Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Mucins are major mucus glycoprotein components that play an essential role in protecting and lubricating the epithelial surface of various tissues, including the respiratory, digestive, and reproductive tracts. Mucins constitute a family of high molecular glycoproteins that are polydispersed and heavily O-glycosylated (1, 2). At least 12 human mucin (MUC) genes have been cloned and completely or partially sequenced (3, 4; review in 5). Of these, at least seven mucinsMUC1, MUC2, MUC4, MUC5AC, MUC5B, MUC7, and MUC8are expressed in the human upper and/or lower respiratory tracts at the message level (6, 7; review in 5). It is not clear whether all of these messages are being translated and whether their gene products are secreted. Three gel-forming mucins, MUC2, MUC5AC, and MUC5B gene products, are expressed by airway epithelium; only MUC5AC and MUC5B have been convincingly demonstrated to be major components of human airway secretions (8). In situ hybridization demonstrated that MUC5AC message is restricted to surface goblet cells, whereas MUC5B message is expressed mainly by submucosal-gland mucous cells (7). It has been recently shown that MUC5B products are also synthesized by goblet cells (11). The discrepancy between studies on MUC5B gene expression has not been accounted for. One possible explanation, as described in this report, is that MUC5B expression is not restricted to the submucosal gland area, and that the expression could be elevated in surface goblet cells, especially under a pathologic condition.

Reverse transcriptase polymerase chain reaction (RT-PCR) has been used to profile mucin gene expression in primary and passage-1 cultures of human airway epithelial cells (13). That study has demonstrated that at least nine MUC genes (MUC1-4, 5AC, 5B, and 6-8) are expressed in cell cultures, and only MUC4, MUC5AC, and MUC5B messages can be detected by using Northern blotting (13). Apparently, messages of other MUC genes are extremely low in cultured cells. It is not clear what controls MUC gene expression in culture. Interestingly, it was demonstrated that retinoids and the specific culture condition, such as an air-liquid interface culture with collagen gel substratum as described here, could elevate MUC gene expression. These results are apparently associated with an increase in the mucous-cell population in these cultures (13). However, the molecular mechanism underlying this elevation is unknown.

To examine the molecular mechanism of the gene expression regulation, it is necessary to elucidate the genomic structure of MUC5B gene. MUC5B and three other MUC genes, MUC6, MUC2, and MUC5AC, have been mapped to 11p15.5 on a single band of 400 kb, and their order has been determined to be tel-MUC6/MUC2/MUC5AC/ MUC5B-cen (14). Desseyn and colleagues (15) used three different genomic clones and various RT-PCR methods to complete most of the genomic organization and complementary DNA (cDNA) sequence of MUC5B gene, including its transcriptional start site. A specific feature of MUC5B gene is an unusually large central exon of 10,713 base pairs (bp), which codes for a 3,571-amino acid peptide (17). This single large exon contains the entire tandem repeat domain, and 19 subdomains have been identified. Some of them are similar, thus creating so-called super- repeats of 528 amino acid residues, which are probably the longest ever determined in mucin genes. A similar large central exon encoding the repetitive domain has also been demonstrated for MUC2 gene (18). For the upstream genomic organization of the large central exon, Desseyn and associates (15) suggested that the remaining MUC5B 5' message, 3,886 bp, was distributed into 29 exons with a genomic size of 15,143 bp. Meanwhile, Offner and coworkers (19) observed a new transcriptional start site upstream to that found by Desseyn and associates (15). When these assertions lay unresolved*, we independently isolated a human genomic clone, Cos-1, which contained sequences of both the 3' end of MUC5AC and the large central exon of MUC5B gene. Thus, this clone contains a genomic insert that flanks the region between the 5' end of MUC5B and 3' end of MUC5AC. Using this clone, we confirmed the results of Offner and coworkers (19) with additional sequence information on the 5'-flanking region of this gene. (The sequence reported in the present work appears in the Genbank/EMBL database with the accession number AF107890.) In the present study we examined whether the 5' flanking region of human MUC5B gene contains cis- elements that are responsible for its basal and cell type- specific activity, and the stimulatory activity by all-trans-retinoic acid (RA) and the specific culture conditions.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Sources of Human Airway Tissues and Cells

Human tracheobronchial and lung tissues were obtained from the University of California at Davis Medical Center or the Anatomic Gift Foundation (Laurel, MD), with consent as described elsewhere (20). The Human Subjects Review Committee of the University of California at Davis approved all tissue procurement procedures. We obtained 11 tracheobronchial and lung tissues from various donors. The patients' information is shown in Table 1. Excised tissues were transported to the lab in an ice-cold minimal essential medium (Sigma, St. Louis, MO). Portions of these tissues were directly fixed in 4% paraformaldehyde at 4°C overnight. The fixed tissues were washed twice by 50% ethanol solution for 20 min each time, followed by two washes with 70% ethanol; tissues were then stored in a 70% ethanol solution at cold temperature until processing for paraffin block and sections (5 µm). The rest of these tissues were processed for airway epithelial cell isolation and culture as described elsewhere (20). Primary tracheobronchial epithelial (TBE) cells were plated on tissue culture dishes in a serum-free Dulbecco's modified Eagle's medium/F12 (1:1) medium supplemented with insulin (5 µg/ml), transferrin (5 µg/ml), epidermal growth factor (EGF) (10 ng/ml), dexamethasone (0.1 µM), cholera toxin (20 ng/ml), bovine hypothalamus extract (15 µg/ml), and RA (30 nM). Once confluent, cells were passaged into tissue culture dishes, collagen gel substratum, Transwell chambers, and collagen gel-coated Transwell chambers (BICG), and confluent primary human TBE cultures under BICG expressed mucociliary differentiation (13, 20, 21).

