Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Pulmonary surfactant represents a biochemically heterogeneous complex of phospholipids and proteins which functions to maintain alveolar stability at low lung volumes through reduction of surface tension at the air-liquid interface (1). The biophysical activity of surfactant is primarily due to its phospholipid components, but organic extracts of isolated surfactant also contain two small hydrophobic proteins, surfactant proteins (SPs) B and C, each of which has been shown to enhance adsorption and surface tension-reducing properties of the lipid monolayer (2).

The alveolar form of rat SP-C (mature SP-C) is a 35- amino acid hydrophobic peptide (SP-C_3.7) that results from extensive post-translational proteolysis of a 194-amino acid precursor (proSP-C₂₁) shown to be synthesized exclusively by the alveolar type II cell (1, 3). Structurally, the mature form of SP-C is positioned internally within the larger SP-C propeptide and is asymmetrically flanked by propeptide domains of 23 amino acids at the NH₂ terminus and 136 amino acids at the COOH terminus (4). Previous studies in perfused rat lungs, isolated rat type II cells, and human fetal lung explants have indicated that the post-translational processing of proSP-C₂₁ requires several discrete proteolytic events to generate SP-C_3.7 (7). Results obtained using epitope-specific antisera have shown that initial processing of proSP-C₂₁ involves removal of COOH flanking propeptide via two distinct cleavage steps to produce an identifiable 16-kD intermediate and several lower molecular weight forms which are composed of vestigial NH₂ flanking domains and the mature sequence. Although the basic steps in COOH-terminal processing have been defined, specific proteolytic events for subsequent processing and removal of the NH₂ terminus to liberate mature SP-C have not been determined.

Similarly, the subcellular location of proSP-C processing events is incompletely understood. Pulse-chase experiments in isolated rat type II cells demonstrate that either Brefeldin A (7) or low temperature block (12) completely inhibits cleavage of the synthesized 21-kD primary translation product. Previous work by Vorbroker and colleagues (12) and Voorhout and associates (14) using immunoelectron microscopy with antisera to recombinant proSP-C detected proSP-C labeling in endoplasmic reticulum (ER), in Golgi, in HA-1-negative coated vesicles, and in multivesicular bodies (mvb). Subcellular fractionation experiments have shown that isolated lamellar bodies (LBs) contain both a 6-kD low molecular weight proSP-C form (11) as well as mature SP-C (15). Thus, the data to date are in accordance with the concept that proSP-C is transported though the biosynthetic pathway from the Golgi via mvb to LBs and suggests that whereas cleavage of the COOH terminus begins as early as the medial Golgi, involvement of both mvb and LBs ("late organelles") in intracellular transport and processing is required for complete processing. The potential importance of these later compartments in SP-C biosynthesis is further highlighted by findings in human patients with inherited SP-B deficiency and in SP-B-null mice in which proSP-C is not processed beyond the 6-kD intermediate and is accompanied by a commensurate absence both of structural LBs and of mature SP-C protein (16).

The extensive post-translational processing of SP-C is reminiscent of post-translational processing of the other hydrophobic surfactant protein, SP-B. Like SP-C, production of mature SP-B protein involves sequential cleavage of a larger propeptide which is first translocated to the ER and routed (together with proSP-C) to distal compartments of the secretory pathway, while undergoing proteolysis of COOH- and NH₂-terminal flanking regions (19- 21). Recently, we have demonstrated that cleavage of SP-B NH₂-terminal propeptide is a two-stage process that first generates a 10-amino acid flanking remnant which is later removed in a final processing step to liberate mature SP-B (19). Further, the NH₂ terminus of proSP-B is absolutely required for correct intracellular routing, although the exact domain remains undefined (20, 21).

In this report, NH₂ epitope-specific proSP-C antibodies, subcellular fractionation, metabolic labeling, immunogold electron microscopy, and chimeric fusion proteins were used to effect a more complete characterization of the proteolysis and function of the NH₂-terminal flanking domain of proSP-C. We hypothesized that NH₂-terminal processing of proSP-C was a multistep process and that the domain(s) in the NH₂ flanking propeptide was required for targeting to post-Golgi processing compartments. Our results extend previous observations of proSP-C biosynthesis by demonstration that removal of the NH₂-terminal domain is a two-step event involving a previously unrecognized cleavage of the proximal portion of the NH₂ terminus (residues M¹-L⁹). Remodeling of the amino terminus begins in the multivesicular body after removal of COOH-terminal domains and concludes after transfer of proSP-C to the LB where mature SP-C is liberated. Further, new functional analysis using heterologous chimeric enhanced green fluorescent protein (EGFP)/SP-C fusion proteins demonstrates that within this vestigial NH₂ fragment of proSP-C, a functional targeting motif (residues Met¹⁰-Thr¹⁸) exists which is essential and sufficient for direction of translocated proSP-C from proximal compartments (ER) to a post-Golgi, lysosome-related organelle in lung epithelial cells, thus offering a potential mechanism for conservation of this motif until the end (distal) of the processing pathway.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

Trans-³⁵S-label was purchased from ICN/Flow, Inc. (Irvine, CA). Protein A-agarose and ¹⁴C molecular weight markers were obtained from Bethesda Research Labs (Gaithersburg, MD). The pcDNA3 eukaryotic expression plasmid was obtained from Invitrogen, Inc. (San Diego, CA). pEGFP-C1 plasmid was purchased from Clontech, Inc. (Palo Alto, CA). Culture medium was produced by the Cell Center Facility, University of Pennsylvania. Except where noted, all other reagents were electrophoretic grade and were purchased from either Sigma Chemical, Inc. (St. Louis, MO); Fisher Scientific, Inc. (Pittsburgh, PA); or Bio-Rad (Melville, NY).

Monoclonal antiserum to early endosomal antigen (EEA)-1 was obtained from Transduction Laboratories, Inc. (Lexington, KY). Anti-ubiquitin monoclonal antibody (mAb) was purchased from Chemicon (Temecula, CA). Polyclonal anti-calnexin NT antiserum was purchased from Stressgen, Inc. (Victoria, BC, Canada). A monoclonal anti-CD-63 antibody was obtained from Immunotech, Inc. (Marseilles, France). Texas Red dye-conjugated mAbs and polyclonal antibodies were obtained from Jackson Immunoresearch Laboratories, Inc. (West Grove, PA).

