Cloning and Characterization of the Gene for a New Epithelial β-Defensin

Mammalian β-defensins are endogenous cysteine-rich peptide antibiotics that are produced either by epithelial cells lining the respiratory, digestive, and urogenital tracts or by granulocytes and macrophages. A growing body of evidence has implicated these peptides in host defense, particularly mucosal innate immunity. We previously reported the cloning of the full-length cDNA for a porcine β-defensin (pBD-1), which was found to be expressed throughout the airway and oral mucosa. Here, we provide the structural organization of the pBD-1 gene, showing that the entire gene spans ∼1.9 kilobases with two short exons separated by a 1.5-kilobase intron. Fluorescence in situ hybridization mapped the pBD-1 gene to porcine chromosome 15q14-q15.1 within a region of conserved synteny to the chromosomal locations of human and mouse α- and β-defensins. We also provide several independent lines of evidence showing that thepBD-1 gene is expressed constitutively during inflammation and infection, despite its resemblance to many inducible epithelial β-defensins in amino acid sequence, genomic structure, and sites of expression. First, stimulation of primary porcine tongue epithelial cells with lipopolysaccharide, tumor necrosis factor-α, and interleukin (IL)-1β failed to up-regulate the expression ofpBD-1 mRNA. Second, pBD-1 gene expression was not enhanced in either digestive or respiratory mucosa of pigs following a 2-day infection with Salmonella typhimurium orActinobacillus pleuropneumoniae. Last, direct transfection of the pBD-1 gene promoter into NIH/3T3 cells showed no difference in reporter gene activity in response to stimulation by lipopolysaccharide and IL-1β. The constitutive expression ofpBD-1 in airway and oral mucosa, which is consistent with a lack of consensus binding sites for nuclear factor-κB or NF-IL-6 in its promoter region, suggests that it may play a surveillance role in maintaining the steady state of microflora on mucosal surfaces.

Mammalian defensins constitute a large family of endogenous peptide antibiotics with broad spectrum activity against various bacteria, mycobacteria, fungi, and viruses (1)(2)(3)(4). Upon microbial invasion, mature active defensins are released quickly by proteolytic processing from precursor peptides either in cytoplasmic granules of most mammalian neutrophils and macrophages or in epithelial cells of the respiratory, gastroenteric, and urogenital tracts (2)(3)(4)(5). Killing of bacteria by defensins appears to involve direct interactions with target cell membranes, followed by insertion and permeabilization of membranes by forming multimeric pores (1-4, 6 -9).
All defensins are polycationic 3-5-kDa peptides and are characterized by the presence of six conserved cysteine residues forming three intramolecular disulfide bonds with a compact triple-stranded ␤-sheet structure (6 -10). Based on the positions of six cysteine residues and linkages of the disulfide bonds, defensins are divided further into two groups: ␣and ␤-defensins. In contrast to most ␣-defensins, which are of myeloid origin and expressed constitutively in granulocytic leukocytes, many ␤-defensins are synthesized inducibly by epithelial cells during inflammation and infection (2)(3)(4)11). Interestingly, despite a lack of similarity in nucleotide and amino acid sequences between these two groups, all ␣and ␤-defensin genes are clustered in close proximity on the same chromosome, as demonstrated in humans (12)(13)(14)(15), mice (16,17), cattle (18,19), and sheep (19,20). This finding suggests that they may have evolved from a common ancestral defense gene by duplication (14,15).
