Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Pulmonary surfactant is a mixture of lipids and specific proteins that is secreted by type II pneumocytes into the alveolar space. Its main function is the reduction of the surface tension at the air-liquid interface in the lung, thus preventing alveolar collapse at end-expiration. Lipids constitute ~ 90% of the pulmonary surfactant. The most abundant lipid in surfactant, dipalmitoylphosphatidylcholine, is also the major surface-active component. The surfactant proteins SP-A, SP-B, and SP-C play an essential role in the metabolism and dynamics of the lipids of pulmonary surfactant (reviewed in Ref. 1). SP-C is the only surfactant-associated protein that is pulmonary surfactant- specific and exclusively expressed in type II pneumocytes of the lung (2). Although SP-C-deficient (/) mice grew normally to adulthood, the decreased stability of captive bubbles indicated that SP-C plays a role in the stabilization of surfactant at low lung volumes (3). In newborn calves of the Belgian White and Blue breed and a human full-term baby, the lack or deficiency of SP-C were associated with a respiratory dysfunction and interstitial lung disease (4, 5). Furthermore, in newborns with a severe respiratory dysfunction due to hereditary SP-B deficiency the abnormal intra-alveolar surfactant is not only characterized by the lack of mature SP-B but also by the intra-alveolar accumulation of aberrantly processed SP-C and a lack of mature SP-C (6, 7).

The hydrophobic surfactant protein SP-C is transported from its site of synthesis at the ribosomes to the lamellar bodies via the endoplasmic reticulum, the Golgi system, and multivesicular bodies. On the route from its site of synthesis to the lamellar bodies, the primary translation product, a proprotein (proSP-C) with a molecular mass of 21 kD, undergoes extensive C- and N-terminal posttranslational processing in multivesicular bodies and lamellar bodies in type II pneumocytes (8). Although various proSP-C processing steps have been described for rat and fetal human lung (9, 11, 12), nothing is known about the sites of proSP-C processing in the adult human lung or the identity of the proteases involved. Due to the acidic pH in multivesicular and lamellar bodies, only a limited set of proteases can catalyze the N- and C-terminal processing.

Cathepsins are lysosomal proteases which are involved in proenzyme activation, antigen processing, and tissue remodeling (13). It has been shown in vitro that cathepsin D or a "cathepsin D-like protease" is involved in the post-translational processing of the hydrophobic surfactant protein B (14), but cathepsin D itself was not detectable in type II pneumocytes and no specific activity was found in isolated lamellar bodies (15, 16). In contrast, the lysosomal protease cathepsin H was localized in lamellar bodies of type II pneumocytes of the rat lung (17).

Cathepsin H (EC 3.4.22.16) is a cysteine protease that belongs to the papain superfamily, which includes plant cysteine proteases and, for example, the mammalian lysosomal proteases cathepsin B, cathepsin K, cathepsin L, and cathepsin S (18). All cysteine proteases contain a cysteine residue in their active site but differ in their distribution, substrate specificity, and other enzymatic properties. From its proteolytic activity, cathepsin H was characterized as an "endoaminopeptidase," which could behave as an aminopeptidase as well as an endopeptidase (18, 19). Various proteins (for example, insulin B-chain, vimentin, and tubulin) are hydrolyzed by cathepsin H using its endoproteolytic activity (18). The aminopeptidase activity makes it unique among the cysteine proteases characterized so far.

To investigate the post-translational processing of proSP-C and the potential role of cathepsin H, we studied the localization of proSP-C and cathepsin H in human as well as proSP-C in rat lungs by immunohistochemistry and immunoelectron microscopy, measured cathepsin H activity in isolated rat lamellar bodies, and examined the in vitro processing of recombinant human proSP-C by purified cathepsin H.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Human Lungs

For the present study, we used eight nontransplanted human single-donor lungs. Left and right donor lungs were separated shortly before transplantation. While one donor lung was used for transplantation, the contralateral donor lung was fixed at the time of transplantation as soon as the clinical procedure allowed. Donor lungs were used for investigation only if they could not be made available for another suitable recipient by The Eurotransplant Foundation Center (Leiden, The Netherlands) or if they were excluded from transplantation by the explanting surgeon for clinical reasons. Fixation was performed by instillation of the fixative via the airways to ensure rapid and uniform fixation as described in detail below.

Human Bronchoalveolar Lavage

Bronchoalveolar lavage (BAL) fluid from three patients with no pathologic findings was used for the detection of cathepsin H.

Rat Lungs

Eight male Wistar rats were obtained from the Zentralinstitut für Versuchstierforschung in Hannover, Germany.

Purified Cathepsin H

Purified cathepsin H was kindly provided by E. Weber (Institute of Physiological Chemistry, Halle, Germany) (18).

