|Home | About | Journals | Submit | Contact Us | Français|
Pseudomonas aeruginosa colonizes and can persist in the lungs of cystic fibrosis (CF) patients for decades. Adaptation of P. aeruginosa to the CF lung environment causes various genotypic and phenotypic alterations in the bacterium that facilitate persistence. We showed previously that isocitrate lyase (ICL) activity is constitutively upregulated in the P. aeruginosa CF isolate FRD1. We show here that high ICL activity in FRD1 contributes to increased hydrogen cyanide (HCN) production by this isolate. Disruption of aceA, which encodes ICL, results in reduced cyanide production by FRD1 but does not affect cyanide production in the wound isolate PAO1. Cyanide production is restored to the FRD1aceA mutant by addition of glyoxylate, a product of ICL activity, or glycine to the growth medium. Conversion of glyoxylate to glycine may provide a mechanism for increased cyanide production by P. aeruginosa growing on compounds that activate the glyoxylate pathway. Consistent with this hypothesis, disruption of PA5304, encoding a putative d-amino acid dehydrogenase (DadA), led to decreased cyanide production by FRD1. Cyanide production was restored to the FRD1dadA mutant by the addition of glycine, but not glyoxylate, to the growth medium, suggesting that loss of the ability to convert glyoxylate to glycine was associated with the dadA mutation. This was supported by increased glycine production from toluene-treated FRD1 cells with the addition of glyoxylate compared to FRD1dadA cells. This study indicates a larger role for ICL in the physiology and virulence of chronic isolates of P. aeruginosa than previously recognized.
Pseudomonas aeruginosa causes chronic pulmonary infections that contribute to high fatality rates in cystic fibrosis (CF) patients. Adaptation of P. aeruginosa to the CF lung plays a role in pathogenesis and is characterized by mutagenic events throughout the infection (26). Copious production of alginate is considered to be one of the most important adaptation phenotypes acquired by P. aeruginosa within the CF lung because of the protective functions of this exopolysaccharide. Alginate neutralizes oxygen radicals; inhibits antibiotic penetration, phagocytosis, and complement activation; and acts as a decoy for antimicrobial peptides (8, 12, 16, 18, 25). Increased biofilm formation and production of hydrogen cyanide (HCN) by CF isolates of P. aeruginosa may also facilitate chronic colonization of the CF lung (31, 37).
HCN, a potent inhibitor of cellular respiration, is produced under microaerophilic growth conditions at high cell densities by P. aeruginosa (4, 19). HCN is a virulence factor for P. aeruginosa during infection of Caenorhabditis elegans, and high levels of cyanide have been detected in CF patients who harbored stable P. aeruginosa infections (6, 22). Moreover, cyanide levels are associated with impaired lung function (21, 22).
HCN synthase (glycine dehydrogenase:cyanide-forming, EC 188.8.131.52) appears to be encoded by three genes producing three subunits bearing homology to known dehydrogenases and is predicted to be a flavoenzyme associated with the cytoplasmic membrane (10, 35). Production of HCN synthase is controlled by the quorum sensing regulators LasR and RhlR (19) and the global transcriptional regulator GacA (20). In addition, a transcriptional regulator of alginate biosynthesis, AlgR, regulates the hcnABC synthase gene cluster to increase cyanide production by mucoid CF isolates of P. aeruginosa (2).
We showed previously that isocitrate lyase (ICL) is required for P. aeruginosa virulence on alfalfa and in rat lungs (11). ICL, encoded by aceA, is one of two enzymes specific to the glyoxylate pathway. This pathway allows some bacteria to grow on certain compounds, such as long-chain fatty acids and acetate, as the sole carbon source. Within the CF lung, expression of aceA is highly upregulated in CF P. aeruginosa strains, possibly indicating the presence of inducible substrates in this environment (27). However, deregulation of aceA accounts for constitutive high ICL activity by FRD1 and other CF isolates of P. aeruginosa (11; unpublished data). Both mechanisms suggest that P. aeruginosa benefits from high ICL activity in the CF lung either by utilization of certain compounds as carbon sources or for optimal alginate production (11). In this study, we show that enhanced production of ICL in P. aeruginosa CF lung isolates leads to increased cyanide production. In addition, we have identified PA5304, predicted to encode d-amino acid dehydrogenase (EC 184.108.40.206), as a gene that appears to play a role in cyanide production via conversion of glyoxylate to glycine, a preferred substrate for HCN synthase.