Several peptide antibiotics, particularly ␤-defensins, have been implicated in chronic bacterial infections in cystic fibrosis patients, where it is hypothesized that they are inactivated by the high-salt airway surface fluid (21)(22)(23)(24)(25), suggesting that epithelial ␤-defensins may play an important role in mucosal host defenses. Based on gene structure, length, and homology of amino acid sequences and sites of expression, epithelial ␤-defensins are classified further into two subgroups (11,26) (Fig.  1). The first group contains precursor peptides of 68 -69 amino acid residues, each with a relatively large gene containing an intron of Ͼ6.5 kb. 1 This group includes human ␤-defensin-1 (22,(27)(28)(29)(30)(31)(32) and mouse ␤-defensin-1 (17,33). Besides the epithelia of respiratory and digestive tracts, kidney and genitourinary epithelia appear to be the major expression sites. The second group contains precursor peptides that are generally 5-6 amino acids shorter, and each has a rather compact gene with a shorter intron of Ͻ2 kb. Bovine tracheal antimicrobial peptide (TAP) (34 -37), lingual antimicrobial peptide (37)(38)(39), and enteric ␤-defensin (40) and human BD-2 (12,24,41) belong to this group. They are expressed abundantly in oral and airway epithelia. Another striking difference between these two groups is that the ␤-defensins of the first group have a constitutive expression pattern, whereas genes in the second group are expressed inducibly during inflammation and infection. Genes for TAP-like inducible ␤-defensins contain several consensus recognition sites for nuclear factor (NF)-B and NF-IL-6 in their promoter regions, which may explain their inducibility upon inflammation and infection (12,35,40). However, the reliability of this classification scheme remains to be tested because significant interspecies differences exist in gene families of many peptide antibiotics (1).
Recently, we cloned the full-length cDNA for a porcine ␤-defensin (pBD-1), which is expressed abundantly in tongue epithelia and to a lesser extent throughout the respiratory and digestive tracts (26). It resembles many inducible epithelial ␤-defensins, including bovine TAP, lingual antimicrobial peptide, and enteric ␤-defensin and human BD-2 in both amino acid sequence and sites of expression (26). Recombinant pBD-1 has potent antibacterial activity against Escherichia coli, Salmonella typhimurium, and Listeria monocytogenes in a salt-dependent fashion and also is synergistic with several porcine neutrophil-derived peptide antibiotics in killing these bacteria (42).
Here, we report the cloning, mapping, and characterization of the pBD-1 gene. The gene mapped to Sus scrofa chromosome 15q14-q15.1, providing the first direct evidence that Homo sapiens chromosome 8p23, the location of human ␣and ␤-defensin genes (12)(13)(14)(15), corresponds to the region on S. scrofa chromosome 15 previously shown to contain a narrow portion of H. sapiens chromosome 8 (43,44). In contrast to the genes for other inducible epithelial ␤-defensins, the pBD-1 gene does not contain consensus binding sites in the promoter region for either NF-B or NF-IL-6, suggesting that it may not be modulated by proinflammatory mediators. Indeed, we provide several lines of evidence showing that the pBD-1 gene is expressed constitutively during inflammation and infection. To our knowledge, this is the first example that contradicts the current classification of epithelial ␤-defensins by showing that TAP-like ␤-defensins do not always exhibit an inducible expression pattern. These findings also support the hypothesis that epithelial pBD-1 may play a role in maintaining the balance of natural microflora on mucosal surfaces under basal conditions.

EXPERIMENTAL PROCEDURES
Cloning of the pBD-1 Gene-A two-step polymerase chain reaction (PCR) was used to screen a porcine genomic bacterial artificial chromo-some (BAC) library 2 for BAC clones containing the pBD-1 gene as described previously (45). Two pairs of primers derived from the cDNA or partial genomic sequence of pBD-1 were used: 5Ј-ATG AGA CTC CAC CGC CTC CTC CT-3Ј (sense) and 5Ј-GCA GCA TTT GAC TTG GGG CAT G-3Ј (antisense) (product size: 1.7 kb); and 5Ј-GGG TTA TGA AGT ATC CAT GTA ACC GAC-3Ј (sense) and 5Ј-GGT CAG CTC CAC TAG GAC AGC AGA G-3Ј (antisense) (product size: 812 bp). Positive BAC clones were purified and digested with BamHI or EcoRI and then subcloned into the pBluescript II KS ϩ plasmid (Stratagene, La Jolla, CA) as described (46). The subsequent clones were screened for the presence of the pBD-1 gene by PCR amplification using the above four primers. The positive plasmids were sequenced directly with genespecific primers using a 33 P-labeled dideoxynucleotide terminator cycle sequencing kit (Amersham Pharmacia Biotech).
