The Lactoperoxidase System Functions in Bacterial ... - ATS Journals

The Lactoperoxidase System Functions in Bacterial Clearance of Airways Cynthia Gerson, Juan Sabater, Mario Scuri, Aliza Torbati, Richard Coffey, Jeffrey W. Abraham, Isabel Lauredo, Rosanna Forteza, Adam Wanner, Matthias Salathe, William M. Abraham, and Gregory E. Conner Department of Cell Biology and Anatomy and Division of Pulmonary and Critical Care Medicine, University of Miami School of Medicine, Miami; and Mt. Sinai Medical Center, Miami Beach, Florida Airway mucus is a complex mixture of secretory products that provides a multifaceted defense against pulmonary infection. Mucus contains antimicrobial peptides (e.g., defensins) and enzymes (e.g., lysozyme) although the contribution of these to airway sterility has not been tested in vivo. We have previously shown that an enzymatically active, heme-containing peroxidase comprises 1% of the soluble protein in sheep airway secretions, and it has been hypothesized that this airway peroxidase may function as a biocidal system. In this study, we show that sheep airway peroxidase is identical to milk lactoperoxidase (LPO) and that sheep airway secretions contain thiocyanate (SCN⫺) at concentrations necessary and sufficient for a functional peroxidase system that can protect against infection. We also show that airway LPO, like milk LPO, produces the biocidal compound hypothiocyanite (OSCN⫺) in vitro. Finally, we show that in vivo inhibition of airway LPO in sheep leads to a significant decrease in bacterial clearance from the airways. The data suggest that the LPO system is a major contributor to airway defenses. This discovery may have significant implications for chronic airway colonization seen in respiratory diseases such as cystic fibrosis.

The airway mucosa provides a sophisticated defense against inhaled toxins and particles, including infectious disease agents. The nonimmunologic airway defense is composed of a physical barrier (epithelial cells and secreted mucus), mechanical clearance (cilia and cough), immunoglobulins, and chemical/biocidal components (e.g., lysozyme and defensins), which are secreted into the mucus. In addition to mucosal products, leukocytes contribute to innate mucosal defenses. Neutrophil myeloperoxidase (MPO) uses H2O2 to oxidize chloride (Cl⫺) or thiocyanate (SCN⫺) to hypochlorite or hypothiocyanite (OSCN⫺), respectively, and both are antibacterial (1). Similarly, in milk, lactoperoxidase (LPO) uses H2O2 to catalyze the conversion of SCN⫺ to OSCN⫺ (2). It has been shown that the airway mucosa secretes a peroxidase, and it was suggested that it is also active in preventing infection of the airway (3–6). This hypothesis was premised on the known functions of other related heme-peroxidases such as MPO and LPO and on histochemical studies of the respiratory tract that showed airway goblet cells and submucosal glands contain a peroxidase localized to the cellular secretory apparatus. (Received in original form October 11, 1999 and in revised form December 14, 1999 ) Address correspondence to: Gregory E. Conner, Dept. of Cell Biology and Anatomy, R124, University of Miami School of Medicine, P.O. Box 016969, Miami, FL 33101. E-mail: [email protected] Abbreviations: complementary DNA, cDNA; colony-forming unit(s), CFU; eosinophil peroxidase, EPO; glyceraldehyde-3-phosphate dehydrogenase, GAPDH; lactoperoxidase, LPO; myeloperoxidase, MPO; messenger RNA, mRNA; hypothiocyanite, OSCN⫺; phosphate-buffered saline, PBS; thiocyanate, SCN⫺; sodium dodecyl sulfate, SDS; thyroid peroxidase, TPO. Am. J. Respir. Cell Mol. Biol. Vol. 22, pp. 665–671, 2000 Internet address: www.atsjournals.org

Our previous biochemical studies of H2O2 scavenging activity of sheep airway mucus showed that a single peroxidase is a major component of secreted tracheal mucus and is responsible for the majority of H2O2 scavenging activity in sheep airway secretions (6, 7). These studies also showed that the secreted sheep airway peroxidase, though different in size from bovine milk LPO, had strong enzymatic resemblance to milk LPO, and thus, by analogy, also suggested that it might be an important contributor to the sterility of healthy airways. In this report, we have determined the molecular identity of sheep airway peroxidase by amino acid sequencing of the purified protein and nucleotide sequencing of complementary DNA (cDNA) clones isolated from sheep tracheal messenger RNA (mRNA). We also examined sheep airway secretions for SCN⫺ that is necessary to complete a functional peroxidase biocidal system in the airway. Finally, we tested the hypothesis that airway peroxidase is directly involved in the maintenance of airway sterility using an in vivo sheep model.

Materials and Methods All materials were obtained from Sigma Chemical Co. (St. Louis, MO) unless otherwise noted.

