Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Pulmonary surfactant performs several essential functions in the lung, is composed of lipids and proteins that coat the alveolar epithelium, and is synthesized by lung epithelial type II cells. Due to its biophysical properties, it plays an important role in lowering the surface tension at the air-water interface of the respiratory epithelium, thus preventing alveolar collapse at end-expiration (1, 2). Furthermore, by interacting with inhaled pathogens, the surfactant system participates in pulmonary host defense. Important components of this innate defense system are the surfactant proteins A and D (SP-A and SP-D) (3), which are synthesized and secreted by type II cells and nonciliated bronchiolar Clara cells in the lung (6). These glycoproteins are members of the collagenous Ca²⁺-dependent lectin family termed "collectins," which also includes the serum proteins mannose binding lectin, conglutinin, CL-43, and the recently reported intracellular collectin CL-L1 (7).

The collectins are characterized by polypeptide chains that are composed of an N-terminal cysteine-rich region and a collagen-like region linked to a carbohydrate recognition domain (CRD) via an -helical coiled-coil "neck" region. The collagen-like domains of three polypeptide chains are folded into a collagen-like triple helix, a process which is initiated by the neck region in the polypeptide chain. The trimeric subunits can associate into higher order oligomers which are stabilized via interchain disulfide bonds between the N-terminal domains. SP-D and conglutinin are organized as tetramers of these collagenous trimers, generating dodecameric cruciform structures (8, 9), whereas SP-A and mannose-binding lectin are found as octadecamers (hexamers of trimers) resembling C1q. SP-D can also form higher order oligomeric "fuzzy ball" complexes (10).

SP-D, like the other collectins, binds to glycoconjugates expressed on the surface of microorganisms in a Ca²⁺- dependent manner. Pathogens recognized by SP-D include bacteria, viruses, yeasts, and fungi (11). Aggregation of these pathogens can lead to receptor-mediated clearance by phagocytes followed by killing (12, 13). Recent data suggest that SP-D may affect the growth of microorganisms (14) and exhibits antiallergic properties (15). SP-D can also act as a scavenger for lipopolysaccharide (16) and modulates the inflammatory response (17, 18). SP-D gene-knockout studies in mice indicate that SP-D might also be important in surfactant homeostasis (19).

Characterization of the surfactant protein components might contribute to a better understanding of the protective roles that these proteins display and makes it possible to study their functional properties in animal models of lung infectious diseases (20, 21). Respiratory infections are not only an important human health problem, but are also regarded as the most important health problem in modern swine production (22). Because the humoral defense system in young animals is not yet fully developed, nonclonal defense systems could be important in the first line defense against pathogens, especially during early development. Considering the potential role for SP-D in porcine pulmonary defense, studies were initiated to characterize porcine SP-D (pSP-D).

Previously, we have deduced the full-length cDNA sequence (23). The predicted protein sequence contains three unique features that are not present in the SP-D of other species previously characterized, including human (24), cow (25), rat (26), and mouse (27). These differences include an extra cysteine in the collagen region, an insertion of three amino acids in the CRD, and the presence of a potential N-glycosylation site in the CRD (Figure 1 and Ref. 23). The consequences of these differences for both the structural properties of pSP-D and binding characteristics of the lectin domain, however, are not yet known.

View larger version (18K): [in this window] [in a new window]   Figure 1.   Alignment of porcine and human SP-D predicted protein sequence. Amino acids are numbered according to the pSP-D sequence, including its leader sequence (lower case). Dots in the hSP-D sequence indicate identical residues compared with pSP-D. N-glycosylation sequence motifs are underlined. Important sequence features only found in pSP-D are in bold. Amino acid sequences obtained after N-terminal amino acid sequence analysis on various SP-D derived digestion products are in italics above (pSP-D derived) or below (hSP-D derived) the full length SP-D sequences. Asterisk indicates the presence of a hydroxylated proline residue; arrow indicates the presence of a modified residue as determined by N-terminal amino acid sequence analysis on the tryptic fragment. Genbank database accession numbers: AF132496 (porcine SP-D) and X65018 (human SP-D).

This paper describes the isolation and characterization of SP-D from porcine bronchoalveolar lavage and reveals that pSP-D migrates as a 50-kD monomer on reducing sodium dodecyl sulfate/polyacrylamide gel electrophoresis (SDS-PAGE), which is 7 kD larger than human and rat SP-D. The difference in size is caused by the presence of a highly sialylated Asn-linked complex type oligosaccharide in the CRD. The carbohydrate moiety was also shown to enhance the hemagglutination (HA)-inhibiting activity of pSP-D against influenza A virus (IAV).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Purification of Porcine and Human SP-D from Bronchoalveolar Lavage Fluid