Chromosomal Localization of the pBD-1 Gene-Metaphase chromosome spreads, prepared from a porcine peripheral blood mononuclear cell culture, were subjected to fluorescence in situ hybridization (FISH) as described previously (47,48). Briefly, BAC DNA was labeled with biotin by nick translation (Roche Molecular Biochemicals) and hybridized to metaphase chromosome DNA. Specific binding was detected with fluorescein isothiocyanate-labeled streptavidin and biotinylated anti-streptavidin antibody (Vector Labs, Inc., Burlingame, CA), followed by propidium iodide staining to simultaneously induce the Rbanding. A two-color FISH was performed using elongated chromosome spreads with a biotin-labeled pBD-1 BAC DNA and a digoxigeninlabeled BAC DNA for SW919, which is a United States Department of Agriculture Meat Animal Research Center marker for porcine chromosome 15. After hybridization, the probes were detected with rhodaminelabeled anti-digoxigenin antibody (Roche Molecular Biochemicals) and fluorescein isothiocyanate-labeled streptavidin/biotinylated antistreptavidin antibody, followed by 4,6-diamidino-2-phenylindole staining to visualize chromosomes.
Primer Extension of pBD-1 mRNA-Poly(A) ϩ mRNA from porcine tongue epithelial cells was isolated using oligo(dT) 30 -containing Oligotex latex beads (QIAGEN Inc., Chatsworth, CA), annealed with the 5Ј-end-labeled primer 5Ј-CTG GCA CAG GTA ACA GGA CCA TGA GG-3Ј, and extended with avian myeloblastosis virus reverse transcriptase (Promega, Madison, WI) as described previously (49). The extension product was heat-denatured and analyzed on a urea-6% polyacrylamide DNA sequencing gel and then subjected to autoradiography. A dideoxynucleotide termination DNA sequence ladder was used as the size standard.
Culture and Stimulation of Porcine Tongue Epithelial Cells-Because pBD-1 mRNA could be detected easily by Northern blot analysis in porcine tongue epithelial cells, primary tongue epithelia were scraped from the filiform papillae of the dorsal side of tongues of 6 -8-week-old healthy outbred pigs and used for stimulation. All experiments were conducted under stringent LPS-free conditions. Low-endotoxin fetal bovine serum was used (Hyclone Laboratories, Logan, UT), and cell culture media were below 0.03 endotoxin units/ml (Limulus amebocyte lysate assay, Sigma). Cells were washed three times with serum-free RPMI 1640 medium and plated in six-well tissue culture plates in complete RPMI 1640 medium (RPMI 1640 medium containing 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin). After 3 h of incubation at 37°C and 5% CO 2 ,
Infection of Pigs with S. typhimurium and Actinobacillus pleuropneumoniae-Healthy outbred pigs at 6 -8 weeks of age were obtained from the Kansas State University Swine Research Unit and housed in an environmentally controlled isolation facility. Infection studies began 1 week after pigs were acclimated to their new environment. Clinical signs for bacterial infections were monitored during the experiments. For S. typhimurium infection, pigs were gavaged with 10 9 colonyforming units of bacteria suspended in growth medium. Samples of jejunum and colon were collected at 12, 24, and 48 h post-infection immediately following euthanasia using three to four pigs/time point. Four control pigs were gavaged with growth medium and euthanized at 6 h. Epithelial tissues were scraped from the collected intestinal samples and used for RNA isolation. For A. pleuropneumoniae infection, pigs were challenged intranasally either with 5 ϫ 10 8 colony-forming units of bacteria in growth medium or with medium alone using four pigs/group. Tracheal epithelial tissues were collected 48 h post-infection and used for isolation of RNA.