Purification and Amino Acid Sequencing of Sheep Airway Peroxidase Ovine tracheal secretions were collected by lavage with phosphatebuffered saline (PBS), pH 7.4, and centrifuged to pellet contaminating cells and any cellular debris, as described previously (6, 7). The fluid was adjusted to pH 8.0, 0.01 M NaPO4, 0.04% Tween 20, and passed sequentially through filters of increasing retention, culminating with 0.45 ␮m pore membrane. The filtrate was applied to S-sepharose and eluted with a gradient of NaCl (0.15 to 0.5 M) in the same buffer. Peroxidase activity fractions were then pooled and loaded onto agarose conjugated with Lens culinaris lectin and eluted with 0.5 M ␣-methyl mannoside in 0.05 M NaPO4, 0.04% Tween 20. Sheep milk LPO was purified by CG-50 ion exchange chromatography. Bovine milk LPO was obtained from Sigma Chemical Co. and canine MPO was a generous gift of Dr. Roger Fenna (University of Miami, Miami, FL) (8). For amino acid sequencing, the purified sheep airway peroxidase was precipitated with 10% trichloroacetic acid and cleaved by overnight incubation in 70% trifluoroacetic acid saturated with cyanogen bromide (CNBr). Peptides were resolved on Tris-Tricine sodium dodecyl sulfate (SDS) gels (9), transferred to Immobilon-PSQ paper (Millipore, Bedford, MA), and sequenced by the Protein Chemistry Core Facility at the University of Florida (Gainesville, FL).

Enzyme and Substrate Assays OSCN⫺ forming activities of purified sheep airway LPO, sheep milk LPO, and bovine milk LPO were assayed in triplicate in 0.3 mM H2O2, 1 mM potassium thiocyanate (KSCN), 50 mM NaPO4, pH 7.0, using 10 mM p-carboxybenzene sulfonamide to trap OSCN⫺. Assays were carried out for 5 min at 37⬚C and stopped by

666

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 22 2000

the addition of catalase before measurement using 5,5⬘-dithio-2nitrobenzoic acid (10). Sheep airway LPO and sheep milk LPO were also assayed by measuring oxidation of 3,3⬘,5,5⬘-tetramethylbenzidine (TMB) in the presence of varying concentrations of NaI, as described previously (6). TMB oxidation product was measured on a Perkin Elmer (Norwalk, CT) Lambda 3b recording spectrophotometer by following increases in absorbance at 652 nm during the initial 2 min of the reaction. Activity was expressed relative to assay values in the absence of NaI. SCN⫺ was assayed by the ferric nitrate method (11).

Isolation of cDNA Clones Mucosa was manually stripped from excised sheep trachea and frozen until mRNA was prepared. Total RNA was purified according to Chomczynski and Sacchi (12), and mRNA was isolated using oligo dT coupled to magnetic beads. Double-stranded cDNA was synthesized using the Superscript kit (Life Technologies, Gaithersburg, MD) and was size-fractionated before directional insertion into the plasmid pSPORT1 (Life Technologies). The two largest size classes each gave over 106 independent transformants from the sheep library, in Escherichia coli DH10B. The bacteria containing this library was screened for cDNA clones representing peroxidase mRNA using degenerate oligonucleotides derived from conserved regions of peroxidase nucleotide and protein sequences (Figure 1). DNA sequencing was performed by the DNA Core, University of Miami School of Medicine. Sequence alignments were performed using the Wisconsin Package (GCG, Madison, WI).

Northern Blots Sheep tissues were excised and frozen at ⫺70⬚C until preparation of total RNA using Ultraspec (Biotech Laboratories, Houston, TX). Three micrograms of total RNA were run on 1% agarose-formaldehyde gels, transferred to nitrocellulose, and sequentially hybridized with probes to ovine airway LPO and ovine glyceraldehyde phosphate dehydrogenase (GAPDH). Probes were prepared using either a 2,000-bp fragment of ovine airway LPO cDNA or a 217-bp fragment of sheep GAPDH that was purified and labeled

by random priming in the presence of [␣32P]deoxycytidine triphosphate (NEN, Boston, MA). Blots were hybridized at 50⬚C in 6⫻ saline sodium phosphate EDTA with 50% formamide, washed in 0.1⫻ saline sodium citrate at 52⬚C and quantitated using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).

