Bps polysaccharide production and biofilm development (15, 16). The activity of the NA degradation pathway will lead to a depletion of the NA pool in B. bronchiseptica and may thus lead to a defect in NAD and NADP synthesis. In addition to B. bronchiseptica, the majority of nic genes and bpsR are intact in B. parapertussis and B. pertussis (Fig. 1). The speciation of B. pertussis and B. parapertussis from B. bronchiseptica has been driven mainly by large gene losses and inactivation events that have considerably reduced the genomic content, altered regulatory network, and modified the metabolic pathways (17, 18). Thus, evolutionarily, the conservation of the nic cluster in B. pertussis and B. parapertussis suggests a physiological function for these genes. It is important to understand the contribution of the nic cluster to the growth and life cycle of Bordetella.
In this report, we demonstrate that B. bronchiseptica exerts tight control over the expression of nic genes by the activity of BpsR. We show that the transcriptional control of Bordetella gene expression by NA in the bacterial cell occurs at concentrations that are below those present in animal and humans. The implications of this regulation for bacterial survival in an environmental niche and for the selection of a pyridine source for NAD biosynthesis are discussed. Finally, this work lays the framework for investigating the regulatory mechanisms linking the control of growth, metabolism, and biofilm formation in B. bronchiseptica and expands the role of BpsR in this process.
RESULTS
Severe growth defect of the ΔbpsR strain can be recovered by transfer to fresh growth medium.In a previous report, we made an in-frame deletion of bpsR in the B. bronchiseptica wild-type (WT) strain RB50. In that paper and further described here, we found that culturing RB50 and its derivative, the ΔbpsR mutant (15), for 24 h in Stainer-Scholte (SS) medium, a simple chemically defined medium used for the routine growth of B. bronchiseptica, led to a decreased growth of the mutant strain (Fig. 2A). This phenotype persisted out to 40 h in SS medium, with the bpsR mutant failing to reach WT growth levels (data not shown). However, when these strains were grown in LB medium, there were no significant differences in growth (Fig. 2B).
Growth yields of RB50 (open bars) and ΔbpsR (closed bars) strains grown in various media. x axes, time in hours. (A) Strains grown in SS medium. (B) Strains grown in LB broth. (C) RB50 grown first in SS medium for 12 h and then washed, resuspended in adjusted spent medium (described in Materials and Methods) harvested from either strain RB50 or the ΔbpsR strain, and grown additionally for the indicated times. (D) Strains grown in regular SS medium for 12 h and then resuspended in fresh SS medium and grown additionally for the indicated times. Each data point represents the average from triplicates from one of two or more independent experiments.
To understand the biochemical mechanism of the growth defect, we first tested the hypothesis that the mutant strain was growing poorly due to the production of an inhibitory factor(s) during growth. Experiments were performed with the spent medium collected at 12 h of growth, a time point when the growth of the ΔbpsR strain has stopped. RB50 cells actively growing for 12 h in regular SS medium were resuspended in reconstituted spent medium (as described in Materials and Methods) from either the RB50 or ΔbpsR strain and grown for an additional 24 h. As shown in Fig. 2C, the WT strain grew to a similar optical density at 600 nm (OD600) in spent medium collected from either of the strains. This suggests that the growth defect of the ΔbpsR strain in SS medium is not due to the production of a growth inhibitory factor(s).
SS medium contains sources for multiple amino acids and vitamins, many of which are essential for robust growth (5). Next, we hypothesized that as the ΔbpsR cells grow, the essential nutrients are consumed and eventually exhausted, leading to a slowing and ultimate cessation of growth. To test this hypothesis, we repeated the growth experiment using fresh medium. Bacterial cells were grown in SS medium for 12 h and then washed and resuspended in fresh SS medium and grown for an additional 24 h. As shown in Fig. 2D, both the WT and ΔbpsR cells continued to grow and reached OD600s of ∼16 and 12, respectively, by 24 h. The lower OD600 of the ΔbpsR strain compared to the WT strain is probably due to a slight depletion of the nutrients. These results suggest that the growth defect of the ΔbpsR strain in SS medium can be recovered by transferring to fresh medium.