Reverse Transcriptase-PCR-Total RNA was extracted using Trizol reagent (Life Technologies, Inc.) following the manufacturer's instructions. One g of each RNA sample was used to perform first-strand cDNA synthesis in a total volume of 20 l with 25 units of Moloney murine leukemia virus reverse transcriptase (Perkin-Elmer) and 10 units of RNase inhibitor (Perkin-Elmer) as described previously (28). One-tenth of each resulting cDNA was used in the subsequent PCR in a 25-l reaction volume with a 0.1 M concentration of each sense and antisense primer and 1.25 units of platinum Taq DNA polymerase (Life Technologies, Inc.). The primers for amplification of pBD-1 cDNA were 5Ј-ATG AGA CTC CAC CGC CTC CTC CT-3Ј (sense) and 5Ј-GCA GCA TTT GAC TTG GGG CAT G-3Ј (antisense). The primers for the housekeeping gene ␤-actin were 5Ј-GGA CTT CGA GCA GGA GAT GG-3Ј (sense) and 5Ј-GCA CCG TGT TGG CGT AGA GG-3Ј (antisense). The PCR program was 2 min at 94°C, followed by 26 (for pBD-1) or 16 (for ␤-actin) cycles of 30 s at 60°C, 30 s at 72°C, and 30 s at 94°C, followed by a final extension for 7 min at 72°C.
Southern and Northern Blot Analyses-For Southern analysis, 10 g of porcine genomic or BAC DNA or 10 l of PCR products were separated on agarose gels, followed by denaturation and overnight capillary transfer onto positively charged nylon membranes (Roche Molecular Biochemicals). For Northern analysis, 10 g of total RNA from tongue epithelial cells were denatured and fractionated on formaldehyde-containing 1.2% agarose gels and then blotted onto positively charged nylon membranes. Southern and Northern blots were hybridized in ExpressHyb hybridization solution (CLONTECH, Palo Alto, CA) for 2 h at 60°C, followed by low-and high-stringency washes as described (28). The full-length cDNA of pBD-1 and the 198-bp cDNA fragment of ␤-actin were labeled randomly with 32 P and used as probes. A 452-bp glyceraldehyde-3-phosphate dehydrogenase cDNA probe, generated by PCR with primers 5Ј-ACC ACA GTC CAT GCC ATC AC-3Ј (sense) and 5Ј-TCC ACC ACC CTG TTG CTG-3Ј (antisense), was used as the internal control in Northern blots.
Plasmid and Cell Transfection-A 1466-bp pBD-1 promoter, which contained the 1384-bp sequence upstream of the pBD-1 transcription initiation site and the 82-bp complete exon 1 sequence, was constructed by PCR amplification from the positive pBluescript plasmid using primers 5Ј-CCG CTC GAG GGT CTG GAT AAC TCG GCT-3Ј (sense) and 5Ј-CCC AAG CTT CTG GCA CAG GTA ACA GGA C-3Ј (antisense). The XhoI and HindIII restriction sites were introduced at the 5Ј-ends of the sense and antisense primers, respectively. The subsequent PCR product was digested with HindIII and XhoI and then fused with a luciferase reporter gene in the pGL3-Basic vector (Promega). The presence of the correct insert was confirmed by direct sequencing and restriction digestion. For transfection, the plasmid was purified, and endotoxin was removed with an EndoFree plasmid purification kit (QIAGEN Inc.). The pNF-B-Luc vector (CLONTECH), which contains four tandem copies of the NF-B response element fused with the herpes simplex virus thymidine kinase minimal promoter and the luciferase reporter gene, was transfected as a positive control.
Murine embryonic fibroblast NIH/3T3 cells were obtained from the American Tissue Culture Collection (Manassas, VA), expanded, and maintained in Dulbecco's modified Eagle's medium supplemented with 10% low-endotoxin fetal bovine serum. Cells were seeded in 24-well tissue culture plates and grown to 60 -80% confluence before transfec-tion. One g of plasmid DNA containing the pBD-1 promoter was introduced into each well using 3 l of Transfast transfection reagent (Promega) in 200 l of serum-free medium in triplicate following the manufacturer's instructions. The pNF-B-Luc vector (0.5 g/well) also was transfected into cells separately. Following 1 h of incubation, 1 ml of serum-free Dulbecco's modified Eagle's medium was added to each well. All cells were incubated for 24 h after transfection, followed by stimulation with LPS (1 and 10 g/ml) and/or IL-1␤ (10 ng/ml) for another 6 h. Luciferase activity was measured in cellular extracts using the luciferase assay system (Promega). Data were normalized against the protein concentration of each cellular extract, which was measured by the micro BCA protein assay reagent (Pierce).