ProSP-C Antisera

The location of the epitopes chosen for production of two NH₂-terminal proSP-C antisera used in this study are schematically shown in Figure 1. Synthetic rat SP-C peptides rSP-C^2-9 (Asp²-Leu⁹) and rSP-C^11-23 (Glu¹¹-Gln²³) based on the published rat SP-C complementary DNA (cDNA) sequence (4) were commercially synthesized on solid-phase resin supports using t-butoxycarbonyl chemistry (Macromolecular Resources, Inc., Fort Collins, CO). The coupling efficiency for each step was greater than 99.9%, indicating that the crude peptides were greater than 90% pure. After purification by reversed-phase high-performance liquid chromatography (HPLC), analytical HPLC of the peptides yielded a single elution peak at 280 nm for each peptide. Mass spectroscopy also confirmed the purity of the peptides.

View larger version (11K): [in this window] [in a new window]   Figure 1.   Epitopes of NH2-terminal proSP-C antibodies. Schematic diagram of epitopic regions of polyclonal proSP-C antisera produced using synthetic peptides. The positions of newly produced NH2-terminal antisera (NPROSP-C2-9 and NPROSP-C11-23) are diagrammed. Amino acid numbers corresponding to full-length rat proSP-C based on published cDNA sequence (4) are shown at top.

Two monospecific polyclonal rat proSP-C antisera were commercially produced using rSP-C^2-9 and rSP-C^11-23 synthetic peptides as immunogens as previously described (11). The resulting antisera, anti-NPROSP-C^11-23 and anti-NPROSP-C^2-9, were analyzed for reactivity against their immunizing peptides by dot-blot immunoblotting protocol performed using successive incubation of samples with primary antisera at dilutions ranging from 1:100 to 1:1,000 for 2 h and secondary goat antirabbit (GAR) horseradish peroxidase (HRP) at a dilution of 1:2,000 for 2 h at room temperature. Reactivity was developed colorimetrically using a Bio-Rad HRP developing kit containing 4-chloronaphthol. Results were confirmed by Western blotting as detailed later.

Recombinant Wild-Type and Mutant rSP-C Expression Constructs

Eukaryotic expression constructs subcloned into the pcDNA3 plasmid backbone were generated for production of ³⁵S-labeled in vitro translation products. All procedures involving oligonucleotide and cDNA manipulations were performed essentially as described by Ausbel and coworkers (22). The following constructs were synthesized.

Wild-type SP-C. A full-length rSP-C cDNA (816-base pair) insert (4) was previously subcloned into the pcDNA3 polylinker at the EcoRI site (pcDNA3-SP-C^wt) (8). The vector contains human cytomegalovirus promoter (early promoter and enhancer region), bovine growth hormone polyadenylation sequence, -lactamase and neomycin resistance genes, and T7 promoters for sense/ antisense in vitro transcription.

ProSP-C COOH- and NH₂-terminal truncations. A mutant cDNA deletion construct containing a nine-amino acid truncation of the COOH-flanking region of proSP-C (Met¹-Leu¹⁸⁵) was previously generated by polymerase chain reaction (PCR) and subcloning into pcDNA3 (8). Two other inserts containing a partial truncation of the NH₂ domain (Met¹⁰-Ile¹⁹⁴) or complete truncation of NH₂ domain and substantial removal of the COOH region (Phe²⁴-Ser⁷²) were each produced using PCR with pcDNA3- SP-C^wt as template.

For Met¹⁰-Ile¹⁹⁴, oligonucleotide primers were 5'-GGGGGTACCATGGAGAGCCCACCGGATTAC-3' and 3'-GCGAGCTCGTACGTAGATCT-5'. For Phe²⁴-Ser⁷², the primers used were 5'-GGGGTACCATGGCATTTCGCATTCCCTGCT-3' and 3'-ACCAGGAACTCTACTCGATTGT GAGCTCCC-5'. Both inserts contained consensus sequences for translational start and restriction sequences for religation into pcDNA3 using sequential digestion with KpnI and XhoI.

Production of propeptides using in vitro translation. Primary translation products of wild-type SP-C (relative molecular mass [M_r] = 21,000) as well as mutant forms containing COOH-terminal and/or NH₂-terminal truncations were prepared from pcDNA3 constructs using a combined sequential in vitro transcription/in vitro translation of plasmid cDNA (TNT Reticulocyte Lysate System; Promega, Inc., Madison, WI) as published (8). Circular plasmid DNA encoding either full-length rSP-C or mutant forms (1 to 2 µg) and Trans³⁵S-label (0.8 µCi/ml) were used with the TNT kit utilizing T7 polymerase. The fidelity of translation products was confirmed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS/PAGE) and autoradiography (see ANALYTICAL METHODS).

Chimeric EGFP Fusion Proteins

Chimeric fusion proteins consisting of EGFP and wild-type rSP-C or mutant SP-C cDNA containing N-terminal or C-terminal deletions were produced using PCR as previously published (10). In all cases, pcDNA3-rSP-C (+) was used as template. To create in-frame fusion proteins, sequences encoding the indicated regions of the proSP-C peptide were amplified by PCR using modified primers. A BspEI site at the 5' end and an XhoI site at the 3' end were introduced for cloning into pEGFP-C1. The oligonucleotide primers are listed in Table 1.

Amplification reactions contained 0.2 µM primers, 1.25 µM deoxynucleotide triphosphate mixture, 1.5 µM MgCl₂, 10 ng template, and 2.5 U Vent_R DNA polymerase (New England Biolabs, Beverly, MA). They consisted of 30 cycles of: denaturation at 95°C for 30 s, primer annealing at 60°C for 30 s, and primer extension at 72°C for 15 s. After the last cycle, the mixture was incubated at 72°C for 7 min. Purified PCR fragments were ligated into pEGFP-C1 after digestion with BspEI and XhoI.

Automated DNA sequencing in both directions was performed at the Core Facility in the Department of Genetics at the University of Pennsylvania. No nucleotide mutations in the coding region of full-length SP-C or any deletional constructs were detected.