Bacterial Challenge of Sheep Airways Aerosol treatments and bacterial clearance was performed as described previously (13). Sheep were pretreated by aerosols with 3 ml of 10⫺3 M dapsone in PBS or with PBS alone. Dapsone is an inhibitor of heme-peroxidases that under some conditions will distinguish MPO from LPO (14) and from airway peroxidase (6). Control sheep and pretreated animals were challenged with Pasteurella haemolytica (American Type Culture Collection [ATCC] no. 29698). P. haemolytica cultures (100 ml) were grown overnight at 37⬚C in a rotary water bath, collected by centrifugation, resuspended in 10 ml of PBS, and 3 ml, containing on average 1.6 ⫻ 109 (standard error [SE] ⫾ 3 ⫻ 108) colony-forming units (CFU), were delivered to sheep by aerosol. Immediately, 30 min, 1 h, and 3 h after challenge, samples of tracheal surface fluid were collected by cytology brush through a bronchoscope. Brushes were vortexed in 1 ml of saline and the saline used for quantitative bacterial cultures to determine the number of CFU (13). To control for variations between inocula, CFU at each time point were normalized to that obtained immediately after challenge. Samples from all sheep at the initial time point averaged 9.1 ⫻ 106/sample (SE ⫾ 3.2 ⫻ 106) CFU. The day before experiments, the tracheas were identically sampled and cultured to rule out prior infection of the airway. The day after experiments, animals were treated with antibiotics to prevent pneumonia. Nonparametric statistics were used to compare differences among groups at each time point by the Kruskal-Wallis test. If the null hypothesis was rejected, a post hoc multiple comparison procedure was used to determine which pairs of treatments were different. A P value of ⬍ 0.05 using a two-tailed test was considered significant.

Results Peroxidase mRNA Expressed in Sheep Airways

Figure 1. Degenerate oligonucleotide primers. Highly conserved nucleotide sequences of LPO, EPO, MPO, and TPO were identified and synthesized in forward (oligos I and II) and reverse (oligos III and IV) directions. Numbering is that of human preproMPO (PIR locus OPHUM). Amino acids that differed from the consensus are underlined.

To determine the molecular identity of airway peroxidase and to further our understanding of the potential variety of other peroxidase transcripts that may be found in airways, peroxidase cDNAs were cloned from a sheep tracheal cDNA library. Degenerate oligonucleotides were made from four regions of high nucleotide sequence similarity between the mammalian heme-containing peroxidases, MPO, milk LPO, eosinophil peroxidase (EPO), and thyroid peroxidase (TPO) (Figure 1). These were used to probe a sheep tracheal mucosa cDNA library. Polymerase chain reactions using the sheep tracheal cDNA and combinations of these oligonucleotides, as well as an SP6 promoter oligonucleotide, amplified fragments of predicted sizes. The nucleotide sequence of each fragment showed high sequence identity (85 to 100%) to bovine LPO. Hybridization screening of the sheep cDNA library using oligo II identified several clones. The two longest cDNAs were sequenced (Figure 2) and were 96% identical to bovine LPO (E.C. 1.11.1.7) and 80% identical to human milk and salivary LPO. Restriction analysis demonstrated that the other cDNAs were fragments of the longest clone. All of the clones were less related to human MPO (55% iden-

Gerson, Sabater, Scuri, et al.: Lactoperoxidase and Bacterial Clearance of Airways

667

Figure 2. Nucleotide sequence of ovine airway peroxidase. The complete sequence of the longest clone is shown (Genbank accession no. AF027970). The underlined protein sequence was confirmed by amino acid sequence analysis. Nucleotide 1957 (bold) was determined to be a C in the other cDNA isolated by hybridization screening.

tity), EPO (56% identity), or TPO (49% identity). Thus, only LPO sequences were detected in the tracheal mucosa cDNA library, strongly suggesting that airway peroxidase was identical to LPO. Comparison of the predicted amino acid sequence of sheep airway LPO to bovine milk, human milk, and salivary LPOs showed that all potential Asn-linked glycosylation sites were conserved between sheep airway LPO and

bovine milk LPO. Comparison of both these sequences to human salivary LPO showed that human LPO contained one less potential Asn-linked glycosylation site. The tissue distribution of LPO mRNA was investigated in the sheep respiratory tract by Northern blot analysis (Figure 3). LPO mRNA was detected in the conducting airways but was greatly reduced in lung parenchyma. Higher levels were seen in the bronchi than in the trachea,

668


Figure 3. Distribution of LPO mRNA in sheep tissues. Northern blots of sheep tissues (trachea, main bronchus, third and fifth generation bronchi, lung parenchyma, lactating mammary gland, small intestine, and lacrimal gland) were hybridized and quantitated as described in MATERIALS AND METHODS. Relative LPO values are expressed as ratios of mammary gland values corrected for loading differences by GAPDH signals. Blots were normalized by setting mammary gland LPO and GAPDH values to 1.0. Values are the mean ⫾ standard deviation of six Northern blots. Asterisk indicates that no standard deviations are shown for mammary as its value is arbitrarily set to 1.0 for normalization.

perhaps reflecting higher levels of nonepithelial tissues in the trachea. The near absence of peroxidase message in lung parenchyma suggests that LPO does not contribute to the peroxidase activity found in alveoli and previously identified by others as glutathione peroxidase (15). The higher variability seen in third through fifth generation bronchi may reflect differences in dissection that lead to inclusion of different amounts of parenchyma with these smaller airways.

