Release of Mucus Glycoconjugates by Pseudomonas aeruginosa Rhamnolipids into Feline Trachea In Vivo and Human Bronchus In Vitro Margaret Somerville, Graham W. Taylor, David Watson, Nigel B. Rendell, Andrew Rutman, Howard Todd, Julia R. Davies, Robert Wilson, Peter Cole, and Paul S. Richardson Department of Physiology, St. George's Hospital Medical School; Host Defence Unit, National Heart and Lung Institute; and Department of Clinical Pharmacology, Royal Postgraduate Medical School, London, United Kingdom
Pseudomonas aeruginosa colonizes the lower respiratory tracts of patients with severe bronchiectasis, including cystic fibrosis, a condition associated with increased airway mucus output. We have shown that an extract containing chloroform-soluble extracellular products of P. aeruginosa releases glycoconjugates into the cat trachea in vivo. This activity was not related to pyocyanin, a major component of the extract, but WdS associated with the rhamnolipids. Purified monorhamnolipid (100 p.glml) released radiolabeled and periodic acid-Schiff (PAS)-reactive glycoconjugates (A3H = +490 ± 70%, A35S = +170 ± 40%, APAS = +8.6 ± 1.7 p.g/min; n = 6, P < 0.02 for each). Dirhamnolipid (200 p.g/ml) was also effective (A3H = +640 ± 70%, A35S = +130 ± 20%, APAS = +9.3 ± 1.5 p.g1min; n = 6, P < 0.02 for each). Monorhamnolipid (100 p.g/ml) also released 35S-labeled and PAS-reactive glycoconjugates from human bronchial tissue in vitro (A35S = +189 ± 47%, APAS = +26.3 ± 8.5 p.g/min; n = 7, P < 0.001 versus control tissues in which no stimulus was given). The cat tracheal glycoconjugates released by the rhamnolipids differed from those released by pilocarpine 50 p.M, in having a higher 3H:35S ratio (P < 0.001). After gel chromatography on a Sepharose CL-4B column, the void volume fractions of the glycoconjugates also had different profiles in a cesium chloride density gradient. Those released by rhamnolipid banded at 1.62 g/ml, while those released by pilocarpine banded mainly at 1.50 g/ml, with some of the higher density material also present. The rhamnolipids released glycoconjugates without causing ultrastructural damage to the cat tracheal epithelium, as shown by transmission electron microscopy. Monorhamnolipid, at a concentration similar to those releasing glycoconjugates in the cat trachea, was present in lung secretions from a patient with cystic fibrosis, colonized by P. aeruginosa, undergoing heart-lung transplantation. The rhamnolipids may contribute to the increased mucus output and respiratory morbidity associated with P. aeruginosa colonization.
Pseudomonas aeruginosa is an opportunist pathogen rarely found in the lower respiratory tract in the general population when host defenses are intact. In contrast, this organism usually colonizes the lower airways of patients with cystic fibrosis and other forms of severe bronchiectasis, and colonization hastens deterioration in such patients (1). P. aeruginosa secretes a large number of extracellular products, among which are the phenazine pigments, exotoxins A and S, rhamnolipids, proteases, and mucoid exopolysaccharide (MEP or (Received for publication May 2, 1991) Address correspondence to: Dr. Margaret Somerville, Department of Biochemistry and Molecular Biology, The Medical School, University of Manchester, Manchester M13 9PT, United Kingdom. Dr. Watson's current address: Department of Clinical Mass Spectrometry, Hospital for Sick Children, Cincinnati, OH. Abbreviations: fast atom bombardment, FAB; 1 hydroxyphenazine, IHP; high-pressure liquid chromatography, HPLC; periodic acid-Schiff, PAS; transmission electron microscopy, TEM; thin-layer chromatography, TLC; void volume, Yo. Am. J. Respir, CeU Mol. BioI.
Vol. 6. pp. 116-122, 1992
alginate) (2). Some of these may aid P. aeruginosa in its colonization of and persistence in damaged respiratory tracts. For example, the phenazines, pyocyaninand 1 hydroxyphenazine (lHP), slow ciliary beating and cause epithelial disruption (3, 4), proteases are implicated in generalized lung damage (5), and alginate secretion by mucoid strains protects the organism from elimination by host defense mechanisms (6). An increased mucus output is a well-recognized clinical association of respiratory bacterial infection, particularly in the chronic conditions that predispose to P. aeruginosa colonization, but the mechanisms involved are obscure. Many agents, such as muscarinic and adrenergic agonists, inflammatory mediators, and some proteases (7, 8) have been shown to stimulate mucus output, as have some bacterial products (9). While studying the effectof the P. aeruginosa broth culture filtrates on ciliary beating (3), we noticed increased amounts of mucus in our in vitro preparations of human nasal cilia (R. Wilson, unpublished observations). We have investigated this observation further by testing some extracellular Pseudomonas products, including the phena-
Somerville, Taylor, Watson et al.: Pseudomonas Rhamnolipids and Airway Mucus Release
zine pigments, for an effect on mucus glycoconjugate release in vivo and in vitro.