Cell Lines and Subcellular Fractions

Rat type II cells. Type II pneumocytes were isolated by elastase digestion of lungs from Sprague-Dawley rats using the method of Dobbs and colleagues (23). The preparation obtained after panning on immunoglobulin (Ig) G-coated plates was greater than 85% type II cells. Freshly isolated type II cells were used immediately for membrane association assays as detailed later, or were placed in suspension culture in Dulbecco's modified Eagle's medium containing 10% fetal calf serum (FCS) at 37°C in 5% CO₂ and used for metabolic labeling studies.

A549 cells. The lung epithelial cell line A549 used in transfection studies was originally obtained through the American Type Culture Collection (Rockville, MD) and made available as a gift of Dr. S. I. Feinstein (University of Pennsylvania, Philadelphia, PA). A549 cells were grown at 37°C and 5% CO₂ in Dulbecco's modified essential medium supplemented with 10% FCS, 100 U/ml penicillin, and 100 µg/ml streptomycin as previously described (8, 10).

Isolation of LBs. To isolate LBs, whole lungs from anesthetized Sprague-Dawley adult rats were cleared of blood by perfusion with 0.9% NaCl, removed from the chest, and homogenized sequentially using a Polytron and Potter-Elvehjem vessel. A LB fraction was isolated using upward flotation on a sucrose density gradient as previously described (11, 24).

Isolation of microsomes. Microsomes were isolated from whole rat lung as previously described (11).

Membrane Association Assays

Integral association of proSP-C peptide forms with subcellular membranes was assessed in purified microsomal and LB fractions by NaCO₃ extraction using the method of Fujiki and associates (25).

Purified LBs (50 to 100 µg) or microsomes (100 to 200 µg) were resuspended in 4 ml of either homogenization buffer alone (0.32 M sucrose-10 mM N-2-hydroxyethyl piperazine-N'-ethane sulfonic acid [Hepes]-2-amino-2[hydroxymethyl]1,3-propanediol [Tris], pH 7.4) or buffer containing 100 mM NaCO₃, pH 11, or 0.1% Nonidet P-40 (NP-40). After four rapid freeze-thaw cycles using liquid N₂, resuspended fractions were centrifuged for 1 h at 100,000 × g. The recovered pellet containing treated membranes was resuspended in 30 µl of homogenization buffer. The supernatant contained soluble proteins which were concentrated by precipitation in 10% trichloroacetic acid. Both fractions were then frozen at 70°C.

Fractions were analyzed for proSP-C by SDS-PAGE and Western blotting using anti-NPROSP-C^11-23 as detailed later in ANALYTICAL METHODS.

Transfection

For transfection studies, A549 cells were grown to 80% confluence in 35-mm² plates (Corning, Inc., Corning, NY). EGFP/SP-C constructs (10 µg/dish) were transiently transfected into A549 cells using calcium phosphate precipitation (0.18 ml of 0.25 M CaCl₂ was added dropwise to 0.18 ml plasmid DNA dissolved in 2× Hepes-buffered saline [50 mM Hepes-280 mM NaCl and 1.5 mM NaPO₄, pH 7.1]) (8, 10). The media were replaced at 24 h and cells were maintained for up to 72 h.

Analytical Methods

PAGE. One-dimensional SDS-PAGE was performed in 16.5% polyacrylamide gels using a Tris-N-(2-hydroxy-1, 1 bis [hydroxymethyl]ethyl) glycine (Tricine) buffer system as modified in our laboratory for surfactant proteins (7, 8, 11).

Immunoblotting. Immunoblotting of transferred samples was performed using primary proSP-C antisera (1:4,000) as previously described (7, 8, 11). Bands were visualized by enhanced chemiluminescence using the ECL kit (Amersham, Inc., Arlington Heights, IL).

³⁵S metabolic labeling. Metabolic labeling studies were performed using a pulse-chase protocol and freshly isolated type II cells as previously described (7, 8, 10).

Immunoprecipitation. Lysates from labeled type II cells and cell culture media were immunoprecipitated using anti-proSP-C antiserum as previously published (7, 8, 10).

Protein determination. Total protein of cell and lung fractions was quantitated by the method of Bradford (26), using bovine Ig as standard.

Immunogold Electron Microscopy

Tissue preparation. Small fragments of lung tissue from 250-g male Wistar rats were fixed by immersion in 4% paraformaldehyde and 0.1% glutaraldehyde in 0.13 M phosphate buffer (pH 7.3) for 2 h. After rinsing in the same buffer, samples were dehydrated avoiding osmium tetroxide post-fixation, and embedded in London Resin White methacrylic resin (Polysciences, Inc. Warrington, PA).

Immunogold labeling. Immunostaining was performed on ultrathin sections using immunogold techniques based on the method of Varndell and colleagues (27). Sections collected on 300-mesh nickel grids were incubated in 1:30 normal goat serum (Dako, Copenhagen, Denmark), in 0.015 M phosphate-buffered saline with 0.2% bovine serum albumin and 0.02% NaN₃ and then in primary polyclonal antibodies: anti-NPROSP-C^11-23 diluted 1:2,000 or anti-NPROSP-C^2-9 diluted 1:1,000-2,000. Primary antibody incubation was carried out overnight at 4°C; negative controls omitting incubation with primary antibody were included. After several washes in buffer solutions, all grids were incubated with GAR secondary antibody conjugated with 20- or 10-nm gold particles (GAR 20 or GAR 10; Bio Cell, Inc., Cardiff, UK) for 1 h at room temperature. Grids were washed again in buffer and counterstained with uranile acetate and lead citrate and observed in a Jeol JEM 1010 (Tokyo, Japan) electron microscope. Detailed examination of a total of 16 grids for anti- NPROSP-C^11-23 and 20 grids for anti-NPROSP-C^2-9 was performed to generate representative micrographs.

Fluorescence Microscopy

Vital fluorescence microscopy for localization of expressed EGFP/SP-C fusion proteins was performed on 35-mm² plastic dishes using an Olympus I-70 inverted fluorescence microscope with filter package High Q fluorescein isothiocyanate (FITC) (excitation [Ex] = 480 nm; emission [Em] = 535/550 nm) obtained from Chroma Technology Corp. (Brattleboro, VT). Fluorescent and phase images were captured using a Hamamatsu 12 bit CCD camera. Image processing was performed using IMAGE 1 software (Universal Imaging Corp., West Chester, PA).

