Introduction

The bovine genome consists of an expanded repertoire of antimicrobial peptides in which β-defensins are the major family (Cormican et al. 2008). β-defensins are a group of multifunctional antimicrobial peptides active against a broad spectrum of microbes including bacteria, viruses, and fungi (Ganz 2003; Kaiser and Diamond 2000). Apart from their antimicrobial activities, they exert chemotactic activity to T lymphocytes, immature dendritic cells (Yang et al. 1999), and monocytes (Garcia et al. 2001). Thus, β-defensins comprise a critical defensive barrier in various mucosa with regard to combating microbial invasion and to link innate and adaptive immune responses (Ganz 2003; Kaiser and Diamond 2000; Luenser and Ludwig 2005). Tracheal antimicrobial peptide (TAP), a peptide with 38 amino acids, was the first mammalian β-defensin identified (Diamond et al. 1991). The peptide was isolated from bovine tracheal mucosa, and induced expression of TAP by LPS was attributed to an NF-κB binding site in the 5′ promoter region of the gene (Diamond et al. 1993, 2000). Lingual antimicrobial peptide (LAP), a peptide with 42 amino acids, was isolated from bovine tongue mucosa, and LAP mRNA was shown to be expressed in tongue epithelium at sites of infection or injury (Schonwetter et al. 1995). Both groups of investigators concluded that the two β-defensins, LAP and TAP are mostly expressed in epithelial cells in response to inflammation or infection (Diamond et al. 1993, 2000; Schonwetter et al. 1995; Stolzenberg et al. 1997). LAP and TAP are the most similar to each other in structure and the closest in evolutionary relationship among the β-defensins that have been identified to date (Luenser and Ludwig 2005). Studies to date have comprehensively addressed the structure, function, and evolution of these β-defensins, as well as the characteristics of their expression. However, the mechanisms underlying the regulation of their expression have not been extensively studied. Although a recent study has shown that the p65 subunit of NF-kB is stimulatory for LAP expression, whereas C/EBPβ is repressive, and mutations of the binding sites for these transcription factors abrogated LAP expression in mammary epithelial cells (Liu et al. 2011). Further understanding of the regulatory mechanisms responsible for stimulation of the expression of the epithelial defensins may also have therapeutic applications in enhancing mucosal immunity.

Infection of the mammary gland (mastitis) is one of the most common and costly diseases affecting dairy cattle worldwide (Bradley 2002). Most of the causal bacterial and fungal pathogens, such as Staphylococcus aureus, Escherichia coli, Pseudomonas aeruginosa, Klebsiella pneumoniae, and Candida ssp. are targeted by β-defensins (Kaiser and Diamond 2000; Roosen et al. 2004). Moreover, several investigations reported that eight β-defensins, including LAP and TAP, were differentially expressed in epithelial tissues of udder and increased in response to mastitis (Roosen et al. 2004; Schonwetter et al. 1995; Swanson et al. 2004). More recent analyses have reported the secretion of LAP into milk (Isobe et al. 2009). While these results indicated that LAP, TAP, and other β-defensins play a role in the innate immune response to mastitis, they also suggested that bovine mammary epithelial cells could provide a controllable cellular environment for investigation of expression and regulation of LAP and TAP genes.

Our study therefore used BME-UV, an immortal cell line derived from bovine mammary epithelial cells (Zavizion et al. 1996a, b), to investigate the transcriptional regulation of TAP and LAP genes. Three issues were addressed: (1) identifying the basal promoter for transcription expression of LAP and TAP, (2) determining the crucial DNA elements and their interaction with transcription factors which are involved in regulation of gene expression, (3) analyzing the differential expression of LAP and TAP and the effects of phorbol 12-myristate 13-acetate (PMA) and lipopolysaccharide (LPS) on β-defensin gene expression.

Materials and methods

Cell culture

Bovine mammary epithelium—University of Vermont (BME-UV; originally described by Zavizion et al. (Zavizion et al. 1996a, b) cells were a gift from Dr. Bruce Schultz at the College of Veterinary Medicine, Kansas State University, USA. Cultures were grown in typical bovine medium (pH 7.2) containing 21.25% Dulbecco’s modified Eagle’s medium, 21.25% Ham F-12 nutrient mixture (Sigma-Aldrich, MO), 25.5% RPMI 1640, 17% NCTC 135 (Gibco, NY), 10% fetal bovine serum (Sigma-Aldrich, MO), 3% newborn calf serum, 2% iron-supplemented newborn calf serum (Hyclone, UT), 1% penicillin/streptomycin (Gibco, NY), 1% insulin-transferring selenite, 1 mg/ml lactalbumin hydrosylate, 10 mg/ml l-ascorbic acid and 3 mM lactose (Sigma-Aldrich, MO). Cells were passaged with trypsin/EDTA (Sigma-Aldrich, MO) and seeded on 6- and 96-well tissue culture plates or 25 cm2 cell culture flasks at a density of 105 cells/ml. Cells were maintained at 37°C in a 95% air–5% CO2-humidified atmosphere. Experiments were conducted with cells from the first ten passages after receipt. For treatments, LPS (Promega, WI) and PMA (Promega, WI) were added to cell cultures at a final concentration of 100 ng/ml for 24 h before harvesting.