Bacterial strains and plasmids used in this study are listed in Table Table1.1. Unless otherwise indicated, bacteria were cultured in L broth or on L agar. Pseudomonas isolation agar (PIA) supplemented with glycine (10 mM) or glyoxylate (0.2 mM) was used for the HCN assay. UV-visible-light absorption spectra were recorded on a Shimadzu UV-1601 spectrophotometer using 1-cm-path-length cells. A 1:1 mixture of L agar and PIA was used to select for P. aeruginosa transconjugants and to counterselect for Escherichia coli following triparental mating. Media were solidified with 1.5% (wt/vol) Bacto agar (Difco; Becton Dickinson, Inc., Franklin Lakes, NJ). Antibiotics were purchased from Sigma-Aldrich (St. Louis, MO) and used at the indicated concentrations in this study: ampicillin, 100 μg ml−1 for E. coli; carbenicillin, 180 μg ml−1 for P. aeruginosa; gentamicin, 20 μg ml−1 for E. coli and 200 μg for P. aeruginosa; kanamycin, 50 μg ml−1 for E. coli and 800 μg ml−1 for P. aeruginosa.
E. coli strain DH10B was routinely used as a host strain for cloning. DNA was introduced into E. coli by electroporation and into P. aeruginosa by conjugation as previously described (29). Plasmids were purified with QIAprep Spin Miniprep columns (Qiagen, Santa Clarita, CA). DNA fragments were excised from agarose gels and purified using the Qiaex II DNA gel extraction kit (Qiagen) according to the manufacturer's instructions. Restriction enzymes and DNA modification enzymes were purchased from New England Biolabs (Beverly, MA). Either Pfu from Stratagene (La Jolla, CA) or Taq from New England Biolabs was used for PCR amplification of DNA. Oligonucleotides were purchased from Integrated DNA Technologies, Inc. (Coralville, IA).
To generate dadA mutants of P. aeruginosa, the suicide plasmid pLS1755 was constructed. Briefly, a DNA sequence containing ~380 bp upstream of dadA and the coding sequence was PCR amplified from FRD1 cells with Pfu and cloned into the SmaI site of pBluescript K(+). The resulting plasmid was digested with XhoI, and the aacCI gene encoding Gmr as a SalI fragment (23) was cloned into the XhoI site. This was followed by introduction of an origin of transfer (oriT) of RP4 on a ~230-bp HindIII fragment (28). pLS1755 was introduced into P. aeruginosa strains by triparental mating, and potential dadA mutants were isolated as gentamicin-resistant but carbenicillin-sensitive clones, indicating a double crossover event. Replacement of the wild-type dadA gene with the dadA101::aacCI allele was verified by PCR analysis.
To complement the dadA mutation, the wild-type coding sequence was PCR amplified from FRD1 using Pfu and cloned into pLS1155 as a BamHI-EcoRI fragment to position the open reading frame for control by the PT7(A1/04/03) promoter present on pLS1155. An moriT isolated from pLS214 was then cloned as a HindIII fragment to generate pLS1793, and the plasmid was introduced into P. aeruginosa dadA mutants via conjugation. The complemented FRD1dadA and PAO1dadA mutants were designated FRD1dadA+ (LS1794) and PAO1dadA+ (LS1796), respectively.