Antisera

Rabbit antisera specific for cathepsins B, H, L, and S and were kindly provided by H. Kirschke and E. Weber (Institute of Physiological Chemistry, Halle, Germany) (20, 21). Antiserum against cathepsin K was kindly given by D. Brömme (Mount Sinai Medical School, NY) (22). A monoclonal antibody (anti-Xpress; Invitrogen, Carlsbad, CA) was used for the detection of the His-tag of recombinant HisproSP-C. The polyclonal antibodies against the epitopes D²-L⁹ (anti-NPROSP-C^2-9), E¹¹-R²³ (anti-NPROSP-C^11-23), and G¹⁶²-G¹⁷⁴ (anti-CPROSP-C^162-174) had been characterized before (10, 11). An antiserum against mature SP-C was kindly provided by W. Steinhilber (23) (Byk Gulden, Konstanz, Germany).

Immunohistochemistry: Tissue Preparation and Immunostaining

For immunohistochemistry, two human nontransplanted single donor lungs were fixed by instillation of 4% buffered formaldehyde. Several samples from different sites were taken and subsequently routine paraffin embedding was performed. Immunostaining was performed using the alkaline-phosphatase method. Sections (4 µm thick) were mounted on poly-L-lysine capillary slides and dried overnight at 37°C. Paraffin sections were dewaxed with xylene, rehydrated in a graded series of alcohol, and finally washed in Tris-HCl (pH 7.6) for 10 min. The following steps were performed at room temperature in an automated staining system (TechMate 500; Dako, Glostup, Denmark). To avoid unspecific staining, sections were blocked with buffer 1 (Dako) for 5 min before incubation with the primary antibody at the appropriate dilution in blocking buffer (Zytomed, Berlin, Germany) for 30 min at room temperature. After several rinses in buffer (Buffer Kit; Dako), the immunoreaction was demonstrated using the APAAP Kit (Dako) according to the specifications of the manufacturer. Fast Red (Dako) was used as alkaline-phosphatase substrate. Finally sections were rinsed in distilled water and counterstained with Mayer's haematoxylin (Dako).

Immunoelectron Microscopy (Immuno-EM): Tissue Preparation and Immunogold Labeling

For immuno-EM, six human nontransplanted donor lungs and four rat lungs were prepared as described in detail recently (24). Briefly, the lungs were fixed by instillation of a mixture of 4% paraformaldehyde and 0.1% glutaraldehyde in 0.2 M HEPES buffer (pH 7.4) into the alveoli via the airways. Sampling of tissue blocks was performed according to the rules of the systematic uniform random sampling by superimposing a transparent grid over lung slices at random distribution (24). Tissue blocks were infiltrated with 2.3 M sucrose for 1 h and frozen in liquid nitrogen. Frozen samples were transferred to 0.5% uranyl acetate in methanol at -90°C for at least 36 h. Temperature was raised to -45°C at a rate of 5°C/h. Samples were washed several times with pure methanol and transferred to Lowicryl HM20 via HM20/ methanol 1:1 and 2:1 for 2 h each. The blocks were polymerized under ultraviolet light for 2 d at -45°C.

Ultrathin sections were labeled according to the following procedures:

1. 1. Sections were mounted on Formvar-coated copper or nickel mesh grids. To block remaining free aldehyde groups and nonspecific binding sites, the grids were floated first on 0.02% glycine in Tris-buffered saline (TBS; pH 7.6) for 15 min and then on blocking buffer containing 5% fetal calf serum/0.2% Tween 20/0.5% albumin in TBS (pH 7.6) for 30 min. Subsequently, the grids were transferred to the primary antisera (anti-NPROSP-C^11-23, anti-CPROSP-C^162-174, or anti-cathepsin H) diluted in blocking buffer for 60 min. Grids were rinsed six times for 5 min with blocking buffer and immunoreactivity was visualized by incubation with a secondary 5- or 10-nm gold-coupled antibody diluted in blocking buffer. The grids were rinsed four times for 5 min with blocking buffer, five times for 5 min with TBS, and three times with distilled water. Finally, the sections were stained on a drop of 4% aqueous uranyl acetate, rinsed quickly three times with distilled water, and then dried overnight at 40°C. Labeled sections were viewed and photographed in a Leo EM 900 (Leo, Oberkochen, Germany) electron microscope at 50 kV.
2. 2. Double-sided labeling was performed for NPROSP-C^11-23 and CPROSP-C^162-174, NPROSP-C^11-23 and cathepsin H, as well as CPROSP-C^162-174 and cathepsin H. Ultrathin sections were picked up from the water trough of the Ultracut (Leitz, Germany) by a "perfect loop" (Science Service, Munich, Germany) and transferred to the first drop without being mounted on a grid. The first labeling was performed as described above without the final staining. The sections were transferred from drop to drop with the "perfect loop." Finally the sections were mounted with the labeled plane down on a Formvar-coated copper or nickel grid and dried overnight at 40°C. The next day the grids were labeled with the second primary antibody as described above. For the visualization of the immunoreactivity of the second labeling a secondary antibody with gold particles of a different size was used.
3. 3. Additionally, we prepared serial sections and labeled alternating sections with anti-cathepsin H or anti-NPROSP-C^11-23 as described above.