BpsR controls the expression of genes involved in diverse cellular processes in B. bronchiseptica.Previously, we reported that BpsR is homologous to the MarR family of transcriptional regulators. BpsR functions as a repressor of transcription of the bpsA-D operon, a gene locus that encodes the enzymatic machinery for the synthesis of the Bps polysaccharide (15). This family of regulators controls diverse cellular processes, including those for the survival of bacterial cells under various environmental stresses (15, 19, 20). To understand the mechanism of the growth defect of the ΔbpsR strain, whole-genome transcriptome analysis was conducted to determine which genes are differentially expressed between the WT and ΔbpsR strains grown in LB broth, a medium chosen to avoid any effects on gene expression due to growth differences. The expression levels of 31 genes were elevated and 33 were reduced in the ΔbpsR mutant relative to that in the RB50 strain (see Table S1 in the supplemental material). The genes highly expressed in the mutant strain included fim2 (BB3674), which encodes one of the several fimbrial subunits expressed by B. bronchiseptica (fimbriae have been implicated in cellular adhesion in Bordetella) (21, 22), a transcriptional regulator of the LysR family (BB0088) (a family of protein regulators involved in oxidative stress responses, cell wall shape, quorum sensing, efflux pumps, secretion, motility, nitrogen fixation, and virulence) (23), 13 cellular metabolism associated genes, 3 genes encoding periplasmic/exported lipoproteins, and 2 genes encoding proteins with folding and chaperone activities. The genes identified to have reduced expression in the ΔbpsR mutant included 8 genes categorized as having roles in metabolism and 10 in substrate transport. Additionally, 3 genes encoding membrane proteins, one gene encoding a transcriptional regulator of the LysR family (BB3731), and 5 genes encoding periplasmic/exported lipoproteins were downregulated in the ΔbpsR mutant (Table S1).
Expression of genes homologous to those involved in aerobic catabolism of NA and biosynthesis of NAD is elevated in the ΔbpsR strain.Because of the differential regulation of multiple genes involved in metabolism, we focused on this group for further analyses. nadC, a gene encoding a quinolinate phsophoribosyl transferase, was expressed at higher levels in the mutant strain. In B. bronchiseptica, nadC is involved in the conversion of quinolinic acid to nicotinic acid mononucleotide, which is a precursor for NAD synthesis (24). The expression of a cluster of genes (BB1772 to BB1783) located adjacent to bpsR (BB1771) (Fig. 1) was elevated (2.14- to 38.22-fold) in the ΔbpsR mutant. Previously, this cluster of genes was designated as homologous to the P. putida nic cluster (13). In P. putida strain KT 24440, the nic cluster is involved in the degradation of NA via the maleamate pathway (13). The first step of NA degradation is its conversion to 6-hydroxynicotinic acid (6-HNA) by a nicotinic acid hydroxylase encoded by the nicA and nicB genes. 6-HNA is then converted to 2,5-dihydroxypyrimidine by a 6-HNA monooxygenase encoded by the nicC gene which is further degraded in a stepwise manner via the maleamate pathway by the involvement of the nicX, nicD, nicF, and nicE gene products (Fig. 1) (13, 25, 26).
The enhanced expression of two genes of the nic cluster (BB1781 and BB1772) and the nadC (BB4286) gene in the ΔbpsR strain was independently confirmed by real-time reverse transcriptase PCR (RT-PCR). The expressions of BB1772, BB1781, and nadC were elevated 20.89-, 14.73-, and 38.13-fold, respectively (Fig. 3). As a positive control, the expression of the bpsA gene (Fig. 1) was tested. Consistent with the previous report (15), the expression of bpsA was 2.41-fold higher in the mutant strain.
Relative mRNA expression of BpsR-regulated genes. Relative expression of nadC, BB1781, BB1772, and bpsA was assessed by real-time RT-PCR in RB50 and ΔbpsR strains. The strains were harvested at mid-log phase of growth. Expression in strain RB50 is set to 1. rpoD was used as the standardization control. Fold changes were calculated using the Pfaffl method. Statistical difference was analyzed by a pair-wise fixed reallocation randomization test, run by Rest2009 software (Qiagen). Average values are presented from two biological duplicates with reciprocal technical triplicates and their respective standard deviations. **, P < 0.01.