Isolation of the pBD-1 Gene and Its Genomic Sequence-
After two rounds of PCR using gene-specific primers, two positive clones, designated 838 and 221, were identified from a total of 103,448 clones from the porcine BAC genomic library. Southern blot analysis of digested genomic DNA and both BAC clones, probed with the full-length pBD-1 cDNA, showed the same restriction pattern with only one or two bands with each of four restriction enzymes (Fig. 2). These findings suggested that both BAC clones contained the entire pBD-1 gene and that pBD-1 is a single-copy gene. An oligonucleotide from the 5Ј-end of the pBD-1 cDNA hybridized to only one of the two bands in EcoRI-, PstI-, and BamHI-digested DNAs (Fig. 2C), indicating that the other band contained the exon 2 sequence. A 3769-bp sequence of the gene for pBD-1 was determined and deposited in GenBank TM . Comparison of this sequence with the pBD-1 cDNA indicated that the pBD-1 gene is composed of two short exons interrupted by a 1535-bp intron (Fig. 3A). Exon 1 encodes the 5Ј-untranslated region and signal sequence of the prepro-pBD-1 peptide, and exon 2 encodes the pro-and mature sequences as well as the 3Ј-untranslated region of the pBD-1 peptide (Fig. 3A). Primer extension revealed a major product of 82 nucleotides (Fig. 4), suggesting that transcription of the pBD-1 gene was initiated primarily from thymine, which was assigned as nucleotide ϩ1 (Fig. 3B).
A 1384-bp fragment upstream from the pBD-1 gene transcription initiation site was scanned for the presence of consensus binding sites for known transcription factors using the Findpatterns program (Genetics Computer Group, Madison, WI). Somewhat to our surprise, no consensus binding sites for NF-B or NF-IL-6 were found in the promoter region of the pBD-1 gene, distinguishing it from other inducible epithelial ␤-defensins (Fig. 3C). A search of the non-redundant Gen-Bank TM gene data base using the gapped BLASTN program

pBD-1 Gene Is Expressed Constitutively during Infection
with the entire pBD-1 gene sequence identified at positions ϩ1972 to ϩ2108 of the intron a short interspersed nuclear element sequence, which is a repetitive DNA sequence in the genomes of many mammalian species (50 -52). Interestingly, TAP and enteric ␤-defensin genes also contained a highly repetitive Alu sequence in their promoter and intron regions (35,40). However, the regulatory role of these repetitive elements remains unclear. No other matches were notable except for some homology in the exon regions to several ␤-defensins.
Chromosomal Localization of the pBD-1 Gene-Two BAC clones that contained the pBD-1 gene were used to map the pBD-1 gene on porcine chromosomes. Based on an inspection of Ͼ100 metaphase chromosome spreads, the pBD-1 gene was demonstrated to be localized on S. scrofa chromosome 15 at 15q14-q15.1 (Fig. 5A). In situ hybridization was also performed simultaneously with both BAC clones 838 and 221 and revealed FIG. 3. Structure of the pBD-1 gene. A, structural organization of the pBD-1 gene and schematic drawing of the pBD-1 cDNA as well as its encoded precursor peptide. The pBD-1 gene contains two short exons interrupted by a large intron with a short interspersed nuclear element sequence between positions ϩ1972 and ϩ2108. The prepro-pBD-1 peptide is composed of a signal sequence, a proregion, and a mature peptide with antibacterial activity. E, EcoRI; P, PstI, B, BamHI; SINE, short interspersed nuclear element; UTR, untranslated region. B, a 1466-bp promoter sequence of the pBD-1 gene. This fragment was used in the transfection studies, except for the intron sequence. The sequence is numbered based on the primer extension results (see Fig. 4), with the transcription start site numbered as ϩ1. Exon 1 sequence is shown in uppercase letters; the start codon is in boldface; and the TATA box is indicated. C, comparison of the core promoter sequence of pBD-1 with those of inducible ␤-defensins. The TATA box as well as NF-B and NF-IL-6 recognition sites are boxed. Dashes were introduced to maximize the alignment, and exon sequences are shown in uppercase letters. Besides the TATA box, the consensus binding sites for NF-B and NF-IL-6 are fairly conserved in the genes for TAP, enteric ␤-defensin, and human BD-2, but are absent in the pBD-1 gene. that they could not be distinguished from each other on chromosome 15, suggesting that both clones reside at the same location (data not shown). To confirm the relative position of the pBD-1 gene on chromosome 15, two-color FISH was performed using BAC clone 838 for pBD-1 and another BAC DNA for SW919 (Fig. 5B). SW919 is a United States Department of Agriculture Meat Animal Research Center marker for S. scrofa chromosome 15 and was assigned to S. scrofa chromosome 15q13.1-q13.2. 3 As shown in Fig. 5B, both BAC clones were colocalized on the same chromosome in close proximity, and the pBD-1 gene (green) was relatively telomeric to SW919 (red).