Two immortalized normal human TBE cell lines, BEAS-2B subclone S and HBE1, were obtained from Drs. J. F. Lechner (Wayne State University, Detroit, MI) and J. Yankaskas (University of North Carolina, Chapel Hill, NC), respectively. These cell lines were maintained in serum-free Ham's F12 medium supplemented with six hormonal supplements, as described elsewhere (20). These supplements were: insulin (5 µg/ml), transferrin (5 µg/ml), EFG (10 ng/ml), dexamethasone (0.1 µM), cholera toxin (20 ng/ml), and bovine hypothalamus extract (15 µg/ml). To induce mucous cell differentiation in these cell lines, RA (30 nM) was added to the medium and cultures were maintained in an air- liquid interface, as described earlier for the primary cultures.

RNA Isolation and Northern Blot Hybridization

At Day 21 after plating, RNA was isolated from cultures by a single-step acid guanidinium thiocyanate phenol-chloroform extraction method (22). For Northern blot hybridization, equal amounts of total RNA (20 µg/lane) were subjected to electrophoresis on a 1.2% agarose gel in the presence of 2.2 mM formaldehyde and transblotted onto Nytran membranes. The RNA was crosslinked to membrane by a UV Stratalinker 2400 (Stratagene, La Jolla, CA). Single-stranded antisense oligonucleotide, corresponding to the tandem repeat unit of MUC5B, was end-labeled with -³²P-adenosine triphosphate by polynucleotide kinase. Membrane prehybridization and hybridization with a ³²P-labeled MUC5B-specific oligonucleotide probe were carried out as described elsewhere (23). The relative abundance of MUC5B message in Northern blot was normalized with 18S ribosomal RNA (rRNA) band, as described elsewhere (23).

In Situ Hybridization

Single-stranded antisense oligonucleotides, corresponding to the tandem repeat unit of MUC5B and MUC5AC, in accordance with previous work (7), were synthesized. These sequences were: 5'-TGT GGT CAG CTT TGT GAG GAT CCA GGT CGT CCC CGG AGT GGA GGA GGG-3' and 5'-AGG GGC AGA AGT TGT GCT CGT TGT GGG AGC AGG GGT TGT GCT GGT TGT-3', respectively. These DNAs (100 pmol each) were end labeled with the digoxigenin oligonucleotide tailing kit (Roche Molecular Biochemicals, Indianapolis, IN), according to the manufacturer's protocol. Sense oligonucleotides corresponding to these sequences were also synthesized and used as a control for the hybridization. In situ hybridization was carried out as per the manufacturer's protocol (Roche Molecular Biochemicals). Briefly, slides were digested with 10 µg/ml Proteinase K in 50 mM Tris-Cl (pH 8.0) and 50 mM ethylenediamenetetraacetic acid for 15 min at 37°C, rinsed twice in 0.2 × saline sodium citrate (SSC), and then postfixed in 4% paraformaldehyde/phosphate-buffered saline for 20 min. Slides were treated twice for 5 min each time with 0.1 M triethanolamine (pH 8.0) and blocked by 0.25% acetic anhydride in 0.1 M triethanolamine. The sections were then dehydrated through the ethanol series. For each section, 0.5 pmol digoxigenin-labeled oligonucleotide probe in 50 µl of hybridization buffer was applied. The hybridization buffer contained 2× SSC, 1× Denhardt's solution, 10% dextran sulfate, 50 mM phosphate buffer (pH 7.0), 50 mM dithiothreitol, 250 µg/ml yeast transfer RNA, 100 µg/ml poly A, and 500 µg/ml salmon-sperm DNA. The section was hybridized at 45°C overnight in a humidified chamber. After hybridization, the section was washed twice for 15 min each time at 37°C with 2× SSC, twice for 15 min each time with 1× SSC, and twice for 15 min each time with 0.25× SSC. After the wash, the slide was reacted with antidigoxigenin primary antibody conjugated with alkaline phosphatase. After several washes, the reacted probes in the slide were color-developed with the Digeoxigenin Nucleic Acid Detection kit from Roche Molecular Biochemicals.

Isolation of Genomic Clone

A DNA probe corresponding to MUC2 amino-terminal and promoter region (24) was used to screen a cosmid genomic library derived from human placenta (Clontech, Palo Alto, CA) by low stringent hybridization (42°C hybridization temperature and high salt wash, etc.). Considering the similarities in the 5' end cystine-rich domains between MUC2 and MUC5B, we expected that this approach would discover genomic clones containing the amino-terminal and promoter region of MUC5B gene. Eight out of 10⁶ cosmid clones screened were seen positive in the hybridization. Subsequently, in Southern blot, only one clone, Cos-1, was positive for the hybridization with MUC5AC cDNA probe (12) under more stringent hybridization conditions (65°C, low salt wash). Initially, DNA sequencing of this cosmid clone from both T3 and T7 ends of cloning sites revealed the presence of the 3' end MUC5AC cDNA sequence (12) and the 5' end cDNA sequence of the large central exon of MUC5B (17), respectively. Inasmuch as the four clustered mucin genes at chromosome 11p15.5 were suggested in the order of 5'-MUC6-MUC2- MUC5AC-MUC5B-3' end (14), Cos-1 clone should contain genomic DNA that spans the region between the 3' end of MUC5AC gene and the 5' end of MUC5B gene.