Addition of NA and nicotinamide restores the growth of the ΔbpsR strain in SS medium.We hypothesized that due to the increased expression of the nic genes in the ΔbpsR strain, NA in the growth medium is depleted earlier than in the growth medium of the WT strain, resulting in the observed slowing/cessation of growth (Fig. 2). If this hypothesis is valid, then the supplementation of SS medium with NA should improve the growth. To test this, RB50 and ΔbpsR strains were grown in either SS medium (base concentration of NA, 0.03 mM) or in media supplemented to contain 0.16, 0.32, and 0.81 mM NA, and the OD600 was monitored at various intervals. In support of our hypothesis, the supplementation of SS medium with NA at 0.16 and 0.32 mM resulted in better growth of the ΔbpsR strain (Fig. 4B and C), whereas 0.81 mM nicotinic acid restored the growth of the mutant strain to that of the RB50 strain (Fig. 4D). As a control, the supplementation of SS media with any of these concentrations of NA did not affect the growth of the WT strain (compare Fig. 4A to D).
Growth yields of RB50 (▲), ΔbpsR (△), and ΔbpsR ΔnicA (○) strains in SS medium supplemented with NA (black) or nicotinamide (gray). (A and E) SS media containing regular concentrations (0.03 mM) of NA. (B to D). SS media supplemented with indicated (top) concentrations of NA or nicotinamide. x axes, time in h. Each data point represents the average from triplicates from one of two or more independent experiments.
In addition to NA, nicotinamide also supports the robust growth of B. bronchiseptica. Thus, we tested if supplementation with nicotinamide would also restore the growth of the bpsR mutant. The addition of nicotinamide also resulted in the restoration of growth of the mutant strain to that of the WT strain. The growth of the bpsR mutant in SS media containing 0.16 and 0.32 mM nicotinamide was better that that in media containing the same concentrations of NA (Fig. 4B and C). This is likely due to the degradation of NA but not nicotinamide by the gene products of the nic cluster.
Deletion of nicA in the bpsR mutant results in the restoration of growth.If the hypothesis that the growth defect of the bpsR mutant is due to an excessive degradation of NA by the activity of genes of the nic cluster is correct, then the prevention of NA conversion in the bpsR mutant should ameliorate the growth defect. To test this, we deleted the nicA gene from the bpsR mutant strain, resulting in the ΔbpsR ΔnicA strain, and compared its growth with that of the ΔbpsR and WT strains. As shown in Fig. 4E, the growth of the ΔbpsR ΔnicA strain was similar to that of the WT strain, further supporting our hypothesis.
Micromolar concentrations of NA induce the expression of nic genes.The induction of catabolic genes involved in a specific pathway in the presence of the metabolite is a common theme in bacteria (27). To test the induction of the nic genes by NA, the levels of β-galactosidase activity produced by RB50 strains carrying single-copy chromosomally integrated nicC-lacZ or nicE-lacZ transcriptional fusions were determined. These reporter strains were grown in SS media supplemented with a range of concentrations of NA. Compared to those in regular SS medium, the expressions of both nicC-lacZ and nicE-lacZ fusions increased by ∼3-fold in SS media containing 0.16 and 0.32 mM nicotinic acid. The addition of higher concentrations of nicotinic acid (0.81 and 1 mM) resulted in further increases (∼4-fold and ∼5-fold, respectively) in β-galactosidase activity (Fig. 5A and B).
Expression of the nicE and nicC lacZ fusions integrated in RB50 strain. Fold changes in the β-galactosidase activity were calculated by dividing β-galactosidase units after growth in SS media supplemented with indicated concentrations of nicotinic acid by β-galactosidase activity units after growth in SS medium supplemented with 0.03 mM NA (SS+NA) or nicotinamide (SS+NM). Each data point represents the average of triplicates from one of two or more independent experiments. The error bars indicate the standard deviations. Statistical significance was calculated from unpaired Student's t tests. *, P < 0.05; **, P < 0.005; ***, P < 0.0005.
The concentration of NA in the SS medium is 0.03 mM. Thus, it was not possible to measure if NA in concentrations lower than 0.03 mM could serve as an inducer. Given that nicotinamide can replace NA for growth (Fig. 4), we determined the lowest concentration of NA that is sufficient for the induction of nic expression. RB50 derivatives harboring the reporter fusions were grown in SS medium either containing 0.03 mM nicotinamide alone or supplemented with different concentrations of NA. There was a slight but statistically significant induction of both the transcriptional fusions at 1 μM NA and higher levels of induction at 5 and 10 μM (Fig. 5). These results suggest that low micromolar concentrations of NA are sufficient to induce nic gene expression.
