The Pseudomonas aeruginosa Secretory Product Pyocyanin Inactivates α₁ Protease Inhibitor: Implications for the Pathogenesis of Cystic Fibrosis Lung Disease

Reagents.

NADH, CuZn superoxide dismutase (SOD), trypsin, porcine pancreatic elastase (PPE), H₂O₂, methionine, histidine, catalase, and dimethylthiourea (DMTU) were obtained from Sigma Chemical Co., St. Louis, Mo. α₁PI was purchased from Calbiochem, La Jolla, Calif. The PPE preparations contain chymotrypsin-trypsin (25 to 100 U of trypsin activity/mg of protein) as an impurity. 5,5 Dimethyl-N-pyrroline 1-oxide (DMPO) was purchased from the Oklahoma Medical Research Foundation, Oklahoma City, Okla.

Ability of α₁PI to form a complex with serine proteases.

α₁PI (10 μg/ml) in H₂O was incubated with either PPE (50 to 200 μg/ml) or trypsin (50 to 200 μg/ml) for 15 to 30 min at 37°C. Aliquots of the reaction mixture were then added to Laemmli solubilizing buffer and subjected to standard sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The results were similar regardless of whether or not the gels were run under reducing conditions. The α₁PI-protease complex is well known to remain associated under these conditions (7, 31, 46). Detection of an α₁PI-protease complex was assessed by either (i) staining the gel with silver and determining the presence or absence of a new band with an apparent molecular weight consistent with a complex of α₁PI and the protease; or (ii) transfer of the SDS-PAGE-separated sample to nitrocellulose. The nitrocellulose was then blocked with 4% bovine serum albumin–Tris-buffered saline for 1 h and then incubated overnight with a 1:100 dilution of rabbit anti-human α₁PI (Sigma). The blots were then washed three times, and immunoreactive protein was determined with antirabbit immunoglobulin G linked to alkaline phosphatase. The α₁PI-protease complex was manifested as the presence of an α₁PI immunoreactive band with an apparent molecular weight higher than that of an α₁PI standard.

α₁PI activity.

α₁PI activity was measured in terms of α₁PI’s ability to inhibit PPE-mediated cleavage of succinyl-l-alanyl-l-alanyl-l-alanyl-p-nitroanilide (Sigma) according to the spectrophotometric assay of Travis and Johnson (47).

Cell culture of human lung epithelial cells.

The A549 human lung carcinoma cell line which resembles type II alveolar epithelial cells was maintained in continuous culture with Dulbecco’s modification of Eagle’s medium (DMEM) obtained through the University of Iowa Cancer Center, Iowa City. The culture medium was routinely supplemented with 10% heat-inactivated fetal bovine serum, 1% penicillin–streptomycin, and 1% glutamine. HBE cells were cultivated in a similar manner, except that the primary culture medium was 50% DMEM and 50% HAM (hepatocyte medium F12; University of Iowa Cancer Center). For each set of experiments, cells from the stock culture were seeded (2.5 × 10⁴ to 5 × 10⁴ cells/well) into 24- or 48-well plates. Cells were maintained at 37°C in 5% CO₂ until they were at least 80% confluent (usually 72 h), at which time they were utilized in the desired experiments.

Purification of pyocyanin.

Pyocyanin was extracted from culture supernatant of P. aeruginosa PAO1 (ATCC 15692; American Type Culture Collection, Rockville, Md.) by serial chloroform extractions followed by sequential extractions with acid and neutral water as previously detailed (9). After the completion of five separation sequences, the pyocyanin was crystallized and dried under vacuum. It was resuspended in water and stored at 4°C until used.

Effect of pyocyanin redox cycling on α₁PI-protease complex formation.

Experiments were performed in which α₁PI (or, in some cases, the target protease) was exposed to either NADH- or cell-mediated redox cycling of pyocyanin. α₁PI (10 μg/ml) was added either to a solution of NADH (6 mM) in H₂O or to an A549 or HBE cell monolayer in Hanks’ balanced salt solution after which the desired concentration of pyocyanin was added. The system was then incubated for 30 min at 37°C. At this point, trypsin or PPE (200 μg/ml) was added for an additional 15 min (37°C). The solutions were assessed for the formation of α₁PI-protease complex as described above.

EPR and spin trapping.