For cyanide measurements, P. aeruginosa cultures were grown overnight in L broth for 15 h from which 0.1-ml samples were plated onto PIA plates and incubated overnight at 37°C for 24 h. Each plate, without the lid, was placed in a quart-sized Ziploc bag along with 1 ml of 4 M NaOH in another lidless container. The bags were sealed and allowed to incubate at 30°C for 4 h. Cyanide was assayed according to the method of Carterson et al. (2) and normalized to CFU of bacteria recovered from each PIA plate. CFU were averaged from three plates at a density of ~100 CFU/plate. Cyanide levels were quantified by comparison to KCN standards using the same protocol and presented as μM/109 CFU. d-Amino acid amination activity was assayed using toluene-treated cells. Briefly, cultures of P. aeruginosa grown to an optical density at 600 nm of ~2.4 in 50 ml of L broth were mixed 1:1 with saline, vortexed, and centrifuged to harvest the cells. The cells were washed once with saline and then resuspended in 1 ml of 50 mM pyrophosphate buffer. Ten microliters was removed and used to generate dilutions to determine CFU. To the remaining cells, 0.2 ml of toluene (30%) was added, and the tubes were inverted several times and incubated at 25°C for 20 min. To some samples, glyoxylate was added to a final concentration of 5 mM. All samples were then incubated for an additional 20 min at 25°C. The reaction mixtures were centrifuged to pellet the cells, and the supernatant was removed and provided to the University of California—Davis Molecular Structure Facility for gas chromatography analysis. The extracts were treated with sulfosalicylic acid to remove protein and then analyzed on a Hitachi L-8900 amino acid analyzer using a lithium citrate buffer system. Glycine concentrations extracted from approximately 1 × 109 to 2 × 109 CFU are presented.
Glyoxylate is a product of ICL activity and can be converted into glycine by several types of enzymes (e.g., aminotransferases and amino oxidases) in various organisms. Because glycine is a substrate for HCN produced by P. aeruginosa, we questioned whether ICL plays a role in production of HCN.
We showed previously that ICL activity is constitutively and highly expressed in FRD1 compared to the non-CF isolate PAO1. Consistent with our hypothesis that glyoxylate can be converted to glycine, disruption of aceA (encoding ICL) in FRD1 led to reduced cyanide production. The loss of aceA did not significantly affect cyanide production by PAO1 presumably because of low aceA expression in this isolate under the conditions tested. Cyanide production was restored to the FRD1aceA mutant when it was complemented with a wild-type copy of aceA from FRD1 in trans (Fig. (Fig.11).
The increased HCN production by FRD1 compared to PAO1 has been noted previously and is partially attributed to increased expression of hcnABC via the transcriptional regulator AlgR (2). AlgR levels are controlled by the alternative sigma factor AlgT (also know as AlgU), which typically becomes activated following adaptation of P. aeruginosa to the CF lung. Although cyanide production increases in PAO1 upon activation of AlgT (2), the levels are lower than those observed for FRD1. This suggests that additional mechanisms are present in FRD1 that contribute to high cyanide. One possible explanation for this observation might be elevated substrate (i.e., glycine) concentrations for cyanide production within FRD1. Taken together, the data suggest that a product of ICL activity impacts cyanide synthesis in the CF isolate of P. aeruginosa.
Cyanide production by P. aeruginosa can be stimulated by the addition of glycine to the growth medium (3). If glyoxylate is converted to glycine within P. aeruginosa, then disruption of aceA would lead to reduced levels of glyoxylate and glycine. Therefore, we tested whether cyanide production could be restored to the FRD1aceA mutant by addition of glyoxylate or glycine to the growth medium. Both compounds increased production of cyanide by FRD1 and the FRD1aceA mutant (Fig. (Fig.2).2). More importantly, cyanide production was restored to wild-type levels in the FRD1aceA mutant by both compounds, which is consistent with the presented model that glyoxylate is an intermediate in the production of cyanide by P. aeruginosa. Taken together, the data suggest that high cyanide production by the CF isolate FRD1 is due to high ICL activity that produces the intermediate glyoxylate.