Construction of luciferase fusion plasmids

The gene sequences of TAP and LAP were extracted from two contigs with GenBank™ accession numbers of NW_944605 and NW_938412, respectively. The two sequences were aligned and analyzed for transcription start site (TSS), using MATINSPECTOR (Cartharius et al. 2005) as shown in Fig. 1. The transcription start sites of the two gene sequences were annotated at the 6- and 13-nt upstream of the translation start codons (TSC), respectively. For convenience of comparison, the 6-nt upstream of their TSC were arbitrarily numbered as +1 in both sequences.

Fig. 1
figure 1

Sequence alignment of TAP (−668 to +154) and LAP (−629 to +154) promoter regions. The bases shared by the two sequences are indicated with asterisks. Putative transcription start sites (TSS) are marked with a +1 sign and gray background. The translation start codes (TSC) are shown with bold-type letters. The putative minimal promoter region of 61 base pairs is underlined. The Oct-1 core sequence is indicated with gray background, and the segment containing predicted NF-κB and C/EBP3 binding site in LAP promoter is framed with dashed line in comparing to the corresponding TAP sequence. PCR primers are marked with bold italic-type letters

Full size image

The predicted promoter regions of TAP and LAP genes are very similar (Fig. 1). The primers used to amplify promoter sequence are listed in Table 1. The TLapLF primer could anneal to both TAP and LAP gene sequences at position −1531 and −629, respectively. Thus, two amplicons with sizes of 1,699 bp (1,531 bp upstream of TSS +154 bp downstream of TSS +14 bp linkers) and 797 bp (629 bp upstream of TSS +154 bp downstream of TSS +14 bp linkers) were obtained using PCR and bovine genomic DNA as templates (Novagen, CA) with primers of TLapLF and TLapR. To create different lengths of DNA fragments upstream of the 5′ end of the LAP and TAP genes, the longer TAP amplicon was used as a PCR template to obtain two TAP fragments with sizes of 836 and 515 bp (using sense primers TapLF and TLapMF and the common antisense primer (LapR)). The LAP amplicon of 797 bp was used to produce a shorter LAP fragment (509 bp) using TLapF and TLapR. PCR was performed using a multiple PCR kit (Qiagen, CA).

Table 1 Oligonucleotides used in the study
Full size table

PCR products were purified with a QIAquick gel extraction kit (Qiagen, CA) and cloned into pGEM®-T plasmids (Promega, WI). The confirmed DNA fragments were digested from T-vector clones with restriction enzymes and transferred into the reporter plasmids of pGL3-basic and pGL3-enhancer (Promega, WI) using NheI and BglII cloning sites. Both vectors have the firefly luciferase reporter gene and the pGL3-enhancer contains a SV40 enhancer. The recombinants were transformed into E. coli (JM109, Promega, WI). Plasmids were extracted from the bacterial cultures derived from identified single colonies screened with the primer pair of RVprimer3 (Promega, WI) and TLapR (see Table 1), and eluted by QIAprep® Spin Miniprep kit (Qiagen, CA). The inserts were re-confirmed by sequencing with RVprimer3/TLapR primers on an ABI 3700 DNA Analyzer at the Kansas State University Sequencing and Genotyping Facility (Manhattan, KS). The structures of recombinant plasmids were nominated based on the location of two ends of an insert relative to the TSS. pGL3e(-668 + 154T), for example, stands for a recombinant pGL3-enhancer having an insert of 822 bp which ends at −668 bp and +154 bp. (Abbreviations: e in middle of the nominee stands for pGL3-enhancer, otherwise b stands for pGL3-basic. T at end of the nominee indicates TAP; L indicates LAP.

