Nutritional Cues Control Pseudomonas aeruginosa Multicellular Behavior in Cystic Fibrosis Sputum

Abstract

The sputum (mucus) layer of the cystic fibrosis (CF) lung is a complex substrate that provides Pseudomonas aeruginosa with carbon and energy to support high-density growth during chronic colonization. Unfortunately, the CF lung sputum layer has been difficult to mimic in animal models of CF disease, and mechanistic studies of P. aeruginosa physiology during growth in CF sputum are hampered by its complexity. In this study, we performed chromatographic and enzymatic analyses of CF sputum to develop a defined, synthetic CF sputum medium (SCFM) that mimics the nutritional composition of CF sputum. Importantly, P. aeruginosa displays similar phenotypes during growth in CF sputum and in SCFM, including similar growth rates, gene expression profiles, carbon substrate preferences, and cell-cell signaling profiles. Using SCFM, we provide evidence that aromatic amino acids serve as nutritional cues that influence cell-cell signaling and antimicrobial activity of P. aeruginosa during growth in CF sputum.

A key concept in bacterial pathogenesis is the ability of invading pathogens to obtain sufficient carbon and energy from the host for in vivo growth. Although Garber originally proposed the host as a growth medium over 40 years ago (12), the nutritional environment of most infection sites is poorly defined and often inadequately modeled by laboratory growth media. This lack of knowledge, combined with the limited utility of many animal models, provides significant challenges for mechanistic studies aimed at examining host nutrients as mediators of colonization and disease. To overcome these challenges, it is critical both to define the nutritional composition of key infection sites and to study bacterial physiology in the context of in vivo-relevant growth substrates.

The heritable disease cystic fibrosis (CF) is an archetype for the development of nutritional models with which to study bacterial pathogenesis. A hallmark of CF disease is the accumulation of large volumes of sputum (mucus) within the lungs, which diminishes the host's ability to clear bacterial infections (17, 31). The viscous CF lung sputum provides bacteria with a nutritionally rich growth environment composed of host- and bacterial-derived factors (17, 38). The opportunistic pathogen Pseudomonas aeruginosa chronically colonizes the CF lung, where it often grows to high cell densities in CF sputum (>10⁹ cells/ml sputum). Although many other bacterial species persist and grow in the CF lung, chronic P. aeruginosa infection is likely the most clinically relevant, as it is correlated with declining lung function (17). Mechanistically, P. aeruginosa colonization and progression to chronic infection is poorly understood, although potential contributing factors are high-density growth and enhanced fitness of P. aeruginosa in CF sputum. P. aeruginosa fitness has been linked to nutritional components in CF sputum (39), thus necessitating the development of a versatile model that allows examination of CF sputum nutritional cues.

The growth environment impacts several clinically relevant phenotypes in P. aeruginosa. For example, individual carbon and nitrogen sources have been shown to modulate P. aeruginosa in vitro biofilm development (6, 23, 53, 56) and surface motility (25). P. aeruginosa cell-cell signaling (quorum sensing) is also influenced by nutritional cues (9, 39, 63). Growth of P. aeruginosa in CF sputum promotes increased synthesis of the cell-cell signaling molecule 2-heptyl-3-hydroxy-4-quinolone (the Pseudomonas quinolone signal [PQS]) (39), and recent studies implicate aromatic amino acids as potential mediators of this phenotype (39). However, definitive studies aimed at examining how individual CF sputum components, such as aromatic amino acids, impact P. aeruginosa physiology are difficult due to the complexity of sputum. As a means of circumventing this problem, we describe the development of a synthetic CF sputum medium (SCFM) that nutritionally mimics CF sputum. Studies using SCFM reveal that the expression of P. aeruginosa nutritionally controlled genes is similar in CF sputum and in SCFM and that CF sputum-specific phenotypes, including increased PQS production, can be recapitulated in SCFM. In addition, carbon consumption analyses provide evidence that specific amino acids support rapid/high-density growth of P. aeruginosa in CF sputum.

MATERIALS AND METHODS

Bacterial strains and media.

P. aeruginosa strain UCBPP-PA14 (44) and the isogenic pqsA transposon insertion mutant were obtained from the MGH-Parabiosys:NHLBI Program for Genomic Applications (http://pga.mgh.harvard.edu/cgi-bin/pa14/mutants/retrieve.cgi). Staphylococcus aureus strain Xen 36 (Xenogen Biosciences), constitutively expressing the luxABCDE genes, was used for antimicrobial studies. P. aeruginosa was routinely cultured on tryptic soy agar plates, and S. aureus was grown in brain heart infusion (BHI) broth/agar. Morpholinepropanesulfonic acid (MOPS) glucose medium was previously described (39). Bacterial growth was assessed by monitoring optical density at 600 nm (OD₆₀₀). All cultures were incubated at 37°C with shaking at 250 rpm unless otherwise noted.

CF sputum sampling and preparation.