The induction of the nic genes by NA was also measured in the ΔbpsR strain. The expressions of both nicC and nicE were higher in the bpsR mutant (Fig. 6A). This result is in agreement with results from microarray and real-time RT-PCR assays (Table S1 and Fig. 3). While the presence of NA in the growth medium of the bpsR mutant resulted in a slight increase (1.7-fold) in the expression of nicC, there was no significant difference in the expression of nicE (Fig. 6B). This result suggests that the induction of nicC and nicE gene expression by NA is primarily dependent on the presence of an intact bpsR gene.
Expression of the nicE and nicC lacZ fusions integrated in RB50 (open bars) or ΔbpsR (closed bars) strains. Fold change was calculated by dividing β-galactosidase units of the ΔbpsR strain with that of RB50 after growth in SS media (A) or β-galactosidase units of the ΔbpsR strain after growth in SS medium supplemented with 1 mM NA by β-galactosidase activity units after growth in SS medium (B). Each data point represents the average from triplicates from one of two or more independent experiments. The error bars indicate the standard deviations. Statistical significance was calculated from unpaired Student's t tests. ***, P < 0.0005; ns, not significant.
BpsR binds to nicC-nicE intergenic region and 6-HNA inhibits its DNA binding activity.We have previously shown that BpsR is a DNA binding protein (15). We hypothesized that BpsR represses the transcription of the nic genes by directly binding to DNA. Additionally, on the basis of the induction of the activity of the transcriptional fusions by NA (Fig. 5), we hypothesized that its presence will interfere with the DNA binding activity of BpsR. To examine these hypotheses, we performed electrophoretic mobility shift assays (EMSAs) with purified BpsR and a DNA fragment spanning the intergenic region between the divergently transcribed nicC and nicE genes. As shown in Fig. 7A and B, purified BpsR bound to this fragment. When an EMSA was conducted with NA present in the reaction buffer, there was no observable effect on the DNA binding activity of BpsR (Fig. 7A). Since in the bacterial cell, the catabolism of NA starts with its hydroxylation to 6-HNA (13), we reasoned that 6-HNA is the actual inducer. The inclusion of 0.5 mM 6-HNA in the EMSA reaction buffer resulted in an observable inhibition of the DNA binding activity of BpsR, whereas 1 mM 6-HNA completely prevented the DNA binding (Fig. 7B).
Electrophoretic mobility shift assays. 32P-labeled DNA fragment spanning the region overlapping the nicC and nicE genes was incubated with 0.4 μg of BpsR (+) and the indicated concentrations of either NA (A) or 6-HNA (B).
6-HNA is the in vivo inducer of nic genes.On the basis of the inhibition of the DNA binding activity of BpsR by 6-HNA, we hypothesized that it is the in vivo inducer of nic gene expression. To test this, the nicC-lacZ transcriptional fusion was introduced into the RB50 ΔnicA strain (Fig. 8). In this strain, since the nicA gene is missing, NA cannot be converted to 6-HNA. Therefore, only 6-HNA should function as the inducer of nic genes. In the WT strain, nicC-lacZ expression was strongly induced by both NA and 6-HNA. In comparison, in the RB50 ΔnicA strain, the expression of nicC-lacZ was induced only by 6-HNA. These results suggest that 6-HNA, the product of the first pathway of NA degradation, serves as the in vivo inducer of nic gene expression.
Expression of the nicC-lacZ fusions integrated in RB50 (open bars) or ΔnicA (closed bars) strains. Strains were grown in SS medium alone (SS) or supplemented with 1 mM NA, 6-HNA, or 0.3% DMSO as a solvent control for 6-HNA. Each data point represents the average from triplicates from one of two or more independent experiments. The error bars indicate the standard deviations. Statistical significance was calculated from unpaired Student's t tests. ***, P < 0.0005; ns, not significant.
DISCUSSION
Nicotinic acid is essential for the growth of B. bronchiseptica, since it serves as a source for NAD synthesis by the salvage pathway. NAD and its reduced and phosphorylated derivatives (NADH, NADP, and NADPH) have important roles in electron transfer and multiple metabolic processes by functioning as hydride acceptors and donors. NA is also a carbon, nitrogen, and energy source for many bacteria, which can degrade this molecule by both anaerobic and aerobic routes (13, 28). B. bronchiseptica harbors the genes orthologous to the nic cluster of P. putida, which is involved in aerobic degradation of NA. The presence of both the salvage and degradation pathways for nicotinic acid suggests that this bacterium has evolved to metabolize NA by one of these two pathways and has the ability to regulate its utilization route. In this report, we show that the transcriptional regulator BpsR represses the expression of the nicC and nicE genes. This discovery was prompted by the finding that the bpsR mutant had a severe growth defect in SS medium. Multiple transcriptional assays showed that the expression of the nic genes was elevated in the bpsR mutant. The supplementation of the SS medium with NA resulted in the recovery of the growth of the bpsR mutant, and the deletion of the nicA gene from this mutant did not result in a growth defect. Thus, the poor growth of the bpsR mutant in regular SS medium containing low levels of NA is likely due to the enhanced degradation of NA as a result of the higher expression and activity of the nic degradation pathway. The supplementation of the growth medium with NA also induced the expression of nic genes in the WT strain. Thus, NA serves both as a substrate and as an inducer of the nic pathway.