Isolation of Rat Lamellar Bodies and Rat BAL Fluid

Lamellar bodies were isolated from rat lung homogenates by upward flotation on a discontinuous sucrose gradient by modification of the method of Duck-Chong (25). Briefly, rat lungs were lavaged several times to remove the intra-alveolar surfactant. Cell-free fluid from the first pass of BAL was used for Western analysis without prior reduction of the samples. The heart, trachea, and large bronchi were dissected and the lungs were chopped into small pieces. The lung tissue was homogenized and density gradients of seven consecutive layers of 0.8 M to 0.2 M sucrose were layered over the homogenate in centrifuge tubes. The tubes were first centrifuged at 1,000 rpm and 7°C to sediment cellular debris and subsequently, without stopping, the speed was increased to 80,000 × g for 180 min. The lamellar body-rich layer was clearly detectable in the upper third of the tubes.

Preparation of Lamellar Bodies for Transmission Electron Microscopy

Isolated lamellar bodies were embedded in low-melting point agarose. The hardened gel was minced into ~ 1 mm³ sized cubes which were fixed with a mixture of 1.5% paraformaldehyde and 1.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4), post-fixed with 1% osmium tetroxide, stained en bloc with 4% aqueous uranyl acetate, dehydrated through a graded series of alcohol, and embedded in the epoxy resin Vestopal. Ultrathin sections were counterstained with lead citrate and examined using a Zeiss EM 900 (Oberkochen, Germany) operated at 80 kV.

Enzyme Assay for Cathepsin H

The assay was performed at 37°C in 0.1 M phosphate buffer containing 1 mM ethylenediaminetetraacetic acid and 10 mM cysteine at pH 6.8 with Arg-AMC as a substrate (18). To document the specificity of the enzymatic substrate cleavage, control assays were performed in the presence of 5 µM trans-epoxysuccinyl-L-leucylamido-(4-guanidino)butane (E64; Sigma, Deissenhofen, Germany), an inhibitor of cysteine proteases, which blocks the substrate cleavage (26). The enzymatic activity in isolated lamellar bodies was measured fluorometrically and the amount of specific enzymatic activity was expressed as units per total protein or volume of lamellar bodies (1 unit is the amount of enzyme that cleaves 1 µmol of substrate per min).

Western Blot Analysis of Isolated Lamellar Bodies and Human BAL Fluid

Aliquots of isolated lamellar bodies and human BAL fluid were separated using 4-12% NuPage Bis-Tris Gel and transferred to nitrocellulose membranes according to the standard procedure of the manufacturer (Novex/Invitrogen, Carlsbad, CA). The immunologic detection of cathepsin H, NPROSP-C^11-23, and SP-C was performed as described below.

Construction of Recombinant Baculovirus

Recombinant HisproSP-C virus was produced as follows: cDNA encoding proSP-C was amplified by polymerase chain reaction (PCR), using a plasmid containing full-length cDNA encoding human proSP-C (27) as a template and 5'-CGGGATCCATGG ATGTGGGCAGCAAAGAGGTC-3' and 3'-GATGTAGATCC TGCGGAGG-CCCTTAAGGC-5' as left- and right-hand primers, respectively. The PCR product was cloned into pBlueBacHis2A transfer vector (Invitrogen) using the BamHI and EcoRI sites present in the primers. This construct was verified by DNA sequencing. Recombinant virus was generated by homologous recombination between the transfer vector and linearized Autographa californica multiple nuclear polyhedrosis virus (AcMNPV) DNA in Spodoptera frugiperda 9 (Sf9) cells using the Bac-N-Blue Transfection kit (Invitrogen) as described by the manufacturer. Recombinant viruses were plaque purified, amplified, and tested for their ability to express HisproSP-C.

Insect Cell Culture and Expression of HisproSP-C

The Sf9 insect cell line was propagated in Grace's insect medium (Invitrogen) supplemented with 10% fetal calf serum and 10 µg/ml gentamicin. The cells were grown in monolayers and infected with recombinant virus at a multiplicity of infection (MOI) of five. Cells were harvested after 72 h, resuspended in 50 mM NaCl, 20 mM Tris (pH 7.4) or 50 mM NaCl, 20 mM phosphate (pH 7.4), frozen and thawed three times, and sonicated 3 × 20 s on ice at 6 watts using a Soniprep 150 ultrasonic disintegrator (MSE Scientific Instruments, Crawley, UK).

In Vitro Processing of Recombinant HisproSP-C with Purified Cathepsin H

Recombinant HisproSP-C was incubated with purified cathepsin H in Na-acetate buffer (0.05 mM, pH 5.5), containing 1 mM cysteine and 10 mM EDTA for 24 h at room temperature. In control reactions recombinant HisproSP-C was incubated without cathepsin H as well as in the presence of cathepsin H and 5 µM E64. After incubation, the reaction mixture was separated by gel electrophoresis on 4-12% NuPage gels (Novex/Invitrogen) and blotted to nitrocellulose membranes (Hybond ECL, Amersham Pharmacia Biotech, Little Chalfont, UK). The membranes were blocked using blocking reagent (Bio-Rad Laboratories, Munich, Germany) for 1 h at room temperature and incubated with one of the following antisera: anti-Xpress, anti-NPROSP-C^2-9, anti-NPROSP-C^11-23, anti-CPROSP-C^162-174, or anti-SP-C. The immunoreaction was detected using polyclonal antisera against rabbit or mouse immunoglobulin G conjugated to horseradish peroxidase (1:20,000; Dianova, Hamburg, Germany). The immunoreaction was visualized using the Western Blot Amplification Module and the Opti-4CN Substrate Kit (Bio-Rad Laboratories).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Distribution of NPROSP-C^11-23 and CPROSP-C^162-174 in the Human Lung and NPROSP-C^11-23 in the Rat Lung