A series of 5′-flanking deletion mutations of pGL3e (−347 + 154-T) and pGL3e (−341 + 154-L) were created by the Erase-a-Base system (Promega, WI). In brief, the two recombinants were digested with KpnI (NEB, MA) to generate exonuclease III resistant ends to prevent digestion of vector sequences, and cut with NheI to provide a 5′-overhang end which can be digested by exonuclease III for different periods of time. After digestion, the ends were filled with the Klenow fragment of DNA polymerase in a 2 mM mixture of all four deoxyribonucleoside triphosphates. These plasmid fragments were re-ligated and transformed into E. coli (JM109, Promega, WI). Locations of the 5′ ends of the insert sites were determined by sequencing and the produced deletion mutations were annotated as described above.

Transient transfection and functional promoter analysis

BME-UV cells were seeded in 96-well tissue culture plates at 24 h prior to transfection and allowed to reach 70–80% confluence. Cell count, viability and proliferation were analyzed using an MTT (3-[4,5-dimethylthiazol-2yl]-2,5-diphenyl-tetrazolium bromide) assay, following manufacturer's instruction (R&D Systems, Cat# 4890-025-K, Minneapolis, MN, USA). For cell transfection, 0.6 μg plasmid DNA was mixed with 1.8 μl transfection reagent (FUGENE6, Roche, Switzerland) in 50 μl of serum-free medium. The mixture was incubated at room temperature for 20 min, and 10 μl was added to each well with 100 μl fresh medium. Plasmids of pGL3-control (Promega, WI) were transfected as a positive control, and empty vectors of pGL3-basic and pGL3-enhancer were used as a negative control. Three replicate wells were used for each transfection sample. A parallel transfection with green fluorescent protein-expressed plasmid, pEGFP-C3 (Clonetech, CA), was conducted for measuring transfection efficiency, which was estimated at 50–60% in above standardized procedure. After incubation for 36 h, the cells were lysed and prepared by Bright-Glo™ Luciferase Assay System (Promega, WI) according to the manufacturer’s instructions. Luciferase activities were measured with a luminometer, Fluoroskan Ascent FL (Thermo Labsystems, MA). Relative luciferase activity was calculated as a ratio of sample chemiluminescent intensity over the background control, in which no plasmid DNA was added to cell culture.

Quantitative RT-PCR

BME-UV cells were seeded into 6-well culture plates and allowed to reach 80% confluence. Twenty-four hours before collection for extracting RNA, cells were treated by addition of LPS (100 ng/ml), PMA (100 ng/ml) or endotoxin-free water as negative controls. Total RNA was extracted from 2 × 106 cells of each treatment using Tri reagent (Sigma-Aldrich, MO). RNA samples were treated with RQ1 RNase-free DNase I (Promega, WI) to remove possible genomic DNA contamination. For the RT-PCR procedure, subsaturation levels of cDNA templates needed to produce a dose-dependent amount of PCR product were defined in initial experiments by testing a range of template concentrations. The RNA samples in identical amounts (20 ng) were used in a 25-μl RT-PCR reaction mixture with a 0.1-μM concentration of each sense and antisense primer (Table 1). RT-PCR was performed using a QIAGEN Onestep RT-PCR kit (Qiagen, CA). After amplification, 10 μl of each reaction mixture was analyzed by 2% agarose gel electrophoresis and visualized with ethidium bromide. The optical density of each band was determined by an AlphaEase® FC software (Version 6.0.0, Alpha Innotech, CA). The relative mRNA expression levels were compared with the level of the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

Electrophoretic mobility shift assays

BME-UV cells were cultured in 25 cm2 flask with/without treatment of LPS and PMA. BME-UV cells (~1 × 107) from each treatment were harvested after 24 h. Their nuclear proteins were extracted using NE-PER® Nuclear and Cytoplasmic Extraction Reagents (Pierce, IL). Concentrations of total protein of nuclear extract were estimated using the BCA protein assay kit (Pierce, IL). Synthetic complementary oligonucleotides (IDT, IA; see Table 1) were 3′-biotinylated using the biotin 3′-end DNA labeling kit (Pierce, IL) according to the manufacturer's instructions. The labeled complementary olignucleotides were annealed (denature at 90°C for 1 min, then slowly cool to 55°C and incubated at the temperature for 30 min) before use.