Materials and Methods Preparations and Analysis of P. aeruginosa Products Pyocyanin was prepared in two ways- by extraction from P. a~~gin~sa growth medium and by synthesis (10). Briefly, a chmcalls.olate of P. aeruginosa (P455) was grown on Kings agar, which was subsequently scraped free of bacteria, chopped, and extracted with chloroform (3). Petroleum ether was added to this initial extract, and the resulting precipitate is referred to as the P455 extract. Both this and the supernatant remaining after its precipitation were tested for an effect on glycoconjugate release in the cat tracheal model (see below). The P455 extract was also passed through a column of Detoxi-Gel (Pierce Chemical Co.) in order to remove any endotoxin. The P455 extract was fractionated by high-pressure liquid chromatography (HPLC) at 2 ml/min on a J-tBondapak C I8 column (30 x 0.8 em), Fractions were eluted at 2 ml/min isocratically (i.e., in the same solvent) with acetonitrile.water (30:70 vol/vol, containing 0.04% trifluoroacetic acid), followed by a 30-min linear gradient to acetonitrile:water (90:10, vol/vol, with 0.04% trifluoroacetic acid). Solvents were removed under vacuum, and fractions were pooled and tested for an effect on mucus glycoconjugate release in the cat trachea. . A sample of mixed P. aeruginosa rhamnolipids was obtaI?ed fro~ Drs. A. T. Hastie (Thomas Jefferson University, Philadelphia, PA) and T. R. Shryock (Indiana State University, Terre Haute, IN). P455 was grown in liquid medium (11) for 72 h on a rotary shaker at .3r C. Bacteria were removed by centrifugation and filtration, and the supernatant was extracted twice with chloroform. This extract was applied to a column of silica gel (30 x 3 ern) and eluted isocratically in chloroform:methanol:acetic acid (16:3:1, vol/vol/vol). The column was monitored by thin-layer chromatography (TLC) and fast atom bombardment (FAB) mass spectrometry (12). Two anthronepositive species, corresponding to monorhamnolipid and dirhamnolipid, eluted from the column and were well resolved. Further purification of each of these species was by reverse-phase HPLC on a ILBondapak column, eluting at 2 ml/min with a 25-min linear gradient of acetonitrile: water (40:60 to 80:20). Partially purified rhamnolipids from P. aeruginosa were also obtained from Bideco (Baden, Switzerland) and purified in a similar manner. FABmass spectra were obtained on a Finnigan 4500 mass spectrometer using an M-scan ion gun with xenon (8 to 10 kV) as the primary ionizing source and glycerol as the matrix. Salt dosing was used to confirm the molecular assignments in the FAB mass spectra. Measurement of Glycoconjugate Output Cat tracheal model. An isolated tracheal segment in the neck of the anesthetized cat was pulse-labeled at the start of each experiment with precursors of glycoconjugates, (lH]glucose and P5S] sulfate, as described elsewhere (13). Briefly, the fluid-filled tracheal segment was flushed through at 15-min intervals with Krebs-Henseleit solution and the washings were collected onto solid guanidinium chloride.
