Cloning and characterization of the cDNA for human airway trypsin-like protease.

Previously we isolated a trypsin-like enzyme designated human airway trypsin-like protease from the sputum of patients with chronic airway diseases. This paper describes the cDNA cloning, characterization of the primary protein structure deduced from the cDNA, and gene expression of this enzyme in various human tissues. We obtained an entire 1517-base pair sequence of cDNA with an open reading frame encoding a polypeptide with 418-amino acid residues. The polypeptide consisted of a 232-residue catalytic region and a 186-residue noncatalytic region with a hydrophobic putative transmembrane domain near the NH2 terminus. The polypeptide was suggested to be a type II integral membrane protein in which the COOH-terminal catalytic region is extracellular. Therefore, this protein is thought to be synthesized as a membrane-bound precursor and to mature to a soluble and active protease by limited proteolysis. It showed 29-38% identity in the sequence of the catalytic region with human hepsin, enteropeptidase, acrosin, and mast cell tryptase. The noncatalytic region had little similarity to other known proteins. In Northern blot analysis a transcript of 1.9 kilobases was detectable most prominently in the trachea among 17 human tissues examined.

Previously we isolated a trypsin-like enzyme designated human airway trypsin-like protease from the sputum of patients with chronic airway diseases. This paper describes the cDNA cloning, characterization of the primary protein structure deduced from the cDNA, and gene expression of this enzyme in various human tissues. We obtained an entire 1517-base pair sequence of cDNA with an open reading frame encoding a polypeptide with 418-amino acid residues. The polypeptide consisted of a 232-residue catalytic region and a 186-residue noncatalytic region with a hydrophobic putative transmembrane domain near the NH 2 terminus. The polypeptide was suggested to be a type II integral membrane protein in which the COOH-terminal catalytic region is extracellular. Therefore, this protein is thought to be synthesized as a membrane-bound precursor and to mature to a soluble and active protease by limited proteolysis. It showed 29 -38% identity in the sequence of the catalytic region with human hepsin, enteropeptidase, acrosin, and mast cell tryptase. The noncatalytic region had little similarity to other known proteins. In Northern blot analysis a transcript of 1.9 kilobases was detectable most prominently in the trachea among 17 human tissues examined.
Many previous investigations have indicated that proteases released from immunoinflammatory cells participate in pathogenesis of several kinds of respiratory diseases. For instance, neutrophil elastase has been shown to be intimately related to the pathologic states of pulmonary emphysema (1,2), cystic fibrosis (3,4), interstitial pneumonia (5), and adult respiratory distress syndrome (6) through destruction of extracellular matrix components, such as elastin, of alveolar and bronchial tissues. Mast cells, which abound in airway mucosa and in alveolar wall, release trypsin-like protease (tryptase) and chymotrypsin-like protease (chymase) into extracellular spaces during degranulation (7). The tryptase has potential to stimulate smooth muscle, fibroblast, and tissue turnover (8). Different substrates for chymase (9 -11) indicate the potential involvement of the enzyme in a variety of processes related to the inflammatory response. Recently it was revealed that chymase from human mast cells selectively converted big endothelins to trachea-constricting peptides (12). These effects of the two mast cell proteases have attracted considerable attention as one of the pathogenic determinants and the therapeutic targets of bronchial asthma and allergic inflammation. Elastase released from alveolar macrophages has also been suggested to contribute to the pathogenesis of pulmonary emphysema by degrading matrix components of alveolar walls (13,14).
However, there are very few reports dealing with the functions and roles of proteases secreted from respiratory tissues, such as secretory glands or surface epithelial cells of the airway. Kido and co-workers (15,16) found a novel trypsin-like protease that is secreted from rat Clara cells, secretory cells localized to the distal airway only. The protease, named tryptase Clara, was shown to enhance the infectivity of influenza and Sendai viruses (17), although its physiological role is unknown.
Previously, we found trypsin-like activity in the sputum of patients with chronic airway diseases and isolated a novel trypsin-like protease from the sputum, designated human airway trypsin-like protease (HAT) 1 (18). Gel filtration studies showed that HAT was a monomeric enzyme with an apparent molecular mass of 27 kDa. Immunohistochemical studies showed that HAT was localized mainly in cells of submucosal serous glands of the bronchi and trachea. These results indicate that HAT is released from the submucosal serous glands onto mucous membrane, at least in patients with chronic airway diseases.