Binding reactions were carried out at room temperature in the presence of 50 ng/μl poly(dI-dC), 0.05% Nonidet P-40, 5 mM MgCl2, 10 mM EDTA, and 2.5% glycerol in 1× binding buffer (LightShift™ chemiluminescent EMSA kit, Pierce, IL) using 20 fmol of biotin-end-labeled target DNA and 4 μg of nuclear extract protein. Unlabeled target DNA (4 pmol) or 2 μl of anti-Oct-1 (2,000 μg/ml, Santa Cruz Biotechnology, CA) or corresponding IgG (400 μg/ml, Santa Cruz Biotechnology, CA) was added to the binding mix at 10 min before adding labeled oligonucleotides. After incubation for 30 min, the reaction mixes were loaded onto native 5% precast polyacrylamide gels (Bio-Rad, CA) pre-electrophoresed for 45 min in 0.5× Tris borate/EDTA. Electrophoresis was then conducted at 100 V for 50 min. The gel was blotted onto a positively charged nylon membrane (Pierce, IL) in 0.5× Tris borate/EDTA at 100 V for 30 min. Transferred DNAs were cross-linked to the membrane at 120 mJ/cm2 and detected using horseradish peroxidase-conjugated streptavidin (LightShift™ chemiluminescent EMSA kit, Pierce, IL) according to the manufacturer's instructions. A Kodak Image Station (4000R, Kodak, NY) was used to capture chemiluminescent images.

RNA interference

Oct-1 siRNA(h) (sc-36119), transfection reagent (sc-29527) and transfection medium (sc-36868) were purchased from Santa Cruz Biotechnology, Inc (CA). AllStars negative Control siRNA (scrambled siRNA) with 3′-end labeled with ALexa Fluor 488 was purchased from QIAGEN (CA) for evaluating transfection efficiency and using as a negative control. BME-UV cells were seeded into 6-well culture plates. After 24-h incubation, the cells (40–50% confluence) were washed with PBS and transfected with 50 nM Oct-1 siRNA or scrambled siRNA following the manufacturer’s instructions. The cells were incubated in the transfection medium for 24 h and then replenished with fresh growth medium. Transfection efficiency was estimated at 50–70% using Alexa Fluor 488 labeled siRNA. The procedures of cell treatments, protein extraction and EMSA were conducted as above.

Statistical analysis

The SAS Program PROC GLM (Version 8.2, SAS Institute, NC) were used to conduct ANOVA analyses and compare values of the experimental data. Results are expressed as mean±SE and P values of <0.05 were considered significant.

Results

Expression of LAP and TAP in BME-UV cells

To examine expression and regulation of LAP and TAP mRNA in the bovine mammary epithelial cell line, BME-UV, RT-PCR was used in cells stimulated with/without LPS or PMA. As shown in Fig. 2, LAP and TAP mRNA in BME-UV cells cultured in growth medium (lane 3) were at levels of 50% and 40% of GAPDH mRNA. Treatment with LPS (lane 1) brought the mRNA levels of LAP and TAP to 80% and 60% of GAPDH mRNA, respectively. PMA treatment (lane 2) brought mRNA levels to 85% and 65% of GAPDH expression. GAPDH mRNA was maintained at a constant level regardless of treatments. This experiment revealed that in BME-UV cells, LAP and TAP are constitutively expressed under normal growth conditions and the expression is increased in response to both LPS and PMA.

Fig. 2
figure 2

The expression of LAP and TAP in a bovine mammary epithelial cell. RNA was extracted from mock-treated cells (vehicle), and cells treated with lipopolysaccharide (LPS) and phorbol 12-myristate 13-acetate (PMA) at 100 ng/ml for 24 h. Semiquantative RT-PCR was performed to determine mRNA abundance using an agarose gel electrophoresis (upper) and the optical density of LAP and TAP bands was quantified by normalization to a housekeeping gene of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (bottom). Results are means±SE from five independent experiments. Values with different superscript letters are significantly different at P < 0.05

Full size image

Basal promoter for LAP and TAP genes

In order to identify the components comprising the basal promoter for both LAP and TAP genes, fusion plasmids of pGL3-basic and pGL3-enhancer with different 5′-flanking length PCR fragments upstream of the two genes were compared. All fusion reporter constructs of pGL3-basic plasmids showed no detectable promoter activity while all fusion reporter constructs of pGL3-enhancer (pGL3e) plasmids showed significant promoter activity (Fig. 3), indicating the requirement for the SV40 enhancer in activating expression. Secondly, pGL3e (−668 + 154-T) showed the highest activity (>11-fold higher luciferase activity over control), but was not significantly different from pGL3e (−1531 + 154-T). The region spanning −1531 to −668 was thus reasoned to play no critical role in the enhanced expression of TAP detected (Fig. 3). Thirdly, fusion plasmid pGL3e (−629 + 154-L), in which the 5′ end of its insert corresponded to the 5′ end of the insert in pGL3e (−668 + 154-T) (Fig. 1), showed an 8-fold increase in activity of the promoter, although significantly lower than pGL3e (−668 + 154-T).