For colocalization of EGFP/SP-C fusion proteins with subcellular markers, A549 cells were grown on uncoated glass coverslips. At 48 h after transfection with EGFP fusion proteins, adherent cells were fixed in 4% paraformaldehyde-Na cacodylate, pH 7.4, and then permeabilized with 0.3% Triton X-100. Immunostaining was done as previously described (10). Unlabeled primary specific antisera were as follows: EEA-1 (1:250) overnight at room temperature, anti-Calnexin-NT (1:100) for 1 h at room temperature, or CD-63 (0.5 µg/ml) for 1 h at room temperature. Texas Red-conjugated secondary antisera was then applied and, after placement of coverslips, sequential images were acquired using High Q FITC and Texas Red filter packages. Image processing for colocalization was done using IMAGE 1 software.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Characteristics and Specificity of ProSP-C Antisera

By dot-blot analysis, rSP-C^2-9 and rSP-C^11-23 peptides used to generate polyclonal antisera each elicited an immunogenic response in two out of two injected rabbits by the fourth week after immunization (not shown). Maximal response was obtained by Wks 8 to 12 and was maintained up to 24 wk with subsequent antigen injection.

Western blot analysis of synthetic proSP-C peptides probed with either anti-NPROSP-C^2-9 or anti-NPROSP-C^11-23 demonstrated that each antiserum recognized only the sequence region in proSP-C that was used for immunization (Figure 2A). Anti-NPROSP-C^11-23 recognizes the region Glu¹¹-Gln²³, which is located adjacent to the amino terminus of mature SP-C. Anti-NPROSP-C^2-9 recognizes the propeptide segment Asp²-Leu⁹. Neither antiserum exhibited cross-reactivity against other synthetic SP-C peptides or against samples containing mature SP-C. A third proSP-C peptide, rSP-C^24-35 (Phe²⁴-Arg³⁵), containing 11 residues from the NH₂-terminal portion of the mature rat SP-C sequence, was synthesized in an identical fashion. Sera from two rabbits obtained up to 18 wk after initial antigen injections with rSP-C^24-35 failed to identify native mature SP-C (data not shown).

View larger version (53K): [in this window] [in a new window]   Figure 2.   Epitope specificity of polyclonal antisera. (A) Western blot analysis. A total of 5 µg each of MATC2, rSP-C2-9, rSP-C11-23 peptides, or human alveolar proteinosis material (PAP) were subjected to 16.5% Tricine/SDS-PAGE under reducing conditions and transferred to nitrocellulose. Immunoblotting performed with 1°C anti-NPROSP-C11-23 or anti-NPROSP-C2-9 (1:4,000) and 2° GAR-HRP (1:10,000) was visualized by enhanced chemiluminescence. (B) 35S-labeled full-length and truncated proSP-C forms were produced by in vitro transcription/ translation of pcDNA3 plasmids containing inserts of full-length rat SP-C cDNA (rSP-C1-194) or truncated constructs (rSP-C1-185, rSP-C24-72, rSP-C10-194) using rabbit reticulocyte lysate. Immunoprecipitation of equal amounts of labeled lysates was performed as detailed in MATERIALS AND METHODS. Autoradiographs of 16.5% Tricine/SDS-PAGE of captured proteins were exposed for 3.5 d. Left panel: Anti-NPROSP-C2-9 immunoprecipitated the primary translation product (lane 1) and a COOH truncated form (rSP-C1-185, lane 2) but failed to recognize NH2-terminal deletions (rSP-C24-72 and rSP-C10-194, lanes 3 and 4). Middle panel: Anti-NPROSP-C11-23 recognized rSP-C1-194 (lane 5), rSP-C1-185 (lane 6), and rSP-C10-194 (lane 8), but not rSP-C24-72 (lane 7). Right panel: Immunoprecipitation of each lysate with nonimmune serum (NIS; lanes 11-14). MW: 14C-BRL low molecular weight markers (lane 9); lane 10 is blank.

Epitope specificity of the antibodies was further characterized using native proSP-C protein by immunoprecipitation of wild-type rat proSP-C and of mutant constructs containing COOH- and NH₂-terminal truncations. Sequential in vitro transcription/translation of full-length rSP-C cDNA using TNT reticulocyte lysate containing [³⁵S] met/cys produced a radiolabeled product of M_r 21,000 not seen in reactions omitting plasmid DNA (not shown). Similarly, pcDNA3 constructs containing mutant forms rSP-C^1-185, rSP-C^10-194, and rSP-C^24-72 expressed smaller ³⁵S-labeled translation products of predicted M_r (not shown). Immunoprecipitation of lysates containing radiolabeled full-length and mutant forms of proSP-C demonstrated that anti- NPROSP-C^2-9 and anti-NPROSP-C^11-23 specifically recognized the appropriate in vitro translation products in a pattern consistent with epitope specificity (Figure 2B). Both antisera immunoprecipitated wild-type (rSP-C^1-194) and rSP-C^1-185 forms. Neither recognized rSP-C^24-72 due to the absence of the respective amino-terminal epitopes (Figure 2B, lanes 3 and 7). As expected, anti-NPROSP-C^2-9 failed to recognize rSP-C^10-194 (Figure 2B, lane 4), whereas immunoprecipitation with anti-NPROSP-C^11-23 yielded a product of M_r 20,000 (Figure 2B, lane 8). No labeled protein was immunoprecipitated using preimmune serum (Figure 2B, lanes 11-14).

Synthetic Processing of ProSP-C: The NH₂ Terminus is Cleaved in Two Steps

Processing events for proSP-C NH₂ terminus were assessed using radiolabeled type II pneumocytes to identify temporal appearance of specific SP-C intermediates. Freshly isolated type II cells in suspension culture were pulse-labeled with [³⁵S]-cys/met for 30 min, chased, and immunoprecipitated. Using anti-NPROSP-C^11-23 (Figure 3A, left), a pattern of proSP-C₂₁ synthesis, time-dependent generation of a 16-kD intermediate, and subsequent appearance then disappearance of lower molecular weight proSP-C forms was consistently observed. A precursor product relationship was apparent between SP-C_7kD and SP-C_6kD (Figure 3B). Immunoprecipitation patterns of the same cell lysates with anti-NPROSP-C^2-9 also showed the proSP-C primary translation product and time-dependent processing to 16- and 7-kD forms (Figure 3A, right). Importantly, using this antiserum, absence of a 6-kD band was a consistent finding. Together, the immunoprecipitation patterns indicate that processing of the NH₂ flanking domain requires two cleavage steps.

