Identification of a putative surfactant convertase in rat lung as a secreted serine carboxylesterase

Frederick Barr, Howard Clark, Samuel Hawgood


In the alveolar lumen, pulmonary surfactant converts from the contents of secreted lamellar bodies to tubular myelin to apoprotein-depleted vesicles during respiration. Using an in vitro system, researchers have reported that the conversion of tubular myelin to vesicles is blocked by inhibitors of serine hydrolase activity and have tentatively ascribed “convertase” activity to a diisopropyl fluorophosphate (DFP)-binding protein in mouse bronchoalveolar lavage (BAL). We purified and sequenced the homologous enzyme from rat BAL fluid. Amino acid sequence from the amino terminus and an internal cyanogen bromide peptide of the purified rat DFP-binding protein perfectly match the sequence of the carboxylesterase ES-2. Although ES-2 was initially cloned from liver, we found a 1.8-kilobase mRNA for ES-2 in decreasing relative abundance in rat liver, kidney, and lung but not in heart or spleen. Although further studies are required to establish the identity between “convertase” and ES-2 or a homologous member of the carboxylesterase family, our results raise the possibility that a protein with esterase/lipase activity plays a role in extracellular surfactant metabolism.

  • diisopropyl fluorophosphate-binding protein
  • in vitro cycling
  • subfractions
  • extracellular surfactant metabolism

pulmonary surfactant, a lipoprotein secretion of the alveolar epithelial type II cell, is important for maintaining the volume stability of the lung (6, 27). Surfactant is synthesized in type II cells and stored in specialized secretory granules, the lamellar bodies (LBs). LBs are secreted into the alveolar liquid lining layer by regulated exocytosis. Once free of the LB limiting membrane, the contents of LBs rapidly expand and coalesce into the ordered tubular aggregate known as tubular myelin (TM) (36). This apoprotein-rich structure is thought to be the precursor of a phospholipid-rich alveolar surface layer (SL) responsible for lowering the surface tension of the lung air-fluid interface (20). Apoprotein-depleted vesicular structures (VSs) are also present in the alveolar liquid lining layer (21). These vesicular forms are heterogeneous in structure, density, and composition (3, 20, 21) and may represent both products and remnants of the disassembly of LBs, TM, or the SL. These findings have led to a working hypothesis that these varied forms of extracellular surfactant are linked in a metabolic cycle that in its simplest form consists of LB contents converting sequentially to TM to SL to VSs (20). As recently suggested (2), there may not be complete mass transfer of all components between these four subfractions, with the subsequent generation of additional remnant particles not yet characterized. The kinetics of such a metabolic cycle and the pool sizes of individual subfractions would presumably be controlled by the demands of regular cyclic expansion and contraction of the alveolar surface during respiration, but the precise factors regulating conversion of one form of surfactant to the next are largely unknown.

Recently, a series of provocative experiments by Gross and Schultz (12) led to the conclusion that a serine protease may participate and, in fact, be required for the generation of the apoprotein-poor VSs in the mouse. Using an in vitro cycling system to model cyclic surface area changes occurring during respiration in the lung, they reported that serine protease inhibitors, including the nonspecific, irreversible serine hydrolase inhibitor diisopropyl fluorophosphate (DFP) but not several other classes of protease inhibitors, inhibit the conversion of dense forms of surfactant to lighter vesicular forms lacking significant surface activity (11). Their observations have now been repeated in other species using other serine hydrolase inhibitors and in a variety of experimental settings (34, 35). An intriguing finding in the studies of Gross and Schultz (11, 12) was the strict dependence of structural conversion on cyclic surface area changes, leading to the suggestion that some form of interfacial activation of either the putative “convertase enzyme” or its substrate was taking place. Gross and Schultz (12) utilized the irreversible binding of [3H]DFP to the active site of most serine hydrolases to tentatively identify a 75-kDa protein in mouse bronchoalveolar lavage (BAL) as the enzyme involved in TM-to-VS conversion (11).