Fig. 3
figure 3

Functional analysis of TAP and LAP promoters which were fused with the pGL3-basic or pGL3-enhancer reporter plasmids. On the left, the numbers related to the reporter constructs indicate the positions of the 5′ end of the TAP and LAP (in parentheses) promoter regions referring to the putative transcription start sites (TSS), which are set as position +1. Reporter constructs were transfected into the BME-UV cells and relative luciferase activity was measured and calculated as a ratio of luciferase activity over the background control, in which no plasmid DNA was added. Results are means±SE from three independent experiments. Values with different superscript letters are significantly different at P < 0.05

Full size image

The pGL3e (−341 + 154-L) construct, which was not statistically different from pGL3e (−347 + 154-T), showed a 4-fold increase in the activity of the pGL3-enhancer. Both of these shorter constructs (pGL3e (−341 + 154-L) and pGL3e (−347 + 154-T)) had significantly lower promoter activity than pGL3e (−629 + 154-L) and pGL3e (−668 + 154-T). The reduction in promoter activity by over 50% indicates that the region spanning −668 to −347 in TAP and −629 to −341 in LAP contain one or more enhancer elements. We refer to this region as enhancer region 1 (ER1). A second enhancer region is evident between −347 to +1 for TAP and −341 to +1 for LAP, as deletion of these regions further significantly diminishes promoter activity (Fig. 3). This region is denoted as enhancer region 2 (ER2).

Focus was then placed on the refinement of the proximal promoter elements detected in ER2 response to cell stimulation. The activities of a series of 5′ deletion mutations of pGL3e (−347 + 154-T) and pGL3e (−341 + 154-L) were tested together with treatments of cells with/without LPS or PMA. Figure 4a, b showed that a minimal proximal promoter, which promoted luciferase reporter activity for approximately 4-fold in pGL3-e system, was between −238 and −172 bp in TAP promoter and between −227 and −137 bp in LAP promoter. Deletion of this region completely abrogated the significant increase in promoter activity under the control or stimulated conditions with PMA and LPS (Fig. 4a, b). Notably, although both PMA and LPS significantly enhanced activity of the two minimal promoters, the response of the TAP promoter was more pronounced (Fig. 4a). The two larger upstream plasmids (−347 and −266, TAP and −341 and −251, LAP) had the same promoter activity as pGL3e (−238 + 154-T) and pGL3e (−227 + 154-L), respectively.

Fig. 4
figure 4

Functional analysis of TAP (a) and LAP (b) minimal promoters and their responses to the stimulation of LPS or PMA. For each construct, the region relative to TAP or LAP promoter is indicated by the nt position referring to the +1 TSS. On the right, relative luciferase activity was measured and calculated as a ratio of luciferase activity over the background control, in which no plasmid DNA was added. Results are means±SE from three independent experiments. Values with different superscript letters are significantly different at P < 0.05

Full size image

Protein-DNA interaction on the minimal promoter

Sequence aligment (Fig. 1) revealed a fragment of 61 bp, spanning −234 to −174 in TAP and −228 to −168 in LAP that potentially interact with cis-element(s). These 61-bp regions were used in subsequent experiments to identify core transcription factor(s). Despite there are only three nucleotide differences between the LAP and TAP genes in this region (Fig. 1), 25 over 21 transcription factors are predicted to bind within this 61-bp region of TAP and LAP genes, respectively. Binding sites for the nuclear factor kappa B/c-rel and Oct-1 transcription factors are predicted to bind within this promoter region of both genes. However, an additional NF-κB transcription factor binding site is predicted within this region of the TAP gene but not the LAP gene; whereas a CAAT/enhancer-binding protein (C/EBP) transcription factor is predicted to bind within this region of the LAP gene but not the TAP gene. Therefore, the difference conferred by these nucleotide differences between TAP and LAP could have functional significance in the regulation of promoter activity.