Figure 4. Sheep airway peroxidase. (Top panel) SDS gel analysis of peroxidases. Purified airway LPO activity corresponded to a homogeneous protein species of 83 kD (lane 1) that was similar in size to sheep milk LPO (lane 2) and different from bovine LPO (lane 3) and canine MPO (lane 4). MPO has two chains formed by proteolytic processing that accounts for the large difference in size when compared with LPOs. (Bottom panel) Amino acid sequence of ovine airway LPO. A CNBr digest of sheep airway peroxidase was resolved on Tris-tricine SDS gels (9), and the sequenced peptide was identical to bovine LPO and distinct from other peroxidases. The underlined methionine is deduced from the CNBr cleavage mechanism. This comparison of sheep airway peroxidase sequence with other heme-peroxidases confirms that airway peroxidase is LPO.

Amino Acid Sequence and Enzymatic Activity of Airway Peroxidase To confirm that the airway peroxidase was LPO, as suggested by the sequences of the isolated cDNAs, purified sheep airway peroxidase (Figure 4) was compared with sheep milk LPO on SDS gels and used for amino acid sequence analysis. Sheep milk LPO comigrated with sheep airway peroxidase, indicating that the previously reported apparent molecular weight (MWapp) difference between sheep airway peroxidase and bovine milk peroxidase was due to species differences. The N-termini of the purified sheep enzyme appeared to be blocked as reported for bovine LPO (16), and peptide fragments were generated by CNBr treatment. One internal peptide was sequenced and was identical to a region of the amino acid sequence deduced from the airway LPO cDNA and to bovine LPO. The sequence of this peptide is shown in the bottom panel of Figure 4. Comparison with the aligned amino acid sequences of other LPOs and with other heme-peroxidases shows that the peptide is most likely derived from LPO. Thus, based on amino acid sequencing and cDNA sequences, airway peroxidase is LPO that is synthesized and secreted by the airway mucosa. The difference in apparent molecular weight between sheep and bovine LPO must be due to differences in postsynthetic modifications as suggested previously for rat LPOs (17) because their amino acid sequences were virtually identical. To demonstrate that airway LPO, like milk LPO, is capable of catalyzing the formation of hypothiocyanous acid (HOSCN)/OSCN⫺, purified sheep airway LPO and sheep milk LPO (Figure 4) were assayed for H2O2-dependent OSCN⫺ formation (Figure 5). The data showed that, in vitro, purified airway LPO was active in the production of this biocidal compound. The specific activity of sheep airway LPO was similar to sheep milk LPO and similar to published values for bovine milk LPO (18). Iodide stimulation of 3,3⬘,5,5⬘ N,N-tetramethylbenzidine oxidation was virtually identical for sheep LPOs (Figure 5) but different when compared with bovine milk LPO and canine MPO (6). The enzymatic differences between sheep airway LPO and bovine milk LPO (6) may be due to species differ-

Figure 5. Enzymatic activity of airway LPO and milk LPO. (A) OSCN⫺forming activities of purified sheep airway LPO (closed circles) and sheep milk LPO (closed triangles) were assayed as described in MATERIALS AND METHODS. (B) Pure sheep airway LPO (closed circles) and sheep milk LPO (closed triangles) were assayed by measuring oxidation of 3,3⬘,5,5⬘-tetramethylbenzidine in the presence of varying concentrations of NaI, as described previously (6). Activity is expressed relative to assay in the absence of NaI.