The evidence supporting the participation of this 75-kDa DFP-binding protein in TM-to-VS conversion is at present circumstantial:1) DFP inhibits surfactant conversion in vitro, 2) the 75-kDa protein is the major DFP-binding protein found in mouse BAL,3) a 75-kDa DFP-binding protein is present in both heavy precursor forms of extracellular surfactant and adenocarcinoma cell lines of presumed type II cell origin (11), and4) a BAL fraction enriched in the 75-kDa DFP-binding protein promotes surfactant conversion in vitro (10). Hall et al. (13) have recently challenged the “convertase” hypothesis by suggesting that the hydrolase inhibitors act primarily as nonspecific inhibitors of surface activity rather than by specific enzyme inhibition. To help define the role of an enzyme in extracellular surfactant metabolism, we isolated and sequenced a homologous candidate convertase from rat BAL. In this report, we show that the major DFP-binding protein in rat BAL is a member of the serine carboxylesterase family. This work is consistent with the recent report that the mouse BAL DFP-binding protein is also a member of the carboxylesterase family (16).


Identification of DFP-binding proteins in rat BAL.

Adult male Sprague-Dawley rats were anesthetized with intraperitoneal pentobarbital sodium and exsanguinated by transecting the abdominal aorta. A tracheotomy was performed using a 16-gauge blunt needle, and the lungs were perfused through the pulmonary artery to the left atrium with phosphate-buffered saline containing 2 mM CaCl2. During perfusion, the lungs were cyclically expanded with air from roughly functional residual capacity to total lung capacity. When the lungs were bloodless, four sequential lavages were performed with 8 ml of lavage buffer consisting of 10 mM tris(hydroxymethyl)aminomethane (Tris), 140 mM NaCl, 2 mM CaCl2, and 2 mM MgCl2, pH 7.4. The pooled lavages were collected on ice and centrifuged at 150 g for 20 min to remove cells. The cell-free supernatant was centrifuged at 1 × 105 g for 2 h at 4°C. The supernatant was used for isolation of DFP-binding proteins.

To identify DFP-binding proteins, 40 ml of BAL supernatant were incubated with 5 μCi of [3H]DFP (DuPont NEN, Wilmington, DE) at 37°C for 30 min. BAL proteins were concentrated by trichloroacetic acid precipitation using sodium deoxycholate as a carrier and separated by one-dimensional (1-D) or two-dimensional (2-D) polyacrylamide gel electrophoresis (PAGE) using standard techniques (17). The gels were stained with silver, soaked in Enhance (DuPont NEN), dried, and placed against film at −70°C.

Isolation of a 68-kDa DFP-binding protein from rat BAL and rat serum.

Labeled rat BAL supernatant or rat serum was passed over a concanavalin A (ConA)-Sepharose affinity column (Sigma, St. Louis, MO) at room temperature. The column was washed with lavage buffer and eluted with lavage buffer containing 50 mM α-methyl mannopyranoside (Sigma). Peak eluate fractions were identified by reading optical density at 280 nm and by scintillation counting. The pooled eluate was dialyzed overnight against 20 mM Tris, pH 7.4, at 4°C to reduce the NaCl concentration before the proteins were fractionated further on an HR5/5 Mono Q anion-exchange column (Pharmacia Biotech, Piscataway, NJ) eluted with a continuous gradient of NaCl from 0 to 1 M over 60 min using a flow rate of 1 ml/min. This and all subsequent high-performance liquid chromatography (HPLC) were performed on a 1080M HPLC system (Hewlett-Packard, Mountain View, CA). Fractions containing DFP-binding protein were identified by scintillation counting and made to 1.6 M NH4SO4. This sample was loaded on an HR5/5 phenyl Sepharose column (Pharmacia Biotech) and eluted at 0.5 ml/min with a continuous reverse gradient of 1.7–0 M NH4SO4. The fractions containing [3H]DFP-binding protein were analyzed by 1-D and 2-D PAGE and by autoradiography.

Amino acid sequencing of the DFP-binding protein from rat BAL.