Restriction Mapping of Cosmid Cos-1

Cos-1 genomic DNA was prepared and digested with KpnI and EcoRI restriction enzyme. Southern blotting hybridization was carried out to determine which DNA fragments belong to MUC5AC gene (12) and MUC5B gene (17), using cDNA probes corresponding to the 3' end of MUC5AC message and the 5' end of MUC5B large central exon, respectively. DNA fragments, which were hybridized to the cDNA probe of MUC5B, were further subcloned by various restriction enzyme digestions and cloning into pGem 4Z (Promega, Madison, WI). These subclones were further mapped by restriction enzyme digestion, and were DNA sequenced.

Characterization of Genomic DNA, Sequencing, and Sequence Analysis

Human genomic DNA in Cos-1 was sequenced by using the ABI Prism Model 377 Automated DNA sequencer (Applied Biosystems, Foster City, CA). Various primers corresponding to different regions of Cos-1 cosmid clone were constructed for DNA sequencing. The sequencing data was analyzed and aligned using LaserGene (DNASTAR, Madison, WI) software. These sequences were then used to verify the restriction map and compared with that obtained from various rapid amplification of cDNA ends (RACE) cDNA clones. The comparison between the genomic and cDNA sequences provided the basis for setting up the exon/intron junction. Both the genomic sequence of the 5' end of MUC5B gene in Cos-1 clone and the cDNA sequences in various RACE DNA fragments were submitted to GenBank with the accession number AF107890.

Synthesis of Oligonucleotide Primers

Additional oligonucleotide primers used in PCR, 5'-RACE, RT-PCR, and sequencing experiments were synthesized by the Genomic Research Institute. Their positions and sequences are described in Table 2.

5'-RACE

The 5'-RACE kit (Roche Molecular Biochemicals) was used to synthesize the first-stranded cDNA from total RNA (3 µg) isolated from human tracheobronchial tissues or primary cultures of human TBE cells, which had been cultured under an air-liquid interface culture condition for more than 21 d (20). Various antisense primers corresponding to different regions of MUC5B message were used to initiate first-strand cDNA synthesis. This was followed up by a 3' tailing with oligo d(A) with terminal deoxynucleotidyl transferase, then a PCR was carried out using the nested primer of 3' end and the 5'-anchor oligo d(T) adapter (Table 2). The PCR products were subcloned into the TA vector (Invitrogen, Carlsbad, CA) for cloning and DNA sequencing.

RT-PCR Amplification

cDNA was synthesized from total RNA (3 µg) by RT with downstream MUC5B gene-specific antisense primer. The resulting single-strand cDNA was used as a template for PCR amplification by a pair of other upstream primers. PCR products were TA cloned and sequenced.

Primer Extension and a Modified 5'-RACE Method to Determine the Transcription Start Site

The experiment was carried out with 50 µg total RNA which was reverse transcribed with a radioactive primer, Pel1, based on the nucleotide (nt) sequence from +123 to +105 of MUC5B message (Table 2) and end labeled with ³²P by polynucleotide kinase. The reverse-transcribed product was analyzed on a 6% polyacrylamide gel alongside a DNA sequencing ladder and a DNA size reference marker (pBR322 DNA digested by MspI; New England Biolabs, Inc., Beverly, MA).

Because the human MUC5B message is quite large in size (15), it is not easy to maintain its integrity during the RNA isolation. As a result, the primer extension signal is weak. To overcome this problem, a modified 5'-RACE method was developed to help determine the transcription start site. An antisense primer, spanning from +250 to +230, was used to generate first-strand cDNA from 3 µg total RNA by an RT. Instead of 3' tailing with only oligo d(A), the first-strand cDNA was also anchored with oligo d(T) by terminal deoxynucleotidyl transferase. After tailing, a PCR amplification was carried out with two 5'-anchor adapters, oligo d(A) and oligo d(T) (Table 2), and the nest antisense primers upstream to +250/+230 (e.g., +234/+214 and +162/ +142; see Table 2). The resulting PCR DNA fragments were cloned into TA clone. Several clones were isolated and their sequences were determined. Because there should be only one common DNA sequence adjacent to oligo d(T) and oligo d(A) adapters, this DNA sequence should be identical to that of the 5'-end message upstream to the +250/+230 primer. A major advantage of this approach is the use of PCR, which allows the amplification of the 5' end of low-abundance message for further characterization.

Construction of Chimeric Promoter Reporter and Transient Transfection

DNA fragments corresponding to different 5'-flanking regions and the region flanking between exons 1 and 2 of the human MUC5B gene were PCR-amplified using appropriate primer pairs, as listed in Table 2. For the preparation of DNA fragments from +7 to 4,169 (MUC5B-b1) and from +7 to 1,098 (MUC5B-b2), their amplification utilized PU1/PL1 and PU2/PL1 primer pairs with NheI/BamHI cloning sites, respectively. For the DNA fragments of MUC5B-il (from 13 to +2,738), the primer pair was PiU1 and PiL1 with NheI/BglII cloning sites. These PCR DNA products were digested with appropriate restriction enzymes and then cloned into a promoterless pGL-3 basic vector (Promega). The chimeric construct clones were verified by DNA sequencing. For transient transfection studies, chimeric DNAs were purified with the Qiagen columns. Fugene 6 (Roche Molecular Biochemicals) was used in the transfection study, according to the manufacturer's instructions. The quantity of 0.5 µg DNA per 35-mm dish with 60 to 80% confluence was used for each transfection. p-Simian virus (pSV)--galactosidase (-gal) reporter gene construct (0.5 µg/transfection) was included in each transfection for the normalization of the transfection efficiency among different culture dishes. After 48 to 72 h of transfection, cultured cells were harvested for reporter gene assays. For studying effects of culture conditions on the promoter-reporter gene activity, primary human TBE cultures grown in a 60-mm dish were transfected with 1 µg of MUC5B promoter-luciferase construct DNA and 0.5 µg pSV--gal reporter DNA. At 1 d after the transfection, cultures were passaged into a 35-mm tissue culture dish or collagen gel-coated Transwell chamber (25 mm) in media supplemented with or without RA (30 nM). For Transwell cultures, chambers were maintained in an air-liquid interface culture condition for an additional 3 d. Cell extracts were prepared from these cultures according to the manufacturer's instructions (Packard Instrument, Meriden, CT). Luciferase reporter gene activity was determined using Luclite (Packard Instrument) and counted in a Luminometer (Packard Instrument). The -gal reporter gene activity was assayed according to instructions provided with the kit. For each transfection, relative luciferase activity was expressed after normalization with -gal activity. The results were averaged from at least triplicate dishes in three separate cultures.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

