Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Antimicrobial peptides are cationic proteins of < 10 kD that are found in a wide range of organisms (1). They are important components of the innate immune response (2), particularly at epithelial surfaces that are vulnerable to colonization by potential pathogens (3).

A prominent group of antimicrobial peptides are defensins, 3.5- to 4.0-kD peptides that share a consensus motif of six invariant cysteines and three intramolecular disulfide bridges. In addition to their activity against various bacteria, fungi, and viruses (4), defensins have other physiologic properties, including mitogenic and chemotactic activity (5, 6). Some defensins also activate calcium channels and may act in control of cell volume (7). The two major groups of vertebrate defensins, - and -defensins, differ in the arrangement of their disulfide bonds. In humans, six -defensins (human defensin [HD] 1 through 6) and two -defensins (human  defensin [hBD] 1-2) have been described to date. HD-1 through -4 are present in the neutrophils, where they constitute 30% to 50% of the total protein in azurophilic granules (8). HD-5 and HD-6 have been identified in the Paneth cells of small intestinal crypts (9, 10) and in female reproductive tissue (11). hBD-1 has been purified from plasma (12) and has been detected in a range of epithelial tissues (13). A second -defensin, hBD-2, has recently been purified from skin and was shown to be expressed in the lung and uterus (14).

-defensins have been implicated in host defense of human airway epithelia. Antibacterial activity associated with low-molecular-weight, salt-sensitive, and heat-stable factors has been detected in the surface fluid of primary cultures from normal airway epithelia (15), and expression of hBD-1 messenger RNA has been detected in human lung (16). An antibacterial activity found in the surface fluid from xenograft models of normal human airway epithelia was largely abolished by pretreatment with specific antisense nucleotides to hBD-1 (17). Hence, hBD-1 has been implicated as a major component of host defense in human airways. Furthermore, inactivation of hBD-1 in the lungs of cystic fibrosis (CF) patients may be an important cause for the onset of chronic lung infection, which is the major cause of morbidity and mortality in this disease.

Although hBD-1 expression has been detected in lung epithelia and this molecule appears to account for antibacterial activity in surface fluid derived from xenografts, the factors responsible for antibacterial activity in human lung airway surface fluid (ASF) in vivo have not been purified. To assess the contribution of different antimicrobial peptides to human airway defense mechanisms, we isolated these molecules from bronchoalveolar lavage fluid (BALF) from human lungs.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Samples

BALF was obtained with permission from patients undergoing investigative bronchoscopy for lung tumors. The included patients met the criteria of not having lung infection or inflammation and not being on antibiotic therapy. Samples that showed evidence of bleeding were not included in this study.

Lavage Technique

The lung regions sampled were lingula or right middle lobe. Prewarmed saline, 100 to 150 ml, was used for two to three cycles of lavage, and the sample was recovered in the third cycle. Between 5 and 10 ml of BALF were processed from each patient for these experiments, representing one-tenth to one-fifth of the final volume recovered. Samples from seven patients were processed.

Antibacterial and Lysozyme Assays

Antibacterial assays were carried out as described previously (18). Briefly, plates were poured with an approximately 1-mm-thick lawn of about 10⁵ cfu/ml of Escherichia coli D21 (CGSC5158) in 0.1 mM Sorensen's phosphate buffer (pH 7.2), 0.02% Tween-20, 100 µg/ml Luria Bertani (LB) broth, and 1% agarose. Wells of 1.5-mm diameter were punched into the lawns, and 1 µl sample was added per well. The plates were incubated at 37°C for 3 h and overlaid with an approximately 1 mm layer of LB broth with 1% agarose. Antibacterial activity was expressed as the diameter of clear zones measured after incubation of the plates at 37°C for 16 h. The antimicrobial peptide Magainin-1 (Sigma, St. Louis, MO) was used as a positive control, at a concentration of 100 µg/ml.

Lysozyme activity was assessed by a similar method, except that plates were poured with lawns of Micrococcus lysodeikticus cell walls (Sigma) at a concentration of 1 mg/ml and incubated for 16 to 20 h at 37°C. Hen egg-white lysozyme (Sigma) was used as a positive control at a concentration of 1 mg/ml.

Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE)

Tris-tricine SDS-PAGE was carried out as described previously (19) using a mini Protean II electrophoresis cell (Bio-Rad, Hercules, CA).

Determination of Protein Concentration

Protein concentration was determined with the Bradford Assay (20), using bovine serum albumin as standard.