EMSA experiments were conducted to identify key element(s) within the minimal promoter region responsible for constitutive gene activation. The 61 bp oligo (oligo T in Table 1) and three shorter oligos (A, B, and C in Table 1) homologous to parts of the 61-bp regions were used to probe nuclear extracts from BME-UV cells grown under normal growth conditions. Figure 5a shows that only the band at the top of lane 7 represented a specific protein-DNA complex although several other nonspecific bands appeared below it. The same specific complex was captured exclusively by oligo C (lane 5, Fig. 5a). Two more oligos (C1 and C2, Table 1), covering oligo C and partially overlapping each other, were then tested for their reaction with the same nuclear extract. Only oligo C1 (lane 4, Fig. 5b) yielded one band matching the specific band detected by oligo C. The band represented a complex produced by the transcription factor Oct-x family and its binding motif in oligo C1 as predicted by internet-based program TFSEARCH (http://www.cbrc.jp/research/db/TFSEARCH.html) and MATINSPECTOR (Cartharius et al. 2005). Transcription factor Oct-1 was shown to be bound to the oligo by comparing the band location with that of oligo O (lane 6, Fig. 5b) which contained a canonical Oct-1 binding motif ATTTGCATA, as well as by comparing the intensity of the two bands. The EMSA band caught by oligo C1 which contains ATTAGCATA, a variant of Oct-1 binding motif, was weaker than that caught by oligo O. Further evidence was provided by competitive experiments (Fig. 6a) in which both unlabelled DNA competitor oligo O (lane 2) and protein competitor Oct-1-specific antibody (lane 6) abolished or weakened significantly the protein-DNA reaction.

Fig. 5
figure 5

Identification of DNA–protein interaction elements by electrophoretic mobility shift assay (EMSA). Biotin-labeled double-stranded oligonucleotides covering the proximal promoter region of TAP (−234 nt/−174 nt) and an oligonucleotide containing the canonical Oct-1 binding motif (oligo O) were incubated with 5 μg of BME-UV cell nuclear extract in the presence and absence of 200-fold molar excess of unlabeled specific oligonucleotides. The Oct-1 binding motif in Oligo T and O are framed with regular line; and in the Oligo T the element containing mutated NF-κB and C/EBP3 binding sites in comparison to LAP promoter is framed with a dashed line (the mutated nt shown in italic). The image presented here is representative of one of three independent experiments that yielded similar results

Full size image
Fig. 6
figure 6

a Verification of transcription factor Oct-1 binding to oligo C1, which contains a predicted variant of Oct-1 binding motif. EMSA was conducted with a cold competitor oligo O containing the canonical Oct-1 binding motif and Oct-1 antibody. The EMSA was conducted by incubating 5 μg BME-UV cell nuclear extract with unlabeled probes, IgG, and Oct-1 antibody, respectively as shown at top for 30 min. Biotin-labeled double-stranded oligo C1 was then added to each mix. b The stimulatory effects of LPS and PMA on the formation of Oct-1-DNA complex. Nuclear extracts were prepared from BME-UV cells pre-treated with PMA, LPS, and vehicle (control). 5 μg nuclear extract was incubated with unlabeled probe shown on the top for 10 min and then biotin-labeled double-stranded oligo C1 was added to each mix. The mixes were incubated for another 30 min. The image presented here is representative of one of three independent experiments that showed similar results

Full size image

Role of Oct-1 in regulation of induced LAP and TAP expression

The experiments of promoter serial deletion (Fig. 4) suggested that the minimal promoter regions of both TAP and LAP genes may respond to LPS and PMA stimulations. To identify the responsible element(s) in the region, nuclear extracts from BME-UV cells treated with LPS and PMA were used for EMSA using the oligos A, B, C, C1, and C2. In response to PMA, oligos C and C1 yielded a band at the location of the Oct-1 DNA–protein complex as in the case of no treatment. However, PMA induced a stronger band of the Oct-1 DNA–protein complex while LPS did not significantly increase the intensity of the band compared with the untreated control (Fig. 6b). It was noted that oligo B, which contains a core NF-κB binding motif as predicted, did not produce any detected band to response to LPS (data not shown).

The effect of PMA and LPS on Oct-1 DNA–protein was further examined by EMSA together with RNA interference. After Oct-1 gene expression was suppressed by siRNA treatment, BME-UV cells were treated with PMA and LPS. Figure 7 shows the results of EMSA of oligo C1 with nuclear extract from the cells with the different treatments. siRNA treatment weakened the band of the Oct-1 protein-DNA complex (lane 2). While PMA was able to recover intensity of the band (lane 6), LPS failed to restore the band intensity (lane 5). Because labeled oligo C1 was overloaded in the each binding reaction, as shown by the dark thick band at the bottom of each lane in Fig. 7, the change of intensity of band of protein-DNA complex should reflect the change in amount of Oct-1 protein in nuclear extract of cells treated. Thus, this result suggested that PMA was able to stimulate an increase of Oct-1 protein amount, while LPS was not.