pBD-1 Gene Is Expressed Constitutively during Inflammation and Infection-Because no consensus binding sites for NF-B and NF-IL-6 were found in the promoter region of the pBD-1 gene, we conducted in vitro and in vivo challenge studies to evaluate the inducibility of this gene. We first stimulated in vitro primary porcine tongue epithelial cells, where the pBD-1 mRNA could be detected easily by Northern blot analysis. As shown in Fig. 6, the pBD-1 mRNA was expressed abundantly in resting porcine tongue epithelia. However, stimulation of the cells with LPS (0.1 and 1 g/ml), TNF-␣ (10 ng/ml), IL-1␤ (10 ng/ml), or a combination of the three proinflammatory mediators failed to further up-regulate the expression of the pBD-1 gene at 9 h (Fig. 6) or 16 h (data not shown) post-challenge. Because of the paucity of information on inducible genes in porcine tongue epithelia, it is not possible to disregard experimental procedure as the reason for the lack of up-regulation of pBD-1. However, electrophoretic mobility shift assays have shown the nuclear translocation of NF-B in these cells when stimulated with LPS. 4 The constitutive expression of the pBD-1 mRNA during infection was also confirmed by two in vivo bacterial challenge studies. We experimentally infected 6 -8-week-old healthy pigs either with 10 9 colony-forming units of an enteric diseasecausing bacterium, S. typhimurium, or with 5 ϫ 10 8 colonyforming units of a respiratory disease-causing bacterium, A. pleuropneumoniae. Typical clinical signs of salmonellosis or pneumonia, including fever and diarrhea, were observed 6 -12 h following oral or intranasal infection with S. typhimurium or A. pleuropneumoniae, but not in control pigs, which were mockchallenged with growth medium alone. Epithelial tissues from intestine or trachea were sampled from each of the control and infected pigs at various times following infection. The expression of pBD-1 mRNA was not increased in jejunum (Fig. 7A), colon (data not shown), or tracheal epithelia (Fig. 7B) at any time examined during a period of 2 days.