Purification of Antibacterial Factors from BALF

Acetonitrile (ACN) and trifluoroacetic acid (TFA) were added to BALF samples to final concentrations of 20% and 0.1%, respectively, and the samples were centrifuged at 13,400 × g for 5 min. The supernatants (3-5 ml per 100-mm³ cartridge) were applied to SepPak C-18 cartridges (Waters, Milford, MA) equilibrated in 20% ACN, 0.1% TFA, and washed with the same buffer and eluted with 2 × 1 ml of 80% ACN, 0.1% TFA. The eluted fractions were vacuum-dried, resuspended in one-fiftieth volume distilled water (dH₂O), and assayed for antibacterial and lysozyme activity. The total level of antimicrobial activity in each sample varied, and the second elution did not always contain antibacterial activity.

Continuous acid-urea PAGE was carried out as described previously (21), using a MiniPrep continuous electrophoresis cell (Bio-Rad). SepPak fractions showing antibacterial activity were pooled to yield samples from four patients per electrophoretic separation (representing an initial volume of 35 ml of BALF). Loading buffer was added to a final concentration of 3 M urea, 5% acetic acid, with methyl green as the tracker dye, and the sample applied to a 7-cm gel comprising 16% acrylamide, 3 M urea that had been pre-run in 5% acetic acid for 3 h. Samples were run at 150 V and eluted with dH₂O at a flow rate of 0.15 ml/min. Fractions of 1.5 ml were collected, vacuum-dried, resuspended in 25 µl dH₂O, assayed for antibacterial and lysozyme activity, and subjected to SDS-PAGE.

Antibacterial fractions that appeared to be associated with peptides of < 6.5 kD (as judged by SDS-PAGE) were pooled from three continuous acid-urea PAGE runs and subjected to reverse-phase high-performance liquid chromatography (RP-HPLC) on a 4.6 × 250-mm Vydac C-18 column (The Separation Group, Sigma). The column was equilibrated with 0.1% TFA and eluted with a gradient of ACN in 0.1% TFA, the shape of which was based on a study by Frohm and colleagues (22) and found to give good resolution of relatively hydrophobic proteins, such as lysozyme and defensins. Fractions were vacuum-dried, resuspended in 15 µl HPLC-grade water (Sigma), assessed for antibacterial and lysozyme activity, and subjected to SDS-PAGE. A fraction with antibacterial activity that was associated with a single band of < 6.5 kD on a silver-stained SDS-PAGE gel was subjected to automated N-terminal amino acid sequencing. The sample was reduced and alkylated with 4-vinyl pyridine, adsorbed to ProSorb (Perkin-Elmer, Norwalk, CT), washed, and sequenced on an Applied Biosystems amino acid sequencer (Perkin-Elmer) for 35 cycles.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Purification of Antibacterial Peptides from BALF

Antibacterial activity and lytic activity against M. lysodeikticus cell walls was detected in BALF samples. Figure 1 shows a continuous acid-urea PAGE profile of BALF after extraction on SepPak C-18 columns, and the corresponding activity profile. An initial peak of lytic activity against M. lysodeikticus cell walls is followed by a broad zone of activity against E. coli D21 (Figure 1). Fractions 25 to 37, indicated by a bar in Figure 1, were associated with elution of peptides of < 6.5 kD, as judged by SDS-PAGE (not shown), and were pooled for RP-HPLC.

View larger version (22K): [in this window] [in a new window]   Figure 1.   Continuous acid-urea PAGE of BALF. Fractions of 1.5 ml were collected, lyophilized, and reconstituted in 1.5 ml dH2O. Protein concentration (single line), antibacterial activity against E. coli D21 (open bar), and lytic activity against M. lysodeikticus cell walls (solid bar) are shown for each reconstituted fraction. The horizontal bar indicates fractions that were pooled for RP-HPLC.

Application of the pooled fractions to RP-HPLC yielded several peaks of antibacterial activity against E. coli D21 (Figure 2). Activity peaks before 80 min elution time (< 32% ACN) were not associated with the elution of visible bands on SDS-PAGE gels. Activity against E. coli D21 at 80 to 84 min was associated with the elution of a single band of < 6.5 kD (not shown). Lytic activity against M. lysodeikticus cell walls at 98 to 105 min (> 40% ACN) was associated with some antibacterial activity against E. coli D21 and coincided with the elution of a band at approximately 14.3 kD, which is likely to be lysozyme. The remaining peaks of antibacterial activity did not coincide with the elution of single bands on SDS-PAGE gels.

View larger version (22K): [in this window] [in a new window]   Figure 2.   RP-HPLC of fractions enriched for antibacterial peptides of < 6.5 kD by acid-urea PAGE. (A) Elution profile and acetonitrile gradient. Buffer B contains 80% ACN in 0.1% TFA. (B) Antibacterial activity against E. coli D21 (open bar) and lytic activity against M. lysodeikticus cell walls (solid bar).