In Situ Hybridization of MUC5B Message on Airway Tissues

In situ hybridization demonstrated that MUC5B message is expressed mainly on submucosal gland cells of tracheobronchial tissues from a patient (H316 as shown in Table 1) with no obvious airway disease and inflammation (Figures 1A and 1C). The enlarged picture of the submucosal gland in Figure 1C supports this conclusion. For surface airway epithelium, the expression was generally very low (Figure 1A), except in some regions (Figure 1B). No MUC5B message could be demonstrated in the distal airway and parenchyma region (data not shown). Similar results were also observed on the tissue section from three other patients (H313, H317, and H311) without diagnosed lung diseases. MUC5B message was elevated in both the surface epithelium and submucosal gland of tissue sections obtained from a patient with usual interstitial pneumonitis (UIP) (H312) (Figures 3A and 3B) and a patient with emphysema (H297) (Figure 3C), respectively. In Figures 3A and 3C, MUC5B message was elevated in both the surface epithelium and the submucosal gland region, in contrast to sections from the "normal" patient (Figure 1). Interestingly, MUC5B message could also be seen in the surface epithelium of the bronchiole region of the patient with UIP (H312) (Figure 3B) and patients with emphysema (data not shown). Consistently, in situ hybridization using three other patients with emphysema and two others with UIP demonstrated the same results. In the airways of all these patients, we observed extensive goblet-cell hyperplasia (or metaplasia) in their airway epithelia (Figures 2B and 2C), in contrast to normal airway that had only a few goblet cells (Figure 2A). The surface expression of MUC5B was exclusively located in those cells. These results pinpointed the association between the expression of MUC5B message by surface epithelial cells and the pathogenic features in the corresponding airway region. Such an association was not seen for the expression of MUC5AC message. One example of such a comparative study on seven lung tissue sections from four patients with emphysema and three with UIP is shown in Figures 3C and 3D. In serial tracheal tissue sections from a UIP patient, MUC5B message could be seen in both the airway surface epithelium and the submucosal glands (Figure 3C), whereas MUC5AC message was seen restrictedly in the airway surface epithelium (Figure 3D) despite an elevated expression.

View larger version (96K): [in this window] [in a new window]   Figure 1.   In situ hybridization of MUC5B message in human bronchial sections. A 48-mer oligonucleotide corresponding to the antisense sequence of human MUC5B tandem repeats was synthesized and labeled, as described in text. Detailed procedures for the in situ hybridization are included in that section of the text. The tissue section was from a patient with no obvious airway disease and inflammation. After the hybridization and the immunohistochemical development (without counterstaining), tissue sections were examined under a light microscope. (A) A cross section of bronchial tissue after in situ hybridization. The section shows prominent MUC5B message in the submucosal gland cells and a relatively low presence in the surface epithelium. (B) Enlarged picture of the surface epithelium in the other region, which has MUC5B message. This type of hybridization is infrequent in this tissue section. (C) Enlarged picture of the submucosal gland region from A (as shown in the rectangle). Original magnifications are ×100 for A and ×400 for B and C. Note: Sense probe showed no reactivity with this tissue section (data not shown).

View larger version (105K): [in this window] [in a new window]   Figure 3.   In situ hybridization of MUC5B message in a human airway section derived from patients with UIP or emphysema. In situ hybridization was carried out as described in Figure 1. (A) A strong, positive hybridization with MUC5B antisense probe was seen in both the surface epithelium and the submucosal gland region in the trachea of a patient UIP. (B) Surface epithelium of the bronchiole region of the UIP patient's lung. MUC5B message level was elevated in both the surface epithelium and the submucosal gland, as compared with the "normal" one (Figure 1). Original magnifications are ×100 and ×400 for A and B, respectively. Similar results were seen in lung tissue sections from other patients with emphysema and UIP (data not shown). In situ hybridization of MUC5B (C) and MUC5AC (D) message in a human tracheal section derived from a patient with emphysema. In situ hybridization was carried out as described in Figure 1 and in MATERIALS AND METHODS. MUC5B message was seen in both the airway surface epithelium and the submucosal gland region (C), whereas in the same serial section, MUC5AC message was seen only in the surface epithelium (D). Original magnification for both C and D is ×100.

View larger version (14K): [in this window] [in a new window]   Figure 5.   Characterization of Cos-1 and genomic organization of the amino-terminal and 5'-flanking region of MUC5B. (A) Genomic organization of Cos-1. The regions corresponding to both MUC5B and MUC5AC messages are shown in filled bars. For MUC5B gene, 22,773 bp were sequenced. (B) Genomic organization of MUC5B gene upstream of the large central exon. The total length of MUC5B gene portion in Cos-1 is 22,773 bp. Arrows indicate the corresponding regions that are included in the study. Open bars and numbers indicate the exon, and the sizes of these bars are relative to their genomic size. Both the TATA box and 5' UTR and ATG are marked by thick arrows.