Sputum samples from 12 adult patients with nonexacerbating CF with P. aeruginosa concentrations of ≤10⁸ bacteria/ml sputum were collected by expectoration into sterile containers and stored at −80°C prior to lyophilization, as previously described (39). For chromatographic analyses, powdered CF sputum was weighed and resuspended in sterile deionized water to a final concentration of 10, 20, or 25% (vol/vol). Fifty milliliters of CF sputum corresponded to approximately 2 g dry weight (39). Resuspended sputum was homogenized using a tip sonicator (Branson Ultrasonics) as previously described (39), and insoluble material was subsequently removed by centrifugation at 16,000 × g for 5 min. The resulting supernatant was filtered through a 0.45-μm-pore-size syringe filter and stored at −20°C prior to analysis.

Chromatographic and enzymatic analyses of CF sputum.

Anion concentrations in CF sputum supernatants were determined by anion-exchange chromatography using a Dionex DX5000 system and an AS-4A column. Cation concentrations were determined by cation-exchange chromatography using a Dionex DX500 system and a CA12A analytical ion-exchange column. Data analysis was performed using Peaknet software (Dionex Corporation).

Free amino acid levels in CF sputum supernatants were determined by cation-exchange chromatography. CF sputum supernatants were prepared for free amino acid analysis as follows. CF sputum supernatant (100 μl) was mixed with 100 nmol of the internal standard β-(2-thienyl)-dl-alanine and an equal volume of 6% trichloroacetic acid. Samples were vortexed for 30 s and incubated at 89°C for 1 min prior to centrifugation for 10 min at 1,500 × g. Supernatants were removed and extracted three times with 600 μl ethyl acetate, and organic fractions were discarded. Aqueous fractions were subsequently filtered through 0.45-μm-pore-size spin filter columns by centrifugation at 16,000 × g for 5 min. Spin filter columns were prewashed with 0.01 N HCl. Filtrates were evaporated to dryness by vacuum centrifugation (Eppendorf) and resuspended in 200 μl 0.01 N HCl. Cation-exchange chromatography and data analysis were performed at the University of Oklahoma Health Sciences Center Molecular Biology-Proteomics Facility, using a Beckman system Gold model 126 high-performance liquid chromatography amino acid analyzer. Lactate levels in CF sputum supernatants were measured with the lactate assay kit A-108 (Biomedical Research Service Center, SUNY at Buffalo, NY). This assay measures both d- and l-isomers of lactate. Glucose levels were measured with a glucose (hexokinase) assay kit (Sigma).

SCFM.

SCFM was developed from the average concentrations of ions, free amino acids, glucose, and lactate in CF sputum samples. For amino acids, the percentage of total free amino acids in each CF sputum sample was determined for individual amino acids. The average percentage of each amino acid was extrapolated to a final total amino acid concentration of 19 mM for SCFM (for raw data, see Table S2 in the supplemental material). Amino acids were maintained as 100-mM stocks in deionized water and stored in the dark at 4°C. Tyrosine, aspartate, and tryptophan were resuspended in 1.0 M, 0.5 M, and 0.2 M NaOH, respectively. Lactate stocks were adjusted to a pH of 7.0 with NaOH.

For SCFM, amino acids were added from 100-mM stocks to a buffered base (6.5 ml 0.2 M NaH₂PO₄, 6.25 ml 0.2 M Na₂HPO₄, 0.348 ml 1 M KNO₃, 0.122 g NH₄Cl, 1.114 g KCl, 3.03 g NaCl, 10 mM MOPS, 779.6 ml deionized water) in the following volumes: l-aspartate, 8.27 ml; l-threonine, 10.72 ml; l-serine, 14.46 ml; l-glutamate·HCl, 15.49 ml; l-proline, 16.61 ml; l-glycine, 12.03 ml; l-alanine, 17.8 ml; l-cysteine·HCl, 1.6 ml; l-valine, 11.17 ml; l-methionine, 6.33 ml; l-isoleucine, 11.2 ml; l-leucine, 16.09 ml; l-tyrosine, 8.02 ml; l-phenylalanine, 5.3 ml; l-ornithine·HCl, 6.76 ml; l-lysine·HCl, 21.28 ml; l-histidine·HCl, 5.19 ml; l-tryptophan, 0.13 ml; and l-arginine·HCl, 3.06 ml. SCFM was adjusted to pH 6.8 and filter sterilized through a 0.2-μm-pore-size filter. After sterilization, the following sterile components were added per liter: 1.754 ml 1 M CaCl₂, 0.606 ml 1 M MgCl₂, and 1 ml 3.6 mM FeSO₄·7H₂O. Three milliliters 1 M d-glucose and 9.3 ml 1 M l-lactate were added for gene expression and carbon consumption experiments. For some experiments, aromatic amino acids were removed from SCFM and replaced with an equimolar amount of serine (referred to as SCFM-aromatics).

Carbon consumption analyses.