For amino acid sequencing, the proteins in the fractions containing [3H]DFP-binding protein from the phenyl Sepharose column were separated by 2-D PAGE using 10% polyacrylamide gels in the second dimension. The proteins were transferred to Problot membrane (Applied Biosystems, Foster City, CA) by electroblotting at 250 mA for 60 min. The membrane was lightly stained at 4°C with Coomassie blue for 20 min, destained in methanol, and placed on film to definitively localize the [3H]DFP-binding protein. The amino-terminal sequence was determined directly from excised Problot membrane by gas-phase microsequencing (Applied Biosystems). To obtain internal protein sequence, the 2-D PAGE gel was lightly stained in Coomassie blue in methanol at 4°C for 20 min, and the appropriate spot (68 kDa, pI 6.0) was excised. Cyanogen bromide (CNBr) digestion of the protein in the gel was carried out according to the method described by Nikodem and Fresco (25). Briefly, the gel fragment was soaked in CNBr and HCl for 2 h at room temperature. After being washed in 0.125 M Tris ⋅ HCl, pH 6.8, and 0.1% sodium dodecyl sulfate (SDS) to remove residual CNBr and HCl, the gel fragment was loaded into the well of a 15% SDS-PAGE gel and overlaid with standard sample buffer, and electrophoresis was performed to resolve peptide fragments. The transfer to Immobilon, staining, and autoradiography were as described above for amino-terminal sequencing. The most abundant labeled band at 16 kDa was selected for gas-phase microsequencing (see Fig. 4).

Antibody production and immunoaffinity chromatography.

A peptide with the sequence CNPPQTEHTEHT, corresponding to the last 11 amino acids of ES-2 (18, 32) and an amino cysteine for purposes of conjugation, was synthesized by standard techniques. This peptide was conjugated to keyhole limpet hemocyanin and used to raise a polyclonal antibody in rabbits using standard immunization protocols. An ES-2 affinity column was prepared by purifying specific anti-ES-2 immunoglobulin G (IgG) from the total IgG fraction using Sepharose 6B (Pierce, Rockford, IL) coupled with the ES-2 peptide and then cross-linking the specific IgG to protein A bound to Sepharose (Pierce) using dimethylpimelimidate (Pierce). Carboxylesterase in rat BAL or serum partially purified by ConA and anion-exchange chromatography was incubated with antibody-coupled beads overnight at 4°C and eluted with 2 M glycine, pH 2. Eluate fractions (1 ml) were collected into tubes containing 100 μl of 1 M Tris base to immediately neutralize the pH.

Esterase assay.

The nonspecific carboxylesterase assay described by Heymann et al. (15) utilizing p-nitrophenyl acetate as a substrate was modified slightly for the measurement of activity after purification of carboxylesterases from the immunoaffinity column. Briefly, a fresh solution of 0.5 mM nitrophenyl acetate in acetone was prepared immediately before use. With the use of temperature-controlled quartz cuvettes in a Gilford response spectrophotometer, the baseline absorbance at 405 nm was established at 30°C, pH 7.4. Purified esterase quantified using the Lowry assay (19) was added at time zero, the cuvette contents were rapidly mixed, and absorbance was recorded every 15 s for 3 min. Esterase activity was calculated from the slope of the absorbance curve using a molar absorbance coefficient of 12,000 l ⋅ mol−1 ⋅ cm−1and was expressed as micromoles per minute per microgram.

Polymerase chain reaction amplification of rat ES-2 from liver and lung RNA.