ences or perhaps to differences in handling during purification. Similar differences have been reported between other secreted peroxidases and bovine milk LPO (19–21) and ascribed to species differences. Thiocyanate in Sheep Airway Secretions To function as a protective mechanism against infection, substrates for LPO must also be present in the airway lumen together with the secreted catalytically active LPO. H2O2 and superoxide have been detected by us and others in airways of several species, including sheep (22), guinea pigs (23), and humans (24). Reports of SCN⫺ in human sputum and bronchoalveolar lavage have been conflicting and previously ascribed to saliva contamination (11). Our assays of sheep airway lavages, in which secretions undergo large dilutions during collection, demonstrated no detectable levels of SCN⫺; however, suctioning of airway secretions from intubated sheep that prevents contamination with saliva showed that undiluted secretions contained 0.16 mM SCN⫺. This concentration of SCN⫺ is within the normal range of human saliva (11, 25), is high enough to serve as a substrate for LPO (18), and is considered bacteriostatic (25). This demonstration of SCN⫺ in airway secretions, together with the detection of H2O2 reported previously, and our identification of airway peroxidase as LPO, shows that all components of the LPO biocidal system are present in airways. Role of Lactoperoxidase in Clearance of Bacteria from Airways The hypothesis that the LPO system functions in vivo to maintain airway sterility was examined using experimental bacterial challenge of the sheep respiratory tract. Sheep were pretreated by aerosol with dapsone, an inhibitor of peroxidases including airway peroxidase (6), or treated with PBS as a control. The animals were then challenged by aerosol with P. haemolytica, a natural pathogen in sheep. Immediately after the challenge and sequentially thereafter, a sample of airway fluid was collected and a quantitative culture of these samples was performed. Sheep pretreated with dapsone had significantly slower clearance compared with control animals at 60 and 180 min after the challenge (P ⬍ 0.05). After 60 min, a slight increase in CFU was noted in dapsone-treated animals, although this was not significantly different from the zero time point. The magnitude of bacterial clearance seen in these experiments is similar to that seen in vitro using human nasal secretions (26). Dapsone inhibition of peroxidase activity caused a 100-fold change in the number of bacteria remaining in the airway after 1 h. This change in bacterial clearance suggested that peroxidase was a significant contributor to bacterial clearance from the airways. Thus, sheep whose airway LPO was inhibited by dapsone were compromised in their ability to clear viable bacteria from their airway (Figure 6). To show that the result of the dapsone treatment was due primarily to inhibition of peroxidase activity and to rule out other effects such as depression of mucociliary velocity, dapsone-pretreated animals were administered bovine milk LPO by aerosol immediately after the bacterial challenge. In dapsone-treated animals, bovine LPO post-

669

treatment reversed most of the effects of dapsone treatment, supporting that the effect of dapsone is primarily inhibition of endogenous peroxidase. This was not due to an unrelated effect of the LPO, as LPO treatment did not improve clearance of the PBS-pretreated control animals (Figure 6). MPO from neutrophils is not expected to be a major contributor to H2O2 metabolism in the sheep airway lumen before bacterial challenge (6, 7). However, the infiltration of leukocytes into the airway lumen could contribute to endogenous peroxidase levels later, after bacterial challenge. Therefore, cell counts were performed on the secretions at the time of collection, and no white blood cells were detected in collected secretions at all time points, in all animals, as expected for the short time course of the experiment. Although dapsone could have an unknown effect on mucociliary clearance, it is not expected to act through inhibition of mucociliary clearance at the time points where our results are most significant because potent inhibitors of mucociliary clearance, e.g., atropine, have no effect on clearance of P. haemolytica from sheep airways during the first 2 h. In addition, the dapsone effect was reversed by adding exogenous LPO, an effect not expected if dapsone acted by inhibition of mucocilary clearance (13). Thus, the major endogenous peroxidase responsible for increasing bacterial clearance is not likely to be MPO derived from leukocytes infiltrating the airway lumen, although some MPO may contribute to the dapsone inhibitable activity.

Figure 6. Experimental bacterial challenge of sheep airways. Sheep were either pretreated with dapsone in PBS (open squares) or with PBS alone (open circles). Control sheep (n ⫽ 6) and pretreated animals (n ⫽ 4) were challenged with P. haemolytica (ATCC no. 29698) in 3 ml of PBS. Immediately after, 30 min, 1 h, and 3 h after challenge, samples of tracheal surface fluid were collected and quantitative bacterial cultures used to determine CFU. Values, normalized to the initial value after challenge, are plotted as means ⫾ SE. Control sheep showed rapid clearance of inhaled bacteria. Dapsone treatment significantly inhibited bacterial clearance at 60 and 180 min (*P ⬍ 0.05). Treatment with 5 mg bovine milk LPO reversed the impaired clearance in dapsone-treated animals (open inverted triangles). As an additional control, treatment with 5 mg of LPO alone did not significantly improve clearance of bacteria (open triangles).

670


These in vivo experiments show that inhibition of airway peroxidase enzymatic activity resulted in impaired clearance of inhaled bacteria that can be restored by addition of exogenous LPO. Thus, the LPO system may be an important component of normal airway mucosal defense.