117
Potential stimuli to glycoconjugate release were diluted in Krebs-Henseleit and placed intrasegmentally for 15 min at ~ourly intervals. Tracheal washings were dialyzed exhaustively, and radiolabel incorporation into macromolecules was determined by liquid scintillation counting. A chemical est.imat~ of gl~coconjugate output was also made using periOdIC. acid Schiff (PAS) assay (14) with bovine submaxillary mucin as standard. All samples were assayed and counted in triplicate. Output rates were expressed as Bq/min (for radiolabels) or ILg of bovine submaxillary mucin/min (for PAS reactivity) of collection. Responses to stimulations were expressed as percent changes (for radiolabels) or absolute changes (for P~S reactivity) in output rate over the preceding control (unstimulated) sample. Human bronchial mucosa. Bronchial tissue, less than 1 ~ thick and free fromtumor, was obtained from lung specimens r~sected for carcinoma and mounted in Ussing chambers, WIth oxygenated Krebs-Henseleit circulating in each half-chamber (15). ['5S]sulfate, 2 mCi, in 12 ml KrebsHenseleit solution remained on the submucosal side of the tissue throughout each experiment. Output of 35S-1abeled glycoconjugates stabilized after 3 h, and potential stimuli to glycoconjugate release were only applied to the luminal side of the tissue after this period. Washings from the luminal side were dialyzed exhaustively, counted by liquid scintillation counting, and their radioactivity and glycoconjugate contents measured as described above for the cat samples. Results were expressed in a similar fashion and compared with those obtained from control tissues, which were exposed to Krebs-Henseleit only. Structural and Functional Assessment of Cat Tracheal Epithelium Transmission electron microscopy (TEM). Portions of trachea were removed at the beginning of four experiments, before exposure to Krebs-Henseleit, Tracheal rings were also removed at the end of the same experiments after exposure to monorhamnolipid 100 ILg/ml for 15 min immediately bef~re sacrifi.c~ and in two more experiments after exposure to dirhamnolipid 200 ILg/ml. Tracheal tissues were fixed in 2.5 % glutaraldehyde with 4 % tannic acid buffered in sodium cacodylate (16) for 2 to 6 hat 4° C and then transferred to 2.5 % glutaraldehyde without tannic acid for at least a further 24 h. Tissue was postfixed in osmium tetroxide and processed through to araldite. Surface epithelial cells in each tissue were scored on a scale of 1 to 5 for each of four criteria o~ .cell damage: projections of ciliated epithelium, loss of cilia, formation of blebs on the apical surface, and mitochondrial damage. A damage score of 1 reflects normal tissue and the higher the score, the greater the amount of damage: The assessment was performed by one observer (A.R.), who was unaware of the treatment received by each tissue. Mucus transport. At the end of all experiments, except those in which tissue was removed for TEM, powdered charcoal was dusted onto the tracheal surface and observed for movement. No estimate of transport rate was made. Analysis of Cat Tracheal Secretions Pilocarpine 50 J-tM (a cholinergic agonist) and the P455 extract were used to stimulate mucus release in six cats, using
118
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992
SH-labeled~1eo
~
IiPJ
!
i
:m
"SIabeIed~1eo
6
I..
t>
~
;
]
.8
~e
!l=
1m
4
r3
10
3}
:II
Pseudomonas aeruginosa colonizes the lower respiratory tracts of patients with severe bronchiectasis, including cystic fibrosis, a condition associated with increased airway mucus output. We have shown that an extract containing chloroform-soluble extracellular products of P. aeruginosa releases glycoconjugates into the cat trachea in vivo. This activity was not related to pyocyanin, a major component of the extract, but WdS associated with the rhamnolipids. Purified monorhamnolipid (100 p.glml) released radiolabeled and periodic acid-Schiff (PAS)-reactive glycoconjugates (A3H = +490 ± 70%, A35S = +170 ± 40%, APAS = +8.6 ± 1.7 p.g/min; n = 6, P < 0.02 for each). Dirhamnolipid (200 p.g/ml) was also effective (A3H = +640 ± 70%, A35S = +130 ± 20%, APAS = +9.3 ± 1.5 p.g1min; n = 6, P < 0.02 for each). Monorhamnolipid (100 p.g/ml) also released 35S-labeled and PAS-reactive glycoconjugates from human bronchial tissue in vitro (A35S = +189 ± 47%, APAS = +26.3 ± 8.5 p.g/min; n = 7, P < 0.001 versus control tissues in which no stimulus was given). The cat tracheal glycoconjugates released by the rhamnolipids differed from those released by pilocarpine 50 p.M, in having a higher 3H:35S ratio (P < 0.001). After gel chromatography on a Sepharose CL-4B column, the void volume fractions of the glycoconjugates also had different profiles in a cesium chloride density gradient. Those released by rhamnolipid banded at 1.62 g/ml, while those released by pilocarpine banded mainly at 1.50 g/ml, with some of the higher density material also present. The rhamnolipids released glycoconjugates without causing ultrastructural damage to the cat tracheal epithelium, as shown by transmission electron microscopy. Monorhamnolipid, at a concentration similar to those releasing glycoconjugates in the cat trachea, was present in lung secretions from a patient with cystic fibrosis, colonized by P. aeruginosa, undergoing heart-lung transplantation. The rhamnolipids may contribute to the increased mucus output and respiratory morbidity associated with P. aeruginosa colonization.