Total RNA was isolated from adult rat tissues by a slightly modified guanidine isothiocyanate method (5). Briefly, organs were removed from a −70°C environment, crushed with a pestle on dry ice, homogenized directly into homogenizing buffer (4.25 M guanidine isothiocyanate, 0.5% sodium lauroyl sakozyl, 5 mM sodium citrate, pH 7, and 0.1 M β-mercaptoethanol) on ice, and centrifuged in the presence of cesium chloride at 30,000 rpm in a SW-41 rotor (Beckman Instruments, Palo Alto, CA) overnight. The RNA pellet was resuspended in diethyl pyrocarbonate-treated water and quantified by absorbance at 260 nm and on analytic formaldehyde gels. A 454-base pair (bp) fragment of rat ES-2 cDNA, corresponding to bases 78–532 of the published esterase cDNA clone 2-1 (32), was generated from total RNA from adult rat lung and liver by reverse transcription-polymerase chain reaction (RT-PCR). With the use of the Ampliwax PCR Gem-mediated hot-start technique (Perkin-Elmer Cetus, Norwalk, CT), 2–3 μg of total RNA from lung and liver were incubated with avian myeloblastoma virus reverse transcriptase (Gibco BRL, Life Technologies, Gaithersburg, MD) at 42°C for 30 min in the presence of antisense primer TAGTCCACCTCCATGGATCC, complementary to nucleotides 512–532 of the rat ES-2 cDNA. This was followed by the addition of upstream sense primer AAGAGCTCTTGTTCTTCCGC from positions 78–98 of rat ES-2 cDNA and subsequent amplification by PCR consisting of 24 cycles of amplification at an annealing temperature of 60°C for 30 s, extension at 72°C for 1 min, and denaturation at 94°C for 1 min. Products were analyzed on a 1.0% agarose gel. PCR products from both lung and liver were subcloned into pCRTMII vector using the TA cloning kit (Invitrogen, San Diego, CA), and their identity was confirmed by automated DNA sequencing of both strands (Biomolecular Resource Center, Univ. of California, San Francisco, CA).

Northern hybridization.

For Northern blots, the 454-bp cDNA probe for ES-2 obtained by RT-PCR of rat lung RNA was labeled with [32P]dCTP (DuPont NEN) by random priming (Gibco BRL, Life Technologies) and purified from unincorporated nucleotide over a Nuctrap push column (Stratagene, La Jolla, CA). Aliquots of total RNA were resolved by electrophoresis on formaldehyde-agarose gels, transferred overnight to a MagnaCharge nylon membrane (Micron Separations, Westboro, MA), and prehybridized in PreHyb (Stratagene) at 42°C for 2 h. Membranes were hybridized with 1–2 × 106 counts/min of probe per milliliter of hybridization solution for 4 h at 60°C, washed in 1× SSC (1× SSC is 0.15 M NaCl and 0.015 M sodium citrate, pH 7.0)-0.1% SDS for 20 min at room temperature, and washed three more times in 0.1× SSC-0.1% SDS for 20 min at 68°C. Autoradiography was performed at −70°C. The quality and quantity of the extracted RNA were confirmed on analytic gels by ethidium bromide staining and ultraviolet absorbance of the resolved ribosomal subunits.


DFP-binding proteins in rat BAL.

After labeling the high-speed supernatant of rat BAL with [3H]DFP, we detected a single DFP-binding protein in rat BAL by 2-D PAGE and autoradiography (Fig. 1). The DFP-binding protein had a molecular mass of 68 kDa and a pI of 6.0 and was quantitatively bound to the lectin ConA and eluted with the hapten sugar α-methyl mannopyranoside (Fig.2 A). Further purification was achieved on sequential nondenaturing anion-exchange and hydrophobic-interaction columns (Fig. 2,B andC). The product off the final column was markedly enriched in DFP-binding protein (Fig.2 D), with a 150-fold increase in specific activity relative to the ConA eluate. Heavy loading of the final fraction onto silver-stained 2-D gels still showed a number of other proteins, some with molecular masses of 60–70 kDa, similar to the molecular mass of the DFP-binding protein (68 kDa) (not shown). Final separation from these proteins for the purpose of amino acid sequencing was achieved by 2-D PAGE.

Fig. 1.

Two-dimensional SDS-polyacrylamide gel electrophoresis (PAGE) of [3H]diisopropyl fluorophosphate (DFP)-labeled rat bronchoalveolar lavage (BAL) proteins. A: silver-stained gel.B: autoradiogram of same gel as shown in A. Arrow, 68-kDa (pI 6.0) DFP-binding protein in rat alveolar lavage.