In this paper, we report the cloning of HAT cDNA, the primary structure of this enzyme and characterization of the polypeptide deduced from the nucleotide sequence of the cDNA, and results of analysis of expression of HAT mRNA in various human tissues. The primary structure of HAT was compared with that of other known serine proteases.

EXPERIMENTAL PROCEDURES
Materials-Human trachea QUICK-Clone TM cDNA, human trachea poly(A) ϩ RNA, human trachea gt10 cDNA library (oligo(dT) and random-primed), 5Ј-RACE kit, human multiple tissue Northern blots, and human ␤-actin cDNA were purchased from CLONTECH Laboratories Inc. (Palo Alto, CA). Taq DNA polymerase was from Promega Corp. (Madison, WI). SureClone TM ligation kit, dNTP, and plasmid vector pUC18 were from Amersham Pharmacia Biotech. Avian myeloblastosis virus reverse transcriptase and RNase inhibitor were from Boehringer Mannheim. Restriction endonucleases, random primer labeling kit, and Escherichia coli JM109 were from Takara Shuzo Co. Ltd. (Otsu, Japan). Nylon membrane Hybond TM -Nϩ for blotting and [␣-32 P]dCTP for probe labeling in hybridization were from Amersham. Denhardt's solution and salmon sperm DNA were from Wako Pure Chemical Industries Ltd. (Osaka, Japan). Qiagen lambda kit for purification of phage DNA was * This work was supported by Grant-in-aid for Developmental Scientific Research 09670616 of the Ministry of Education, Science, Sports, and Culture, Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB002134.
§ To whom correspondence should be addressed. DNA Amplification by Polymerase Chain Reaction (PCR)-PCR was performed according to the procedure described by Sambrook et al. (19). Oligonucleotides used as PCR primers were synthesized by a DNA/RNA synthesizer (Applied Biosystems Inc., model 394) and purified by oligonucleotide purification cartridge column. Unless otherwise stated, PCR was carried out by adding 15 pmol of each primer and an appropriate amount of template DNA to 20 l of PCR buffer (10 mM Tris-HCl, pH 9.0, 50 mM KCl, 1.5 mM MgCl 2 , 1% Triton X-100) containing 0.5 units of Taq DNA polymerase and 0.2 mM dNTP. The reaction using a DNA thermal cycler (Perkin-Elmer Corp.) was carried out for 35 cycles of 1-min denaturation at 94°C, 1.5-min annealing at 57°C, and 2-min extension at 72°C.
Subcloning of DNA Fragments-To subclone DNA fragments that were amplified by PCR, SureClone TM ligation kit was used. DNA fragments were blunted by Klenow fragment, inserted into the SmaI site of plasmid vector pUC18, and introduced into E. coli JM109 by Hanahan's method (20). On the other hand, for subcloning of insert DNA of gt10 phage clone, the insert DNA was excised by EcoRI from phage DNA, which was purified using Qiagen lambda kit and inserted into the EcoRI site of plasmid vector pUC18. E. coli JM109 was transformed as described above. Plasmid DNA was isolated from each transformant by the alkaline lysis procedure (21) with minor modifications.
Analysis of DNA and Amino Acid Sequence-The nucleotide sequence of the DNA inserted into plasmid vector pUC18 was analyzed by an automated DNA sequencer (Applied Biosystems Inc., model 373) using the Dye-Deoxy TM terminator cycle sequencing kit. Both strands of all clones were completely sequenced. Hydropathy of amino acid sequence was analyzed (22) with the Genetyx program package (Software Development Co. Ltd., Tokyo, Japan). A computer survey of the National Biomedical Research Foundation (Washington, D.C.) and SWISS-PLOT (European Bioinformatics Institute, Geneva, Switzerland) data banks for similarity of amino acid sequences between HAT and other known proteins was carried out using MPsrch program, which was modified from the method of Smith and Waterman (23) with Teijin Systems Technology Ltd. (Yokohama, Japan).