Fig. 7
figure 7

PMA, not LPS, reversed Oct-1 siRNA silencing effects on the formation of Oct-1-DNA complex. Nuclear extracts were prepared from BME-UV cells which were pre-treated with Oct-1 siRNA, scrambled siRNA, and vehicle solution for 24 h, and then treated with LPS and PMA and vehicle solution for another 24 h; 5 μg nuclear extract was incubated with binding mix solution for 10 min and then Biotin-labeled double-stranded oligo C1 was added to each mix. The mixes were incubated for another 30 min. The image presented here is representative of one of three independent experiments that showed similar results

Full size image

Discussion

Antimicrobial peptides are key innate immune effectors having direct microbiocidal activity and immunomodulatory functions (Durr and Peschel 2002; Ganz 2003; Yang et al. 1999). Of bovine defensins, LAP and TAP gene expression was previously reported to be constitutive and inducible in all tested bovine epithelial tissues ex vivo including uterine (Chapwanya et al. 2009; Davies et al. 2008) and mammary gland tissue (Goldammer et al. 2004; Roosen et al. 2004; Schonwetter et al. 1995; Swanson et al. 2004). However, limited information is available about their promoter activity and the regulation of their expression. During our manuscript preparation, Liu et al. (2011) reported the identification of a NF-κB and C/EBP interacting cis-element in the LAP promoter region, which was responsive to the stimulation of heat-killed E. coli. Here, we show that both LAP and TAP promoters contain an Oct-1 interacting element, which primarily responds to PMA not LPS stimulation.

In this study, BME-UV, an established bovine mammary cell line was found to constitutive express LAP and TAP under normal growth conditions, but expression was increased in response to exogenous stimuli, including LPS and PMA (Fig. 2). In order to isolate and characterize the basal promoter and key elements for transcriptional expression and regulation of both genes, the DNA regions upstream of the TSS were aligned. The two sequences, particularly their regions from −558 in TAP and −527 in LAP to their TSS, were highly similar, suggesting that the two β-defensins evolutionarily share a common ancestry (Luenser and Ludwig 2005) that may account for their similar patterns of promoter activity detected; 5′ deletion analysis of promoter sequences identified two enhancer regions (ER) in each gene (TAP-ER1 −668 to 348 and ER2 −347 to +1 and LAP-ER1 −629 to −342 and ER2 −341 to +1), which significantly regulated the promoter activity shown by serial deletion analyses (Fig. 3). Sequence alignment the proximal promoter regions of TAP (−347 to +154) and LAP (−341 to +154) revealed a 61 bp-element in common for TAP and LAP, which was deemed responsible for their minimal promoter activity (Fig. 4a, b). However, as promoter activation required the presence of an SV40 enhancer in the vector, the 61-bp fragment by itself is insufficient for basal promoter activity. It seems likely that at least one enhancer, the DNA fragment of 61 bp or some element(s) in it, and the TATA box may act in concert to initiate and maintain basal transcription expression of LAP and TAP. The combination of multiple regulatory elements for basal gene expression is of biological significance in the restriction of gene expression to appropriate tissue and physiological conditions (He and Rosenfeld 1991; Levine and Tjian 2003; Struhl 1991). As multifunctional innate immune effectors, defensins are regulated through multiple pathways (Ganz 2003). Our studies clearly show that PMA, an established inflammation inducer in bovine mammary tissues (Lauzon et al. 2005), regulates TAP and LAP expression through Oct-1 element in the proximal promoter. A study using a similar LAP promoter-reporter system but without SV40 enhancer identified a NF-κB and C/EBP interacting elements (shown in Fig. 3) in LAP promoter to be responsible for LPS induction in a murine mammary epithelial cell line (Liu et al. 2011). Although the activity of the NF-κB and C/EBP interacting element in the LAP promoter has not been detected and focused in our system in the bovine mammary epithelial cells, we cannot exclude the role of the NF-κB and C/EBP interacting element in regulation of TAP and LAP expression because our cis-element elucidation is primarily based on TAP promoter (Figs. 5, 6, and 7), which has a predicted NF-κB and C/EBP interacting element (GGGGTTTTCCCCA) different from that in LAP promoter (GGGGATTTCCACA) (Fig. 1).

It is widely accepted that mRNA expression is generally initiated and regulated by the binding of transcription factor(s) to a specific DNA motif within the promoter region (Mitchell and Tjian 1989; Oliver and Mitchell 1983). Therefore, EMSA was conducted to investigate the interaction of transcription factors and DNA motifs within the minimal promoter region of 61 bp. The results from EMSA (Fig. 5), competition EMSA with cold competitor and antibody (Fig. 6a), and siRNA (Fig. 7) revealed the functional significance of the Oct-1 binding motif. Because the Oct-1-DNA complex was the only one detected in the minimal promoter region, the formation of the complex could be considered a key event for the initiating and maintaining of transcription of TAP. In addition, LAP has a core sequence of Oct-1 binding motif similar to TAP and a DNA context similar to TAP (Fig. 1), as well as the same pattern of promoter activity generated by promoter constructs using a luciferase reporter assay. We deduce that the Oct-1 binding site in the minimal promoter region of LAP should function similarly to its counterpart in TAP, but we cannot conclude for the NF-κB and C/EBP interacting element based on our and studies of others (Liu et al. 2011).