View larger version (111K): [in this window] [in a new window]   Figure 2.   Alcian-Blue (AB)-PAS staining of normal and diseased airway. (A) Normal trachea showing very little staining in some areas. (B) Significant increase of positive cells in the trachea of a patient with UIP. (C) Appearance of a large number of AB-PAS staining-positive cells in the bronchiole region of a patient with UIP. In addition, the lumen was filled with mucus-like secretion. Similar results were seen in lung tissue sections from other patients with UIP or emphysema (data not shown).

Expression of MUC5B Message in TBE Cultures

To further elucidate the regulation of MUC5B gene expression in vitro, primary and passage-1 cultures of TBE cells derived from these human tissues (Table 1) were studied. These cells were either plated on tissue culture (TC), collagen gel-coated (CG), or cultured in a Transwell chamber (BI) or in a collagen gel-coated Transwell chamber (BICG). For the first two culture conditions, cells were immersed in the culture medium; and for the Transwell chamber conditions (BI and BICG) they were maintained biphasically in an air-liquid interface. As shown in Figure 4A, passage-1 human TBE cells from a "normal" patient expressed MUC5B message, which could be detected by Northern blot. MUC5B messages in TC and CG cultures were very low and appeared unaffected by RA. However, we observed that the message was further elevated in cultures under BI and BICG conditions and showed strong response to RA. This is consistent with reports from previous publications (20, 21). Thus, MUC5B message in culture was affected not only by RA but also by the culture condition, with an order of BICG > BI >> CG > TC. These Northern blot results were not changed with different primary and passage-1 cultures derived from these 11 diseased human tissues. The expression of MUC5B gene was also studied in two commonly used human TBE immortalized cell lines (HBE1 and BEAS-2B). These cultures were also maintained under the same culture conditions as those primary and passage-1 human TBE cells. A similar result was found in the HBE1 cell line, except that the level of expression was lower than those of primary and passage-1 TBE cultures (Figure 4B). For the SV-40 immortalized TBE cell line, BEAS-2B subclone S, the level of expression was undetectable at the Northern blot level under all four conditions, as described earlier (Figure 4B).

View larger version (77K): [in this window] [in a new window]   Figure 4.   Northern blot analysis of MUC5B message in passage-1 human TBE cultures as well as HBE1 cells and BEAS-2B cells. (A) Passage-1 TBE cultures were plated on tissue culture dishes (TC), collagen gel-coated dishes (CG), Transwell chambers (BI), or collagen gel-coated Transwell chambers (BICG) with (+A) or without (A) RA as described in MATERIALS AND METHODS. For TC and CG cultures, cells were immersed in the medium, whereas BI and BICG chambers were lifted up and maintained in an air-liquid interface culture condition after being immersed in a culture condition for 7 d. RA (30 nM) was added to half of these cultures as indicated. Cultures at 21 d after plating were harvested for RNA preparation and Northern blot hybridization analysis. Hybridization with 18S rRNA cDNA probe was used as a reference for the equality of RNA loading in each gel well. (B) Northern blot hybridization of MUC5B message in various airway cultures. Passage-1 TBE, HBE1, and BEAS-2B (S clone) cells were plated on BICG culture condition as described in MATERIALS AND METHODS. RA (30 nM) was added to these cultures. At 21 d later, cultures were harvested for RNA and Northern blot hybridization, as described earlier.

Isolation of Cos-1 Genomic Clone and Characterization of cDNA Sequence of the 5'-End MUC5B

Because of the conserved nature of the mucin gene family, our initial goal was to use MUC2 amino terminal region (24) as a probe to isolate other mucin genes from a human cosmid genomic library derived from placenta. Under a low-stringency condition, one of the selected cosmid clones, Cos-1, contained the nucleotide sequence identical to the 3' end of MUC5AC gene at one end of the clone while at the other end it contained the nt sequence of the large central exon of MUC5B gene (Figure 5A). Therefore, on the basis of the 5' cen-MUC5AC-MUC5B-tel 3' genomic organization (14), the insert of Cos-1 must have contained the genomic DNA fragment flanking the region between the 3' end of MUC5AC and the 5' end of MUC5B gene. A complete genomic DNA sequence of the 22,773-bp size upstream of the large central exon of MUC5B was completed (GenBank accession number AF107890). Most of this data was identical to the published 5'-end sequence reported by Desseyn and coworkers (15) and Offner and associates (19).