Amplification of a Partial cDNA Fragment-In a previous report (18), we showed that the sequence of the 20 NH 2 -terminal amino acids of native HAT purified from the sputum of patients with chronic airway diseases was ILGGTEAEEGSWPWQVSLRL (amino acids 187-206 in Fig. 1). Based on this amino acid sequence, we designed and synthesized two kinds of degenerate PCR primers, namely 5Ј-ATCYTNGGRG-GNACNGAGGC-3Ј 2 (sense) and 5Ј-ARKCKMAGGCTSACYTG-3Ј 2 (antisense) to obtain the 59-bp cDNA fragment encoding the front 19 residues of the NH 2 -terminal amino acid sequence by PCR. PCR was carried out in the reaction mixture containing 5 pmol of each primer and 1 ng of cDNA derived from human trachea (QUICK-Clone TM cDNA). The amplified DNA fragment was then subcloned and sequenced as described above. The analysis of the sequence showed that a 59-bp DNA fragment encoding the 19-residue amino acid sequence corresponding to the NH 2 terminus of the purified HAT was produced by this PCR.
Amplification of cDNA by 3Ј-Rapid Amplification of cDNA Ends (RACE)-To obtain a cDNA that had a nucleotide sequence in the downstream side of the 59-bp DNA fragment, we employed the 3Ј-RACE method developed by Frohman et al. (24). Two kinds of sense primers were used to amplify the cDNA specifically and effectively. These primers were designed and synthesized based on the nucleotide sequence of the 59-bp cDNA fragment. At first, single-stranded cDNAs were synthesized by reverse transcription at 42°C for 60 min in 20 l of reaction buffer (50 mM Tris-HCl, pH 7.6, 60 mM KCl, 10 mM MgCl 2 , 1 mM dithiothreitol) containing 10 ng of human trachea poly(A) ϩ RNA, 115 pmol of (dT) 17 -adapter primer 5Ј-GACTCGAGTCGACATCGA(dT) 17 -3Ј, 25 units of RNase inhibitor, 1 mM dNTP, and 40 units of avian myeloblastosis virus reverse transcriptase. One-tenth of the reaction mixture was used as a template in the first-round PCR in which 5Ј-ATCTT-GGGGGGCACGGAGGCTGA-3Ј and the adapter primer 5Ј-GACTC-GAGTCGACATCGAT-3Ј were used as the sense and antisense primers, respectively. For further amplification of the cDNA, the second-round PCR was carried out using one-fortieth of the first-round PCR reaction mixture as the template with 5Ј-GAGGCTGAGGAGGGAAGCTGGC-CGT-3Ј (nucleotides 635-659 in Fig. 1) and the (dT) 17 -adapter primer described above as the sense and antisense primers, respectively. The cDNA amplified by 3Ј-RACE was then subcloned and sequenced.
Screening of cDNA Library-Plaque hybridization against human trachea cDNA library was performed according to the standard procedure (19). The DNA fragment obtained by 3Ј-RACE was labeled by the random prime method (25) using [␣-32 P]dCTP and random primer labeling kit. Using this probe, 1 ϫ 10 6 plaques derived from human trachea gt10 cDNA library were screened by hybridization as follows. The blots for the plaques were hybridized with the probe at 65°C overnight (16 -20 h) in a solution containing 5ϫ SSPE buffer (0.75 M NaCl, 50 mM NaH 2 PO 4 , 5 mM EDTA, pH 7.4), 5ϫ Denhardt's solution, 0.1% SDS, and 100 g/ml denatured salmon sperm DNA. These blots were then washed twice at 65°C for 20 min with 0.1ϫ SSPE buffer containing 0.1% SDS. Five positive clones were selected and plaquepurified, and the insert DNAs of these clones were then subcloned and sequenced.