The induction of the nic operon by NA required BpsR, since the induction of the nicE-lacZ fusion by NA was lost and only a slight induction of the nicC-lacZ fusion occurred in the ΔbpsR strain. Purified BpsR bound to the intervening region of DNA between nicC and nicE genes, suggesting that it represses the transcription of the nic cluster by binding to the promoter regions. The DNA binding activity of BpsR was prevented by 6-HNA and not by NA. Consistent with this result, in the WT strain harboring a deletion of the nicA gene, only 6-HNA but not NA induced the expression of the nicC-lacZ fusion. In the bacterial cell, the catabolism of NA begins with its hydroxylation to 6-HNA by a nicotinic acid hydroxylase encoded by the nicA and nicB genes (13). Thus, we suggest that 6-HNA is the actual in vivo inducer of the nic operon. We propose a regulatory model of the Bordetella nic cluster (Fig. 9). In this model, in the absence of NA, BpsR functions as a repressor of the Bordetella nic genes by binding to the nicE and nicC intergenic region. 6-HNA, after conversion inside the bacterial cell from NA, interacts with BpsR. This interaction then prevents the binding of BpsR to the DNA and alleviates the repression, resulting in the induction of expression of the nic cluster.
Model for repression of the Bordetella nic cluster by BpsR. The DNA sequence of the intergenic region between nicC and nicE genes along with the putative −10 and −35 promoter elements and translational codons (underlined and bolded) are shown. (Top rectangle) When NA is absent or its levels are low, BpsR (designated by oval) binds to the intergenic region of nicC and nicE and prevents transcription. (Bottom rectangle) When NA is present at sufficient levels, it enters the cell and is converted to 6-HNA, which then interacts with BpsR. BpsR cannot bind DNA in the presence of 6-HNA, and transcription of the genes of the nic cluster occurs.
Results very similar to ours were recently reported by Brickman and Armstrong (29). However, these authors did not report a restoration of the growth of the bpsR mutant due to the deletion of the nicA gene, the binding of BpsR to the nic promoter, and the effect of NA and 6-HNA on this binding.
It was previously shown that the presence of NA in the medium markedly increases nadC gene (a quinolinate phosphoribosyl transferase) expression and that this induction is under the control of the NadQ regulator (24), but the regulation of nadC by BpsR has not been reported. In addition to the genes of the nic cluster, we found that the expression of nadC also increased in the ΔbpsR strain, suggesting that BpsR represses nadC expression. It needs to be determined if the NA-mediated induction of nadC is also dependent on bpsR. Additionally, the results of our study and those reported previously by Brickman et al. (24) raise the possibility of an overlap and cross-regulation of Bordetella genes by NadQ and BpsR and require further study.
In light of the absolute requirement of NA for growth, it is reasonable to ask why B. bronchiseptica would maintain the expression of genes involved in its degradation and evolve a regulatory mechanism to control such a process. B. bronchiseptica can survive for extended periods in the environment. RB50, the strain of B. bronchiseptica used here, can grow efficiently on soil extracts (30). We propose that because of the high levels of NA available in the soil, the BpsR-mediated repression of the nic cluster is relieved and the bacterium can then utilize NA as a nutrient source. This then leads to its transient survival in soil of animal farms until it is able to infect its relevant hosts.