For spin trapping experiments, the spin trap DMPO (100 mM) and diethylenetriaminepentaacetic acid (DTPA [0.1 mM]) were included in the reaction mixture of interest. After the desired time of incubation, the reaction mixture was transferred to an electron paramagnetic resonance (EPR) quartz flat cell and placed into the cavity of the EPR spectrometer (ES 300; Brüker, Karlsrühe, Germany). EPR spectra were then obtained at 25°C with the following parameters: 4.00 × 10⁹ gain, 335.544-s sweep time, 100-kHz modulation frequency, 0.501-G modulation amplitude, 80-G sweep width, 9.76-GHz frequency, and 20 mW of power. For detection of the pyocyanin radical, 40 μM pyocyanin was added to a 44 μM solution of NADH and 0.1 mM DTPA, which had been sparged with N₂ for 20 min. The pyocyanin and NADH were then allowed to react at 25°C in the continuous presence of N₂ and subjected to EPR spectroscopy with a Varian E-104A EPR spectrometer (Varian Associates, Inc., Palo Alto, Calif.) with a 1-s time constant, 8-min scan time, 2.5 × 10⁴ gain, 100-kHz modulation frequency, and 20 mW of power.

Pyocyanin-NADH blocks α₁PI-protease complex formation.

Several laboratories, including our own, have shown that addition of NADH to pyocyanin leads to the reduction of pyocyanin to the pyocyanin radical (4, 11, 22, 24–26). Under aerobic conditions, the pyocyanin radical will transfer an electron to O₂, leading to the formation of O₂^·− (4, 11, 22, 24–26). Consistent with these earlier data (4, 11, 22, 24–26), when NADH and pyocyanin were incubated under conditions in which O₂ was previously depleted from the system by N₂ bubbling, an EPR spectrum indicative of the pyocyanin radical was detected (data not shown). When the same reaction was performed under aerobic conditions, the pyocyanin radical was no longer seen, but O₂.⁻ production was readily detected by DMPO spin trapping (data not shown).

Given the susceptibility of α₁PI to inactivation by oxidant species (6, 17, 41, 52), the ability of NADH-mediated redox cycling of pyocyanin to decrease α₁PI activity was assessed. As shown in Fig. 1 and 2, the presence of NADH and pyocyanin decreased the ability of α₁PI to form a stable complex with either trypsin or PPE. Concentrations of pyocyanin required to completely prevent detectable complex formation varied from experiment to experiment, ranging from 25 to 125 μM. In the PPE experiments, a major protease-α₁PI complex (top) and a somewhat fainter protease-α₁PI complex (bottom) were routinely observed. The lower bands are likely due to the presence of chymotrypsin and/or trypsin, which routinely contaminate commercial PPE preparations and also form a complex with α₁PI. Neither NADH nor pyocyanin alone prevented formation of the α₁PI-protease complex (data not shown), indicating that reduction of pyocyanin was required. These results were similar, regardless of whether the α₁PI-protease complex was detected by silver staining or immunoblot analysis.

An additional feature frequently noted when the NADH-pyocyanin combination was present was a decrease in the apparent quantity of α₁PI detectable on the gel and evidence of proteolytic degradation products of α₁PI (Fig. 1 and 2). The fact that these were detected by anti-α₁PI immunoblotting indicates they indeed reflect degradation of α₁PI. Lower-molecular-weight bands reactive with anti-α₁PI were not observed with NADH-pyocyanin-treated α₁PI unless an active protease was also added to the reaction mixture (data not shown [n = 3]). These data are consistent with the hypothesis that NADH-pyocyanin exposure leads to a modification of α₁PI, such that the antiprotease is converted to a legitimate substrate for PPE or trypsin.

Since the incubation conditions described above resulted in exposure of trypsin and PPE to NADH-pyocyanin as well as α₁PI, experiments were performed so that we could be certain that the effect of NADH-pyocyanin was on α₁PI and not the protease. α₁PI-protease complex formation was compared under conditions in which α₁PI or the protease was first exposed to NADH-pyocyanin for 30 min, after which, the protease or α₁PI was then added to the system, respectively, and α₁PI-protease complex formation was assessed (Fig. 3). When α₁PI was first incubated with NADH-pyocyanin and then PPE or trypsin was added, no α₁PI-protease complex was generated (Fig. 3). In contrast, when trypsin or PPE was the component incubated initially with NADH-pyocyanin, after which α₁PI was added, the same magnitude of α₁PI-protease complex was observed as in the non-NADH-pyocyanin-treated control (Fig. 3). These results indicate that the effect of NADH-pyocyanin is on α₁PI rather than the protease.

Biochemical evidence of NADH-pyocyanin inactivation of α₁PI.