Pseudomonas aeruginosa is an opportunist pathogen rarely found in the lower respiratory tract in the general population when host defenses are intact. In contrast, this organism usually colonizes the lower airways of patients with cystic fibrosis and other forms of severe bronchiectasis, and colonization hastens deterioration in such patients (1). P. aeruginosa secretes a large number of extracellular products, among which are the phenazine pigments, exotoxins A and S, rhamnolipids, proteases, and mucoid exopolysaccharide (MEP or (Received for publication May 2, 1991) Address correspondence to: Dr. Margaret Somerville, Department of Biochemistry and Molecular Biology, The Medical School, University of Manchester, Manchester M13 9PT, United Kingdom. Dr. Watson's current address: Department of Clinical Mass Spectrometry, Hospital for Sick Children, Cincinnati, OH. Abbreviations: fast atom bombardment, FAB; 1 hydroxyphenazine, IHP; high-pressure liquid chromatography, HPLC; periodic acid-Schiff, PAS; transmission electron microscopy, TEM; thin-layer chromatography, TLC; void volume, Yo. Am. J. Respir, CeU Mol. BioI.
Vol. 6. pp. 116-122, 1992
alginate) (2). Some of these may aid P. aeruginosa in its colonization of and persistence in damaged respiratory tracts. For example, the phenazines, pyocyaninand 1 hydroxyphenazine (lHP), slow ciliary beating and cause epithelial disruption (3, 4), proteases are implicated in generalized lung damage (5), and alginate secretion by mucoid strains protects the organism from elimination by host defense mechanisms (6). An increased mucus output is a well-recognized clinical association of respiratory bacterial infection, particularly in the chronic conditions that predispose to P. aeruginosa colonization, but the mechanisms involved are obscure. Many agents, such as muscarinic and adrenergic agonists, inflammatory mediators, and some proteases (7, 8) have been shown to stimulate mucus output, as have some bacterial products (9). While studying the effectof the P. aeruginosa broth culture filtrates on ciliary beating (3), we noticed increased amounts of mucus in our in vitro preparations of human nasal cilia (R. Wilson, unpublished observations). We have investigated this observation further by testing some extracellular Pseudomonas products, including the phena-
Somerville, Taylor, Watson et al.: Pseudomonas Rhamnolipids and Airway Mucus Release
zine pigments, for an effect on mucus glycoconjugate release in vivo and in vitro.
Materials and Methods Preparations and Analysis of P. aeruginosa Products Pyocyanin was prepared in two ways- by extraction from P. a~~gin~sa growth medium and by synthesis (10). Briefly, a chmcalls.olate of P. aeruginosa (P455) was grown on Kings agar, which was subsequently scraped free of bacteria, chopped, and extracted with chloroform (3). Petroleum ether was added to this initial extract, and the resulting precipitate is referred to as the P455 extract. Both this and the supernatant remaining after its precipitation were tested for an effect on glycoconjugate release in the cat tracheal model (see below). The P455 extract was also passed through a column of Detoxi-Gel (Pierce Chemical Co.) in order to remove any endotoxin. The P455 extract was fractionated by high-pressure liquid chromatography (HPLC) at 2 ml/min on a J-tBondapak C I8 column (30 x 0.8 em), Fractions were eluted at 2 ml/min isocratically (i.e., in the same solvent) with acetonitrile.water (30:70 vol/vol, containing 0.04% trifluoroacetic acid), followed by a 30-min linear gradient to acetonitrile:water (90:10, vol/vol, with 0.04% trifluoroacetic acid). Solvents were removed under vacuum, and fractions were pooled and tested for an effect on mucus glycoconjugate release in the cat trachea. . A sample of mixed P. aeruginosa rhamnolipids was obtaI?ed fro~ Drs. A. T. Hastie (Thomas Jefferson University, Philadelphia, PA) and T. R. Shryock (Indiana State University, Terre Haute, IN). P455 was grown in liquid medium (11) for 72 h on a rotary shaker at .3r C. Bacteria were removed by centrifugation and filtration, and the supernatant was extracted twice with chloroform. This extract was applied to a column of silica gel (30 x 3 ern) and eluted isocratically in chloroform:methanol:acetic acid (16:3:1, vol/vol/vol). The column was monitored by thin-layer chromatography (TLC) and fast atom bombardment (FAB) mass spectrometry (12). Two anthronepositive species, corresponding to monorhamnolipid and dirhamnolipid, eluted from the column and were well resolved. Further purification of each of these species was by reverse-phase HPLC on a ILBondapak column, eluting at 2 ml/min with a 25-min linear gradient of acetonitrile: water (40:60 to 80:20). Partially purified rhamnolipids from P. aeruginosa were also obtained from Bideco (Baden, Switzerland) and purified in a similar manner. FABmass spectra were obtained on a Finnigan 4500 mass spectrometer using an M-scan ion gun with xenon (8 to 10 kV) as the primary ionizing source and glycerol as the matrix. Salt dosing was used to confirm the molecular assignments in the FAB mass spectra. Measurement of Glycoconjugate Output Cat tracheal model. An isolated tracheal segment in the neck of the anesthetized cat was pulse-labeled at the start of each experiment with precursors of glycoconjugates, (lH]glucose and P5S] sulfate, as described elsewhere (13). Briefly, the fluid-filled tracheal segment was flushed through at 15-min intervals with Krebs-Henseleit solution and the washings were collected onto solid guanidinium chloride.