Discussion Previous descriptions of peroxidase in airway goblet cells and submucosal glands (3–6) led to the proposal that endogenous airway peroxidase served as a defense against infection. Our experiments tested this hypothesis and showed that, in sheep, inhibiting the LPO system impaired bacterial clearance from the airway. The data also showed that sheep airway peroxidase is identical to LPO, that the LPO substrate SCN⫺ is present in sheep airways, and that airway peroxidase catalyzes the formation of OSCN⫺ in vitro. Taken together, these studies indicated that the LPO system is an important component of airway defense against infection. The sequence similarity of sheep airway peroxidase to bovine and human milk LPO was within the range expected for species differences (80 to 95%). The amino acid sequence of airway LPO and the failure to detect non-LPO peroxidase sequences in tracheal mucosa cDNA demonstrated that airway peroxidase was LPO and was the major heme-containing peroxidase message in normal airways. Given the sequence identity of bovine and sheep LPO, the 5-kD difference in MWapp between bovine and sheep LPO can only be due to species-specific differences in postsynthetic modifications, probably terminal sugars of N-linked oligosaccharides, and are not tissue-specific differences as sheep milk LPO was identical to sheep airway LPO. The previously reported minor differences in enzymatic activity of sheep airway and bovine milk LPOs (6) are not due to major differences in amino acid sequence and probably reflect the structural differences resulting from postsynthetic modifications or purification procedures. Human salivary peroxidase has been cloned and shown to be identical to human milk LPO (27), and our high stringency hybridization of sheep lacrimal gland cDNA with LPO probes indicates that lacrimal peroxidase is also likely LPO. The expression of LPO in airway, salivary, and lacrimal secretions suggests that LPO is a common feature of these epithelial defenses. That its expression appears to be constitutive in these mucosa, in contrast to the inducible expression of LPO during mammary gland milk production, suggests that LPO expression can be controlled by at least two different regulatory mechanisms. Studies by Kinbara and coworkers (5) indicate that airway peroxidase (i.e., LPO) may be regulated by still other stimuli, such as bacterial adherence or bacterial products, whereas studies by Watanabe and Harada (28) suggest that airway peroxidase is regulated by stimulation through beta adrenergic receptors. Whether airway LPO levels can be altered in response to hormone levels as seen in mammary tissue is not clear. In addition to LPO protein expression and secretion, the regulation of LPO enzymatic levels must also depend on adequate levels of its substrates: SCN and H2O2. The regulation of airway concentrations of these substrates is

poorly understood and clearly will be important to appreciate control of enzyme activity. Our previous studies (6) reported high levels of peroxidase activity in sheep airway secretions, and it was expected that this high level of catalytic capability could have a major impact on metabolic events in the airway lumen. The identification of airway peroxidase as LPO and demonstration of enzyme activity indistinguishable from same species milk LPO suggested that airway peroxidase’s potential functions and activities are those previously described by others for LPO, including that of an anti-infective system (2, 25). The magnitude of bacterial clearance seen in these experiments is similar to that seen in vitro using human nasal secretions (26). At the 1-h time point, dapsone inhibition of peroxidase activity caused a 100-fold change in the number of bacteria remaining in the airway after 1 h. This change in bacterial clearance suggested that peroxidase was a significant contributor to bacterial clearance from the airways. Dapsone was used in the in vivo studies because it is an inhibitor of peroxidases. Dapsone will inhibit neutrophil MPO catalyzed Cl⫺ oxidation at the expected neutral pH found in airways, and for this reason our experiments were limited to 3 h after challenge with bacteria, where no neutrophils were detected in the samples. Thus, the immediate effect of dapsone on bacterial clearance from the airways is expected to be inhibition of endogenous airway peroxidase that we have shown to be primarily LPO. Dapsone itself can be antibacterial most notably in the case of mycobacteria. Any direct growth inhibition of the bacteria caused by dapsone would lower the number of bacteria in the airway and would appear as faster clearance than normal, not slower clearance as the data showed. In order to impact the significance of the bacterial clearance data, dapsone would have to act as a growth stimulant for P. haemolytica, an effect not previously reported for P. haemolytica or any other organism using dapsone or other related compounds. P. haemolytica is catalase-positive and may provide the organism with protection from oxidizing effects of H2O2. Bacterial catalase is restricted to the cytoplasm or periplasm and thus would not be expected to significantly alter the availability of H2O2 for use as a substrate by airway LPO; nor would the bacterial catalase provide protection against the OSCN⫺ produced by LPO catalysis. Some bacteria do produce H2O2, and these organisms would perhaps increase the effectiveness of the LPO system by increasing available substrate, thus suggesting an organismspecific spectrum of LPO antibacterial activity. LPO anti-infective activity is well investigated (2, 25) and has been shown to be effective against viruses (29–31) and fungi (32, 33), as well as bacteria. Its effects in saliva are well studied, and several groups have shown that it is an effective contributor to oral defenses (34, 35). We hypothesize that the LPO system is a significant contributor to the maintenance of airway sterility and only after this first line of defense is overcome does the organism recruit the more damaging activity of neutrophil MPO. This hypothesis is based on (1) the broad spectrum of LPO activity against infectious diseases; (2) the less damaging though less biocidal product of LPO catalysis whose continuous presence is


more easily tolerated by epithelial cells than is the HOCl; and (3) its apparently constitutive expression in the airway. The importance of chronic bacterial colonization of the airway in cystic fibrosis has focused attention on normal airway defense systems. Recently, defensin-like and ␤-defensin activities have been shown to contribute to bacterial killing by secretions of airway epithelial cells cultured at an air-liquid interface (36) and in implanted tracheal chambers (37) and that these activities are compromised in patients with cystic fibrosis (37). The in vivo existence of the airway LPO system shown here complements these antibacterial systems and provides additional protection against viral (30, 31, 38) and fungal infections (32, 33). Acknowledgments: Portions of this work were supported by a grant from the American Lung Association of Florida, Inc. (G.E.C.) and by NIH grant HL20989 (A.W.). The authors’ Genbank accession number is AF027970.