Carbon consumption was examined during P. aeruginosa growth in CF sputum and in SCFM. For CF sputum, P. aeruginosa was grown overnight in SCFM and diluted to an OD₆₀₀ of 0.1 in fresh SCFM. At an OD₆₀₀ of 0.4, cells were pelleted by centrifugation and washed twice in a carbon-free MOPS-buffered medium (39). Washed cells were used to inoculate 20% CF sputum medium (39) to an OD₆₀₀ of 0.05. Sputum samples from two individuals with CF were independently examined. At the time of inoculation and at multiple points during growth, samples were removed for carbon analysis. Samples were centrifuged at 16,000 × g for 5 min and filtered through 0.45-μm-pore-size syringe filters immediately after harvesting. Amino acid, glucose, and lactate analyses were performed as described above for CF sputum. All incubations in CF sputum were performed at 37°C with shaking at 250 rpm. For SCFM, P. aeruginosa was grown overnight in SCFM and diluted to an OD₆₀₀ of 0.05 in fresh SCFM. At an OD₆₀₀ of 0.5, cells were pelleted by centrifugation and washed twice in carbon-free SCFM. Cells were incubated in carbon-free SCFM for 30 min and then used to inoculate fresh SCFM to an OD₆₀₀ of 0.04. Samples were removed at the time of inoculation and at multiple points throughout growth for carbon analysis, as performed for CF sputum. Incubations in SCFM for carbon consumption experiments were carried out at 37°C with shaking at 160 rpm.

Global expression profiling.

P. aeruginosa cultures grown in SCFM were harvested at an OD₆₀₀ of 0.1 to 0.2 and mixed 1:1 with the RNA stabilizing agent RNAlater (Ambion). RNA was isolated using RNeasy minicolumns (QIAGEN), and cDNA was prepared for hybridization to Affymetrix GeneChip microarrays as previously described (34, 39, 45, 52). DNA contamination of RNA samples was assessed by PCR amplification of the P. aeruginosa rplU gene with the primers rplU-for (5′-CGCAGTGATTGTTACCGGTG-3′) and rplU-rev (5′-AGGCCTGAATGCCGGTGATC-3′). The rplU gene is commonly used to monitor DNA contamination in P. aeruginosa RNA samples (27, 29, 68). Agarose gel electrophoresis was used to assess RNA integrity. Washing, staining, and scanning of GeneChips was performed at the University of Iowa DNA core facility, using an Affymetrix fluidics station. Transcriptome data for SCFM-grown P. aeruginosa were compared to that of CF sputum medium-grown bacteria and MOPS-glucose-grown bacteria (39). GeneChip analyses were performed in duplicate (SCFM and MOPS-glucose medium) or triplicate (CF sputum medium) for each condition tested, and data were analyzed using GeneChip operating software version 1.4. Results were reported as differentially regulated based on pairwise comparisons of all GeneChips (P ≤ 0.05).

Quinolone and pyocyanin analyses.

To quantify PQS, P. aeruginosa was grown to an OD₆₀₀ of 0.1 in SCFM or SCFM-aromatics (SCFM in which aromatic amino acids had been replaced with an equimolar amount of serine). Cultures were extracted three times with an equal volume of acidified ethyl acetate (600 μl glacial acetic acid/4 liters ethyl acetate), and the organic layer was removed and evaporated under a continuous stream of N₂. Extracts were analyzed by thin-layer chromatography as previously described (11, 35, 40). Synthetic PQS was used as a standard and n-fold differences in PQS levels were determined by spot densitometry using a FluorChem 8900 gel imager (Alpha Innotech). To quantify other quinolones, P. aeruginosa was grown as outlined above to an OD₆₀₀ of 0.8, and 40-ml volumes were extracted with an equal volume of acidified ethyl acetate. Organic extracts were dried by rotoevaporation, and quinolones were quantified by liquid chromatography-mass spectrometry as described previously (30, 35) at the University of Oklahoma Health Sciences Center Molecular Biology-Proteomics Facility. Briefly, dried extracts were resuspended in 500 μl of a 30% dimethylformamide-30% acetonitrile-1% acetic acid solution, and 125 μl was applied to a Michrom Paradigm model high-performance liquid chromatography system. Elution was performed with a 30 to 100% gradient of a 97% acetonitrile-2% water-1% acetic acid solution over 40 min with a flow rate of 40 μl/min. Quinolone compounds were quantified with a precursor ion scan of 172 (30, 35). To quantify pyocyanin, P. aeruginosa was grown for 24 h in SCFM or SCFM-aromatics. Samples (10 ml) were removed and extracted with 5 ml chloroform, and chloroform fractions were subsequently extracted with 1 ml 0.01 N HCl. Pyocyanin levels in the aqueous phase were measured spectrophotometrically (A₅₂₀) as previously described (8, 35, 65).

S. aureus antimicrobial assays.

For antimicrobial assays, P. aeruginosa culture supernatants were prepared by growing bacteria at 37°C for 18 h in SCFM or SCFM-aromatics. Bacteria were then removed by centrifugation at 6,000 × g for 15 min followed by filtration through 0.22-μm-pore-size syringe filters. Exponential S. aureus Xen-36 was diluted to an OD₆₀₀ of 0.1 in BHI, and 50-μl samples were added to wells of a 96-well plate. Plates were incubated for 15 min at 37°C before addition of 150 μl of P. aeruginosa culture supernatants, 3% H₂O₂, or sterile BHI. S. aureus luminescence was monitored at 2-min intervals for 10 min, using a FLUOstar luminometer (BMG). The internal temperature within the luminometer was maintained at 37°C, and plates were shaken at 150 rpm for 15 s prior to luminescence measurement.