We also performed 3Ј-rapid amplification of cDNA ends using normal and bacteria-challenged epithelial tissues from tongue, colon, and trachea as described (28). A degenerate upstream primer, 5Ј-ATG AG(A/G) CTC CA(C/T) C(A/G)C CT(C/G) CTC CT-3Ј, was designed based on the highly conserved region in the signal sequences of porcine, bovine, and sheep ␤-defensins. This would allow identification of possible isoforms of porcine ␤-defensins. However, only the pBD-1 cDNA sequence was found after sequencing eight independent plasmid clones derived from each of the above epithelial tissues. 4 Furthermore, Southern hybridization of the rapid am- FIG. 4. Primer extension of pBD-1 mRNA. Three g of tongue epithelial mRNA were annealed to a pBD-1-specific oligonucleotide, reverse-transcribed, and then subjected to urea-6% polyacrylamide gel electrophoresis and autoradiography. A DNA dideoxynucleotide termination DNA sequence ladder was used as the size standard on the left. A major product is 82 nucleotides (nt) in length. 5. Chromosomal mapping of the pBD-1 gene. Chromosomal fluorescence in situ hybridization mapped the pBD-1 gene to chromosome 15 at 15q14-q15.1. Arrows indicate the hybridization signals. A, biotin-labeled BAC DNA containing the pBD-1 gene was hybridized to metaphase chromosome spreads and then detected with fluorescein isothiocyanate-labeled streptavidin and biotinylated anti-streptavidin antibody. B, a two-color FISH was also performed using a biotin-labeled pBD-1 BAC DNA and a digoxigenin-labeled BAC DNA for SW919, which is a United States Department of Agriculture Meat Animal Research Center marker for porcine chromosome 15. After hybridization, the probes were detected with rhodamine-labeled anti-digoxigenin antibody and fluorescein isothiocyanate-labeled streptavidin/biotinylated anti-streptavidin antibody. Specific signals for pBD-1 and SW919 are in green and red, respectively. Results are representative of Ͼ100 metaphase chromosome spreads examined. plification of cDNA ends products from infected tissues with full-length pBD-1 cDNA revealed signals comparable to those of normal tissues. These data suggest that, in contrast to humans and cattle, pigs have no inducible isoforms of epithelial ␤-defensins and that the pBD-1 gene is likely to be the only epithelial ␤-defensin present.

FIG.
PBD-1 Gene Promoter Has No Inducible Activity-To directly test if pBD-1 gene expression could be modulated by proinflammatory agents, transient transfections were performed using a 1466-bp pBD-1 promoter-luciferase construct. This promoter contained a 1384-bp sequence upstream of the pBD-1 gene transcription initiation site and the 82-bp complete exon 1 sequence (Fig. 3B), where several consensus NF-B-and NF-IL-6-binding sites were generally conserved in other inducible epithelial ␤-defensins (Fig. 3C). Compared with the negative control, transfection of murine fibroblast NIH/3T3 cells with the pNF-B-Luc vector resulted in 5-8-fold increases in luciferase activity when cells were stimulated with LPS and/or IL-1␤ for 6 h (Fig. 8A). However, stimulation of the cells with up to 10 g/ml LPS, 10 ng/ml IL-1␤, or a combination of these two agents for 6 h failed to further enhance the expression of the reporter gene under the control of the 1466-bp pBD-1 promoter (Fig. 8B), suggesting that the pBD-1 promoter has little or no inducible activity in response to proinflammatory signals. This finding was also consistent with our in vivo and in vitro challenge data (Figs. 6 and 7). DISCUSSION Skin and epithelial tissues are exposed continuously to a variety of microorganisms and thus provide the first line of defense between the host and the environment (53,54). Disruption of this barrier commonly leads to microbial invasion and subsequent inflammation and infection. Indeed, skin and mucosa not only serve as physical barriers, but also produce a vast array of endogenous peptide antibiotics, including two major types, namely cathelicidins (55,56) and ␤-defensins (2,3,11).
Human LL-37/CAP-18 is the only cathelicidin that has been found to be produced by skin keratinocytes and lung epithelia in response to injury or infection (23,57). In contrast, ␤-defensins are the best characterized epithelium-derived peptide antibiotics and many have been implicated in mucosal immunity (11). To date, most ␤-defensins in humans, mice, cattle, sheep, and pigs have been shown to be expressed by either skin keratinocytes or epithelial cells throughout the respiratory, digestive, and urogenital tracts.
