INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

LUNG SURFACTANT, a complex of specific apoproteins and lipids that is essential for maintaining alveolar stability, exists in the alveoli in several structural isoforms that are in sequential relation to each other, namely, lamellar bodies (LBs), tubular myelin (TM), a surface film, and a small vesicular (SV) form (19). LBs are synthesized in type II alveolar cells and secreted into the alveolar space, where they evolve into the other forms in sequence. LBs and TM are highly surface-active forms, whereas the SV form is poorly surface active and is, in part, destined for reuptake by type II alveolar cells and recycling (20). The mechanisms of conversion of surfactant from one structural subtype to the next are not well understood. However, one step, the conversion of TM to the SV form, requires the action of an enzyme that has a serine active site on the basis of its inhibition by diisopropyl fluorophosphate (DFP) (5). Gross and Schultz (6) identified a single DFP-binding protein in the lungs and alveolar washings of mice that was found to have this activity and therefore called it "surfactant convertase" (6).

This protein with a molecular mass of 72 kDa was recently purified from mouse alveolar lavage and shown to be capable of converting surfactant from the TM form to the SV form in vitro (10). From the purified protein, four overlapping partial amino acid sequences (68 residues) were obtained, and the protein was shown to have homology with a previously sequenced mouse liver microsomal carboxylesterase and was therefore a novel member of the carboxylesterase multigene family (EC 3.1.1.1). Carboxylesterases of this family contain a serine active site consensus sequence and a histidine consensus sequence analogous to those of serine proteases (3) and are believed to employ the same serine-active catalytic triad (4). They hydrolyze a variety of ester, thioester, and amide bonds in aromatic and aliphatic substrates including fatty acids in vitro (9). Those that are found in liver microsomes are believed to hydrolyze and inactivate toxins, but the precise in vivo physiological function of any member of this family is not known.

We report here the isolation and characterization of a cDNA that putatively encodes convertase of the mouse lung.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

All chemicals were obtained from either Sigma or Fisher Scientific unless otherwise stated. All experiments were performed on female 15- to 20-wk-old CF-1 mice (Charles River Laboratories, Wilmington, MA). Mice were killed by intraperitoneal injection of 10 mg of pentobarbital sodium and transection of the abdominal aorta.

Isolation of mouse convertase cDNA. Two PCR primers (5' primer of bp 41-62, 5'-GGGCTTCTCTTGCTGTTTGCCCA-3', and 3' primer complementary to bp 235-257, 5'-GCATTCTTCACGAAGCTCCAGGG-3') were synthesized with the use of a previously published mouse liver carboxylesterase cDNA sequence (13) to amplify mouse genomic DNA. The PCR-amplified 400-bp DNA fragment was radiolabeled and used to screen a mouse lung cDNA library (Stratagene, La Jolla, CA). Approximately 3 × 10³ plaques were screened. The filters carrying recombinant phages were prehybridized with 2× PIPES buffer, 50% deionized formamide, 0.5% (wt/vol) SDS, and sonicated salmon sperm DNA (500 µg/ml) at 65°C for 2 h. Hybridization was carried out at 65°C for 12 h with a ³²P-labeled random-primed probe with the same buffer. The filters were washed with 0.1× saline-sodium citrate (SSC) and 0.1% (wt/vol) SDS twice at room temperature and twice at 65°C for 30 min each. A total of four positive clones were isolated. The clone showing the largest size insert was selected for further characterization. Preparation of genomic DNA, manipulation of cDNA, subcloning, and other related techniques were carried out as described in Ref. 16. The sequence of this cDNA was determined for both strands by the dideoxy chain termination method with a Sequenase version 2.0 sequencing kit (Amersham Life Science, Cleveland, OH). DNA sequence analysis was performed with DNASIS software (Hitachi Software, San Bruno, CA).