To elucidate the 5'-end sequences of MUC5B messenger RNA, we carried out various 5'-RACE, RT-PCR, and primer extension experiments. Primers used in these studies are listed in Table 2. As this study started, the sequence of the large central exon domain was published (17). Subsequently, 5' upstream of this large central domain was also elucidated (15, 19). Sequences generated from our 5'-RACE DNA fragments were almost consistent with these studies. There are discrepancies in several locations between our study and other reports, which might be due to errors in sequencing or genetic polymorphism. The farther the 5'-end sequence of the message was extended, the more additional sequences were found in our study which were not found in earlier reports (15, 19). The nucleotide sequences generated from our 5'-RACE DNA fragments were compared with those from the Cos-1 genomic clone, which helped us to determine the junction between intron and exon. When these intron/exon junctions were determined, an additional exon at the 5' end could be demonstrated (Figure 5B). Thus, the overall exon number from the 5' end of the large central exon of MUC 5B gene is 30 instead of 29, contrary to what had been suggested previously (15). We have noticed that the recent report (25) from the same research group also demonstrates an additional exon upstream of the "exon 1" described in their previous report (15). Thus, the "exon 1" in ref. 15 is the "exon 2" in ref. 25. However, this discrepancy is not addressed in ref. 25. The newly defined exon 1 in ref. 25 is largely consistent with our findings described in the present report, which has further strengthened the notion that there are 30 instead of 29 exons upstream of the large central exon of human MUC5B. This new exon (exon 1) contains the putative transcriptional start site, the 5' untranslated region (UTR), and the portion of amino terminal sequence with a putative translation start codon ATG embedded in a Kozak consensus sequence (26). Further, on the basis of the deduced amino acid sequence, the amino terminal peptide contains a classic putative secretory signal sequence. This feature is consistent with the secretory nature of this mucin gene in airway and various organs. The transcription start site of MUC5B gene was determined by a primer extension experiment (Figure 6) and a modified RACE protocol (see MATERIALS AND METHODS). Both approaches yielded the same conclusion, suggesting that the transcription start site is 18,604 bp upstream of the large central exon. This putative transcription start site is one nt upstream of that reported by Offner and colleagues (19) and two nts upstream of that reported recently by Van Seuningen and associates (25).

View larger version (101K): [in this window] [in a new window]   Figure 6.   Primer extension study. Lane 1: RNA sample from human trachea tissue. Lane 2: RNA sample from human primary tracheobronchial epithelial cells. Lanes 3-6: Sequencing ladder. Radiolabeled dephosphorylated molecular weight pBR322/MspI (New England Biolabs) are shown on the right.

Characterization of the 5'-Flanking Region and the Promoter Activity

Among the 22,773 bp sequenced, 4,169 bp corresponded to the 5'-flanking region of MUC5B (Figure 7). Nucleotide sequence revealed the presence of a TATA box at 30 nt position relative to the transcription start site. Several putative motifs for various transcription factor bindings were also identified, including binding of c-Myc at 101, activator protein (AP)-2 at 1,155, Hoxd9/10 at 1,189, and glucocorticoid response element at 1,978. In addition, there are two putative motifs for binding of nuclear factor (NF)-B (at 237 and 371) and AP-1 (at 497 and 2,000). These findings are consistent with the recent report by Van Seuningen and associates (25), except with the additional 1,051-bp nucleotide sequence information.

View larger version (74K): [in this window] [in a new window]   Figure 7.   Nucleotide sequence of part of the 5'-flanking region and first exon of MUC5B gene; 2,007 nt out of 4,169-nt 5'-flanking region of MUC5B is shown. Various putative motifs are underlined. The transcription start site is indicated by an arrow. The deduced amino acid sequence of the coding region is underlined and the corresponding amino acid is represented by a capital letter.

To determine whether this 5'-flanking region contains cis-elements that are involved in the regulation of the expression of MUC5B gene, various promoter-reporter chimeric constructs were prepared (Figure 8A). In addition, the transcriptional start site previously determined by Desseyn and coworkers (15) has been shown to be located in exon 2 by this study as well as by the recent report from Van Seuningen and coworkers (25). Thus, to determine which sequences could be the potential promoter region for MUC5B gene transcription, the DNA fragment (MUC5B-i1, 13 to +2,738) upstream from the transcriptional start site reported by Desseyn and associates (15) was also included in the promoter-reporter gene transfection assay. The reporter gene activity in MUC5B-b1 and MUC5B-b2 transfected cells was 2- to 5-fold higher, respectively, than those transfected with the promoterless construct, pGL-3. However, such an above-basal activity was not seen in the transfection with MUC5B-il construct (Figure 8B). These results indicate that the first intron DNA fragment has no promoter activity (Figure 8B).

View larger version (29K): [in this window] [in a new window]   Figure 8.   Promoter-reporter gene transfection study. (A) A schematic illustration of chimeric constructs used in this study. Chimeric constructs containing b1, b2, and il DNA fragments of MUC5B gene and the reporter gene luciferase were prepared as described in MATERIALS AND METHODS. Passage-1 primary TBE cells were plated into 35-mm tissue culture dishes and transfection was carried out according to the Fugene 6 method. (B) Relative reporter gene activity expressed in transfected cultures, and carried out as described in MATERIALS AND METHODS.

On the basis of the study described earlier, we used MUC5B-b2 construct to characterize the specificity of the promoter activity. As shown in Figure 9, the promoter activity was seen in transfected HBE1 and primary TBE cultures, but not in the BEAS-2B culture. This result was consistent with the Northern blot data, which suggest a cell type-specific gene expression for MUC5B gene. We also observed that RA-stimulated promoter activity was influenced by the culture condition. As shown in Figure 10, when transfected cells were plated on a tissue culture dish, the reporter gene activity, after being normalized to the -gal activity, was not affected by RA. In contrast, the reporter gene activity was elevated 5-fold by RA treatment when transfected cells were maintained under a BICG condition. This culture condition-dependent, RA-stimulated promoter activity was consistent with the Northern blot data, which show that culture conditions can influence RA-dependent MUC5B gene expression.