To localize proSP-C and processing intermediates in the human lung, we used epitope-specific antibodies developed against antigenic sequences within the C- and N-terminal ends of proSP-C. In the human lung, the immunohistochemical staining of anti-NPROSP-C^11-23 and anti-CPROSP-C^162-174 was restricted to type II pneumocytes. At the light microscopic level, the red reaction product appeared as a fine and granular cytoplasmic staining (not shown). We next looked at the distribution of NPROSP-C^11-23 and CPROSP-C^162-174 in freeze-substituted HM20 embedded human lung specimens using immunogold EM. At the EM level, we distinguished between small electron light and dense vesicles as well as electron light and dense multivesicular bodies according to the morphological criteria described by M. C. Williams (28) as well as S. L. Young and coworkers (29). We detected NPROSP-C^11-23 in small electron light vesicles, in many electron light as well as dense multivesicular bodies, and in composite bodies throughout the cytoplasm (Figure 1c-d, 4b-d). In some electron dense multivesicular bodies the inner vesicles appeared as discs and sometimes small membrane stacks (Figure 1d). Furthermore, we found NPROSP-C^11-23 in a few lamellar bodies (Figure 1e), but not in the intraalveolar space. The CPROSP-C^162-174 epitope was also localized in small electron light vesicles and multivesicular bodies (Figure 1a), but only in a few electron dense multivesicular bodies and not in composite bodies (Figure 1a-b). To further investigate the distribution of NPROSP-C^11-23 and CPROSP-C^162-174 in human lungs, double labeling experiments were performed. A colocalization of NPROSP-C^11-23 and CPROSP-C^162-174 was found only in a few electron dense multivesicular bodies (Figure 1e), while in most of the electron dense multivesicular bodies, the composite bodies and a few lamellar bodies only NPROSP-C^11-23 was detected (Figure 1e). In the human BAL fluid only mature SP-C was identified (not shown).

View larger version (125K): [in this window] [in a new window]   Figure 1.   Localization of CPROSP-C162-174 and NPROSP-C11-23 in type II pneumocytes of human lungs by immuno EM. Ultrathin lowicryl HM20 sections of freeze-substituted human lung specimens were labeled with antisera against CPROSP-C162-174 and NPROSP-C11-23. Some vesicles and a few electron-dense multivesicular bodies (dmvb) in type II pneumocytes were positive for CPROSP-C162-174 (10 nm secondary antibody-gold complex) (a-b), whereas many electron-dense multivesicular bodies (dmvb) were positive for NPROSP-C11-23 (10 nm secondary antibody-gold complex) (c-d). In some dmvb* membrane stacks ( ) were detected (d). In a few electron-dense multivesicular bodies (dmvb) CPROSP-C162-174 (5 nm secondary antibody-gold complex) and NPROSP-C11-23 (10 nm secondary antibody-gold complex) were colocalized, whereas a small number of lamellar bodies (lb) contained only NPROSP-C11-23 (e).

In isolated lamellar bodies (Figure 2c), but not in rat BAL fluid, we found a 6-kD proSP-C processing intermediate detected by NPROSP-C^11-23 (Figure 2a). Anti-SP-C identified monomeric (~ 4-kD) as well as dimeric (~ 8-kD) mature SP-C in the lamellar bodies (Figure 2b) and monomeric mature SP-C in the rat BAL fluid (not shown). To correlate the Western blot data with the ultrastructural localization of NPROSP-C^11-23, rat lung specimens were prepared according to the same procedure as described for the human lung. At the EM level, we found NPROSP-C^11-23 in many electron dense multivesicular bodies, composite bodies, and a few lamellar bodies (Figure 2 d-f).

View larger version (111K): [in this window] [in a new window]   Figure 2.   Identification of a 6-kD proSP-C processing intermediate and mature SP-C in isolated rat lamellar bodies by Western blot analysis and localization of SP-C precursors in the rat lung. In isolated lamellar bodies, only anti-NPROSP-C11-23 antiserum identified a 6-kD proSP-C processing intermediate (a). Anti-SP-C detected monomeric (~ 4-kD) as well as dimeric (~ 8-kD) mature SP-C and reacted weakly with the 6-kD proSP-C processing intermediate (b). After isolation, the rat lamellar bodies appear as loosely coiled membranes (lb) (c). At the ultrastructural level, NPROSP-C11-23 (10 nm secondary antibody-gold complex) was detected in many electron-dense multivesicular bodies (dmvb) (d and e), composite bodies (cb) (  f  ), and also in a few lamellar bodies (d and g) in type II pneumocytes of rat lungs. Most of the NPROSP-C11-23 labeling in electron-dense multivesicular bodies was found over the electron-dense matrix (e).