In plants and some animals, conversion of glyoxylate to glycine is catalyzed primarily by amino acid (alanine, serine, or glutamate):glyoxylate aminotransferases (14) by transamination. Alternatively, glycine is formed by the reductive amination of glyoxylate by glycine dehydrogenase, also known as glycine oxidase (30). Open reading frames bearing homology to genes encoding both categories of enzymes are present in the P. aeruginosa genome, but most of them have yet to be characterized. PA5304 is predicted to encode a d-amino acid dehydrogenase and has been designated dadA. Because expression of dadA is upregulated in P. aeruginosa growing in CF sputum and also upregulated by AlgT (17, 27, 36), we were interested in whether dadA plays a role in cyanide production. As shown in Fig. Fig.3,3, the FRD1dadA mutant was severely reduced in cyanide production and could not be restored for activity by the addition of glyoxylate to the growth medium. However, the addition of glycine increased cyanide production approximately threefold. Taken together, the phenotype is consistent with the expectation that glyoxylate is not efficiently converted to glycine in the absence of this enzyme. The PAO1dadA mutant did not differ significantly from the parental strain for cyanide production under the conditions tested. This is consistent with the expectation that low levels of glyoxylate are present in PAO1 under noninducing growth conditions for aceA expression. Complementation of the dadA mutants with a wild-type copy of dadA from FRD1 under the control of a regulatable promoter restored cyanide production to the FRD1dadA mutant but did not significantly affect cyanide production by the PAO1dadA mutant (Fig. (Fig.33).
To verify whether dadA plays a role in conversion of glyoxylate to glycine, we treated P. aeruginosa with toluene and determined glycine concentrations with and without the addition of glyoxylate. Extracts prepared from cells in the absence of additional glyoxylate contained relatively low concentrations of glycine (0.926 and 0.892 nmol of glycine for strains FRD1 and FRDdadA, respectively), while the addition of glyoxylate to toluene-treated cells increased glycine concentrations (32.904 and 8.963 nmol of glycine for strains FRD1 and FRD1dadA, respectively). However, FRD1 accumulated approximately 3.67-fold more glycine than did the FRD1dadA mutant, suggesting that disruption of dadA prevented full conversion of glyoxylate to glycine in the mutant. However, the approximately 10-fold increase in glycine induced by the addition of glyoxylate to the FRD1dadA mutant suggests that additional enzymes other than d-amino acid dehydrogenase may participate in the conversion reaction.
Characterization of the alterations acquired by P. aeruginosa within the CF lung are beginning to shed light on P. aeruginosa physiology during chronic infection. ICL appears to impact P. aeruginosa physiology in three primary ways within the CF lung. First, high-level expression of aceA by P. aeruginosa within CF sputum suggests that ICL is required for catabolism of specific carbon sources in this environment (27). Second, we showed previously that ICL is required for optimal production of alginate by FRD1. Finally, in this study we show that upregulation of ICL activity leads to increased cyanide production by increasing the supply of intermediates for HCN synthase (Fig. (Fig.4).4). While glycine is known to stimulate cyanide production by P. aeruginosa, this is the first report to show that glyoxylate, the product of ICL activity, is also stimulatory.
It is unclear whether cyanide facilitates chronic infection of the CF lung by P. aeruginosa. Although P. aeruginosa produces cyanide in the CF lung and production correlates with impaired lung function, the necessity of cyanide for virulence in a mammalian model of infection has not been demonstrated (21). Suggested functions for cyanide in P. aeruginosa pathogenesis include liberation of nutrients from host cells, control of various microbial populations in the CF lung, and possibly energy generation (1, 34).
AlgT is emerging as an activator that controls the switch from acute infection to chronic infection. Enhanced production of alginate and cyanide is associated with chronic isolates of P. aeruginosa and is controlled by AlgT. AlgT becomes activated in P. aeruginosa isolates harbored by CF patients via a common mutation acquired in the anti-sigma factor mucA. Activation of AlgT leads to increased cyanide production in CF P. aeruginosa by induction of AlgR, which in turn induces the hcnABC gene cluster (2). In addition, AlgT positively regulates dadA (36), which we show here appears to play a role in the conversion of glyoxylate to glycine.