Isolation of total RNA and 5'-primer extension. Total RNA from mouse lung, liver, and kidney was isolated by RNeasy midi columns (Qiagen, Chatsworth, CA) following the vendor's recommendations. For primer extension, a synthetic oligonucleotide complementary to the convertase cDNA sequence (bp 2-24, 5'-AGAACATGGAGCCCACATCCCGGG-3'; Fig. 1) was synthesized. It was end-labeled with ³²P with T4 polynucleotide kinase (16). The oligonucleotide was annealed to 5 µg of equivalent total RNA at 60°C for 10 min and mixed with the reaction solution [in mM: 50 KCl, 10 Tris · HCl, pH 8.3, 1.5 MgCl₂, and 1 3'-deoxynucleoside 5'-triphosphates (dNTPs)] and 100 U of avian myeloblastosis virus reverse transcriptase (Promega, Madison, WI) at 42°C for 2 h. The extension product was analyzed on a 5% sequencing gel along with a sequencing reaction of an M13 DNA fragment used as a size marker. The transcription start site was calculated by counting nucleotides from the end of the primer to the reaction products.

View larger version (89K): [in this window] [in a new window]   Fig. 1.   Nucleotide (top lines) and derived amino acid sequences (bottom lines) of mouse lung convertase cDNA. Open boxes, conserved cysteine residues; shaded boxes, conserved active-site serine (Ser) and histidine (His) residues; lowercase letters, 5'- and 3'-untranslated regions or endoplasmic reticulum targeting signal sequence; underlined box, polyadenylation signal sequence. Five potential N-glycosylation sites are underlined. Boldface amino acid sequences were determined by protein sequence. Convertase nucleotide sequence differences from liver carboxylesterase are shown with dot above respective nucleotides.

Northern blot analysis. A nylon membrane containing 2 µg of poly(A)⁺ RNA of heart, brain, spleen, lung, liver, skeletal muscle, kidney, and testis from mouse was purchased from Clontech (Palo Alto, CA). A ³²P end-labeled oligonucleotide complementary to the convertase cDNA (bp 2-24, 5'-AGAACATGGAGCCCACATCCCGGG-3'; Fig. 1) was used as a probe. With the use of ExpressHyb solution (Clontech), hybridization was carried out with the end-labeled oligonucleotide probe at 37°C for 12 h. The blot was washed with 2× SSC and 0.05% SDS twice at room temperature for 40 min each and with 0.1× SSC and 0.1% SDS once at room temperature for 30 min and once at 68°C for 30 min. The membrane was exposed to X-ray film at 70°C to detect the signal. Under these experimental conditions, the probe hybridized only to the convertase cDNA and not to the mouse liver carboxylesterase cDNA (data not shown).

Deglycosylation. Convertase was purified from alveolar lavage and labeled with [³H]DFP as previously reported (6, 10). Approximately 5 µg of the purified [³H]DFP-labeled convertase from mouse alveolar lavage was incubated with endoglycosidase F in 80 mM sodium citrate buffer (pH 5.5) at 37°C for 16 h. Treated and untreated proteins were handled identically, apart from inclusion of enzyme, and analyzed by 10% SDS-PAGE followed by autoradiography. All autoradiographic films were scanned with a Hewlett-Packard Scan Jet IIC scanner, labeled with Adobe PhotoShop 4.0, and printed on Kodak Ektatherm paper.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

cDNA sequence. Because the partial amino acid sequence obtained from the lung protein with convertase activity (10) was similar to the previously published amino acid sequence of a mouse liver carboxylesterase (13), we amplified with PCR a 400-bp DNA fragment from mouse genomic DNA with primers designed from the liver carboxylesterase sequence and used this fragment as a probe to screen a mouse lung cDNA library. The cDNA library screen yielded four positive clones in which DNA sequences were determined from their poly(A)⁺ tails and found to be identical to each other. They showed several substitutions by comparison with the mouse liver carboxylesterase cDNA. The DNA-derived amino acid sequence matched exactly with the peptide sequences obtained from the purified lung convertase protein (10). After restriction digest analysis of these cDNA clones, the clone with the largest insert (1.8 kb) was chosen to determine the complete DNA sequence. The DNA sequence and its predicted amino acid sequence are shown in Fig. 1.

The overall nucleotide sequence and the amino acid sequence of the putative convertase are 98% homologous with the mouse liver carboxylesterase cDNA and its derived amino acid sequences (13). The clone contains an open reading frame terminating with a stop codon followed by an untranslated region that includes a single polyadenylation signal consensus sequence AATAAA and a long stretch of the poly(A)⁺ tail. Upstream from the first in-frame ATG codon, a 7-nucleotide (nt)-long untranslated region is also present.