Fig. 2.

Steps in purification of 68-kDa (pI 6.0) DFP-binding protein from rat alveolar lavage. A: results of concanavalin A (ConA) affinity chromatography of [3H]DFP-labeled BAL proteins. Autoradiogram of a 15% SDS-PAGE gel shows column load consisting of total rat BAL proteins in lane 1, flow through in lane 2, column wash in lane 3, and fractions eluted with 50 mM α-methyl mannopyranoside in lanes 4 and5. B: chromatogram of Mono Q purification of pooled ConA elute fractions. Chromatography conditions are given inmethods. Absorbance profile was recorded at an optical density of 280 nm. Fractions containing3H are noted by bars.C: chromatogram of phenyl Sepharose purification of pooled Mono Q fractions. Chromatography conditions are given in methods. Absorbance profile was recorded at an optical density of 280 nm. Fractions containing3H are noted by bars.D: silver-stained SDS-PAGE gel of sequential steps in purification of [3H]DFP-binding protein from rat BAL. Lane 1, total BAL proteins;lane 2, ConA elute proteins;lane 3, Mono Q fractions;lane 4, phenyl Sepharose fractions (additional proteins were detected in this fraction with heavier loading). Nos. at left indicate molecular mass.

Amino acid sequence of 68-kDa DFP-binding protein from rat BAL.

There was a low repetitive yield from amino acid sequencing of the intact 68-kDa protein from the polyvinylidene difluoride membrane, suggesting that the amino terminus was partially blocked. The partial sequence obtained was XPXXPPVVDTTKG, where X is indeterminate. This sequence matched the amino-terminal sequence of the translated product of a rat liver cDNA previously characterized by two studies (18, 32) independently attempting to clone the major intracellular carboxylesterase in rat liver (Fig. 3). Direct protein sequence matching this cDNA clone has not been previously reported. To further assess the relationship between the BAL DFP-binding protein and this carboxylesterase, we cleaved the 68-kDa protein with CNBr. Several digestion products were generated (Fig.4). A [3H]DFP-labeled product of 16 kDa gave the unambiguous sequence of SQDGVVGKLLADML, an identical match to the same carboxylesterase identified by the amino-terminal sequence (Fig. 3). Although a standard nomenclature for the carboxylesterases has not been established in the literature, we will use the system introduced by Robbi and colleagues (28-31), in which the carboxylesterase homologous to these sequences is called ES-2. ES-2 is identical to clone 2-1 of carboxylesterase E1 (32), rat liver carboxylesterase (18), carboxyesterase Sec (24), and carboxylesterase S (39).

Fig. 3.

Amino acid sequence of rat carboxylesterase ES-2 (18, 32). Underlined sequences correspond to 2 peptides of [3H]DFP-binding protein sequenced. Boxed sequence at carboxy terminus represents sequence used to raise specific anti-ES-2 antibody. * Putative active-site serine.

Fig. 4.

Autoradiogram of peptides generated by cyanogen bromide digestion of gel-purified [3H]DFP-binding protein. Arrow, 16-kDa peptide sequenced to give sequence aligned to residues 70–83 in Fig. 3. Nos. atleft indicate molecular mass.

Further evidence for a close relationship between the DFP-binding protein in rat BAL and ES-2 was obtained using an antibody previously reported to be specific for ES-2 (24). We produced an antibody to the 11 most carboxy-terminal amino acids. This peptide distinguishes ES-2 from all intracellular esterases (29). Western blot analysis of serum and BAL detected a protein of 68 kDa, identical in size to the DFP-binding protein. Although higher molecular-mass aggregates of ES-2 were also detected on overloaded blots (Fig.5), the 68-kDa DFP-binding protein eluted as a monomer from a Sephadex-75 size-exclusion column (Pharmacia Biotech) in Tris-buffered saline, suggesting that these were nonbiological aggregates. Furthermore, the antibody quantitatively immunoprecipitated the [3H]DFP-labeled product of rat BAL, suggesting that the carboxy terminus of the rat DFP-binding protein and ES-2 were closely related. Direct protein sequence and immunologic methods therefore suggest that the rat BAL DFP-binding protein and ES-2 are closely related. Because all rat carboxylesterases share closely conserved sequences, we cannot formally exclude the possibility that the BAL DFP-binding protein is another member of the carboxylesterase family, homologous but not identical to ES-2.