RESULTS

Construction of a defined medium that nutritionally mimics CF sputum.

To create a defined CF sputum medium, we collected and analyzed CF sputum samples from multiple individuals for levels of free amino acids, cations, anions, glucose, and lactate. CF sputum samples utilized for these analyses were obtained from adults by expectoration and contained P. aeruginosa concentrations of ≤10⁸ bacteria/ml sputum. From these analyses, a defined medium referred to as SCFM was devised (Table 1). A detailed description of SCFM is provided in Materials and Methods and in Tables S1 and S2 in the supplemental material. As observed in previous studies (49), CF sputum chloride levels measured in this study varied over an approximate fivefold range; consequently, this anion was used to balance the salts base of SCFM. Therefore, the final chloride concentration in SCFM (79.1 mM) is higher than the average determined in this study (54.6 mM), although still within the range of CF sputum chloride levels (Table 1 and see Table S1 in the supplemental material). An initial pH of 6.8 was chosen for SCFM, since the average pH of submucosal gland fluid in CF patients is ∼6.6 to 7.0 (19, 55), measurements of airway mucus in explanted CF lungs have detected mucus pH values ranging from ∼6.0 to 6.9 (69), and the pH of lyophilized CF sputum samples resuspended in deionized water varies from 6.8 to 7.4 (data not shown). Since SCFM does not possess the native buffering system of the lung, it was supplemented with 10 mM MOPS to supply additional buffering capacity.

TABLE 1.

Composition of SCFM

Components	Concn (mM)a
Ions
Na+	66.6
K+	15.8
NH4+	2.3
Ca2+	1.7
Mg2+	0.6
Cl−	79.1
NO3−	0.35
PO42−	2.5
SO42−	0.27
Amino acids
Serine	1.4
Threonine	1.0
Alanine	1.8
Glycine	1.2
Proline	1.7
Isoleucine	1.1
Leucine	1.6
Valine	1.1
Aspartate	0.8
Glutamate	1.5
Phenylalanine	0.5
Tyrosine	0.8
Tryptophan	0.01
Lysine	2.1
Histidine	0.5
Arginine	0.3
Ornithine	0.7
Cysteine	0.2
Methionine	0.6
Other
Glucose	3.2
Lactate	9
FeSO4 (μM)b	3.6

^aSCFM component concentrations are based on the average concentrations measured in CF sputum samples (see Tables S1 to S3 in the supplemental material for raw data). Amino acid values do not include those present within peptides but instead represent “free” levels in CF sputum. Values in this table have been rounded for clarity.

^bThe iron level is adapted from previous studies (6, 46, 57, 58).

P. aeruginosa growth and gene expression in SCFM.

P. aeruginosa displays distinct phenotypes during growth in CF sputum (39, 66), and nutritional cues within sputum have been proposed to mediate several of these phenotypes. The goal of this study was to develop a versatile nutritional model for P. aeruginosa growth in CF sputum. The utility of SCFM as a model for CF sputum necessitates that the growth and expression of nutritionally controlled genes in SCFM are similar to those in CF sputum but distinct from common laboratory media. Examination of growth in SCFM and in CF sputum reveals that P. aeruginosa grows well in SCFM with a doubling time (∼32 min) similar to that observed for CF sputum (Fig. 1). For gene expression analyses, the transcriptome of SCFM-grown P. aeruginosa was compared to that of glucose-grown and CF sputum-grown bacteria. The comparison to glucose-grown bacteria provides an evaluation of nutritionally regulated genes similar to that performed previously for CF sputum (39), and the comparison to CF sputum-grown bacteria allows identification of nutritional genes differentially regulated in CF sputum and SCFM. Previous studies of CF sputum and glucose-grown P. aeruginosa revealed 147 genes differentially regulated during growth in CF sputum (39). Of these 147 genes, 80 (54%) were also differentially regulated during growth in SCFM compared to that in glucose (Table 2i also see Table S3 in the supplemental material). Importantly, genes involved in amino acid catabolism and production of the quorum sensing signaling molecule PQS were significantly up-regulated during growth in SCFM (Table 2), to levels that were similar to those observed during growth in CF sputum (39). Of the 67 remaining genes, most were iron regulated or involved in chemotaxis/flagellar motility (see Table S3 in the supplemental material). Since each of these processes has been shown to be impacted by specific host factors in the CF lung, including host chelators and neutrophil elastase (7, 20, 46, 54, 57, 58, 62), it is not surprising that their expression patterns in CF sputum were distinct from those in SCFM. Comparisons of CF sputum and SCFM transcriptomes revealed that 137 genes were differentially regulated at least fivefold (see Table S6 in the supplemental material). Of these genes, few have been implicated in P. aeruginosa nutrient acquisition and catabolism. The acetyl-coenzyme A synthetase (acsA) gene was down-regulated fivefold during growth in SCFM compared to that in CF sputum and is important for P. aeruginosa acetate catabolism (26). The PA3758, PA3759, and PA3761 genes were also down-regulated in SCFM and may be involved in catabolism of N-acetylglucosamine and glucosamine (41). Finally, the dehydroorotase pyrQ (pyrC2) gene was down-regulated in SCFM; however, this gene plays a redundant role in de novo pyrimidine biosynthesis and has no known role in nucleotide catabolism (3).