Amplification of cDNA by 5Ј-RACE-To obtain a cDNA that had a nucleotide sequence in the upstream side of the cDNA coding for native HAT, amplification of the cDNA was carried out using 5Ј-RACE kit (24). Single-stranded cDNAs were synthesized by reverse transcription of 2 g of human trachea poly(A) ϩ RNA using the antisense primer 5Ј-ACGTGGCAATCCAGTCACGAGGATT-3Ј (nucleotides 785-761 in Fig.  1). The single-stranded cDNAs were purified using glass powder in 5Ј-RACE kit after alkaline hydrolysis of RNA in the reaction mixture. Using T4 RNA ligase, AmpliFINDER TM anchor was ligated to the 3Ј-ends of the single-stranded cDNAs. PCR amplification (0.75 min at 94°C, 0.75 min at 57°C, and 2 min at 72°C) was then carried out using 0.01 of the ligation mixture as template, with anchor primer 5Ј-CTG-GTTCGGCCCACCTCTGAAGGTTCCAGAATCGATAG-3Ј and 5Ј-TGA-GCTGCTGTCAGGATCCACATGT-3Ј (nucleotides 741-717 in Fig. 1) as the sense and antisense primers, respectively. The cDNA amplified by 5Ј-RACE was then subcloned and sequenced.
Expression and Purification of Recombinant HAT-A 1.3-kb BamHI-HindIII fragment containing the entire HAT cDNA was cloned into the transfer vector pBlueBacIII (Invitrogen, San Diego, CA) to generate pBacPHAT1. Recombinant HAT-expressing viruses were generated after co-transfection of Sf9 cells with pBacPHAT1 and wild-type AcMNPV DNA essentially as described by the manufacturer (Invitrogen). For baculovirus/insect cell expression (26), 800 ml of Tn5 (27) cells were then infected with the high titer lysate for 72 h and harvested by centrifugation. The cell pellet was treated with 1% Triton X-100 for 1 h on ice and was centrifuged at 100,000 ϫ g for 1 h at 4°C. From this infected cell lysate, the recombinant HAT was isolated by sequential chromatographic procedures of the native HAT purification described previously (18). SDS-polyacrylamide gel electrophoresis, immunoblotting, and degradation of fibrinogen by HAT were done as described (18) Northern Blot Analysis-The expression level of HAT mRNA in various human tissues was examined by Northern blot analysis. To prepare the probe for the analysis, the full-length cDNA for HAT was 32 P-labeled by random priming (25) and hybridized as follows. Northern blots of various human tissues, which contained 2 g of poly(A) ϩ RNA derived from various tissues in each lane, were probed under the same conditions as the library screening described above (except that the concentration of SDS was 0.5%) and then washed. In the case of the blot for trachea, 2 g of human trachea poly(A) ϩ RNA was resolved by 1% agarose-formaldehyde gel electrophoresis (28), and transferred onto Hybond TM -Nϩ blotting membrane and UV-cross-linked. X-ray films were exposed to the probed blots for 4 days at Ϫ80°C with an intensifying screen, and the presence of HAT mRNA in each human tissue was evaluated. These blots were then stripped of the HAT cDNA probe by boiling in 0.5% SDS for 10 min and re-probed with 32 P-labeled human ␤-actin control probe as an internal standard for the amounts of RNA loaded.

RESULTS AND DISCUSSION
Cloning of HAT cDNA-Using a pair of highly degenerate oligonucleotide primers, the partial 59-bp cDNA fragment for HAT, which contained a nucleotide sequence coding for the NH 2 -terminal 19-residue amino acid sequence of the native HAT, was obtained by PCR amplification from human trachea cDNA. To stretch this cDNA sequence to the 3Ј-end, a 3Ј-RACE reaction was carried out. The resulting 0.9-kb amplified product was shown to encompass the entire nucleotide sequence of the 3Ј region, including the poly(A) tail of HAT cDNA (nucleotides 635-1517 in Fig. 1). The amino acid sequence deduced from this 0.9-kb fragment was shown to exactly contain the 15amino acid sequence (amino acids 192-206 in Fig. 1) of the NH 2 -terminal 20-amino acid sequence of the native HAT. With this 0.9-kb cDNA fragment as a probe, 1 ϫ 10 6 clones of a human trachea gt10 cDNA library were screened. Five of 28 independent positive clones were then subcloned and sequenced. The largest insert was shown to contain a 1323-bp sequence of cDNA (nucleotides 133-1455 in Fig. 1) but was considered not to contain the entire nucleotide sequence of the 5Ј region of HAT cDNA. To obtain the missing sequences in the 5Ј region of HAT cDNA, 5Ј-RACE reaction was carried out. The 5Ј-RACE procedure produced a 741-bp cDNA fragment (nucleotides 1-741 in Fig. 1). This product had a 609-bp nucleotide sequence overlapping (nucleotides 133-741 in Fig. 1) with the 5Ј-end of the largest insert of cDNA clone obtained by the cDNA library screening.