References 1. Klebanoff, S. J. 1991. Myeloperoxidase: occurence and biological functions. In Peroxidases in Chemistry and Biology. J. Everse, K. E. Everse, and M. B. Grisham, editors. CRC Press, Boca Raton, Florida. 1–35. 2. Thomas, E. L., P. M. Bozeman, and D. B. Learn. 1991. Lactoperoxidase: structure and catalytic properties. In Peroxidases in Chemistry and Biology. J. Everse, K. E. Everse, and M. B. Grisham, editors. CRC Press, Boca Raton, FL. 123–142. 3. Christensen, T. G., G. C. Blanchard, G. Nolley, and J. A. Hayes. 1981. Ultrastructural localization of endogenous peroxidase in the lower respiratory tract of the guinea pig. Cell Tissue Res. 214:407–415. 4. Christensen, T. G., and J. A. Hayes. 1982. Endogenous peroxidase in the conducting airways of hamsters: morphologic evidence of synthesis and secretion. Am. Rev. Respir. Dis. 125:341–346. 5. Kinbara, M., T. Ueda, and K. Hirai. 1992. Expression of peroxidase activity in rat tracheal epithelial cells associated with Mycoplasma pulmonis. Am. J. Physiol. 262:L92–L99. 6. Salathe, M., M. Holderby, R. Forteza, W. M. Abraham, A. Wanner, and G. E. Conner. 1997. Isolation and characterization of a peroxidase from the airway. Am. J. Respir. Cell Mol. Biol. 17:97–105. 7. Salathe, M., P. Guldimann, G. E. Conner, and A. Wanner. 1995. Hydrogen peroxide-scavenging properties of sheep airway mucus. Am. J. Respir. Crit. Care Med. 151:1543–1550. 8. Zeng, J., and R. E. Fenna. 1992. X-ray crystal structure of canine myeloperoxidase at 3 A resolution. J. Mol. Biol. 226:185–207. 9. Schagger, H., and G. von Jagow. 1987. Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal. Biochem. 166:368–379. 10. Thomas, E. L., and M. Fishman. 1986. Oxidation of chloride and thiocyanate by isolated leukocytes. J. Biol. Chem. 261:9694–9702. 11. Dacre, J. C., and I. R. Tabershaw. 1970. Thiocyanate in saliva and sputum: relationship to smoking and industrial exposures. Arch. Environ. Health 21:47–49. 12. Chomczynski, P., and N. Sacchi. 1987. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162:156–159. 13. Whiteside, M. E., I. Lauredo, G. A. Chapman, K. R. Ratzan, W. M. Abraham, and A. Wanner. 1984. Effect of atropine on tracheal mucociliary clearance and bacterial counts. Bulletin Europeen de Physiopathologie Respiratoire 20:347–351. 14. Thomas, E. L., M. M. Jefferson, R. E. Joyner, G. S. Cook, and C. C. King. 1994. Leukocyte myeloperoxidase and salivary lactoperoxidase: identification and quantitation in human mixed saliva. J. Dent. Res. 73:544–555. 15. Jenkinson, S. G., R. A. Lawrence, and W. Y. Tucker. 1984. Glutathione peroxidase, superoxide dismutase, and glutathione S-transferase activities in human lung. Am. Rev. Respir. Dis. 130:302–304.