Porcine SP-D was purified using a modified protocol described for the isolation of human SP-D (24). Lungs from 12- to 16-wk-old pigs (ID-Lelystad, Lelystad, The Netherlands) were washed three times with 15 ml lavage buffer/kg body weight. Lavage buffer consisted of 20 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 10 mM ethylenediaminetetraacetic acid (EDTA) and was kept at 4°C. EDTA was added to solubilize any aggregated SP-D. After 1 h on ice, the bronchoalveolar lavage (BAL) fluid (BALF) was centrifuged at 150 × g for 10 min to remove cells. After removal of the surfactant fraction by centrifugation (20,000 × g for 2 h, 4°C), Tris-HCl (pH 7.4) was added to a final concentration of 50 mM, and CaCl_2, Tween-80, sodium azide, and phenylmethylsulfonylfluoride were added to final concentrations of 15 mM, 0.1% (vol/vol), 0.02% (wt/vol), and 0.1 mM, respectively. The pH was then readjusted to 7.4. Mannan-sepharose (Sigma, St. Louis, MO), equilibrated in 50 mM Tris-HCl (pH 7.4), 5 mM CaCl₂, and 0.05% (vol/vol) Tween-80, added to the recalcified BAL (1 ml bed volume of mannan-sepharose per liter BALF), and the mixture was stirred overnight at 4°C. Sepharose beads were packed into a column (8 × 30 mm) and washed with five bed volumes washing buffer (50 mM Tris-HCl [pH 7.4], 5 mM CaCl₂, 500 mM NaCl, 0.05% Tween-80). This washing procedure was repeated using washing buffer without Tween-80. The SP-D, bound to the mannan in a Ca²⁺-dependent manner, was eluted with a 50-mM Tris-HCl (pH 7.4) buffer containing 5 mM EDTA. SP-D-positive fractions were pooled and further purified by gel filtration chromatography using a Bio-Prep 1,000/17 column (8 × 300 mm, separation range: 10 000-1 000 000; Bio-Rad Laboratories, Hercules, CA). After equilibration with 20 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 5 mM EDTA (flow rate: 0.4 ml/min), 1 ml of affinity-purified SP-D eluate (0.2-0.4 mg/ml) was applied to the column. Fractions obtained were analyzed by SDS-PAGE and Western blot analysis. Results showed that SP-D eluted as a discrete peak at the void volume of the column. Pooled fractions containing purified pSP-D were stored in aliquots at 20°C. To identify purified SP-D and the major peptides obtained after degradation of SP-D by the enzymatic treatments described, amino acid sequencing was performed on samples which were, after separation by SDS-PAGE, blotted onto polyvinylidene difluoride membrane (Merck, Darmstadt, Germany) and detected by Coomassie Blue staining. Sequence analysis using Edman degradation was performed on the Perkin Elmer/Applied Biosystems Protein Sequencing system (type 476A; Perkin Elmer/Applied Biosystems, Shelton, CT).

Human SP-D was isolated from BALF collected from alveolar proteinosis patients and was purified by the same procedure used to purify SP-D from porcine lung lavage, except that mannan-sepharose was replaced by maltosyl-agarose.

Electrophoresis and Western Blot Analysis

Proteins (0.1-1 µg/lane) were analyzed by SDS-PAGE as described by Laemmli (28) using 9% (wt/vol) or 11% (wt/vol) polyacrylamide gels. Protein bands were visualized by silver staining. For immunoblot analysis or detection of glycoconjugates, the proteins were transferred electrophoretically from the gels onto nitrocellulose membrane. Immunostaining was performed using polyclonal antibodies raised in rabbit against rat SP-D, pSP-D, or hSP-D. The anti-porcine SP-D antibody was affinity purified using a Sepharose-porcine SP-D column. Primary antibodies were detected by horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG) (Nordic Immunological Laboratories, Tilburg, The Netherlands). The presence of glycoconjugates was detected by DIG Glycan staining (Roche Diagnostics GmbH, Mannheim, Germany).

Ultrastructural Analysis

Low-angle rotary shadowing of proteins was performed as adapted from Shotton and coworkers (29) and Tyler and Branton (30). Samples were prepared using the method as described by Fowler and Erickson (31). Briefly, a pSP-D solution of ~ 0.1 mg/ ml was dialyzed against 0.1 M ammonium formate (pH 7.4) and mixed with an equal volume of glycerol. This mixture was sprayed onto freshly cleaved mica and dried in vacuo (0.1-1 mPa) and rotary-shadowed with platinum/tungsten at an angle of 7°. Carbon deposition was performed at an angle of 90°. Replicas were examined in a Philips CM10 electron microscope (Philips, Eindhoven, The Netherlands) and photographed at a nominal magnification of ×150,000.

Enzymatic Digestion

CollagenasePurified porcine or human SP-D was incubated with collagenase (type VII from Clostridium histolyticum, Sigma; 1 U/2 µg SP-D) in 20 mM Tris-HCl (pH 7.2), 150 mM NaCl, and 10 mM CaCl₂ for 16 h at 37°C. After incubation, the collagenase-resistant fragment (CRF) derived from pSP-D (pSP-D-CRF) or hSP-D (hSP-D-CRF) was purified by adding mannan-sepharose or maltose-agarose, respectively, to the incubation mixture (0.5 ml bed volume per 0.5 ml incubation mixture). The matrix-bound CRF fraction was isolated as described for the isolation of intact pSP-D, with the exception of gel filtration chromatography. Purified CRF, after N-terminal sequencing, was aliquoted, lyophilized, and stored in aliquots at 20°C.