FIG. 1.

Growth of P. aeruginosa in SCFM and in CF sputum medium. Growth was monitored by dilution plating for viable cell counts, and data are shown as CFU/ml. Bacteria were harvested at ∼10⁸ CFU/ml for Affymetrix GeneChip analyses. Representative growth curves are shown. The CF sputum growth curve data are from Palmer et al. (39).

TABLE 2.

Gene expression profiles in CF sputum and SCFM^a

ORF function	ORF	Gene	Function or class	Fold change in gene expression
				CF sputum versus glucoseb	SCFM versus glucosec
Amino acid transport	PA0782	putA	Proline dehydrogenase PutA	4.3	18
and degradation	PA0865	hpd	4-hydroxyphenylpyruvate dioxygenase	66	45
	PA0866	aroP2	Aromatic amino acid transport protein	13	9
	PA0870	phhC	Aromatic amino acid aminotransferase	9	6
	PA0871	phhB	Pterin-4-α-carbinolamine dehydratase	5	4.3
	PA0872	phhA	Phenylalanine-4-hydroxylased	32	23
	PA0897	aruG	Arginine/ornithine succinyltransferase AII subunit	3	3.7
	PA0898	aruD	Succinylglutamate-5-semialdehyde dehydrogenase	2.7	3.4
	PA2001	atoB	Acetyl-coenzyme A acetyltransferase	16	9
	PA2007	maiA	Maleylacetoacetate isomerase	8	8
	PA2008	fahA	Fumarylacetoacetase	9	9
	PA2009	hmgA	Homogentisate 1,2-dioxygenase	11	10
	PA2247	bkdA1	2-oxoisovalerate dehydrogenase, α-subunit	20	7
	PA2248	bkdA2	2-oxoisovalerate dehydrogenase, β-subunit	19	6
	PA2249	bkdB	Branched-chain α-keto acid dehydrogenase	13	5
	PA2250	lpdV	Lipoamide dehydrogenase-Val	19	6
	PA3766		Probable aromatic amino acid transporter	2.8	3.6
	PA4470	fumC1	Fumarate hydratase	6	3.2
	PA5302	dadX	Catabolic alanine racemase	9	10
	PA5304	dadA	d-amino acid dehydrogenase, small subunit	20	15
Glucose transport and	PA2322		Gluconate permease	−5.5	−20
metabolism	PA2323		Probable glyceraldehyde-3-phosphate dehydrogenase	−3.8	−35
	PA3181		2-keto-3-deoxy-6-phosphogluconate aldolase	−3	−6
	PA3186	oprB	Carbohydrate outer membrane porin	−2.7	−6
	PA3195	gapA	Glyceraldehyde-3-phosphate dehydrogenase	−3.2	−6
Flagellar synthesis and	PA1092	fliC	Flagellin type B	−21	NC
chemotaxis	PA2867		Probable chemotaxis transducer	−8	NC
	PA4307	pctC	Chemotactic transducer PctC	−8	NC
	PA4310	pctB	Chemotactic transducer PctB	−23	NC
Pseudomonas quinolone	PA0996	pqsA	Probable coenzyme A ligase	18	6
signaling	PA0997	pqsB	β-keto-acyl-acyl-carrier protein synthase	17	8
	PA0998	pqsC	β-keto-acyl-acyl-carrier protein synthase	19	7
	PA0999	pqsD	3-oxoacyl-[acyl-carrier-protein] synthase III	17	6
	PA1000	pqsE	Quinolone signal response protein	19	5
	PA1001	phnA	Anthranilate synthase component I	22	7
	PA1002	phnB	Anthranilate synthase component II	14	3.9

^aOpen reading frame (ORF), gene name, and function data were obtained from the P. aeruginosa genome website (www.pseudomonas.com).

^bRegulation (n-fold change) of genes differentially expressed during P. aeruginosa growth in CF sputum medium compared to that in glucose; a positive number indicates an up-regulation of the gene during growth in sputum. Data are from Palmer et al. (39).

^cRegulation (n-fold change) of genes differentially expressed during P. aeruginosa growth in SCFM as compared to growth in glucose; a positive number indicates an up-regulation of the gene during growth in SCFM. NC indicates no change in mRNA levels as determined by GeneChip operating software version 1.4.

^dPhenylalanine-4-hydroxylase is involved in the synthesis of tyrosine and the degradation of phenylalanine.