Since B. bronchiseptica can infect and cause disease in both animals and humans, what is the physiological relevance of the observed regulation of nic genes in the mammalian hosts? It is reasonable to postulate a tight control over the catabolism of NA because of its utilization for NAD synthesis. We propose that by the BpsR-mediated negative regulation of the nic cluster, B. bronchiseptica prevents unwarranted depletion of NA, thereby ensuring the synthesis of NAD and optimal growth. While the concentration of NA in the mammalian respiratory tract is unknown, its concentration in the whole blood of various animal species (dog, pig, sheep, and horse) and humans is 40 to 70 μM (31, 32). B. bronchiseptica can utilize NA from the plasma transduced into the respiratory epithelium. We show that the induction of nic gene expression by NA occurs at concentrations lower than those present in animals and humans. Thus, it is likely that nic genes are expressed in host organs. Consequently, the degradation of NA by the activity of the nic cluster may pose a survival problem, since sufficient NA may not be available for NAD synthesis. The repression of the nadC gene by BpsR and its subsequent induction by NA may provide Bordetella with an alternative mechanism to produce NAD, if and when the nic cluster is induced in the presence of NA. While the concentration of quinolinic acid in human blood is relatively low, stimulation of the immune system is associated with increased concentrations of quinolinic acid in various organs, including the lungs (33). Additionally, activated macrophages increase the production of quinolinic acid (34). We propose that by exerting control over both biosynthetic and degradative processes involved in NAD synthesis, B. bronchiseptica controls the utilization of pyridines for ensuring optimal growth. In comparison to that in B. bronchiseptica strain RB50, many of the nic orthologs in B. pertussis strain Tohama I and B. parapertussis strain 12822 share a very high sequence identity (≤98%), but some of the nic genes are pseudogenes (Fig. 1). Determining if the nic pathway is functional in B. pertussis and B. parapertussis will reveal if it is required for their growth.
The presence of millimolar concentrations of NA in Bordetella growth medium results in phenotypic modulation, which is characterized by the loss of many of the virulence traits and requires the BvgAS signal transduction system (8–11). When modulated, Bordetella is in the Bvg− phenotypic phase, and in this phase, the expression of the majority of the virulence factors is inhibited, resulting in an attenuation of bacterial virulence. It has been shown that Bordetella can exist in the Bvg− phase in vivo, and a Bvg− phase gene, brtA, is expressed in the rat trachea (35–37). Although as stated above, plasma concentrations of NA in animals and humans are only in the micromolar range, given that NA concentrations in the respiratory tract are not known, one cannot exclude the possibility that the local concentration of NA at this site is in the millimolar range and results in the NA-mediated modulation of gene expression and the switching of Bordetella to the Bvg− phase.
In addition to the genes involved in NAD synthesis and NA degradation, the gene expression analyses reported herein demonstrate that BpsR can activate and repress a number of other genes. These bpsR-regulated genes are predicted to encode proteins serving a diversity of cellular functions. Some of these proteins are periplasmic/membrane/exported proteins and others have function in pathogenesis, transcriptional regulation, cellular metabolism, the binding and transport of substrates, and electron transport. These results suggest that in addition to the previously identified role in regulating polysaccharide production and biofilm formation (15), BpsR controls the expression of genes critical for the growth and pathogenesis of B. bronchiseptica. Continued studies will expand our understanding of the molecular mechanisms by which BpsR mediates the control of diverse cellular processes in Bordetella.
On the basis of the induction of the nic genes by NA, it can be hypothesized that it will also regulate the expression of other bpsR-regulated genes and phenotypes. We previously showed that BpsR represses the expression of the Bps polysaccharide and biofilm formation in B. bronchiseptica (15). Bps is critical for biofilm formation and respiratory tract colonization by both B. bronchiseptica and B. pertussis (38–40). Cummings et al. previously reported that high concentrations of NA led to only a moderate induction in the expression of bpsA, the first gene of the bpsA-D locus (9). Millimolar concentrations of NA can also result in increased biofilm formation in B. bronchiseptica (41, 42). Experiments are under way to determine if NA can regulate Bps polysaccharide levels and influence BpsR-dependent regulation of biofilm formation and the impact of the degradation of NA by the nic cluster in these processes.
MATERIALS AND METHODS
Bacterial strains, plasmids, media, and growth conditions.For growth on solid media, Bordet-Gengou ([BG] Becton Dickinson Microbiology Systems) agar supplemented with 7.5% defibrinated sheep blood and LB agar were used for Bordetella and Escherichia coli, respectively. For growth in liquid media, Bordetella and E. coli strains were grown in Stainer-Scholte (SS) (5) and LB media, respectively. Media were supplemented with the following antibiotics when necessary: ampicillin (100 μg/ml), kanamycin (25 μg/ml), gentamicin sulfate (25 μg/ml), and streptomycin (100 μg/ml). Nicotinic acid and 6-hydroxynicotinic acid solutions were prepared in water and dimethyl sulfoxide (DMSO), respectively. The strains and plasmids used in this study are listed in Table 1. A list of the primers used in this study is included in Table 2. NA and nicotinamide at the indicated concentrations were added to the medium when necessary.