Distribution of cysteine proteases in human lungsImmunohistochemistry revealed that only cathepsin H was localized in type II pneumocytes (Figure 3a). We found no expression of the cathepsins B, K, L, or S in these cells. As an internal positive control, the alveolar macrophages showed a positive staining for all cathepsins (not shown). At the ultrastructural level, we detected cathepsin H in electron dense vesicles, electron dense multivesicular bodies, composite bodies, and in lamellar bodies over the projection core and along the outer limiting membrane (Figure 3b-d). By immuno EM we also found cathepsin H in the alveolar space. On Western blots of human BAL fluid, anti-cathepsin H detected a polypeptide with a molecular mass of ~ 28 kD, corresponding in size to cathepsin H (not shown).

View larger version (129K): [in this window] [in a new window]   Figure 3.   Localization of cathepsin H in human lungs. By means of immunohistochemistry, cathepsin H was localized in type II pneumocytes (*) and alveolar macrophages (AM) (a). At the ultrastructural level, electron-dense multivesicular bodies (dmvb), composite bodies (cb), and the projection core (pc) as well as the outer limiting membrane ( ) of the lamellar bodies (lb) of type II pneumocytes were positive for cathepsin H (10 nm secondary antibody-gold complex) (b-d).

Since it is known that multivesicular bodies represent a heterogeneous population, and the colocalization of the enzyme and substrate is a prerequisite for the processing in vivo, we performed double-labeling experiments using anti-cathepsin H and anti-NPROSP-C^11-23 as well as anti-cathepsin H and anti-CPROSP-C^162-174 using serial sections and double-sided labeling of a single section. We identified electron dense vesicles labeled with anti-cathepsin H or electron light vesicles/multivesicular bodies with anti-NPROSP-C^11-23 and anti-CPROSP-C^162-174. Anti-cathepsin H and anti-CPROSP-C^162-174 labeling were colocalized only in a few electron dense multivesicular bodies (Figure 4a), while in most of the electron dense multivesicular bodies and composite bodies we found a colocalization of anti-cathepsin H and anti-NPROSP-C^11-23 (Figure 4b-f). Within electron dense multivesicular bodies, the anti-CPROSP-C^162-174 and anti-NPROSP-C^11-23 staining were preferentially localized over the electron dense matrix in close proximity to anti-cathepsin H (Figure 4a-c). Only in a few electron dense multivesicular bodies, an anti-NPROSP-C^11-23 labeling was also found along the lumenal surface of the inner vesicles (Figure 4c).

View larger version (140K): [in this window] [in a new window]   Figure 4.   Colocalization of cathepsin H and proSP-C in type II pneumocytes of human lungs. Ultrathin sections were immuno double-labeled with anti-cathepsin H and anti-CPROSP-C162-174 as well as anti-cathepsin H and anti-NPROSP-C11-23 using double-sided labeling technique (a-d) or serial sections (e and f  ). Cathepsin H (15 nm secondary antibody-gold complex) was identified in electron-dense multivesicular (dmvb), composite (cb), and lamellar bodies (lb) in type II pneumocytes, but only in a few electron-dense multivesicular bodies (dmvb) was it colocalized with CPROSP-C162-174 (10 nm secondary antibody-gold complex) (a). Many electron-dense multivesicular bodies (dmvb) and composite bodies (cb) were labeled for both NPROSP-C11-23 (5 nm secondary antibody-gold complex) and cathepsin H (15 nm secondary antibody-gold complex) (b-d), whereas electron-light multivesicular (lmvb) bodies were labeled only for NPROSP-C11-23 (b). Furthermore, electron-dense vesicles were identified that were labeled only for cathepsin H (c). Using serial sections immuno double-labeling for cathepsin H (10 nm secondary antibody-gold complex) (e) and NPROSP-C11-23 (10 nm secondary antibody-gold complex) (  f  ) also identified a vesicle labeled only for NPROSP-C11-23 (vesicle) and an electron-dense multivesicular body labeled for both cathepsin H (dmvb) and NPROSP-C11-23 (dmvb*).

Enzymatic cathepsin H activitySince it is known that cathepsin H is synthesized as an inactive precursor and antisera directed against cathepsin H do not differentiate between the inactive precursor and the mature enzyme, we measured cathepsin H activity in isolated lamellar bodies from four rat lungs. The specific cathepsin H activity was 8.6 ± 3.3 units/g total protein (mean ± S.D.)  0.1 to 0.2 µU/µl and was completely blocked by the cysteine protease inhibitor E64.