P. aeruginosa exhibits similar carbon preferences during growth in SCFM and in CF sputum.

Since carbon catabolism impacts P. aeruginosa virulence and biofilm formation (39, 43, 47, 50, 53), carbon utilization profiles were examined in SCFM and in CF sputum medium to determine the primary carbon sources consumed by P. aeruginosa during growth in these media. For these experiments, concentrations of individual carbon sources were monitored throughout P. aeruginosa growth in CF sputum medium and in SCFM. Results from these experiments revealed that P. aeruginosa carbon consumption profiles are remarkably similar in CF sputum medium and in SCFM (see Tables S4 and S5 in the supplemental material), with six carbon sources utilized first in both medium types, i.e., proline, alanine, arginine, glutamate, aspartate, and lactate (Table 3). It was not possible to distinguish a preference among these six carbon sources, and the results suggest they are consumed concomitantly. These studies indicate that P. aeruginosa carbon preferences are similar in CF sputum and in SCFM and that SCFM serves as an in vivo-relevant medium with which to mechanistically evaluate carbon substrate utilization by P. aeruginosa.

TABLE 3.

Preferred carbon sources in CF sputum and SCFM^a

Carbon source	% ± SD of remaining carbon after growth in:
	CF sputum	SCFM
Proline	18 ± 0	20 ± 9
Alanine	26 ± 13	31 ± 10
Arginine	37 ± 3	0
Lactate	45 ± 3	54 ± 8
Glutamate	46 ± 20	41 ± 9
Aspartate	51 ± 10	24 ± 9

^aThe six carbon sources quantitatively consumed first by P. aeruginosa during growth in 20% CF sputum medium and in SCFM. Shown are the percentages ± standard deviations (SD) of the initial carbon source remaining after P. aeruginosa growth in CF sputum (2 h) and in SCFM (3 h). A later time point was used for SCFM since it contains 5 times more carbon than 20% CF sputum. See Materials and Methods for experimental details and Tables S4 and S5 in the supplemental material for data on all available carbon sources.

P. aeruginosa signaling is similar in CF sputum and in SCFM.

We previously observed increased PQS production during growth in CF sputum (39). Since genes involved in PQS biosynthesis were significantly up-regulated during growth in SCFM (Table 2), we anticipated that PQS levels of SCFM-grown P. aeruginosa would resemble those observed for CF sputum-grown bacteria. Analysis of PQS levels revealed that, as in CF sputum, P. aeruginosa produces high levels of PQS during growth in SCFM (Fig. 2A).

FIG. 2.

PQS production is enhanced during growth in SCFM and in CF sputum. (A) PQS levels were assessed for P. aeruginosa grown in MOPS-glucose laboratory medium (39), SCFM, and CF sputum medium. Bacteria were sampled in exponential phase at an OD₆₀₀ of 0.1. (B) P. aeruginosa was grown in SCFM and in SCFM without aromatic amino acids to an OD₆₀₀ of 0.1, and PQS levels were determined. Error bars represent standard deviations.

The SCFM-mediated increase in PQS biosynthesis suggested that nutritional components within CF sputum enhance PQS biosynthesis. Since SCFM is a defined medium, the impact of individual nutritional components on PQS biosynthesis can be assessed by simply removing them from the medium. Since previous studies implicated aromatic amino acids as modulators of PQS biosynthesis (9, 39), we began by examining their role in PQS biosynthesis. Removal of aromatic amino acids from SCFM and replacement with equimolar levels of serine, which does not impact PQS production (39), resulted in a significant decline in PQS production (Fig. 2B), to levels similar to those observed for glucose-grown bacteria. In addition, levels of several other quinolone molecules that share biosynthetic components with PQS were also significantly reduced upon removal of aromatic amino acids from SCFM (Fig. 3A). It is important to note that no differences in growth rate were observed upon removal of aromatic amino acids (data not shown), and samples were removed for analyses at equivalent bacterial densities.

FIG. 3.

Aromatic amino acids impact quinolone and pyocyanin production in P. aeruginosa. (A) P. aeruginosa cultures grown to an OD₆₀₀ of 0.8 in SCFM or SCFM lacking aromatic amino acids were extracted with acidified ethyl acetate, and quinolone levels were analyzed by liquid chromatography-mass spectrometry. Shown are 4-hydroxy-2-heptylquinoline N-oxide (HQNO), 4-hydroxy-2-heptylquinoline (HHQ), 4-hydroxy-2-octylquinoline (HOQ), 4-hydroxy-2-nonylquinoline N-oxide (NQNO), and 4-hydroxy-2-nonylquinoline (HNQ). (B) Pyocyanin levels in P. aeruginosa cultures grown for 24 h in SCFM or in SCFM lacking aromatic amino acids. Pyocyanin levels were determined after extraction by monitoring A₅₂₀ as outlined in Materials and Methods. Error bars represent standard deviations.

Due to enhanced PQS biosynthesis, CF sputum-grown P. aeruginosa increases the production of several PQS-controlled factors, including the secondary metabolite pyocyanin (39). Based on these results, we reasoned that the presence of aromatic amino acids would significantly enhance pyocyanin production during growth in SCFM. As observed with CF sputum (39), P. aeruginosa produced high levels of pyocyanin during growth in SCFM, and removal of aromatic amino acids decreased pyocyanin levels by approximately 2.5-fold (Fig. 3B).