671

16. Dull, T. J., C. Uyeda, A. D. Strosberg, G. Nedwin, and J. J. Seilhamer. 1990. Molecular cloning of cDNAs encoding bovine and human lactoperoxidase. DNA Cell Biol. 9:499–509. 17. De, P. K. 1992. Tissue distribution of constitutive and induced soluble peroxidase in rat: purification and characterization from lacrimal gland. Eur. J. Biochem. 206:59–67. 18. Pruitt, K. M., B. Mansson-Rahemtulla, D. C. Baldone, and F. Rahemtulla. 1988. Steady-state kinetics of thiocyanate oxidation catalyzed by human salivary peroxidase. Biochemistry 27:240–245. 19. Mazumdar, A., R. Chatterjee, S. Adak, A. Ghosh, C. Mondal, and R. K. Banerjee. 1996. Characterization of sheep lacrimal-gland peroxidase and its major physiological electron donor. Biochem. J. 314:413–419. 20. Mansson-Rahemtulla, B., F. Rahemtulla, and M. G. Humphreys-Beher. 1990. Human salivary peroxidase and bovine lactoperoxidase are crossreactive. J. Dent. Res. 69:1839–1846. 21. Pruitt, K. M., D. N. Kamau, K. Miller, B. Mansson-Rahemtulla, and F. Rahemtulla. 1990. Quantitative, standardized assays for determining the concentrations of bovine lactoperoxidase, human salivary peroxidase, and human myeloperoxidase. Anal. Biochem. 191:278–286. 22. Liberman, H., A. T. Mariassy, D. Sorace, S. Suster, and W. M. Abraham. 1995. Morphometric estimation of superoxide generation in allergeninduced airway hyperresponsiveness. Lab. Invest. 72:348–354. 23. Kinnula, V. L., K. B. Adler, N. J. Ackley, and J. D. Crapo. 1992. Release of reactive oxygen species by guinea pig tracheal epithelial cells in vitro. Am. J. Physiol. 262:L708–L712. 24. Dohlman, A. W., H. R. Black, and J. A. Royall. 1993. Expired breath hydrogen peroxide is a marker of acute airway inflammation in pediatric patients with asthma. Am. Rev. Respir. Dis. 148:955–960. 25. Reiter, N., and J. P. Perraudin. 1991. Lactoperoxidase: biological function. In Peroxidases in Chemistry and Biology, Vol. I. J. Everse, K. E. Everse, and M. B. Grisham, editors. CRC Press, Boca Raton, Florida. 143. 26. Cole, A. M., P. Dewan, and T. Ganz. 1999. Innate antimicrobial activity of nasal secretions. Infect. Immun. 67:3267–3275. 27. Kiser, C., C. K. Caterina, J. A. Engler, B. Rahemtulla, and F. Rahemtulla. 1996. Cloning and sequence analysis of the human salivary peroxidaseencoding cDNA. Gene 173:261–264. 28. Watanabe, K., and H. Harada. 1990. Beta-adrenoceptor control of peroxidase synthesis in nasal glands. Ann. Otol. Rhinol. Laryngol. 99:581–585. 29. Courtois, P., D. van Beers, M. de Foor, I. M. Mandelbaum, and M. Pourtois. 1990. Abolition of herpes simplex cytopathic effect after treatment with peroxidase generated hypothiocyanite. Journal de Biologie Buccale 18:71–74. 30. Pourtois, M., C. Binet, N. Van Tieghem, P. Courtois, A. Vandenabbeele, and L. Thiry. 1990. Inhibition of HIV infectivity by lactoperoxidase-produced hypothiocyanite. Journal de Biologie Buccale 18:251–253. 31. Yamaguchi, Y., M. Semmel, L. Stanislawski, A. D. Strosberg, and M. Stanislawski. 1993. Virucidal effects of glucose oxidase and peroxidase or their protein conjugates on human immunodeficiency virus type 1. Antimicrob. Agents Chemother. 37:26–31. 32. Lenander-Lumikari, M. 1992. Inhibition of Candida albicans by the peroxidase/SCN-/H2O2 system. Oral Microbiology and Immunology 7:315–320. 33. Popper, L., and D. Knorr. 1997. Inactivation of yeast and filamentous fungi by the lactoperoxidase hydrogen peroxide thiocyanate system. Nahrung 41:29–33. 34. Thomas, E. L., K. P. Bates, and M. M. Jefferson. 1981. Peroxidase antimicrobial system of human saliva: requirements for accumulation of hypothiocyanite. J. Dent. Res. 60:785–796. 35. Tenovuo, J., K. M. Pruitt, and E. L. Thomas. 1982. Peroxidase antimicrobial system of human saliva: hypothiocyanite levels in resting and stimulated saliva. J. Dent. Res. 61:982–985. 36. Smith, J. J., S. M. Travis, E. P. Greenberg, and M. J. Welsh. 1996. Cystic fibrosis airway epithelia fail to kill bacteria because of abnormal airway surface fluid [published erratum appears in Cell 1996 Oct 18;87(2):following 355]. Cell 85:229–236. 37. Goldman, M. J., G. M. Anderson, E. D. Stolzenberg, U. P. Kari, M. Zasloff, and J. M. Wilson. 1997. Human beta-defensin-1 is a salt-sensitive antibiotic in lung that is inactivated in cystic fibrosis. Cell 88:553–560. 38. Mikola, H., M. Waris, and J. Tenovuo. 1995. Inhibition of herpes simplex virus type 1, respiratory syncytial virus and echovirus type 11 by peroxidasegenerated hypothiocyanite. Antiviral Res. 26:161–171.