In our attempt to define the complete pathway for the production of cyanide from glyoxylate, we examined the common bacterial pathways for the production of glycine from glyoxylate. To this end, we demonstrated that disruption of the dadA gene led to reduced cyanide production by the CF isolate FRD1 but not by PAO1, a wound isolate. This is consistent with elevated levels of glyoxylate being produced by FRD1 compared to PAO1 under the given growth conditions and with a role for d-amino acid dehydrogenase in conversion of glyoxylate to glycine, the substrate for HCN synthase. Furthermore, cyanide production was restored to the FRD1dadA mutant with glycine, indicating that cyanide synthase was intact. However, cyanide production could not be restored by the addition of glyoxylate, indicating loss of glyoxylate-to-glycine conversion activity.
dadA has not previously been implicated in glycine metabolism; rather, its role is believed to be in the degradation pathway for l-alanine and/or d-amino acids. Some bacteria, including Escherichia coli and Salmonella strains, possess a two-gene operon containing dadA and dadX which is involved in catabolism of l-alanine (13, 33). Degradation of l-alanine requires conversion to d-alanine via the catabolic alanine racemase encoded by dadX (5). d-Alanine can then be directly oxidized by d-amino acid dehydrogenase encoded by dadA (32). While the P. aeruginosa genome contains dadA and dadX, the role of these genes in catabolism of d-amino acids has not been fully characterized.
The inability of the FRD1dadA mutant to produce cyanide levels comparable to those of FRD1 when provided with exogenous glycine may reflect a close interaction between d-amino acid dehydrogenase and HCN synthase at the cytoplasmic membrane. Alternatively, it could be a reflection of high intracellular concentrations of glyoxylate competing with glycine for binding to HCN synthase. In either case, the source of the amino group required to produce glycine from glyoxylate by DadA has not been determined. This amino group may arise from the ammonium ion present in CF sputum (7). It is unclear why disruption of dadA in the PAO1 background had little effect on cyanide production. Perhaps glycine in this background comes from another source, e.g., the presence of a nonammonium-requiring enzyme capable of converting glyoxylate to glycine. This is supported by increased glycine production following addition of glyoxylate to toluene-treated cells prepared from the FRD1dadA mutant. Indeed, the PAO1 genome encodes an uncharacterized glycine oxidase (PA1028) that may play a role in conversion of glyoxylate to glycine.
Glycine stimulates HCN production from P. aeruginosa when added at 10 mM or higher to complex medium (1, 5). However, we observed reduced HCN production, or none, from FRD1 when glyoxylate was added at similar concentrations to the growth medium, even though the glyoxylate had no effect on growth rate (data not shown). The mechanism by which glyoxylate inhibits HCN production is unclear. However, we suspect that glyoxylate may compete with glycine for binding to HCN synthase. If so, the effect of external glyoxylate on HCN production would be more pronounced in FRD1 if high ICL activity results in elevated internal levels of glyoxylate. Furthermore, we were unable to detect HCN production following growth of P. aeruginosa on minimal medium amended with various carbon sources (data not shown). This prevented a more detailed analysis of carbon source effect on HCN production. Based on our data, it is clear that HCN production in P. aeruginosa is more complicated than previously shown.
Our study indicates that central metabolism is critical not only for carbon source utilization but also for the production of potential virulence determinants in P. aeruginosa. In previous studies, we showed that ICL and glucose-6-phosphate dehydrogenase (Zwf) are deregulated in CF P. aeruginosa (11, 24). Like ICL, Zwf is required for optimal production of alginate and cyanide by P. aeruginosa (6, 24). The lack of phenotypes associated with disruption of aceA and dadA for cyanide production in PAO1 indicates that the physiology of CF isolates of P. aeruginosa is highly altered following adaptation to the CF lung. Moreover, it highlights the need for continued characterization of CF P. aeruginosa physiology to better understand how adaptation facilitates persistence in this important human niche.
This research was supported by funds from the Auburn University Biogrant awarded to L. A. Silo-Suh.
We thank Sang-Jin Suh for critical reading of the manuscript.
Published ahead of print on 21 August 2009.