View larger version (40K): [in this window] [in a new window]   Figure 3.   Pulse-chase analysis of synthetic processing of SP-C. Freshly isolated type II cells in suspension culture were starved in cys/met-deficient medium for 1 h, pulsed for 30 min with 35S-Trans-label (150 µCi/ml), and chased for up to 4 h as described in MATERIALS AND METHODS. Lysate samples were immunoprecipitated using NH2-terminal proSP-C antisera and recovered proteins were analyzed by 16.5% Tricine/SDS-PAGE. (A) Left panel: Autoradiograph of immunoprecipitation with anti- NPROSP-C11-23. Right panel: Autoradiograph of immunoprecipitation using anti-NPROSP-C2-9. Lysate samples demonstrate early appearance of primary SP-C21 and subsequent time-dependent appearance of proSP-C intermediates. Arrows denote major forms. MW: position of 14C molecular weight markers; NIS: control immunoprecipitation containing 2-h cell lysate analyzed by substitution of nonimmune serum for proSP-C antisera. (B) Autoradiograph of immunoprecipitation of 35S-labeled cell lysate at chase periods ranging from 15 min to 4 h. Using anti-NPROSP-C11-23, a precursor product relationship between SP-C7 and SP-C6 can be observed (curved arrow).

Characterization of ProSP-C Forms in Lung Subcellular Compartments

Because expression of SP-C propeptides is limited to type II cells, analysis of fractions prepared from whole rat lung reflects the subcellular distribution of proSP-C in type II cells (11). Steady-state levels of SP-C propeptides were evaluated in LBs and microsomes using Western blotting with two proSP-C antisera (Figure 4). In microsomal fractions, both antisera detected the primary translation product (21 kD). In LBs, anti-NPROSP-C^11-23 showed a single band with M_r 6,000 whereas anti-NPROSP-C^2-9 failed to detect any proSP-C form in this fraction. ProSP-C species were not detected in extracellular (surfactant) fractions with either antiserum (not shown). On the basis of epitope specificity, this pattern indicates the presence of a LB form of proSP-C consisting of mature SP-C plus residues Glu¹¹-Gln²³.

View larger version (30K): [in this window] [in a new window]   Figure 4.   Western blot analysis of subcellular lung fractions for proSP-C. Microsomes (Mic) and LBs were separated in 16.5% Tricine/SDS-PAGE and probed with proSP-C antisera as indicated. Gel lanes contain 40 µg of microsomes or 15 µg of LBs (total protein). Microsomes contain only a single band at Mr = 21,000 (thick arrow). A 6-kD form (bracket, right) was markedly enriched in the LBs but is recognized by anti-NPROSP-C11-23, not NPROSP-C2-9 (striped arrow).

ProSP-C Remains Membrane-Associated throughout the Secretory Pathway

Using the method of Fujiki and associates (25), subcellular lung fractions were analyzed for membrane-associated proSP-C. Under control conditions, Western blotting of membrane pellets from microsomes were found to contain the 21-kD proSP-C primary translation product (Figure 5A, lane 1) which remained integrally associated with the membrane despite treatment with NaCO₃ (Figure 5A, lane 2). Likewise, the SP-C₆ intermediate, which completely partitioned with the LB membrane (Figure 5C, lanes 1 and 2), was also resistant to NaCO₃ extraction. As a positive control, proSP-C forms in both membrane fractions were extractable by inclusion of the nonionic detergent NP-40 in the buffer. These results demonstrate that during synthetic processing, proSP-C behaves as an integral membrane protein in both early and late compartments of the type II cell biosynthetic pathway.

View larger version (42K): [in this window] [in a new window]   Figure 5.   ProSP-C is an integral membrane protein. Membrane association assays were performed and analyzed by Western blot analysis using anti-NPROSP-C11-23 as described in MATERIALS AND METHODS. (A) A single 21-kD band is present in a microsomal membrane pellet prepared by centrifugation of 200 µg of whole microsomes at 100,000 × g × 1 h (lane 1) and is not extractable with 0.1 M NaCO3 (lane 3). Treatment with 1% NP-40 removed almost all proSP-C21 from membrane pellets (lane 2). (B) Identical analysis showing pellet fractions (lanes 1-3) and supernatants (lanes 4-6) after centrifugation of 70-µg LBs at 100,000 × g in the absence (MEM) or presence of 0.1 M NaCO3 or 0.1% NP-40. The 6-kD form partitions completely with the LB membrane (lane 1) and is not removed with NaCO3 treatment (lane 2), but is entirely extractable into the supernatant by NP-40 (lane 6).

Localization of ProSP-C Forms in Lung Subcellular Compartments

The spatial expression of proSP-C forms within the secretory pathway of type II cells was defined using immunogold electron microscopy (Figure 6). Staining of ultrathin lung sections with anti-NPROSP-C^11-23 identified proSP-C in Golgi, within mvb, and at the limiting membrane of LBs (Figure 6a). Although most of the LB contents were extracted by the processing method, proSP-C labeling of the residual LB matrix was also observed (Figure 6b, arrowhead). In contrast, anti-NPROSP-C^2-9 labeling in LB was consistently absent, indicating that the NH₂ epitope Met¹-Leu⁹ was removed earlier (Figure 6c). Coupled with Western blot analysis (Figure 4), these results demonstrate that cleavage of NH₂ terminus of proSP-C is a two-step process occurring late in the secretory pathway in separate compartments of the type II cell, and that final processing of proSP-C occurs in LBs before exocytosis.