View larger version (20K): [in this window] [in a new window]   Figure 9.   Effects of different cell types on the promoter-reporter activity. Passage-1 TBE, HBE1, and BEAS-2B (S clone) cells were plated onto 35-mm tissue culture dishes. Cells were transfected with the b2 MUC5B-luciferase reporter construct as described in MATERIALS AND METHODS. At 2 d later, cells were harvested for luciferase assay, as described in MATERIALS AND METHODS. The luciferase reporter gene activity in each transfected culture was normalized with the -gal activity and the results are expressed as relative activities after standarization. Triplicate dishes were used for each assay and the experiment was carried out independently with two different primary cultures (for TBE cells) or cell passages (for HBE1 and BEAS-2B cells). BEAS-2B cells, filled bars; HBE1 cells, striped bars; TBE cells, open bars.

View larger version (7K): [in this window] [in a new window]   Figure 10.   Effects of RA and culture conditions on the promoter-reporter gene activity. Primary human TBE cells in a 60-mm tissue culture dish were transfected with the b2 MUC5B-luciferase reporter construct, as described in MATERIALS AND METHODS. At 1 d after the transfection, cultures were passaged and plated onto 35-mm tissue culture dishes (TC) and collagen gel-coated Transwell chambers (BICG). At 1 d after plating, RA (30 nM) was added to half of these cultures and the Transwell chambers were maintained in an air-liquid interface culture condition. At 3 d later, both TC and BICG cultures were harvested for enzymatic assays, as described in MATERIALS AND METHODS. The luciferase reporter gene activity in each transfected culture was normalized with the -gal activity and the results are expressed as fold increase using RA-untreated cultures as 1. Triplicate dishes were used for each assay and the experiment was carried out independently with two different primary TBE cultures. RA-treated, filled bars; RA-untreated, open bars.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this report, we focused on the expression of MUC5B gene in human airway epithelium and the characterization of the genomic clone that contains the putative regulatory elements for the control of MUC5B gene expression. In the first part, three significant results were described. First, we demonstrated that MUC5B gene expression is more versatile than the other major mucin gene, MUC5AC, in airway epithelium. In situ hybridization demonstrated that MUC5B message is present predominately in the submucosal gland cells of airway tissue sections with no obvious disease and associated inflammation. This finding is consistent with the study reported by another laboratory (7). MUC5B message is present in the surface goblet-cell population of lung tissues obtained from patients with emphysema and UIP, in addition to a general elevated expression in the submucosal gland cells. An increase in the expression of MUC5B in the airway diseases was previously implied by biochemical approaches, which had shown that MUC5B gene product was one of the major components in mucus obtained from patients with asthma (27) and cystic fibrosis (28). In contrast to MUC5B gene expression, the expression of MUC5AC message is restricted in the airway surface epithelium regardless of the source of the tissue. These results suggest a closer correlation between aberrant MUC5B gene expression and the pathogenesis of airway disease than for MUC5AC.

It is difficult to explain why MUC5AC gene expression is restricted in the surface epithelium whereas MUC5B is not in these tissues with airway diseases. Two possible explanations can be put forward. One relates to the origin of the stem cells that are responsible for the repopulation of epithelium in injured tissue; the other relates to the mechanism of transdifferentiation of epithelial cells. According to the first explanation, submucosal gland cells are presumably the "stem" cells that are responsible for the regrowth of epithelium on the injured area in these diseased cells. Under this assumption, MUC5B gene expression will occur naturally in surface epithelium once the gland cells migrate into the injured surface. In contrast, the MUC5AC message in airway surface epithelium would be derived from the remaining surface epithelial cells. This would explain why MUC5AC gene expression is restricted in the surface layer. However, evidence to support the "stem" cell nature of the submucosal gland cells is lacking. Further study is needed to verify this possibility.

Under the second explanation, it is proposed that airway epithelial cells can change their differentiated function according to the microenvironments surrounding them. This transdifferentiation mechanism has been recognized before (29). It is possible that a transdifferentiation occurs on these airway disease tissues. Previously, we demonstrated an immediate loss of differentiated features in protease-dissociated rat airway epithelial cells after plating on a culture surface (30). However, this loss of differention is transient, and redifferentiation occurs in culture when these cells are properly propagated (20). This phenomenon is consistent with the "plastic" nature of TBE cells (29). A similar transdifferentiation mechanism might occur in the expression of MUC5B gene in the surface epithelium of disease airways. In this case, MUC5B gene expression in surface epithelium can serve as a biomarker for transepithelial cell differentiation associated with the pathogenesis of airway diseases.

The second significant finding in the first part of this study is the demonstration of the expression of MUC5B message in cultured TBE cells regardless of the origin of obtained tissue. Our limited culture study from 11 different airway tissue sources, including four nondiseased airway tissues, demonstrated that the level of MUC5B message in culture is independent of the status of the tissue source. It is also necessary to point out that the protease-based dissociation method is less likely to dislodge cells from the submucosal region than it is those from the surface (31, 32). Although we cannot absolutely exclude the possibility that some glandular cells were dislodged by the protease treatment and that these cultures were contaminated with these cells, the percentage of these cells was too small to have a significant impact on the observed level of MUC5B gene expression. Therefore, it is very likely that the cultured cells that differentiated into a MUC5B- expressing cell type came from surface TBE cells. The expression of MUC5B gene in these surface epithelial cell- derived cultures is also consistent with the transdifferentiation mechanism described earlier. Our in vivo and in vitro data strongly support the notion of MUC5B gene expression as a biomarker associated with the "plasticity" of the airway epithelial cells. The elucidation of the molecular events that trigger the expression of MUC5B on the surface epithelium of the airway may have therapeutic applications in the diagnosis and treatment of patients with various airway diseases.