Fig. 5.

Anti-ES-2 immunoblot analysis of rat BAL and serum.Lane 1, 30 μg of rat serum;lane 2, 30 μg of rat BAL;lane 3, 30 μg of ConA column flow through; lane 4, 10 μg of ConA eluate. Blot was blocked with 5% milk, probed with anti-ES-2 antiserum at a 1:500 dilution, and developed with a horseradish peroxidase-linked goat anti-rabbit IgG at a 1:5,000 dilution using Amersham enhanced chemiluminescence system. No. at leftindicates molecular mass.

Identity between liver and lung cDNA sequences.

To determine whether mRNA for ES-2 was present in the lung and to further establish the relationship between the liver and lung carboxylesterases products, we performed RT-PCR of liver and lung mRNA using primers derived from the reported ES-2 cDNA sequence (18, 32). The primers were selected to specifically amplify ES-2 and not the most abundant intracellular lung carboxylesterase ES-10 (31), which is identical to liver hydrolase A (38). The expected 454-bp product was amplified from both liver and lung (Fig.6). Complete sequencing of these two PCR products confirmed their identity with each other and with the reported cDNA sequence of ES-2. The analysis was restricted to the 5′ half of the message so that variation in sequence between liver and lung in the 3′ region, as suggested by Krishnasamy et al. (16), cannot be excluded.

Fig. 6.

Reverse transcription-polymerase chain reaction (PCR) of nucleotides 78–532 of ES-2 from rat liver and lung RNA. Lane 1, PCR control plasmid; lane 2, liver; lane 3, lung. Arrow, expected 454-base pair (bp) product. St, standard DNA ladder.

Northern hybridization.

The probe generated by RT-PCR was used for Northern analysis of total RNA from adult rat liver, lung, kidney, and heart. Stringent hybridization and wash conditions were employed to minimize cross-hybridization with other homologous carboxylesterase sequences. As expected, the liver contained abundant message for ES-2 with the predicted size of 1.8 kilobases (kb). The kidney also contained ES-2 mRNA of the expected size but at a much lower abundance than the liver. Lung RNA hybridized weakly with the ES-2 probe, supporting the PCR result and suggesting a low level (<1% of the liver signal) of expression in the lung (Fig. 7). No signal was detected in the heart under these conditions.

Fig. 7.

Northern blot for ES-2 mRNA. Lane 1, lung (100 μg of total RNA); lane 2, spleen (30 μg); lane 3, heart (30 μg); lane 4, kidney (30 μg);lane 5, liver (30 μg). Arrow, expected mRNA of 1.8 kilobases (kb).

Esterase activity of the 68-kDa DFP-binding protein.

Nondenatured carboxylesterase was isolated from rat BAL or serum by sequential ConA, Mono Q, and immunoaffinity chromatography. The final preparation was markedly enriched for ES-2 by immunochemical and electrophoretic analysis (Fig. 8) but still contained low levels of other unidentified proteins. The greater abundance of immunoreactive protein from rat serum compared with BAL allowed us to characterize the enzyme further. The enriched fraction cleaved the nonspecific esterase substratep-nitrophenyl acetate with a specific activity of 125 μmol ⋅ min−1 ⋅ mg−1, within the range of 5–130 μmol ⋅ min−1 ⋅ mg−1reported for other esterases (15). The esterase activity was not inhibited by the anti-peptide antibody or the serine protease inhibitor human α-1-antitrypsin but was inhibited by heating the enzyme to 65°C.

Fig. 8.