Transcription initiation. Primer-extension analysis was carried out to determine whether the 5'-untranslated region of the cDNA contained more than the already observed 7 bp and whether RNA initiation starts at the same site in each of the mouse organs (lung, kidney, and liver) where the mRNA is expressed (see Tissue-specific expression). With the use of an end-labeled 24-nt-long synthetic primer, an extension reaction was carried out with total RNA prepared from each organ, and the product was analyzed on a DNA sequencing gel. The result for lung RNA (Fig. 2) shows only one extension product. On the basis of the size of the extended product, we determined that transcription starts 29 bp upstream of the presumed ATG start codon and that our cDNA clone lacks only 22 bp. The transcription initiation site for all three different organ RNAs used in this assay (lung, liver, and kidney) was identical (data for liver and kidney not shown).

View larger version (59K): [in this window] [in a new window]   Fig. 2.   Transcription initiation site analysis of convertase mRNA. 5' Primer-extension reaction with mouse lung total RNA was carried out with a 24-nucleotide (nt)-long synthetic primer complementary to mRNA (details in MATERIALS AND METHODS). Extended product was separated on a sequencing gel, along with a sequencing ladder of M13 DNA as a size marker, and is depicted as adenine (A), cytosine (C), guanine (G), and thymine (T). Arrow, primer-extended product of a single band.

Tissue-specific expression. Northern blot analysis was carried out with poly(A)⁺ RNA isolated from mouse heart, brain, spleen, lung, liver, skeletal muscle, kidney, and testis using an end-labeled primer derived from the 5'-end of cDNA where the lung sequence shows very poor homology with the previously published liver cDNA sequence (13). Consistent with the cDNA size, the probe hybridized to a 1,800-nt-long transcript from lung, kidney, and liver and not from any other tissues that we tested (Fig. 3). Expression of the message was apparently greater in liver than in lung or kidney.

View larger version (45K): [in this window] [in a new window]   Fig. 3.   Expression of mRNA encoding putative convertase in various mouse tissues analyzed by Northern blotting as described in MATERIALS AND METHODS. Blots containing 2 µg of poly(A)+ RNA isolated from different tissues were hybridized with a 32P end-labeled 24-mer oligonucleotide specific for convertase. Probe hybridized to transcripts of expected size (arrow) in mouse lung, liver, and kidney.

Convertase deglycosylation. Analysis of the cDNA sequence of convertase (Fig. 1) suggests five potential N-glycosylation sites. To verify the extent of glycosylation of the mature protein, we subjected the purified [³H]DFP-labeled protein to deglycosylation and compared the relative molecular mass of the product with that of the untreated protein using SDS-PAGE and autoradiography. We observed a reduction in molecular mass of 14 kDa (Fig. 4) after deglycosylation of convertase protein, consistent with glycosylation of the mature protein at all five potential sites.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Changes in electrophoretic mobility of purified 72-kDa convertase protein after deglycosylation with endoglycosidase F. Aliquots of 5 µg of [3H]diisopropyl fluorophosphate-labeled denatured convertase were incubated 16 h in absence (middle lane) or presence (right lane) of endoglycosidase F and subjected to SDS-PAGE with 14C-labeled protein standards (left lane) followed by autoradiography. Mass of deglycosylation products corresponds closely to prediction (59.14 kDa) from deduced amino acid sequence.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

As shown in Fig. 1, the nucleotide sequence of the cDNA encoding a putative convertase is unique and matches exactly to the previously reported 68-residue amino acid sequence obtained from the mouse lung DFP-binding protein with convertase activity (GenBank accession no. AF034435). Features of this cDNA sequence and the derived amino acid sequence include the presence of a highly hydrophobic 18-residue endoplasmic reticulum-targeting signal peptide at the NH₂ terminus that ends with a small residue (Gly), which is characteristic of several secretory proteins (18); serine and histidine consensus sequences around Ser²²¹ and His⁴⁵⁵ (14); three highly conserved cysteine residues; five potential N-glycosylation sites (13); and the COOH-terminal secretory His-Thr-Glu-His-Lys (HTEHK) sequence (Fig. 1) (1, 17). The presence of a putative serine-active catalytic triad is consistent with the ability of convertase to bind DFP (2). The change in relative molecular mass of the mature protein after deglycosylation with endoglycosidase F supports its extensive glycosylation (Fig. 4), like that of rat serum carboxylesterase (21). The exact role of glycosylation in the transportation, activity, and stability of convertase has yet to be elucidated.