View larger version (143K): [in this window] [in a new window]   View larger version (111K): [in this window] [in a new window]   View larger version (118K): [in this window] [in a new window]   Figure 6.   Ultrastructural immunocytochemical localization of SPs in whole rat lung. Immunogold electron microscopy performed on ultrathin rat lung sections embedded in London Resin White methacrylic resin labeled with either primary anti-NPROSP-C11-23 (a and b) or anti-NPROSP-C2-9 (c) polyclonal SP-C antisera as described in MATERIALS AND METHODS. (a) Gold particles, 20 nm, indicative of immunoreactivity of Golgi and vesicles (g), mvb (arrows), and LB membrane (arrowhead) with anti-NPROSP-C11-23; original magnification, ×34,500; scale bar = 500 nm; nucleus (n). (b) At higher magnification (×64,500), anti-NPROSP-C11-23 labeled internal vesicles of mvb (arrow) and the protein core of LB that remained after extraction (arrowhead); 10-nm gold particles; scale bar = 500 nm. (c) A multivesicular body immunoreactive for anti-NPROSP-C2-9 (arrow) and a LB core devoid of labeling (arrowhead); 10-nm gold particles; scale bar = 500 nm. Original magnification, ×48,000. Pictures are representative of 16 to 20 grids prepared with each primary antibody.

The NH₂ Flanking Domain Contains a Targeting Motif

To investigate the role of the NH₂ flanking peptide, targeting of EGFP fusion proteins containing truncations of the proSP-C NH₂ and COOH flanking peptides was assessed by transfection into A549 cells. Within 48 h of introduction of plasmid cDNA, expression of EGFP-C₁ and EGFP-C₁/ SP-C^1-194 fusion proteins was readily detected in transiently transfected cells. EGFP-C₁ was distributed in a diffuse pattern with fluorescence throughout the cell (not shown). In contrast, EGFP-C₁/SP-C^1-194 was spatially expressed in cytoplasmic vesicles (Figures 7A and 7E) and partially colocalized with CD-63 (Figures 7G and 7H). Importantly, expression of EGFP-C₁/SP-C^1-194 was not observed on the plasma membrane and failed to colocalize with EEA-1, a marker of early endosomes (Figures 7B- 7D). Vesicles containing EGFP-C₁/SP-C^1-194 expression also failed to stain for ubiquitin (data not shown).

View larger version (35K): [in this window] [in a new window]   Figure 7.   EGFP/SP-C1-194 chimeric protein is directed to EEA-1-negative, CD 63-positive vesicles. A549 cells grown on glass coverslips at 80% confluence were transfected with 10 µg of EGFP-C1/SP-C1-194 using CaPO4 as described in MATERIALS AND METHODS. At 48 h after transfection, cells were fixed, permeabilized, and stained with either primary polyclonal EEA-1 antiserum (A-D) or primary monoclonal anti-CD-63 antiserum (E-H) and IgG-specific secondary Texas Red-labeled antisera. Images were acquired by video fluorescence microscopy using a High Q FITC filter (Ex = 480 nm; Em = 535/550 nm) and a Texas Red filter package (Ex = 560/555 nm; Em = 645/675 nm) and are representative of three separate experiments and > 50 cells for each construct: (A) endogenous EGFP/SP-C1-194 fluorescence visualized via FITC channel; (B) phase image; (C) EEA-1 staining visualized with Texas Red channel; (D) double-label color image assembled with green for EGFP and red for EEA-1, demonstrating lack of colocalization; (E) endogenous EGFP/SP-C1-194 fluorescence visualized via FITC channel; (F) corresponding phase image; (G) CD-63 staining visualized with Texas Red filter; (H) double-label color image assembled with green for EGFP and red for CD-63, showing partial colocalization of CD-63 and proSP-C.

Deletion analysis was then used to identify targeting domains within the proSP-C NH₂ terminus. Removal of the entire NH₂ flanking domain (EGFP-C₁/SP-C^24-194) resulted in continued translocation of fusion protein from the cytoplasm but retention of expressed mutant protein in early compartments (Figure 8A). EGFP fluorescence colocalized with anti-calnexin, indicating that EGFP-C₁/SP-C^24-194 was restricted to the ER (Figure 8B). Although retained in the ER, EGFP-C₁/SP-C^24-194 fusion protein was not ubiquinated (Figures 8C and 8D). In control experiments using a construct containing a deletion of the entire COOH flanking domain (EGFP-C₁/SP-C^1-58), transfection resulted in expression of EGFP fluorescence in cytoplasmic vesicles, indicating that the COOH flanking peptide was not required for trafficking (Figures 9A and 9E).

View larger version (87K): [in this window] [in a new window]   Figure 8.   EGFP/SP-C24-194 chimera is retained in ER and is non-ubiquinated. A549 cells grown on glass coverslips at 80% confluence were transfected with 10 µg of EGFP-C1/SP-C24-194 using CaPO4 as described in MATERIALS AND METHODS. (A and C). At 48 h after transfection, cells were fixed, permeabilized, and stained with either primary polyclonal calnexin NT antisera (B) or anti-ubiquitin antisera (D) and secondary Texas Red-labeled antibody. Sequential images acquired by digital fluorescence microscopy using a High Q FITC filter (Ex = 480 nm; Em = 535/ 550 nm) (A and C) and a Texas Red filter package (Ex = 560/555 nm; Em = 645/675 nm) (B and D) are representative of over 50 cells for each construct.

View larger version (83K): [in this window] [in a new window]   Figure 9.   The amino terminus of proSP-C contains targeting motifs for export from the Golgi. A549 cells at 80% confluence were transfected with 10 µg each of either EGFP-C1/SP-C1-194 (A), EGFP-C1/SP-C10-194 (B), EGFP-C1/SP-C15-194 (C), EGFP-C1/SP-C19-194 (D), EGFP-C1/SP-C10-18 (E), or EGFP-C1/ SP-C1-58 (F) using CaPO4 as described in MATERIALS AND METHODS. At 48 h after transfection, images were acquired by digital fluorescence microscopy using a High Q FITC filter (Ex = 480 nm; Em = 535/550 nm) and are representative of over 50 cells for each construct.