Oct-1 is a ubiquitously expressed protein that is present in a variety of tissues and cell types in mammals (Pankratova and Polanovsky 1998). Its DNA binding domain, the POU domain (a highly charged amino-acid region of sequence similarity contained within three mammalian transcription factors, Pit-1, Oct-x, and unc-86 (Sturm et al. 1988)), has been reported to function critically in regulation of cell-type-specific gene expression (Bhat et al. 1996; dela Paz et al. 2007; Wu et al. 1997). This study is the first report of the involvement of Oct-1 as a transcription factor in the control of β-defensin expression. Interestingly, the Oct-1 binding motif in the minimal promoter of LAP and TAP was not only responsible for initiating transcription but was also involved in their responsiveness to PMA stimulation, an inflammation inducer in bovine mammary tissues (Lauzon et al. 2005) (Fig. 6b). Oct-1-specific siRNA suppressed basal Oct-1 expression in BME-UV cells and led to a weaker Oct-1-DNA complex (Fig. 7), which was restored by applying PMA, which probably highly induce Oct-1 expression to overcome the siRNA suppression. A similar PMA-induced expression pattern was reported in relation to the transcription of the hNHE gene by induced ERG-1 (Malakooti et al. 2006).

Several inflammatory agents and infection factors reportedly regulate the expression of β-defensins (Diamond et al. 1993, 2000; Kaiser and Diamond 2000; Schonwetter et al. 1995). However, this is the first report of PMA, an inflammation and tumor-promoting agent (Lebkowski et al. 1987), stimulating the expression of TAP and LAP (Figs. 2 and 4) in an Oct-1-mediated manner. In contrast, LPS stimulation, which was also able to enhance activity of the minimal promoter of both LAP and TAP (Fig. 4a, b) and increased their mRNA expression, showed no detectable effect on the amount of Oct-1 complex regardless of level of the Oct-1 in nuclear extracts (Figs. 6b and 7). This result indicates that other transcription factors, such as the NF-κB and C/EBP interacting element within this region may be responsible for the LPS-induced activation of β-defensin expression (Chapwanya et al. 2009; Davies et al. 2008).

Studies on human genes have shown that the C/EBP and Oct-1 binding sites overlap, and Oct-1 binding acts as a repressor in unstimulated cells (Wu et al. 1997). Conversely, in stimulated cells, the Oct-1 molecule is displaced by the C/EBP molecule, which then forms part of a transcriptional enhancesome that dramatically elevates transcription of the gene transcript (Hoffmann et al. 2002). Of interest is that the LAP promoter is predicted to have a binding site for the C/EBP factor (Fig. 1) (Mukaida et al. 1990), which overlaps with the Oct-1 binding site but which is absent from the TAP gene due to a nucleotide variant. A relevant study which showed that reduced Oct-1 binding in tumor cell lines caused an increase in gene expression (Zhang et al. 1999), also suggested that Oct-1 functions by blocking the binding of transcription factor activators, implying that the cellular ratio of Oct-1 to transcriptional activators such as C/EBP or NF-κB is crucial (Sibbet et al. 1995). Liu et al. (2011) found that both basal and LPS-induced expression levels of LAP in primary mammary cells were abrogated when the NF-kB or C/EBPβ was mutated but they did not assess Oct-1 function. It would be of interest to investigate if displacement of the transcriptional mediator Oct-1 by either C/EBP or NF-κB is necessary for maximal expression of β-defensins in response to infection. The positive regulation of Oct-1 on LAP and TAP gene expression revealed in this study increases our understanding about this gene regulatory mechanism co-opted by several transcription factors.

In summary, this study demonstrates that LPS and PMA positively regulate both LAP and TAP expressions. Reporter assays identified positive regulatory elements within the proximal promoter region of both genes. Furthermore, using EMSA and siRNA, we have demonstrated a specific role for Oct-1 in regulating the promoter response to PMA stimulation in mammary epithelial cells. The responses detected are likely to be cell-type or stimulant (PMA or LPS) specific and, therefore, may not comprehensively reflect the response in more complex tissue types and pathogen infections in vivo. However, understanding the regulation of β-defensin expression is an important prerequisite to elucidation of their comprehensive biological functions and potential applicability as novel therapeutics to target infectious disease in cattle.