Bacterial strains and plasmids used in the study
List of primers
Bacterial growth.Bordetella strains were grown for 12 to 14 h in either LB or SS medium at 37°C at 90 rpm in a roller drum. Overnight cultures were diluted in fresh medium to yield an OD600 of 0.005. The OD600 was measured at the indicated time points. For medium changes and growth in spent medium, cells of RB50 and ΔbpsR strains were collected by centrifugation (8,000 rpm for 3 min) at 4°C. For growth in spent medium, the supernatant was collected after centrifugation, pH adjusted to 7.5, and filtered through a 0.22-μm syringe filter. For the replenishment of supplements containing essential components, three volumes of adjusted spent medium was mixed with one volume of 4× fresh medium containing the supplement. This medium was designated the reconstituted spent medium. For medium changes, the cells were resuspended in the original volume of fresh SS medium and grown for the indicated times.
Deletion of nicA gene.An in-frame deletion of the nicA gene (BB1777) was constructed using a previously published allelic exchange method (43). A 503-bp KpnI-HindIII fragment including the 5′ region and the first 7 codons of nicA was amplified from the RB50 chromosome using primers MGnicAdnF1Hind and MGnicAdnR1Kpn (Table 2). A 589-bp XbaI-HindIII region corresponding to the 3′ region and the last 27 codons of nicA was amplified using MGnicAupF2Xba and MGnicAupR2Hind (Table 2). These fragments were then ligated into the vector pRE112 to yield the nicA deletion vector pMG26. The resulting plasmid was mobilized from pRK2013 into the RB50 strain and the ΔbpsR strain, and exconjugants were selected on BG agar plates containing streptomycin and chloramphenicol. The loss of the plasmid following a second recombination event was confirmed by sucrose counter selection, and the sensitivity to chloramphenicol was determined as described previously (39, 43, 44). The genotype of the ΔnicA deletion strain was confirmed by PCR.
lacZ transcriptional fusions and β-galactosidase assays.To construct nicC-lacZ and nicE-lacZ transcriptional fusions, DNA fragments encompassing 328 bp upstream and 44 bp downstream of the annotated translational start site of nicC and 344 bp upstream and 49 bp downstream of annotated translational start site of nicE were amplified from the RB50 strain using primers MGRBNicCprom5′ and MGRBNicCprom3′ (nicC) and MGRBNicEprom5′ and MGRBNicEprom3′ (nicE), respectively. The PCR fragments were digested with EcoRI and BamHI and cloned into the suicide lacZ reporter plasmid pEGZ (44), resulting in pMF2 (nicE-lacZ) and pMF3 (nicC-lacZ). This scheme results in the cloning of the respective promoter fragments upstream of the promoterless lacZ gene. The transcriptional fusion constructs were then integrated into the genomes of RB50 and ΔbpsR strains by single crossovers at the nicC and nicE loci, as described previously (10, 45), resulting in the strains RB50nicC-lacZ, RB50nicE-lacZ, ΔbpsRnicC-lacz, and ΔbpsRniE-lacz. β-Galactosidase assays were performed as previously described (10, 45). Two-tailed Student's t tests were used for testing the statistical significance of the data.
RNA extraction.RNA for real-time RT-PCR was prepared from log-phase (A600 of ∼1.0) cultures using a Qiagen RNeasy kit according to the manufacturer's instructions. RNA for microarray analysis was isolated from log-phase cultures using a TRIzol (Life Technologies) extraction according to the manufacturer's instructions.
Real-time RT-PCR.RNA for real-time RT-PCR was prepared from log-phase (A600 of ∼1.0) cultures using a Qiagen RNeasy kit following manufacturer's instructions. Purified RNA was treated with RQ1 DNase I (Promega), and cDNA was synthesized with random hexamers and SuperScript III reverse transcriptase (Invitrogen). Real-time PCR was then performed using a SYBR green system (Applied Biosystems). Differential gene expression was analyzed by means of the Pfaffl method (46) using REST2009 software (Qiagen). The rpoD gene was included as a housekeeping gene for normalization. Real-time PCR analysis was performed with two biological and three technical replicates.