When the region Met¹⁰-Gln²³ was reconstituted with EGFP-C₁/SP-C^24-194 resulting in the construct EGFP-C₁/ SP-C^10-194, the fusion product was exported to cytoplasmic vesicles (Figure 9B). Similar results were obtained using EG- FP-C₁/SP-C^15-194 (Figure 9C) but not EGFP-C₁/SP-C^19-194, which was distributed predominantly in an ER location (Figure 9D). Next, we studied the effect of selective deletion of the region Met¹⁰-Thr¹⁸ from the NH₂ flanking peptide on trafficking. The resulting construct, EGFP/SP-C^10-18, contained 14 residues (Met¹-Asp⁹/Gly¹⁹-Gln²³) as a natural "spacer" between EGFP and the mature SP-C domain. When transfected, EGFP/SP-C^10-18 was retained in the ER (Figure 9E). Together, results using these fusion constructs indicate that within the vestigial NH₂ propeptide flanking domain, a functional targeting motif for direction of proSP-C from the proximal compartments can be localized to residues Met¹⁰-Thr¹⁸.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Although its extreme hydrophobicity and avidity for lipid membranes impart properties important for the biophysical activity of lung surfactant, SP-C represents a structurally and functionally challenging protein that must be synthesized and trafficked through the regulated secretory pathway of the type II cell (3). In this study, epitope-specific proSP-C antisera, immunogold electron microscopy, metabolic labeling, and chimeric fusion proteins were used to further elucidate the biosynthetic processing of rat proSP-C. Previous studies from this laboratory and others have defined early events in targeting and processing of the SP-C primary translation product leading to the development of a model pathway for the intracellular synthesis and processing that includes two initial COOH-terminal cleavages which occur after export of the primary translation product from the Golgi (7, 8, 10). The current report extends prior findings with new data demonstrating that, after removal of its COOH flanking domain, the subsequent processing of the proSP-C NH₂ terminus is also a multistep process that requires two proteolytic cleavages that occur in distinct compartments of the distal secretory pathway. Additionally, ultrastructural studies and subcellular fractionation place a low molecular weight proSP-C intermediate (M_r 6,000) in the LB, indicating that generation of mature SP-C_3.7 via final cleavage of a NH₂ vestigial peptide fragment takes place in this compartment as the terminal event before LB exocytosis. Further, functional data using chimeric EGFP fusion proteins shows that, within the NH₂ flanking domain (M¹-Q²³), a minimal motif mediating trafficking to cytoplasmic vesicular compartments resides within the region M¹⁰-T¹⁸.

Using immunogold electron microscopic staining with anti-NPROSP-C^11-23, we identified proSP-C in both mvb and in LBs of type II cells (Figures 6a and 6b). Previous ultrastructural localization of proSP-C using ultrathin cryomicrotomy and immunogold labeling with polyclonal antisera labeled the ER, Golgi, and transport vesicles of the late secretory pathway (including mvb) but failed to detect proSP-C in LB (12, 14, 28). The differences in LB staining between these studies and our results may be due to differences either in tissue sample preparation and/or in characteristics of the antisera used. The antisera used in other studies were generated using recombinant proSP-C proteins (12, 14, 28) and may not recognize epitopes present in LB forms of proSP-C. Also, in the earlier studies, frozen sections prepared by ultrathin cryomicrotomy preserved antigenicity but morphologic preservation of LB matrix was suboptimal, raising an issue of extraction of phospholipid and consequently lipid-associated proSP-C during fixation. In our studies, we preserved antigenicity and were also able to retain a limited amount of LB matrix that stained for proSP-C (Figure 6b).

The differential labeling with two antisera, as indicated by the failure to detect proSP-C in LBs with anti- NPROSP-C^2-9, is consistent with cleavage of the M¹-L⁹ domain before packaging in the LB. Importantly, isolated LB fractions probed with the two antisera yielded similar differences in recognition of proSP-C forms (Figure 4). Whereas immunoblots with anti-NPROSP-C^11-23 of purified LBs consistently identified a single band of M_r 6,000, anti-NPROSP-C^2-9 failed to detect any proSP-C forms in this fraction. Failure of anti-NPROSP-C^2-9 also to detect nondenatured proSP-C₆ in immunoprecipitation studies (Figure 3B) eliminates the possibility that this is an artifact of protein conformation. Thus, the 6-kD band is the only band attributable to LBs and indicates that this intermediate is composed of the Glu¹¹-Gln²³ epitope and mature SP-C domain but lacks at least the first nine amino acids. Assignment of exact cleavage sites awaits isolation and sequencing of the intermediates.

Pulse-chase labeling studies analyzed with anti-NPROSP-C^11-23 demonstrated qualitative precursor-product relationships between SP-C₂₁ and SP-C₁₆, SP-C₁₆ and SP-C₇, and SP-C₇ and SP-C₆, confirming that four cleavages (two COOH and two NH₂) must occur (Figure 3). An additional band (M_r 9,000) identified in immunoprecipitates with both NH₂-terminal proSP-C antisera, appeared after the formation of proSP-C₇, was not observed with preimmune serum, and did not demonstrate a precursor-product relationship with SP-C₁₆. Thus, this form does not appear to correspond to a specific cleavage event and also cannot represent a COOH-terminal fragment. Because of the hydrophobicity of smaller proSP-C intermediates, this likely represents a non-disulfide mediated combination of proSP-C₇ and other peptides such as mature SP-C_3.7 or an unrelated protein dimerized (heterodimer) or aggregated during sample preparation.

Using transfected A549 cells, wild-type proSP-C was shown to target EGFP fusion protein to cytoplasmic vesicles. Previously, using vital stains, we had shown that these organelles are Texas Red-ceramide negative (non-Golgi), acidic vesicles (10). Further characterization of this compartment using immunolocalization reveals that these vesicles are also EEA-1-negative but that a portion of this population is CD-63-positive (Figure 8). Together, colocalization studies of EGFP/SP-C^1-194 expression demonstrate that the export of proSP-C from proximal compartments to CD-63-positive, acidic compartments of A549 cells does not appear to involve indirect routing to the cell surface and early endocytic pathway but rather uses "direct" targeting to an acidic, lysosome-related, cytosolic compartment. Previous ultrastructural, immunochemical, and biochemical studies have demonstrated that in type II cells, the LB is a CD-63-positive (28), lysosome-associated membrane protein (LAMP)-1-positive (29), acidic vesicle (24) that also contains lysosomal hydrolases. Thus, the LB falls into the category of a "lysosome-related organelle" and appears to functionally resemble similar organelles found in other cells, including melanosomes, basophilic granules, platelet-dense granules, etc. This concept has recently been reviewed by Dell'angelica and associates (30). Because of difficulties both in maintaining cell phenotype and in efficient transfection of primary adult type II cells, the current study used A549 cells as a surrogate cell for evaluation of trafficking motifs required for export of proSP-C from the Golgi. Originally derived from a type II cell, the data presented here indicate that although lacking true lamellated organelles, the A549 cell appears to contain similar lysosome-related organelles. In addition, we have previously shown that targeting of transfected wild-type SP-C to these vesicles results in an orderly proteolytic pattern of precursors-products, demonstrating that this compartment is biosynthetic rather than degradative (10). Therefore, the current approach of using A549 cells to study SP-C processing appears comparable to similar strategies used by others in the evaluation of trafficking motifs of non-surfactant related proteins to other lysosomal-like organelles in non-lung cell systems (30).