N-terminal Amino Acid Sequencing of Antibacterial Peptides from BALF

N-terminal amino acid sequencing of the peak that eluted at 81 to 82 min on the HPLC yielded three sequences corresponding to human neutrophil defensins (HD) 1-3. Five of the six cysteines present in HD-1 through -3 were resolved after alkylation of the sample with 4-vinyl pyridine, resulting in complete sequences of the first 29 N-terminal amino acid residues (Figure 3). The most abundant sequence was HD-1, followed by HD-2 and HD-3, respectively, at proportions of 5.0:2.0:1.5. The total yield of the peak, as judged by an initial yield of 190 pmol and an average molecular mass of 3445 D, was 3 to 5 µg, representing a minimum concentration of 86 to 143 ng/ml of HD-1 through -3 in BALF.

View larger version (11K): [in this window] [in a new window]   Figure 3.   Sequence of the first 29 N-terminal amino acids of HD 1-3, eluted detected at 81 to 82 min of HPLC. The -defensin consensus sequence, except for the last cysteine, is shown in bold.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

hBD-1, and possibly other antimicrobial factors derived from airway epithelial cells, have been suggested to be primary components of antibacterial activity in human (15, 17). We purified defensin-like molecules from BALF to evaluate their contributions to airway defense mechanisms. The purification protocol that we used is designed for the isolation of cationic and hydrophobic antimicrobial peptides. Continuous acid-urea PAGE has been used by Harwig and associates (21) for the purification of rabbit and human neutrophil defensins and by Lee and coworkers (23) for the purification of clavainins, antimicrobial peptides from tunicate hemocytes. RP-HPLC is widely used for the final purification of antimicrobial peptides. The gradient described by Frohm and associates (22) was found to provide the best resolution of peaks at 30% to 40% ACN, the concentration at which most hydrophobic antimicrobial peptides elute from C-18 columns. Frohm and colleagues (22) used this gradient to resolve factors in human wound fluid, including lysozyme and neutrophil defensins. Because these authors used an identical make of column, this gradient also provided a useful reference for the elution of defensin-like factors in BALF.

We identified HD-1 through -3, but did not detect hBD-1, hBD-2, or additional antimicrobial peptides in BALF. HD-1 through -3 have previously been found at high concentrations in purulent sputum from patients suffering from CF (24) or chronic obstructive pulmonary disease (25), conditions that are associated with markedly elevated neutrophil counts. However, patients included in the present study did not suffer from persistent lung infections or obvious signs of inflammation.

The molar ratio of HD-1 through -3 identified in BALF is in accordance with that reported previously (24) for the much-elevated levels in purulent sputum from CF patients, in contrast to their nearly equimolar ratios in neutrophil granules (8). This phenomenon may be attributable to differences in degradation between these molecules outside neutrophil granules, or to preferential binding of HD-2 and -3 to ASF proteins. The concentrations of HD-1 through -3 in BALF reported here indicate that these molecules may occur in micromolar concentrations, which is sufficient to exert antibacterial activity (26). Importantly, at concentrations of < 1 µg/ml, these molecules are mitogenic for airway epithelial cells in vitro (24); hence, they may act as growth factors as well as in antimicrobial defense.

In conclusion, the data reported here suggest that neutrophil defensins, together with lysozyme, are the most prominent antimicrobial factors in ASF derived from alveoli and bronchioles. It is likely that hBD-1 and hBD-2, together with as yet unidentified factors, contribute to antibacterial defense in airway epithelia, but these peptides are not present in BALF at sufficiently high concentrations to be readily isolated. The epithelial defensins hBD-1 and hBD-2 may be effective at much lower concentrations than the neutrophil defensins; alternatively, they may be relatively unstable peptides, though there is no biochemical evidence to support either of these hypotheses. Smith and coworkers (15) worked with cultured airway epithelia, and Goldman and colleagues (17) used a xenograft model to investigate airway defensins. It may be that neither of these systems accurately reflects the situation in the human lung because they select predominantly for epithelial cell products, whereas many other cell types contribute to lung defenses in vivo. It is also possible that hBD-1 and hBD-2 are expressed at higher levels in the ASF of larger airways, which would not be assayed in BALF. However, because CF airway disease starts in the small airways, this would seem the relevant place to look for antibacterial factors. It has been suggested that inactivation of hBD-1 due to elevated salt concentration in the lungs of CF patients contributes to chronic lung infection (17). Neutrophil defensins such as hBD-1 are salt-sensitive (26), and if ASF salt levels are indeed elevated in CF, antimicrobial activity provided by these peptides would still be impaired.

It is likely that in the lung, HD-1 through -3 act synergistically with each other or with other antimicrobial factors, including lysozyme. However, the importance of factors such as hBD-1 and hBD-2 for antibacterial defense of human airway in vivo remains to be elucidated.