In vitro processing of recombinant proSP-C We next investigated whether purified cathepsin H was able to process recombinant human HisproSP-C in vitro. For the characterization of the processing products, antisera against the His-tag (anti-Xpress), the N-terminal epitopes NPROSP-C ^2-9 and NPROSP-C^11-23, the human C-terminal epitope CPROSP-C^162-174 and mature SP-C were used. All antisera against proSP-C and the His-tag identified HisproSP-C migrating at 29 kD (Figure 5a-d). After 18 h of incubation with 150 ng/µl  10 µU/µl purified cathepsin H, three novel bands migrating at 20 kD, 15 kD, and 14 kD were detected by the NPROSP-C^11-23 antiserum (Figure 5c). The 20-kD, 15-kD, and 14-kD processing products were not recognized by either the anti-X-press or anti-NPROSP-C^2-9 consistent with a cleavage of the M¹-L⁹ domain of the NH₂-terminus (Figure 5a-b). The CPROSP-C^162-174 antiserum identified only the 20-kD processing product (Figure 5d), suggesting that this product was processed at the N-terminus but not at the C-terminus. In contrast, the 15-kD and 14-kD processing products were not detected by anti-CPROSP-C^162-174, indicating a proteolytic cleavage step in the C-terminal propeptide region within or before the G¹⁶²-G¹⁷⁴ domain. Since cathepsin H has strong aminopeptidase but only weak endopeptidase activity and isolated lamellar bodies contained only  0.1 to 0.2 µU/µl aminopeptidase activity, we also performed in vitro processing experiments with different cathepsin H concentrations (Figure 6, lane 1: 150 ng/µl  10 µU/µl, lane 2: 15 ng/µl  1 µU/µl, lane 3: 5 ng/µl  0.3 µU/µ, and lane 4: 1.6 ng/µl  0.1 µU/µl) . The 15-kD and 14-kD processing products were detected only at high (Figure 6, lane 1 and lane 2), but not at low concentrations (Figure 6, lane 3 and lane 4). The bands at 20 kD, 15 kD, and 14 kD could be competed off by pretreatment of antisera with human HisproSP-C. Sometimes, an additional nonspecific band at 22 kD was detected only by the NPROSP-C^11-23 antiserum (Figure 5c, 6). This nonspecific band did not compete away after preincubation of the antisera with human HisproSP-C. No mature SP-C was found (not shown). In each experiment, the processing of HisproSP-C could be blocked by the addition of the E64.

View larger version (11K): [in this window] [in a new window]   Figure 5.   In vitro processing of proSP-C by purified cathepsin H. His-tag-proSP-C (lane 1) was incubated with 150 ng/µl purified cathepsin H (lane 2) as well as purified cathepsin H and E64 (lane 3). For the Western blotting, antisera against the His-tag (anti-X-press) (a), the N-terminal epitopes NPROSP-C2-9 (b) and NPROSP-C11-23 (c), as well as the C-terminal epitope CPROSP-C162-174 (d) were used. After 18 h of incubation, a 20-kD processing intermediate was detected by the NPROSP-C11-23 (c, lane 2) and the CPROSP-C162-174 antisera (d, lane 2). The 15- and 14-kD processing intermediates were identified exclusively by the NPROSP-C11-23 antiserum (c, lane 2). Addition of the E64 completely blocked processing of proSP-C (a-d, lane 3). The 20-kD, 15-kD, and 14-kD processing intermediates were recognized neither by anti-X-press (a, lane 2) nor anti-NPROSP-C2-9 (b, lane 2). Sometimes, an additional nonspecific band at 22 kD was detected only by the NPROSP-C11-23 antiserum (c, lane 2). This nonspecific band did not compete away after preincubation of the antiserum with human HisproSP-C.

View larger version (85K): [in this window] [in a new window]   Figure 6.   In vitro processing of proSP-C by purified cathepsin H. His-tag-proSP-C was incubated with different concentrations of purified cathepsin H (lane 1: 150 ng/µl  10 µU/µl total volume; lane 2: 15 ng/µl  1 µU/µl total volume; lane 3: 5 ng/µl  0.3 µU/µl total volume; and lane 4: 1.6 ng/µl  0.1 µU/µl total volume). Although NPROSP-C11-23 also detected a 20-kD processing intermediate after incubation of His-tag-proSP-C with low cathepsin H concentrations (lanes 3 and 4), the 15- and 14-kD processing intermediates were found only with high cathepsin H concentrations (lanes 1 and 2).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Human surfactant protein C is synthesized as a 197 amino acid proprotein (proSP-C) that is sorted and processed as an integral membrane protein by multiple proteolytic cleavages. The generation of mature SP-C from the primary translation product involves the addition of palmitic acid chains, the cleavage of 136 amino acid residues from the C-terminus, and the removal of the N-terminal 23 amino acid residues. In addition, complete proteolytic processing of proSP-C requires targeting of the propeptide through the regulated pathway to the lamellar bodies of the type II pneumocyte. Previously, in studies of rat lungs the lysosomal cysteine protease cathepsin H was detected in lamellar bodies of type II pneumocytes and in BAL fluid (30). In the present investigation, we extended these findings to the human lung and showed that the electron-dense multivesicular bodies in human type II pneumocytes are fusion products of electron-dense vesicles containing cathepsin H and electron-light vesicles or multivesicular bodies containing proSP-C. Additionally, we measured specific cathepsin H activity in isolated lamellar bodies and demonstrated with in vitro assays that cathepsin H is involved in the processing of proSP-C. Thus, the current study characterizes the first protease involved in the post-translational processing of proSP-C.