Preparation of labeled cDNA and microarray analysis.RNA isolated from strains RB50 and the ΔbpsR mutant was used in microarray experiments. A 2-color hybridization format was used for microarray analysis. For each biological replicate, RNA extracted from the RB50 strain was used to generate Cy5-labeled cDNA, and RNA extracted from the ΔbpsR mutant was used to generate Cy3-labeled cDNA. Additionally, dye-swap experiments were performed analogously, in which the fluorescent labels were exchanged to ensure that an 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 reaction, followed by buffer exchange, purification, and concentration as described previously (47, 48). The two differentially labeled reactions to be compared were combined and hybridized to a B. bronchiseptica strain RB50-specific long-oligonucleotide microarray (47, 48). The 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 ratios across each array was 1.0. Spots of low quality were identified and filtered out prior to analysis. Ratio data from the biological replicates were compiled and normalized on the basis of the total Cy3% intensity and Cy5% intensity to eliminate slide-to-slide variability. Gene expression data were then normalized to 16S rRNA. The statistical significance of the gene expression changes was assessed using the significant analysis of microarrays (SAM) program (49). A one-class unpaired SAM analysis using a false discovery rate of 0.04% was performed.
BpsR protein expression and purification.The gene fragment encoding the BpsR protein was inserted into a modified pET19 expression vector (Novagen) which encodes an N-terminal polyhistidine tag followed by a rhinovirus 3C protease cleavage site, which permits the removal of the affinity tag (PreScission protease; GE Healthcare). The pET19-bpsR vector was transformed into E. coli C41(DE3) cells (50) for expression. One liter of LB broth supplemented with 50 μg/ml of ampicillin was inoculated with 10 ml of an overnight culture of the C41 cells containing the pET19-bpsR vector. The cells were grown at 37°C to an OD600 of 0.5 and induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at 16°C for 20 h. Prior to the induction with IPTG, the cells were rapidly cooled on ice to 20°C to bring the temperature of the culture close to the induction temperature. The induction of the cells at low temperature was necessary for protein solubility during overexpression. Overexpressed BpsR was purified from cells by a combination of nickel affinity and ion-exchange chromatography (51). After applying the protein to the Ni-nitrilotriacetic acid (Ni-NTA) column, BpsR was treated with PreScission protease according to the manufacturer's directions to remove the His tag. The purity of the peak fractions after each step was verified by SDS-PAGE. Purified BpsR was dialyzed against a buffer containing 50 mM Bis-Tris pH 6.5, 150 mM NaCl, and 5% glycerol.
Electrophoretic mobility shift assays.A 390-bp fragment spanning −328 and +43 relative to the translational start site of the nicC open reading frame (ORF) was generated by PCR using the primers BB1778-79F and BB1778-79R. End-labeling of the PCR products and EMSAs were conducted as previously described (15, 51, 52). Briefly, a [γ-32P]ATP-labeled fragment was incubated with 0.4 μg of purified BpsR protein alone or in the presence of indicated concentrations of NA or 6-HNA at 37°C for 15 min. The samples were then electrophoresed in 4% polyacrylamide gels and visualized by autoradiography.
Data availability.All microarray data are available in Table S1 in the supplemental material and have been deposited in ArrayExpress under accession number E-MTAB-2881.
ACKNOWLEDGMENTS
This project was funded in part with federal funds from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, Department of Health and Human Services by grants 1R21AI123805-0 and R01AI125560.
We thank Dan Wozniak for critical reading of the manuscript, Tracy Nicholson for conducting the microarray experiments and sharing the data with us, and Sandra Armstrong for discussions during the preparation of the manuscript.
FOOTNOTES
- Received 29 November 2017.
- Accepted 19 March 2018.
- Accepted manuscript posted online 26 March 2018.
- Address correspondence to Rajendar Deora, Rajendar.deora{at}osumc.edu.
Citation Guragain M, Jennings-Gee J, Cattelan N, Finger M, Conover MS, Hollis T, Deora R. 2018. The transcriptional regulator BpsR controls the growth of Bordetella bronchiseptica by repressing genes involved in nicotinic acid degradation. J Bacteriol 200:e00712-17. https://doi.org/10.1128/JB.00712-17.
Supplemental material for this article may be found at https://doi.org/10.1128/JB.00712-17.
REFERENCES
- Copyright © 2018 American Society for Microbiology.