Sequence and Structural Features of HAT cDNA-Analysis of the cDNA clones obtained by the successive procedures in-cluding 3Ј-RACE, cDNA library screening, and 5Ј-RACE showed a 1517-bp nucleotide sequence up to the poly(A) region (Fig. 1), which represented the HAT cDNA sequence. This nucleotide sequence was also shown to contain one open reading frame, and the polypeptide deduced from the cDNA included the 20-residue amino acid sequence of the NH 2 terminus of the native HAT (amino acids 187-206 in Fig. 1). The molecular mass of the polypeptide, including the NH 2 terminus of the 20 residues to the COOH terminus deduced from the stop codon TAG (nucleotide-1316), was estimated to be 25,308 Da. This value is similar to the apparent molecular mass (27 kDa) estimated by gel filtration of the native HAT protein purified from sputum (18).
In the 5Ј-flanking region of this cDNA, one in-frame stop codon TAG was located at nucleotide 26. Four in-frame ATG codons were detectable between this stop codon and the region encoding the native HAT, but none of these ATG codons satisfied the criteria for a Kozak consensus sequence (29). Therefore we could not determine the translational initiation site in the cDNA from the nucleotide sequence. To determine the initiation site, we expressed recombinant HAT in a baculovirus/ insect cell system using the HAT cDNA. The recombinant virus containing the HAT cDNA was isolated, and the insect cell Tn5 was infected with the virus and then cultured. The lysate obtained by 1% Triton X-100 treatment of the infected cells was analyzed by immunoblotting with a rabbit antibody against a peptide corresponding to the NH 2 -terminal 19-amino acid sequence of the native HAT (18) as primary antibody, and the immunoblotting indicated that the infected cells biosynthesized a protein with a molecular mass of 48 kDa as a main product (Fig. 2). The molecular mass of each polypeptide, deduced from the nucleotide sequence initiating from each of 4 ATG codons in the cDNA, was 46,263, 32,933, 31,436, and 30,107 Da, respectively. The molecular mass of 46,263 Da is the most similar to that of the recombinant protein expressed in the insect cells, suggesting that the ATG located nearest the 5Ј-end (at nucleotide 62) is the initiation codon of HAT.
To demonstrate that the cloned enzyme has the same activity as the native HAT, the recombinant HAT that was expressed in the baculovirus/insect cell system was isolated in its active form. The minor product in Fig. 2, lane 3 was isolated selectively as the active recombinant HAT from the infected cell lysate by sequential chromatographic procedures of the native FIG. 1. Nucleotide sequence of HAT cDNA and its deduced amino acid sequence. The nucleotide sequence of the HAT cDNA is shown along with the deduced amino acid sequence beginning with the first ATG codon. A stop codon (TAG) at the terminus of the translation sequence is marked with an asterisk. Nucleotides are numbered at the right margin and amino acids on the left. The NH 2 -terminal sequence obtained from the purified enzyme is underlined. The boxed amino acid sequence represents a potential transmembrane domain. HAT purification (18). The purified recombinant enzyme has the molecular mass of 28 kDa on SDS-polyacrylamide gel electrophoresis and the identical 10 NH 2 -terminal residues to the native HAT. Immunoblotting also showed the purified recombinant enzyme as same size as the native HAT (Fig. 2). The recombinant HAT had an enzymatic activity degrading fibrinogen, especially the ␣-chain (Fig. 3), similar to the native HAT. From these results, it was established that the isolated cDNA clone encodes HAT.
Based on these results, the nucleotide sequence of the cDNA for HAT (Fig. 1) was summarized as follows. The cDNA includes 1254 nucleotides coding for 418 amino acids and two untranslated nucleotide sequences composed of 61 and 185 nucleotides at the 5Ј-end and 3Ј-end, respectively. In the 3Јuntranslated region, there is a polyadenylation signal sequence, ATTAAA, at nucleotides 1478 -1483, 17 nucleotides distant from the poly(A) tail.