117
Potential stimuli to glycoconjugate release were diluted in Krebs-Henseleit and placed intrasegmentally for 15 min at ~ourly intervals. Tracheal washings were dialyzed exhaustively, and radiolabel incorporation into macromolecules was determined by liquid scintillation counting. A chemical est.imat~ of gl~coconjugate output was also made using periOdIC. acid Schiff (PAS) assay (14) with bovine submaxillary mucin as standard. All samples were assayed and counted in triplicate. Output rates were expressed as Bq/min (for radiolabels) or ILg of bovine submaxillary mucin/min (for PAS reactivity) of collection. Responses to stimulations were expressed as percent changes (for radiolabels) or absolute changes (for P~S reactivity) in output rate over the preceding control (unstimulated) sample. Human bronchial mucosa. Bronchial tissue, less than 1 ~ thick and free fromtumor, was obtained from lung specimens r~sected for carcinoma and mounted in Ussing chambers, WIth oxygenated Krebs-Henseleit circulating in each half-chamber (15). ['5S]sulfate, 2 mCi, in 12 ml KrebsHenseleit solution remained on the submucosal side of the tissue throughout each experiment. Output of 35S-1abeled glycoconjugates stabilized after 3 h, and potential stimuli to glycoconjugate release were only applied to the luminal side of the tissue after this period. Washings from the luminal side were dialyzed exhaustively, counted by liquid scintillation counting, and their radioactivity and glycoconjugate contents measured as described above for the cat samples. Results were expressed in a similar fashion and compared with those obtained from control tissues, which were exposed to Krebs-Henseleit only. Structural and Functional Assessment of Cat Tracheal Epithelium Transmission electron microscopy (TEM). Portions of trachea were removed at the beginning of four experiments, before exposure to Krebs-Henseleit, Tracheal rings were also removed at the end of the same experiments after exposure to monorhamnolipid 100 ILg/ml for 15 min immediately bef~re sacrifi.c~ and in two more experiments after exposure to dirhamnolipid 200 ILg/ml. Tracheal tissues were fixed in 2.5 % glutaraldehyde with 4 % tannic acid buffered in sodium cacodylate (16) for 2 to 6 hat 4° C and then transferred to 2.5 % glutaraldehyde without tannic acid for at least a further 24 h. Tissue was postfixed in osmium tetroxide and processed through to araldite. Surface epithelial cells in each tissue were scored on a scale of 1 to 5 for each of four criteria o~ .cell damage: projections of ciliated epithelium, loss of cilia, formation of blebs on the apical surface, and mitochondrial damage. A damage score of 1 reflects normal tissue and the higher the score, the greater the amount of damage: The assessment was performed by one observer (A.R.), who was unaware of the treatment received by each tissue. Mucus transport. At the end of all experiments, except those in which tissue was removed for TEM, powdered charcoal was dusted onto the tracheal surface and observed for movement. No estimate of transport rate was made. Analysis of Cat Tracheal Secretions Pilocarpine 50 J-tM (a cholinergic agonist) and the P455 extract were used to stimulate mucus release in six cats, using
118
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992
SH-labeled~1eo
~
IiPJ
!
i
:m
"SIabeIed~1eo
6
I..
t>
~
;
]
.8
~e
!l=
1m
4
r3
10
3}
:II