In addition to undergoing processing throughout the secretory pathway, membrane association assays using NaCO₃ show that proSP-C intermediates are also integrally associated with type II cell membranes (Figure 5). From the cDNA sequence, modeling of the secondary structure of proSP-C (M¹-I¹⁹⁴) had indicated the polyvaline stretch contained within the hydrophobic domain of the proprotein primary sequence of L³⁶-L⁵⁸ to be highly -helical and capable of spanning phospholipid bilayers in the liquid crystalline phase (31). ProSP-C₂₁ detected in purified lung microsomal membranes was resistant to removal by NaCO₃ treatment but extractable with non-ionic detergents (Figure 5A). This is the first demonstration that proSP-C₂₁ is integrally associated with native subcellular lung membranes and extends earlier in situ studies by both Keller and colleagues (32) and Vorbroker and associates (13), in which translation of recombinant wild-type proSP-C in the presence of canine pancreatic microsomes resulted in membrane integration. Of further importance is the observation that other intermediates, including the 6-kD LB proSP-C form (Figure 5B), were also resistant to NaCO₃ extraction. Taken in toto, the data demonstrate that proSP-C₂₁ as well as proSP-C intermediates likely behave as bitopic transmembrane proteins that remain associated with intracellular membranes during trafficking and proteolytic cleavage. In contrast, prohormones as well as other secretory proteins are typically translocated to the ER and subjected to proteolytic processing within the lumen of early and/or late vesicular organelles (33). In this regard, SP-C appears to be unique among secreted proteins.

Thus, although a secreted protein, the combination of membrane association and direct trafficking to acidic, post-Golgi compartments requires consideration that the biosynthesis of proSP-C is also analogous to trafficking of bitopic membrane proteins such as LAMP-1 and LAMP-2 (34). Like proSP-C, both LAMP-1 and LAMP-2 are bitopic membrane proteins. Structurally, they differ from proSP-C in that they are oriented in a type III fashion with a short cytosolic COOH domain and large intralumenal NH₂ terminus (34). In type II cells, LAMP-1 (29), as well as proSP-C (Figure 6b), appears to be sequestered in a multivesicular body intermediate compartment and in LBs. In other cell types, it has been shown that the preferred pathway by which LAMP-1 and -2 reach the lysosome is via direct trafficking from the trans-Golgi network through late endosomes which is mediated by a tyrosine-based sorting signal containing a COOH-terminal hydrophobic residue in the cytosolic tail (34, 35). In contrast to LAMP-1 and -2, modeling studies as well as in situ data indicate that the transmembrane orientation of proSP-C₂₁ is type II, resulting in the placement of the short, 23-amino acid, NH₂ flanking domain in the cytoplasm (Figure 10). Membrane integration assays by Keller and associates (32) have shown that the NH₂ terminus of proSP-C is positioned in the cytoplasmic side of ER membranes, making it a likely candidate to contain a targeting motif to facilitate export to distal compartments. In addition, on the basis of data in this report, it is this region that is the final flanking domain to be removed from the mature SP-C protein during biosynthesis. Using the A549 cell model, we have shown that targeting of proSP-C is dependent upon a motif residing within the NH₂-terminal propeptides (M¹⁰-T¹⁸; see Figure 9). The sequence and position of the NH₂-terminal targeting domain for rat proSP-C is illustrated in Figure 10 and demonstrates the 100% homology between the rat and human isoforms which also exists among other species (4). Although the region M¹⁰-T¹⁸ is highly conserved, BLAST search failed to identify similarity to known targeting domains such as dileucine or tyrosine-based sorting signals. Thus, this domain appears to contain a novel targeting motif.

View larger version (32K): [in this window] [in a new window]   Figure 10.   Structural and spatial localization of proSP-C targeting domain. (Top) Schematic diagram structure of rat proSP-C21, which contains an -helical domain within the mature peptide flanked on both sides by propeptide regions. Boxed region highlights the NH2-terminal targeting domain D11- S17 that is 100% homologous with the human isoform (4). (Bottom) On the basis of data from Keller and colleagues (32), when translocated, the primary translation product is vectorially inserted into the ER membrane in a type II orientation (NH2 in cytoplasm), allowing for potential interaction of the NH2 region with cytosolic components. The COOH flanking peptide remains in the ER lumen and can undergo disulfide-mediated folding.

In summary, the data presented here demonstrate that NH₂-terminal processing of proSP-C is a multistep event in which a functional targeting motif in the NH₂ flanking domain is contained within a previously unrecognized processing intermediate. The results obtained indicate that proSP-C represents a hybrid molecule in which post-translational processing incorporates features both of lumenal proteins such as SP-B and of bitopic lysosomal membrane proteins such as LAMP-1 and -2. Like proSP-C, the processing of proSP-B is a multistep process including two separate cleavages of NH₂-terminal propeptide (19); however, proSP-B appears to be initiated and completed earlier in the biosynthetic proSP-C, beginning in the medial Golgi and completed before LB packaging (36). As for proSP-C, in vitro data demonstrate that the NH₂ terminus of proSP-B is absolutely required for correct intracellular routing (20, 21); recently, a putative targeting motif contained within the 1-kD vestigial sequence of the NH₂ flanking peptide of proSP-B which is cleaved in later compartments has been described (37). Thus, a basic paradigm for cleavage of both hydrophobic surfactant proproteins has emerged; however, additional questionssuch as the enzymes responsible for the proteolytic events, the regulation of the processing and targeting, and any potential interaction of the two proteins in the secretory pathwayremain to be clarified.