Previous studies showed that the processing of the 21-kD proSP-C through 16-, 7-, and 6-kD proSP-C intermediates to mature SP-C is a four-step process requiring at least two distinct cleavages of the C-terminal propeptide followed by at least two cleavages of the N-terminal propeptide (Figure 7) (10). To study the processing at the ultrastructural level in human as well as rat lungs, we used antisera against specific epitopes of the C- and N-terminal propeptide of proSP-C (10, 11). Only within a few electron-dense multivesicular bodies did we find a colocalization of NPROSP-C^11-23 and CPROSP-C^162-174. In human as well as rat type II pneumocytes, NPROSP-C^11-23 was detectable in many electron dense multivesicular bodies, composite bodies, and a few lamellar bodies (Figures 1a-1e and 2e-2g). In good correlation with the immuno EM, we found a 6-kD proSP-C processing intermediate detected by anti-NPROSP-C^11-23 and mature SP-C in isolated rat lamellar bodies (Figures 2a-2c). In the human and rat BAL fluid, only mature SP-C was identified. These findings are consistent with the occurrence of the first N-terminal processing step and the cleavage of the C-terminus of proSP-C in electron-dense multivesicular bodies and the final removal of the N-flanking domain E¹¹-R²³ in the lamellar bodies before secretion of mature SP-C into the alveolar space (Figure 7).

View larger version (21K): [in this window] [in a new window]   Figure 7.   Model of intracellular localization of SP-C processing (adapted from Refs. 10-12). SP-C is synthesized by type II pneumocytes as a 21-kD propeptide (proSP-C) which is proteolytically processed to a 4.2-kD dipalmitoylated protein on the route from its site of synthesis to the lamellar bodies (lb). In vivo, the proteolytic processing of the 21-kD proSP-C through 16-, 7-, and 6-kD proSP-C intermediates to mature SP-C is a four-step process requiring at least two distinct cleavages of the C-terminal propeptide followed by at least two cleavages of the N-terminal propeptide (11, 12). In the human lung, cathepsin H is involved in the first N-terminal processing steps of proSP-C in electron dense multivesicular body (dmvb) in type II pneumocytes after the fusion of electron-dense vesicles containing cathepsin H and electron-light vesicles or multivesicular bodies (lmvb) containing proSP-C.

In previous immuno EM microscopy studies, no proSP-C was found in lamellar bodies (9, 31). SP-C is an extremely hydrophobic protein, inserted as a transmembrane protein into the lamellae of the lamellar bodies in type II pneumocytes (32). The stabilization of the phospholipid-rich multilamellar bodies in type II pneumocytes is the critical step in the preparation of lung samples for EM studies. Without uranyl acetate treatment, the multilamellar bodies appear as empty vacuoles in resin-embedded lung specimens as well as in ultrathin cryosections (9, 31, 33). To preserve the lipid structures, we performed a freeze substitution in uranyl acetate in methanol combined with Lowicryl HM20 embedding at low temperature (34). Compared with published data, differences in the staining pattern observed may be due to differences either in the sample preparation or the antisera used.

In type II pneumocytes, the normal intracellular processing of proSP-C depends upon an acidification of the multivesicular and lamellar bodies (35). Most cathepsins have a pH optimum between 5.5 and 6.8 and are irreversibly inactivated above pH 7 (18, 19, 36). The pH optimum fits well with the acidic pH in the multivesicular and lamellar bodies (39, 40). In isolated lamellar bodies, we measured specific cathepsin H activity. Using immunohistochemistry, we found a strong cathepsin H but no cathepsin B, K, L, or S expression in type II pneumocytes, suggesting a specific role of cathepsin H in these cells.

Because the colocalization of substrate and enzyme in vivo is a prerequisite for the processing, we performed double labeling for cathepsin H and proSP-C on serial sections and modified the double-sided labeling technique developed by Bendayan (41). We found a colocalization of anti-cathepsin H and anti-NPROSP-C^11-23 labeling in electron-dense multivesicular and composite bodies (Figures 4b-4f), whereas only in a few electron dense multivesicular bodies a colocalization of anti-cathepsin H and anti-CPROSP-C^162-174 staining was detected (Figure 4a). We further identified electron-dense vesicles labeled with anti-cathepsin H or electron light vesicles/multivesicular bodies labeled with anti-NPROSP-C^11-23 and anti-CPROSP-C^162-174. These results indicate that the electron-dense multivesicular bodies in human type II pneumocytes are fusion products of electron dense vesicles containing cathepsin H and electron-light vesicles or multivesicular bodies containing proSP-C.

Most data indicate that proSP-C is an integral membrane protein inserted into membranes in a type II transmembrane configuration with the C-terminus of proSP-C inside the lumen of the ER and the Golgi vesicles (10, 42, 43). The trafficking of proSP-C in the regulated secretory pathway is similar to the trafficking of internalized bitopic membrane proteins such as the epidermal growth factor receptor in the endocytic pathway. It has been shown that the transmembrane orientation of these receptors is reconfigured (inverted) during fusion events with the target compartment (44). Our data indicate that the formation of electron-dense multivesicular bodies occurs by a fusion of electron-dense vesicles containing cathepsin H and electron-light vesicles or multivesicular bodies containing proSP-C which might result in an orientation of the C-terminus of proSP-C outside of the internal vesicles of the multivesicular bodies. Based on this orientation, the C-terminus is exposed to cathepsin H, which was found to be localized over the electron-dense matrix between the outer limiting membrane and the inner vesicles of the multivesicular bodies (Figure 3b).