These epithelial ␤-defensins have been classified into two groups based on the differences in homology of amino acid sequences, gene structure, and the sites of expression (11,26). As shown with epithelial ␤-defensins of humans, mice, and cattle, TAP-like genes are up-regulated in response to infectious or inflammatory agents (12,(35)(36)(37)(38)(39)(40)(41), whereas human BD-1-like genes have a constitutive expression pattern regardless of inflammation or infection (17,28,31,33). Based on its amino acid sequence, pBD-1 could be placed into the group of TAP-like genes with an expected inducible expression pattern. However, several independent lines of evidence indicated that the pBD-1 gene was not up-regulated in response to inflammatory mediators or during infectious states. This finding may reflect in- Epithelial tissues from S. typhimurium-challenged jejunum (A) and A. pleuropneumoniae-challenged trachea (B) were scraped and used for RNA isolation from either control (CON) or infected pigs (see "Experimental Procedures" for challenge protocols). One g of total RNA was used in reverse transcriptase-PCR, and the subsequent PCR product (10 l) was subjected to Southern analysis. Twenty-six cycles were used in PCR for pBD-1 and 16 cycles for ␤-actin. Results are representative of two replicates of three to four pigs/time point .   FIG. 8. The pBD-1 gene promoter has no inducible activity in response to LPS and IL-1␤. The 1466-bp pBD-1 promoter, which contained a 1384-bp sequence upstream of the pBD-1 transcription initiation site and the 82-bp complete exon 1 sequence, was ligated into a luciferase reporter vector (pGL3-Basic). NIH/3T3 cells were transfected with a 0.5 g/well concentration of the pNF-B-Luc vector (A) or with a 1 g/well concentration of plasmid DNA containing the pBD-1 promoter (B) in 200 l of serum-free medium in triplicate. Serum-free medium (1 ml) was added to each well following 1 h of incubation. All cells were cultured for 24 h after transfection, followed by stimulation with 1 or 10 g/ml LPS, 10 ng/ml IL-1␤, or a combination of 10 g/ml LPS and 10 ng/ml IL-1␤ for another 6 h. Luciferase activity was measured in cellular extracts and then normalized against the protein concentration of each cellular extract. Values are means Ϯ S.D. of three wells. Similar results were obtained from three independent experiments. terspecies variation, which is observed commonly in gene families of many peptide antibiotics. The pBD-1 gene appears to be the only example, examined to date, that is contrary to the current classification in that it exhibits a constitutive expression pattern, although resembling TAP-like genes. Thus, classification and generalization of epithelial ␤-defensins must be made with caution, and more precise grouping awaits the identification and characterization of more family members.
In our studies, stimulation of porcine tongue epithelial cells with LPS, TNF-␣, and IL-1␤ did not enhance the expression of the pBD-1 mRNA. These agents are all potent inflammatory mediators and are encountered commonly by epithelial cells of the oral or airway mucosa. Following experimental infections with either S. typhimurium or A. pleuropneumoniae, pigs failed to increase pBD-1 gene expression to counteract the bacterial invasions into the intestinal or tracheal mucosa. These findings are in agreement with our previous data showing no increase in pBD-1 protein after porcine tongue was subjected to localized microinjury (42). Scanning the promoter region of the pBD-1 gene revealed a lack of consensus binding sites for NF-B and NF-IL-6, which are major transcription factors involved in the acute-phase response and inflammation (58 -61) and also have been demonstrated to be rather well conserved in other inducible epithelial ␤-defensins. Consistent with this observation, direct transfection of a 1.4-kb pBD-1 gene promoter showed no difference in the reporter gene activity in response to challenge with up to 10 g/ml LPS.
Our physical localization of the pBD-1 gene to S. scrofa chromosome 15q14-q15.1 provides support for previous ZOO-FISH data indicating that a narrow region of H. sapiens chromosome 8 corresponds to S. scrofa chromosome q15 (43,44). Moreover, it adds valuable information to the comparative map by providing the first direct evidence that H. sapiens chromosome 8p23, the location of human ␣and ␤-defensin genes (12)(13)(14)(15), is homologous to this region of S. scrofa chromosome 15. Thus, it provides an entry point for testing whether other loci on H. sapiens chromosome 8p23 may also map to this region of S. scrofa chromosome 15.
In summary, we have cloned and characterized an additional gene for epithelial ␤-defensins, namely the pBD-1 gene, and have physically mapped it to S. scrofa chromosome 15q14-q15.1. Although it resembles many inducible epithelial ␤-defensins in genomic structure, amino acid sequence, and sites of expression, the pBD-1 gene is expressed constitutively by epithelia of porcine tongue, intestine, and trachea during inflammation and infection. This pattern of expression and disposition of the pBD-1 peptide in airway and oral mucosa suggests that it may play a role in maintaining the steady state of natural microflora on mucosal surfaces.