Although the amino acid sequence derived from the cDNA sequence shown in Fig. 1 corresponds exactly to the partial amino acid sequence obtained from the putative convertase obtained from alveolar lavage, it will be necessary to express it and assay the product for convertase activity in vitro, which has not yet been accomplished.

The nucleotide sequence of the putative convertase (Fig. 1) is highly homologous with the previously published mouse liver carboxylesterase cDNA (13). Of the 21 nucleotide substitutions in the coding region (Fig. 1) between lung convertase and mouse liver carboxylesterase, 10 are in the third position, resulting in no amino acid changes; 5 are in the second position, resulting in 5 amino acid changes; and 6 are in the first position, resulting in 4 amino acid changes (Fig. 1). Mouse lung convertase differs from the liver carboxylesterase by nine amino acids (Fig. 5). Both proteins have the COOH-terminal His-Thr-Glu-His-Lys secretory sequence.

View larger version (59K): [in this window] [in a new window]   Fig. 5.   Amino acid sequence comparison of carboxylesterases. Deduced amino acid sequence of putative mouse lung (mLung) surfactant convertase carboxylesterase is aligned with carboxylesterases of mouse liver (mLiv) (13), rat liver (rLiv) (17), rat serum (rSer) (1), and human macrophage (hMac) (12). Dots, amino acids matching mLung; dashes, missing amino acids; boldface, 9 amino acid changes in putative convertase protein compared with mouse liver carboxylesterase; open box, COOH-terminal amino acid sequence His-Thr-Glu-His-Thr (HTEHT) involved in protein secretion; shaded box, nonsecretory sequence His-Xaa-Glu-Leu (HXEL).

The 5'-untranslated region of convertase cDNA has several nucleotide substitutions by comparison with the liver carboxylesterase cDNA (Fig. 1). This difference was exploited in the design of oligonucleotides for primer-extension analysis and for the determination of differential organ-specific expression. Northern blot analysis of several mouse organ RNAs shows expression of convertase or related transcripts in lung, liver, and kidney but not in heart, brain, spleen, skeletal muscle, or testis (Fig. 3). A greater magnitude of expression in liver than in lung was unexpected. This raises the possibility that some of the protein found in lung alveoli might be derived from the liver and delivered to the pulmonary alveoli via serum as is the case for other proteins such as albumin and ₁-antitrypsin, both of which are of similar molecular weight to convertase.

Members of the carboxylesterase family are expressed in many mammalian tissues, particularly liver (8, 11). The NH₂-terminal amino acid sequence of a protein that was purified from rat lung lavage by methods similar to those described in this study and that is homologous to a rat serum carboxylesterase has recently been reported (2). Most carboxylesterases are retained intracellularly, having a unique COOH-terminal (His-Xaa-Glu-Leu) endoplasmic reticulum retention sequence (15). Those carboxylesterases present in liver microsomes with this COOH-terminal retention sequence are believed to hydrolyze and inactivate toxins. A few members lack this retention sequence and instead have a secretory-type (HTEHK) COOH-terminal sequence and are secreted (1, 15, 21). Except for the last amino acid (Fig. 5), the COOH terminus of the putative convertase sequence (HTEHK) is very similar to the secretory form of rat serum carboxylesterase (1). We presume that convertase also belongs to this category of secretory carboxylesterases.

A question that still remains is whether the cDNA described in this study that putatively encodes convertase represents a unique gene or a strain difference. The putative convertase cDNA sequence is highly homologous with a previously described mouse liver carboxylesterase (13) and has a similar expression pattern (Fig. 3). The 5'-untranslated region of the putative convertase cDNA (Fig. 1) shows little homology with known carboxylesterase sequences, suggesting a unique gene. Comparison of the cDNA sequences of liver, lung, and kidney was required to determine whether the convertase transcript present in these three tissues is identical or only closely related. The functional relevance for the presence of the carboxylesterase transcript in liver and kidney is presently unknown.

Although the functions of previously described members of this family are unknown, they hydrolyze a variety of ester substrates in vitro and have been assumed to be involved in detoxification (9). If, as we suggest, the cDNA we report here corresponds to surfactant convertase, it will be the first member of the family of carboxylesterases with a defined physiological action. We presume that the substrate of convertase is one or more of the phospholipids contained in the mature surfactant film or its extruded fragments. Further experiments are required to define its molecular action on the basis of its identity as a member of the carboxylesterase superfamily.