Besides the processing of proSP-C, the proteolytic processing of proSP-B to mature SP-B takes place in multivesicular bodies (45). Mature SP-B leads to a micellization of vesicles and a fusion of discs into lamellar stacks as shown by Hawgood and coworkers (46). This process is accompanied by a loss of the integrity of the inner vesicles of the multivesicular bodies and the exposure of their inner surface to the electron-dense matrix. Indeed, we found a NPROSP-C^11-23 labeling in electron-dense multivesicular bodies of the human lung associated with membrane discs and small stacks resulting from disturbed inner vesicles. In good agreement with this, in electron-dense multivesicular bodies of the human as well as rat lung we found most of the NPROSP-C^11-23 labeling over the electron dense matrix, whereas only a weak staining was localized along the inner lumenal surface (Figures 2e and 4c). In the electron-dense multivesicular bodies, the N-terminus is in close proximity to cathepsin H (Figure 4c). Therefore, the first N-terminal cleavage step of proSP-C may occur after the C-terminal cleavage, the maturation of SP-B, and the micellization of the inner vesicles of the multivesicular bodies (Figure 7). This hypothesis is further supported by the observation that mice lacking SP-B seem unable to normally convert multivesicular bodies to organized lamellae and to completely process proSP-C (47).

Because we found a colocalization of cathepsin H and SP-C precursors in electron-dense multivesicular, composite, and a few lamellar bodies, we performed in vitro processing assays using recombinant human proSP-C and different concentrations of purified cathepsin H. To produce human proSP-C we used the baculovirus expression system. This eukaryotic system has the advantage that it is capable of many post-translational modifications, including palmitoylation. Previously, the expression of correctly folded, dipalmitoylated, and active SP-C has been described in this system (48). In our in vitro assay, purified cathepsin H cleaved the his-Tag (8 kD) and the M¹-L⁹ domain (1 kD) of the N-terminus of human HisproSP-C (29 kD  20 kD) as determined by immunoblotting using epitope-specific antibodies (Figures 5a-5d). This cleavage corresponds well to the first N-terminal processing step (7 kD  6 kD) of proSP-C in electron-dense multivesicular bodies that is the third step in the overall proSP-C processing cascade. This processing step is characterized by the cleavage of the M¹-L⁹ domain of the N-terminus of proSP-C in vivo (Figure 7) (10). However, purified cathepsin H was not able to cleave the E¹¹-R²³ domain. This domain of the N-terminus contains two proline residues in positions 13 and 14. The N-terminal proSP-C processing is in agreement with previous studies that cathepsin H preferentially releases N-terminal amino acid residues with large hydrophobic or basic side chains by its aminopeptidase activity but is unable to hydrolyze peptides with a proline either in the N-terminal or penultimate position (49). After incubation with high concentrations of purified cathepsin H, another cleavage site was detected in the C-terminal propeptide of proSP-C (conversion of the 20-kD processing product to the 15-/14-kD products) (Figure 5). However, this cleavage step was not catalyzed if the aminopeptidase activity in the reaction mixture was adjusted to that found in isolated lamellar bodies (Figure 6). In contrast to the broad specificity and high aminopeptidase activity of cathepsin H, the endopeptidase activity is much lower and shows a rather strict specificity, preferring hydrophobic residues at the P2 and P3 sites (amino acids in positions 2 and 3 of the substrate next to the cleavage site) (50). Despite the fact that the sequence of the processing events in vivo is different than in vitro due to the reconfiguration of the transmembrane orientation of proSP-C in electron-dense multivesicular bodies, the N-terminal cleavage catalyzed by purified cathepsin H in vitro corresponds well to the first N-terminal processing steps described in vivo (Figure 7) (10). Taking into account that the C-terminal cleavage was catalyzed in vitro only at cathepsin H activities ~ 50-100 times higher than in isolated lamellar bodies, it might not be relevant under in vivo conditions. Furthermore, the first C-terminal cleavage step of proSP-C is non-type II pneumocyte specific (Refs. 51 and 52, and M. F. Beers, unpublished observations) and could not be inhibited in isolated human type II pneumocytes by E-64 (S. Guttentag, unpublished observation).

In conclusion, our data indicate that cathepsin H is involved in the first N-terminal processing step of proSP-C in electron dense multivesicular bodies in type II pneumocytes after the fusion of electron-dense vesicles containing cathepsin H and electron-light vesicles or multivesicular bodies containing proSP-C. We provide three lines of evidence for the involvement of cathepsin H in the processing of proSP-C: (i) SP-C precursors and cathepsin H are colocalized in electron-dense multivesicular, composite, and a few lamellar bodies; (ii) isolated lamellar bodies contain enzymatically active cathepsin H; and (iii) using physiologic enzyme activities, purified cathepsin H is able to cleave the N-terminal M¹-L⁹ domain of recombinant human proSP-C in vitro.

We speculate that at least one additional enzyme must be involved in the C-terminal processing in electron-dense multivesicular and the final remodeling of the N-terminus in composite and lamellar bodies. Cleavage of peptides by a combination of different proteases is known from the processing of the major histocompatibility complex class II invariant chain (53).