Antibody affinity chromatography of ES-2 from rat serum.Top: silver stained.Bottom: immunoblotted with anti-ES-2 antiserum. Samples are same in each case. L, load of fraction containing ES-2 partially purified from rat serum by ConA and Mono Q chromatography (see methods); FT, flow-through fraction; W, column wash; E1–E3, eluate fractions. Nos. at left indicate molecular mass.


Similar to results obtained in the mouse (11, 12), we identified a single predominant DFP-binding protein in the soluble fraction of rat BAL. Characterization of this protein indicates that it is a member of the carboxylesterase family of serine hydrolases highly homologous to rat ES-2, first identified as the putative product of a cDNA cloned from rat liver (18, 32) and subsequently shown to be the major circulating carboxylesterase in rat serum (1, 24, 29). We do not yet have definitive proof that the rat BAL DFP-binding protein and ES-2 are identical, as distinct from closely related gene products. During the progress of our work in the rat, the murine BAL DFP-binding protein was reported also to be a carboxylesterase (16), homologous but not identical to the major mouse serum carboxylesterase Es-N (26). Partially purified mouse BAL carboxylesterase was active in an assay of surfactant conversion (16). The precise genetic and functional relationships between rat ES-2, mouse Es-N, and the rat and mouse BAL DFP-binding proteins remain to be determined.

The assignment of the BAL DFP-binding protein as a carboxylesterase is, at present, based entirely on sequence homology with other members of the large esterase-lipase superfamily (7, 8, 14, 33) and not yet on a determination of the natural substrate of the enzyme. In this study, we have shown that immunoaffinity-purified carboxylesterase has the nonspecific esterase activity predicted by the sequence-based assignment, but we cannot exclude the possibility that the BAL carboxylesterase has other hydrolase activity, including longer chain lipase-like or protease-like activity. Indeed, comparison of the several lipase and esterase crystal structures available suggests that it may be hazardous to assign substrate specificity on the basis of primary structure alone (8). The apparent dependence of convertase activity on the cyclic expansion and contraction of an interface makes the possibility of an interfacially activated lipase-like activity particularly intriguing.

Although previous reports suggest that the expression of secreted carboxylesterases is restricted to the liver in the rat (39), we show here that ES-2 or a closely related secreted carboxylesterase is also expressed at low levels in the lung. This result is consistent with the finding by Ovnic et al. (26) that Es-N, a secreted mouse carboxylesterase, is expressed at low levels in the mouse lung. The recent work by Krishnasamy et al. (16) suggests that more than one secreted carboxylesterase is present in the mouse. Complete clones of the rodent BAL DFP-binding proteins will be needed to establish the precise genetic relationships between ES-2, Es-N, and the rat and mouse BAL DFP-binding proteins. Even with complete clones, care will be needed in the analysis of the expression patterns of the carboxylesterases in both the rat and mouse because several highly homologous cDNAs have been identified in both species already (26, 37-39). Each carboxylesterase appears to have a relatively specific tissue distribution, but considerable overlap exists, and contradictory results are present in the literature for several of the rat and mouse carboxylesterases (23, 26, 28, 37, 38). Some of this variability is probably due to differences in the sensitivity of the techniques used to detect mRNA and some to problems of cross-hybridization.

To date, ES-10 is the most abundant carboxylesterase identified in rat lung (31, 38). We therefore focused on excluding cross-hybridization between our cDNA and ES-10. The cDNAs for ES-2 and ES-10 are 93% identical. To try to minimize the likelihood of cross-hybridization, we chose a probe corresponding to a region with lesser homology (87%) between the two sequences and used stringent hybridization and wash conditions. This ES-2 probe recognized a 1.8-kb mRNA in liver, lung, and kidney but not in heart or spleen. This tissue distribution is not consistent with that reported for ES-10, which is not found in kidney but is found in the heart (23). Greater specificity was designed into the RT-PCR experiments, and the results also support lung expression of ES-2 or a closely related mRNA. In these experiments, we were able to amplify ES-2 mRNA from lung and liver but did not amplify ES-10 mRNA despite its reported greater abundance in the lung and liver (31, 38). The upstream primer used was specific for the ES-2 sequence and did not hybridize to a purified plasmid containing ES-10 cDNA. ES-10 mRNA could be readily amplified from lung, liver, and the ES-10 plasmid by substituting an ES-10-specific upstream primer (not shown). Additionally, in recent experiments, we have localized a secreted carboxylesterase mRNA and protein to lung macrophages and, at lower abundance, to type II cells (H. W. Clark, E. Collins, L. Allen, and S. Hawgood, unpublished observations), a pattern quite distinct from the Clara cell localization reported for intracellular ES-10 (38). Further work is required to determine whether additional secreted carboxylesterases not recognized by our probes are expressed in the lung.