Analysis of Deduced Amino Acid Sequence of HAT-
The open reading frame of HAT cDNA was thought to encode a polypeptide consisting of 418 amino acid residues, thus having the molecular mass of 46,263 Da. The NH 2 -terminal 20-amino acid sequence of the native HAT extends from Ile 187 to Leu 206 in the sequence of the deduced polypeptide (Fig. 1). This result indicates that the Arg 186 -Ile 187 peptide bond in the HAT polypeptide should be cleaved for activation of HAT. This type of cleavage has been shown to be a relatively common step for activation of many known serine protease zymogens (30,31). Therefore it is likely that the HAT gene product is synthesized as a precursor protein that consists of a noncatalytic region with 186 amino acid residues (20,955 Da, amino acids 1-186 in Fig. 1) and a catalytic region with 232 amino acid residues (25,308 Da, amino acids 187-418 in Fig. 1) and that the precursor is converted to an active enzyme by limited proteolysis like trypsinogen to trypsin in the small intestine (32). In this noncatalytic region, there were two potential N-linked glycosylation sites, namely Asn-Asn-Ser and Asn-Pro-Ser, at Asn 144 and Asn 152 , respectively.
A hydropathy plot (22) of the predicted amino acid sequence of HAT precursor (Fig. 4) showed that a typical NH 2 -terminal signal sequence (33-35) is not present, but a single obvious hydrophobic region (amino acids 13-43 in Fig. 1) is present near the NH 2 terminus. This hydrophobic region consisting of 31 amino acid residues does not contain any charged amino acids and is flanked by charged amino acids (Arg 12 and Asp 44 ). This internal hydrophobic region is thought to correspond to a transmembrane domain that anchors the protein to the cell membrane (36). A generalized rule in the eucaryotic transmembrane proteins (37,38) suggests that the difference in total charge between 15-residue sequences on either side of the membrane-spanning hydrophobic region determines the orientation of the protein, with the more positive side facing the cytosol. As for the precursor polypeptide deduced from HAT cDNA, the NH 2 -terminal side of the hydrophobic region had a net charge of ϩ3, whereas the opposite side had that of ϩ1. The charge on the NH 2 -terminal side was ϩ2, as positive as that on the COOH-terminal side. This result suggests that HAT precursor has an intracellular NH 2 -terminal tail region consisting of 12 amino acid residues facing the cytosol and an extracellular COOH-terminal region consisting of 375 amino acid residues and containing the catalytic region. Therefore, the HAT precursor can be classified as a type II integral membrane protein (39,40) and is thought to be synthesized as a membrane-bound precursor protein translocated to the cell surface, processed to a soluble form, and released.
Because cDNA Cloning of Human Airway Trypsin-like Protease has been isolated and characterized, it is unknown whether or not the membrane-bound HAT is active on the cell surface. The mechanisms of expression and activation of many serine proteases have been clarified. The predicted maturation process of HAT precursor described above is similar to that of the Bacillus amyloliquefaciens subtilisin (41). The subtilisin is synthesized as a membrane-associated precursor (preprosubtilisin) and released outside the cell after it is autocatalytically converted to an active form (42). Only mature subtilisin has been detected extracellularly (41). Active HAT contained in sputum samples was also detected extracellularly.