Aromatic amino acids enhance P. aeruginosa antimicrobial activity in SCFM.

Along with increased production of PQS and PQS-controlled factors, recent studies in our laboratory revealed that P. aeruginosa exhibits increased lysis of the CF lung coinhabitant S. aureus during growth in CF sputum (39). P. aeruginosa lysis of S. aureus requires a functional quinolone biosynthesis system (34), and it is likely mediated by multiple factors, including pyocyanin and antimicrobial quinolones. Given the impact of aromatic amino acids on quinolone and pyocyanin levels (Fig. 2 and 3), we hypothesized that high levels of P. aeruginosa antimicrobial activity would be observed for SCFM and that this activity would be dependent on the presence of aromatic amino acids. To test this hypothesis, we examined the antistaphylococcal activity of P. aeruginosa supernatants during growth in SCFM in the presence and absence of aromatic amino acids. For these experiments, P. aeruginosa supernatants were added to a luminescent S. aureus strain constitutively expressing luxABCDE, and the impact of these supernatants on light production was examined. Light production has previously been used as a marker for antimicrobial activity (1, 42, 60, 61). In this assay, decreases in light production correlate with increased antimicrobial activity. Supernatants derived from the SCFM-grown P. aeruginosa exhibited significant antimicrobial activity, similar to that observed for the antimicrobial H₂O₂; however, removal of aromatic amino acids from SCFM resulted in complete loss of this activity (Fig. 4).

FIG. 4.

Aromatic amino acids enhance the antimicrobial activity of P. aeruginosa cultures. P. aeruginosa was grown in SCFM or in SCFM lacking aromatic amino acids, and culture supernatants were applied to the luminescent S. aureus strain Xen 36. S. aureus luminescence was followed over time using a luminometer. Decreases in luminescence correlates with antimicrobial activity. Data are shown relative to the luminescence of a control to which fresh medium (BHI) had been added. Hydrogen peroxide (3%) was used as a positive control, and supernatants from the SCFM-grown P. aeruginosa pqsA mutant, which does not exhibit lysis of S. aureus, was used as a negative control. Error bars represent standard deviations and in some cases are too small to be seen.

DISCUSSION

While it is clear that the nutritional environment significantly impacts P. aeruginosa physiology and gene expression, little is known about the CF lung bacterial growth environment. In general, studies utilizing laboratory media supplemented with CF sputum have sought to identify P. aeruginosa virulence genes differentially expressed in response to sputum (39, 64, 66). Two studies have provided crude, nondialyzed sputum as the sole source of carbon and energy for P. aeruginosa (38, 39); however, for nutritional studies, this approach is hampered by the complexity of CF sputum. Other groups have developed surrogate in vitro growth media to mimic CF sputum. A complex medium designed to foster P. aeruginosa mucoidy with components either directly measured or presumed to be in CF sputum was developed by Ghani and Soothill (14) and further modified by Sriramulu et al. (56). Although these media have utility for assessing P. aeruginosa phenotypes such as microcolony formation (56), the undefined nature of these media limits their use for assessing the roles of specific nutritional cues.

This study describes the development of a defined medium that nutritionally mimics CF sputum. Levels of specific components, including sodium, chloride, potassium, calcium, ammonium, and magnesium resemble those previously reported (13, 21, 24, 48, 49). Although concentrations of individual amino acids, glucose, and lactate have not previously been reported for CF sputum, the total free amino acid concentrations measured in this study are comparable to those previously published by Thomas et al. (59). More importantly, P. aeruginosa exhibits gene expression profiles, carbon consumption patterns, and cell-cell signaling profiles during growth that are similar in SCFM and in CF sputum.

P. aeruginosa carbon preference is poorly understood, despite the fact that the carbon source has profound effects on virulence-associated phenotypes, including toxin production and formation of antibiotic-resistant biofilms (4, 5, 15, 23, 37, 50, 53). Our results clearly show that P. aeruginosa exhibits preferences for carbon substrates that are similar during growth in SCFM and in CF sputum (Table 3i also see Tables S4 and S5 in the supplemental material). The mechanism of carbon source preference is not known, although novel factors are likely involved, since mutation of the carbon repression control (crc) gene, encoding a protein critical for small organic acid preference (16, 32), has no impact on P. aeruginosa carbon preference in SCFM (data not shown). It could be argued that since CF sputum contains a range of potential carbon sources not present in SCFM (such as mucin, lipids, nucleotides, and intact proteins), our studies may not accurately reflect carbon consumption in vivo. Although we cannot completely discount this possibility, several results support similar carbon preferences in CF sputum and in SCFM: growth rates in CF sputum and SCFM are similar (Fig. 1), the expression levels of carbon catabolism genes are similar in CF sputum and in SCFM (Table 2i also see Table S6 in the supplemental material), and the disappearance of specific carbon sources is temporally and quantitatively similar in both CF sputum and SCFM (see Tables S4 and S5 in the supplemental material). Together these points suggest that although other carbon-containing compounds are present in CF sputum, they likely contribute little to P. aeruginosa growth since growth rates and carbon consumption profiles are not drastically different in the presence (CF sputum) and absence (SCFM) of these alternative carbon sources. In addition, genes involved in pathways such as fatty acid β oxidation (fadD1 and fadD2 [2]), ribose catabolism (rbsK [18]), and nucleic acid catabolism (dht [22] and PA5019 [10]) were not differentially expressed when transcriptomes of SCFM-grown and CF sputum-grown cultures were compared (see Table S6 in the supplemental material). In fact, the only differentially expressed carbon catabolism genes between SCFM-grown and CF sputum-grown bacteria are putatively involved in the degradation of N-acetylglucosamine to acetyl-coenzyme A (PA3758-9, PA3761, and acsA [www.pseudomonas.com]). Thus, while we do not exclude the possibility that P. aeruginosa metabolizes these substrates for carbon and energy in CF sputum, our data support the conclusion that SCFM contains the primary carbon sources utilized by P. aeruginosa in vivo.