Muramaki et al. (24) have noted that the presence of a consensus sequence for retention in the endoplasmic reticulum at the extreme carboxy terminus of carboxylesterases can be used to distinguish secreted from intracellular forms of the enzyme. In contrast, antibodies raised against the whole protein are known to show considerable and unpredictable cross-reactivity between esterases (37-39). This is not surprising, given that the six cloned members of the rat carboxylesterase protein family share at least 68% amino acid sequence identity. Several studies have convincingly demonstrated that the carboxy-terminal region of the carboxylesterases can be used to generate antibodies that specifically recognize only secreted carboxylesterases (22, 24). We took advantage of this unique carboxy-terminal region to raise an antibody specific to secreted carboxylesterases and demonstrated that ES-2 or a closely related protein sharing this epitope is present in BAL. Importantly, this antibody immunopreciptated the 68-kDa [3H]DFP-binding protein from rat BAL fluid, strongly supporting the direct amino sequence data in identifying the rat BAL 68-kDa DFP-binding protein as closely related to ES-2. It remains possible that the BAL protein is highly homologous to ES-2 but represents a different gene product.

The same carboxy-terminal antibody functioned effectively to immunopurify a 68-kDa DFP-binding protein from rat BAL and serum. Despite preenrichment of the sample by lectin affinity chromatography and ion-exchange chromatography, at least three additional proteins consistently bound to the immunoaffinity column. At present, their identity is unknown. Because the DFP-binding protein eluted from a size-exclusion chromatography column as a monomer, we do not have any evidence for a direct interaction between DFP-binding protein and any other protein in our final purified fraction or for oligomerization of DFP-binding protein as reported for some other esterases (9). The immunoaffinity-purified fraction, greatly enriched in DFP-binding protein, had heat-inactivateable nonspecific esterase activity, but this activity was not blocked by the carboxy-terminal antibody. With the assumption that the structure of this carboxylesterase resembles other esterase-lipase structures (7), the carboxy-terminal antibody binding site would be spatially distant from the active site of the enzyme, potentially explaining this lack of inhibition.

At this time, we are confident in identifying the 68-kDa DFP-binding protein found in rat BAL as a secreted monomeric carboxylesterase capable of short-chain ester cleavage. The enzyme is at least closely related to ES-2 and is expressed, albeit at low levels, in the lung. The finding that the murine BAL esterase differs by at least three residues from the only reported serum carboxylesterase in the mouse emphasizes the need for caution in distinguishing between members of this large and homologous family (16). Possible substrates for ES-2 in surfactant include phospholipids, triglycerides, cholesterol esters, and possibly even the thioester linkage in surfactant protein C. Further studies will be necessary to determine the precise role of secreted carboxylesterases in extracellular surfactant metabolism.


We acknowledge the contributions of Dr. N. Gross for suggesting the use of ConA chromatography for initial protein purification and Lennell Allen for assistance with the figures and the advice and support of Dr. J. A. Clements throughout this work. F. Barr and H. Clark contributed equally to this work.


  • Address for reprint requests: S. Hawgood, Cardiovascular Research Institute, Univ. of California, San Francisco, CA 94143-0130.

  • This work was supported by National Heart, Lung, and Blood Institute Program Project Grant HL-24075 and training grants from the California Lung Association and Wyeth Pediatrics (to H. Clark).

  • Portions of this study have been published in abstract form (4).


View Abstract