As confirmation that exposure to NADH-pyocyanin leads to loss of α₁PI functional activity, the ability of NADH-pyocyanin-treated α₁PI to inhibit PPE enzymatic activity was assessed. α₁PI exposed to NADH-pyocyanin exhibited an impaired ability to inhibit PPE cleavage of succinyl-l-alanyl-l-alanyl-l-alanyl-p-nitroanilide. In the absence of α₁PI, cleavage of succinyl-l-alanyl-l-alanyl-l-alanyl-p-nitroanilide by PPE yielded a change in A₄₁₀ of 0.132 ± 0.012 U/min (mean ± standard error [SE]; n = 10). Addition of α₁PI decreased this to 0.040 ± 0.005 U/min (mean ± SE; n = 10). When α₁PI was first exposed to NADH-pyocyanin for 10 min, its ability to inhibit PPE activity was significantly decreased (P < 0.001 by analysis of variance) because A₄₁₀ was 0.076 ± 0.009 U/min (mean ± SE; n = 10). NADH-pyocyanin had no effect on PPE activity as assessed in this assay (data not shown).

Epithelial cell-mediated redox cycling of pyocyanin inactivates α₁PI.

In vivo redox cycling of pyocyanin in the P. aeruginosa-infected airway would likely occur via epithelial cell-mediated reduction of the compound rather via its interaction with extracellular NADH or NADPH. We have previously shown that incubation of monolayers of A549 (3, 12, 37) or HBE cells (12) with pyocyanin results in the formation of O₂^·−, as detected by spin trapping. Since initial reduction of pyocyanin would likely occur intracellularly, it was unclear whether cell-mediated redox cycling of pyocyanin could inactivate extracellular α₁PI. In order to assess this, A549 (or HBE) monolayers were incubated with α₁PI (10 μg/ml) ± pyocyanin (10 to 200 μM) for 30 min. PPE or trypsin was then added, and after 15 min of additional incubation, formation of α₁PI-protease complex in the extracellular milieu was assessed by immunoblot analysis. As shown in Fig. 4 and 5, the presence of pyocyanin resulted in a marked decrease in the subsequent ability of α₁PI to form a complex with either trypsin or PPE. This was dependent on the concentration of pyocyanin present. As with NADH-pyocyanin, supernatants from pyocyanin-treated, but not control, epithelial cells often exhibited evidence of α₁PI proteolysis (Fig. 4).

Role of reactive oxygen intermediates in pyocyanin-associated inactivation of α₁PI.

The ability of neutrophils to inactivate α₁PI is linked to oxidation of the protein by reactive oxygen species, particularly those arising from the reaction of H₂O₂ and myeloperoxidase (MPO) (6, 14, 34, 52). Accordingly, we postulated that oxidant species resulting from pyocyanin-induced O₂^·− formation were responsible for NADH-pyocyanin-mediated inhibition of α₁PI activity. In order to test this, we determined if the presence of SOD, catalase, or the H₂O₂-hydroxyl radical scavenger DMTU blocked NADH-pyocyanin-mediated inactivation of α₁PI. Surprisingly, none of these agents alone nor the combination of SOD and catalase exhibited any ability to block the effect of NADH-pyocyanin (Fig. 3). SOD and catalase are reasonably large proteins, raising the possibility that they were unable to adequately reach the site of α₁PI which required protection. However, we found that methionine, which blocks MPO-mediated inactivation of α₁PI (6) via its ability to spare methionine 358 within the active site of α₁PI (6, 29, 40, 48), did not protect α₁PI from the effects of NADH-pyocyanin (data not shown). In addition, since O₂^·− is not generated as a consequence of pyocyanin reduction under anaerobic conditions, we assessed the ability of O₂ depletion on the process by bubbling the experimental system with N₂ prior to the addition of pyocyanin. NADH-pyocyanin retained its ability to inactivate α₁PI under O₂-depleted conditions (Fig. 6). That bubbling of N₂ effectively depleted the system of O₂ was confirmed with the anaerobic indicator resazurin (33).

These data do not suggest a role for O₂^·−, or subsequently generated oxidants such as H₂O₂, in the ability of reduced pyocyanin to inactivate α₁PI. Instead, they raise the possibility that pyocyanin radical, a reducing species formed by the initial electron transfer from NADH to pyocyanin, could be inactivating α₁PI via a reducing rather than an oxidizing reaction. In support of this, exposure of α₁PI to the reducing agent β-mercaptoethanol resulted in an inability of the α₁PI to subsequently bind trypsin (Fig. 7). This is in contrast to reports in the literature (7, 31, 46) and or own experience (data not shown) that, once formed, the α₁PI-trypsin complex does not dissociate upon exposure to β-mercaptoethanol.