It is possible that the membrane-bound HAT or the portion remaining in the membrane after release of the soluble HAT may be involved in some important physiological processes on the cell surface through interaction with ligands, other proteins, or the surface. Recent reports have shown that some viruses and a bacterial toxin utilize cell surface proteases as receptors (43)(44)(45)(46)(47), indicating other usage in addition to intrinsic roles of these proteins. Homology of Amino Acid Sequence of HAT with Other Proteases-To find any similarity in the primary structure between HAT and known proteins, we surveyed publicly available data banks. Previous investigators have shown that the serine protease family has a common catalytic site consisting of cDNA Cloning of Human Airway Trypsin-like Protease three amino acid residues, His, Asp, and Ser, joined by hydrogen bonds to display catalytic action as a catalytic triad, although they are located apart from each other in the primary structure of the enzyme (48). Based on these established facts, the catalytic site of HAT is thought to consist of amino acid residues His 227 , Asp 272 , and Ser 368 (Fig. 5). In comparison of the amino acid sequence of HAT with those of other serine proteases, the most striking similarity was found around this putative catalytic triad as shown in Fig. 5. Six of seven cysteine residues in the catalytic region of HAT were at identical positions as those of other serine proteases (Fig. 5). Nine cysteine residues were contained in the deduced polypeptide of HAT precursor, and the Cys 20 was located in the predicted transmembrane domain. Based on the locations of the known disulfide bridges in other serine proteases (49), it is postulated that the other eight cysteine residues may form four disulfide bonds, which are located at cysteine pairs 212/228, 337/353, and 364/ 393 in the catalytic region and at 173/292 between the noncatalytic region and the catalytic region.
It was shown that the amino acid sequence of the catalytic region of HAT was homologous to that of the other human serine proteases: 38% identity with hepsin (50), 32% with enteropeptidase (51), 30% with acrosin (52), and 29% with mast cell tryptase (53). Hepsin, of which the catalytic region shows the highest similarity with that of HAT in this survey, is a cell surface protease widely expressed in various tissues including liver and is suggested to play a role in cell growth and maintenance of cell morphology (54).
On the other hand, the amino acid sequence of the noncatalytic region of HAT showed no significant similarity with those of other proteins and had neither kringle nor an EGF-like domain, which are found in some kinds of proteases relating to the blood coagulation, fibrinolysis, and complement cascades (55). The function or roles of this unique and relatively long noncatalytic portion of HAT precursor are unknown.
Northern Blot Analysis-Previously, we showed immunohistochemically that HAT protein was expressed in the cells of submucosal serous glands of human bronchi and trachea (18). Serous glands are widely distributed in various human tissues. Therefore multiple tissue Northern blot analysis was carried out to confirm that HAT mRNA was expressed in the human lower airway and also to clarify whether or not HAT mRNA was expressed in human other tissues. As shown in Fig. 6, a 1.9-kb transcript was detectable in only the trachea blot among the 17 different types of tissues examined, such as heart, brain, pancreas, lung, and liver. The mRNA size is in fairly good accordance with that (1517 bp) of the HAT cDNA established in the present work. In addition to the 1.9-kb mRNA, 3.0-kb and 0.9-kb signals were weakly detectable in the trachea and pancreas blot, respectively. These two transcripts may appear as result of an alternative splicing/polyadenylation process or represent a cross-hybridizing mRNA, but the nature of these transcripts is unknown. These results strongly suggest that HAT mRNA is more actively expressed in the lower airway including trachea than in the other tissues examined and support our previous result that HAT is localized in cells of submucosal serous glands of trachea and bronchi.
Although the native HAT was found in the sputum of patients with chronic airway diseases, HAT mRNA is thought to be expressed in the normal tissues of healthy subjects, because the trachea poly(A) ϩ RNA subjected to the Northern blot was obtained from the normal trachea tissues of three white male subjects who died of trauma or acute heart failure. It will be useful to compare expression levels of mRNA and protein of HAT in the patients with airway diseases with those in healthy subjects to clarify the physiological and pathophysiological sig-nificance of HAT in the airway. In the airway, various kinds of proteins such as lysozyme (56), secretory IgA (57), and secretory leukocyte protease inhibitor (58) are secreted from the submucosal serous glands onto mucous membrane and become constituents of airway mucous or bronchial secretions (59). These proteins play important roles in the host defense system of airways together with respiratory mucous glycoproteins, which are secreted from mucous glands cells and goblet cells (59). HAT may be released from the serous glands with these proteins and play some biological role in the host defense system on the mucous membrane independently of or in cooperation with other substances in airway mucous or bronchial secretions.
In summary, it was confirmed through the present work that HAT is a novel trypsin-like serine protease by analyzing the primary structure of the polypeptide deduced from the nucleotide sequence of its cDNA. However, the mechanism of activation of the HAT precursor to mature enzyme, the physiological role of the enzyme, and biological significance of the noncatalytic region of the precursor remain to be resolved.