An intriguing phenotype of CF sputum-grown P. aeruginosa is enhanced production of the cell-cell signal PQS (39). Previous studies suggested that this phenotype is nutritionally mediated (9, 39), and our results with SCFM provide evidence that the presence of high levels of phenylalanine and tyrosine in CF sputum likely enhances PQS production (Fig. 3B). Recent studies indicate that high (high-micromolar to millimolar) levels of tryptophan also increase PQS production (9); however, our CF sputum chromatographic analyses revealed very little “free” tryptophan (10 μM) in CF sputum (Table 1), and this level was insufficient to induce PQS production (data not shown). In fact, tryptophan levels were often below the limit of detection (see Table S2 in the supplemental material); thus, it is unlikely that “free” tryptophan is critical for enhanced PQS production in CF sputum. Although the impact of enhanced PQS production in vivo is unknown, we speculate that this phenotype may be important during the initial stages of P. aeruginosa colonization. P. aeruginosa normally colonizes the CF lung after other bacteria such as S. aureus have established infections (17). Upon entering the CF lung, P. aeruginosa competes with and displaces the current bacterial inhabitants. Several factors likely contribute to this displacement, including production of antimicrobial factors by P. aeruginosa (33, 39). Our results suggest that aromatic amino acids are important inducers of antimicrobial compounds, as P. aeruginosa exhibits no antimicrobial activity against S. aureus in the absence of aromatic amino acids (Fig. 4).

Not all previously reported CF sputum-specific phenotypes were observed for SCFM-grown P. aeruginosa. Several studies have shown that iron-regulated genes are differentially expressed during growth in CF sputum (39, 64), indicating that CF sputum is an iron-limited environment. Although SCFM contains a physiologically relevant concentration of iron (7, 46, 57, 58), readily available iron is likely high in SCFM and low in CF sputum, since iron is bound by host/bacterial chelators in the CF lung. Also, genes involved in flagellar motility were previously reported to be differentially expressed by CF sputum-grown P. aeruginosa (39, 66). Recent studies indicate that the host factor neutrophil elastase mediates the loss of flagella and the reduction of P. aeruginosa flagellar gene expression observed during growth in CF sputum (20, 54). SCFM could easily be modified with the addition of host iron chelators or neutrophil elastase to study these phenotypes in a nutritionally relevant context.

SCFM is presented as a tool to examine the impact of the CF lung nutritional environment on P. aeruginosa physiology and gene expression. One important utility of this medium is the capacity for manipulation of components used in this study to demonstrate the impact of aromatic amino acids on P. aeruginosa cell-cell signaling. The observation that nutritionally controlled CF sputum phenotypes are observed in SCFM provides advantages over complex commercially available laboratory media such as Luria-Bertani and tryptic soy broth. This is an important point, as discrepancies in experimental results, particularly in regard to P. aeruginosa quorum sensing and biofilm formation, are often attributed to differing media conditions (23, 51-53, 63, 65).

While biofilm formation was not examined in this study, SCFM may have utility as a model growth substrate for P. aeruginosa biofilms. The carbon growth substrate has been shown to significantly impact P. aeruginosa biofilm formation (53); thus, SCFM provides a tool for examining the roles of specific CF sputum nutritional parameters in biofilm formation. One potential caveat to these studies is that CF sputum has an inherent viscosity provided by both high mucin content and dehydration of CF mucus (36) that is not present in SCFM. It is within this viscous mucus that P. aeruginosa forms microcolony structures in vivo (28, 67); thus, SCFM biofilm studies would necessitate mucin addition to mimic the natural viscosity.

It is critical to note that SCFM was developed from sputum samples provided by patients with nonexacerbating CF; thus, this medium represents a specific stage of CF lung infection. It is possible that the CF sputum nutritional environment may vary depending on the stage of disease and the clinical treatment methodologies. However, the approach of examining CF sputum nutritional content can be extended to early childhood, when P. aeruginosa infections are intermittent, or for individual patients before, during, and after exacerbation. Information from such in vitro studies will provide new information regarding the roles of nutritional cues in distinct stages of infection and allow for the design of focused studies to examine mechanistic details in vivo.