The third significant result in our research is determining the cell type specificity of MUC5B gene expression in culture. Northern blot hybridization demonstrated the expression of MUC5B message in primary and passage-1 TBE cultures and, to a lesser extent, in the HBE1 cell line. No message was found for RNA prepared from the SV-40 T antigen-immortalized human TBE cell line BEAS-2B, regardless of the culture condition. For primary and passage-1 human TBE cells and the HBE1 cell line, the expression was dependent on RA supplement and was further influenced by the culture condition. The effect of RA on MUC5B gene expression is consistent with the notion that vitamin A and its derivatives are essential mediators for airway mucociliary cell differentiation (20). Upon plating cells on a tissue culture surface under an immersed culture condition, cell differentiation is not extensive (30). However, MUC5B gene expression was significantly elevated in both BI and BICG conditions. This is the first demonstration that MUC5B gene expression is influenced by various culture conditions with an order of BICG > BI >> CG > TC. The effect of the BICG condition was consistent with other reports (13, 21). However, neither Koo and colleagues (21) nor Bernacki and associates (13) reported the individual contribution of collagen gel substratum and air-liquid interface on mucin gene expression. Our present data clearly demonstrate the importance of the air-liquid interface in regulating MUC5B gene expression. Our results on the Northern blot analysis of MUC5B message are also consistent with the extent of mucous cell differentiation in these cultures (data not shown). Thus, the regulation of MUC5B gene expression is cell type-specific and related to in vitro mucous cell differentiation.

The nature of transepithelial cell differentiation is unknown. One possible approach is to study the transcriptional regulation of mucin genes. To elucidate such a mechanism, the isolation and characterization of the 5'-flanking region of MUC5B gene is needed. Our cloning and sequencing data suggest that the size of the message upstream of the unusually large central exon of MUC5B gene is 4,032 bp, and the genomic sequence corresponds to 17,788 bp. This region is composed of 30 exons and encodes 1,324 amino acid residues. Exon sizes vary from 44 to 262 bp, and intron sizes range from 87 to 2,492 bp. Most of the data are consistent with recent publications from Offner and colleagues (19) and Van Seuningen and associates (25), except for the different transcriptional start sites among all three reports. The nature of this discrepancy is unknown.

We demonstrated the basal promoter activity when a 4,169/+7 (b2) or 1,097/+7 (b1) DNA fragment of the 5'-flanking region was introduced to the promoterless luciferase expression clone. These two constructs demonstrated 2- or 5-fold promoter activity, respectively, compared with the promoterless construct. Van Seuningen and colleagues (25) recently suggested that the basal promoter area was 956 bases upstream of the transcriptional start site. However, we were able to show additional basal promoter activity when a larger DNA size, b2, was used. This discrepancy may be due to the different cell typescolon cancer cells versus airway epithelial cellsused in these experiments. Alternatively, there may be a strong negative cis-element in the region between 956 and 1,097 that can downregulate the promoter activity of MUC5B. Further study is needed to elucidate this difference.

Using this large 5'-flanking region of MUC5B gene, we demonstrated that the basal promoter activity was cell type-specific. The reporter gene activity was seen in transfected primary TBE cells and the HBE1 cell line. The activity in the HBE1 cell line was lower than that in primary cells. No activity above the control vector transfected cells was seen in transfected BEAS-2B cells. These results at the promoter level are consistent with MUC5B Northern blot data in these cells.

We also demonstrated that basal promoter activity could be further stimulated by RA in culture. These results suggest that the 5'-flanking region contains cis-elements that are responsible for RA stimulation. The classic RA pathway involves the interactions between the nuclear receptors, retinoid acid receptors (RARs) and retinoid X receptors (RXRs), and the binding sites in the 5' regulatory region of RA-regulated genes. However, a search of the transcriptional factor database with the 4,169-bp 5'-flanking region of MUC5B did not reveal any typical RA- responsive element. Thus, RA may not induce MUC5B promoter activity directly by the classic mechanism. Perhaps an indirect mechanism, through the induction of other factors by RA, is involved in this RA-dependent MUC5B gene expression. It has been observed by others (33), as well as by us, that RA could alter the NF-B activity. Such an alteration may influence the promoter activity because there are at least two putative NF-B binding sites in the 5'-flanking region. In addition, the observation that RA-stimulatory effects on MUC5B promoter activity in TBE cells could be demonstrated only in BICG but not in TC condition suggests that factor(s) other than RAR and RXR might be involved. Cotransfection with RAR and RXR expression constructs was unable to demonstrate RA- stimulated promoter activity on the TC condition (data not shown). These results further rule out the possibility of direct interactions between these nuclear receptors and Retinoic acid response elements involved in the upregulation of MUC5B gene expression. Considering the consistent findings at both the level of the promoter activity and the level of message RNA abundance in response to RA treatment and specific culture conditions, we suggest that at least part of the regulatory mechanism in the RA and culture condition-dependent MUC5B gene expression may occur at the transcriptional level.

Searching the transcriptional factor binding site database with the 4,169-bp DNA fragment of MUC5B 5'-flanking region reveals the presence of several putative motifs for binding transcriptional factors, such as AP-1, AP-2, NF-B, Glucocorticoid receptor, c-Myc, etc. The functional nature of these motifs is still unknown. Inasmuch as MUC5B gene expression is elevated in a variety of airway diseases and infections, some of these motifs may become active after binding the appropriate transcriptional factors in the induction of MUC5B gene expression. Both the elevation of NF-B and AP-1 transcription factors have been implicated in the pathogenesis of airway diseases and inflammation (34). Interestingly, there are two NF-B and AP-1 binding sites in the 5'-flanking region of MUC5B gene. This finding further affirms the significance of these motifs in the regulation of aberrant mucin gene expression. Experiments to elucidate the functional role of these motifs are currently underway.