Endo-/N-glycosidase FN-deglycosylation was performed using a mixture of Endoglycosidase-F and N-Glycosidase-F (NGF; from Flavobacterium meningosepticum, Roche Diagnostics GmbH). Purified SP-D (10 µg/15 µl) or CRF was mixed with an equal volume of 0.1 M sodium phosphate (pH 7.0), 50 mM EDTA, 0.1% (vol/vol) Triton X-100, and 1 mM 2-mercaptoethanol, followed by the addition of 0.5 U enzyme mixture. After incubation at 37°C for 16 h, the sample was lyophilized and stored at 20°C.

TrypsinTo obtain the N-glycosylated peptide from the CRD of pSP-D, 100 µg of purified pSP-D was digested with 1 µg trypsin (modified, sequence grade; Roche Diagnostics GmbH) in a final volume of 0.5 ml 100 mM Tris-HCl (pH 8.5). The reaction was stopped after 16 h by adding 50 µl 1% (vol/vol) trifluoroacetic acid, and the tryptic fragments were separated by reverse phase high performance liquid chromatography as previously described for rat SP-D (32) using an RP-C8 column (250 × 4 mm, lichrosphere 10 µm; Merck). Bound components were eluted with a linear gradient of 0-100% (vol/vol) acetonitrile in 0.1% (vol/vol) trifluoroacetic acid. Fractions were analyzed by SDS-PAGE, and the presence of carbohydrates was analyzed by dot blotting followed by carbohydrate staining using DIG Glycan staining. Glycan-positive fractions were lyophilized and subjected to N-terminal amino acid sequencing.

Fluorophore-Assisted Carbohydrate Electrophoresis Analysis of the Asn³²³-Linked Oligosaccharide

The N-linked carbohydrate moiety present in the CRD of pSP-D was analyzed by Glyko, Inc. (Novato, CA) using fluorophore- assisted carbohydrate electrophoresis (FACE) analysis by submitting the purified oligosaccharide to various glycosidase digestions. The Asn³²³-linked oligosaccharide chains were released from pSP-D-CRF by N-Glycanase treatment (Glyko, Inc.) and, after removal of the protein by ethanol precipitation, labeled at their reducing termini with 8-aminonaphtalene-1,3,6-trisulfonate (ANTS). The glycan chain structures were studied by performing enzymatic digestions on the ANTS-labeled oligosaccharides using the following glycosidases: endoglycosidase H, sialidase, and -galactosidase. After electrophoresis of the digestion products on a PAGE Oligo Profiling Gel, alongside an ANTS-labeled glucose polymer marker as a measure for the degree of polymerization (DP) of the oligosaccharide, the gel images were analyzed using the FACE imager SE2000 and software.

Isoelectric Focusing

Two-dimensional resolution of proteins was performed by isoelectric focusing in the first dimension using Ampholine PAG precast gels (pH range 3.5-9.5; Amersham Pharmacia Biotech Europe GmbH, Freiburg, Germany) or Immobiline DryStrips, pH 3-10 linear (Amersham Pharmacia Biotech). When using Ampholine gels, dialyzed and lyophilized samples were dissolved in 7 M urea/2% (vol/vol) 2-mercaptoethanol and focused at 50 mA for 1.5 h. Samples for Immobiline DryStrips were dissolved in rehydration buffer containing 8 M urea, 2% (wt/vol) 3-(3-cholamidopropyl)diethyl-ammonio-1 propane sulfonate, 2% (vol/vol) immobilized pH gradients buffer pH 3-10L (carrier ampholyte mixture), and 3 mg/ml dithiothreitol and applied to the strips. After reswelling for 16 h the strips were transferred to the Multiphor II IEF system (Amersham Pharmacia Biotech) and focused at increasing voltage: 30 min 150 V, 1 h 300 V, 1 h 1,500 V, and 1 h 3,000 V. Analysis in the second dimension was performed by SDS-PAGE according to Laemmli (28) using 11% (wt/vol) acrylamide gels.

Anion Exchange Chromatography

Purified SP-D was applied to a 5 mm × 5 cm mono-Q anion exchange column (Amersham Pharmacia Biotech). After an initial wash with 10 ml 10 mM Tris-HCl (pH 8.0), 0.1 mM EDTA and 0.1% (wt/vol) octylglucopyranoside (Sigma), the column was eluted with a nonlinear NaCl gradient at a flow rate of 1.0 ml/ min. Elution was monitored using UV light absorption at 280 nm. Samples of 1 ml were collected and analyzed using an enzyme-linked SP-D mannan-binding assay. Fractions from the same peaks in the elution pattern were pooled and the pooled fractions were analyzed by 2D IEF/SDS-PAGE and subjected to Western blot as described above, using anti-porcine SP-D IgG.

Enzyme-Linked Mannan Binding Assay

Bovine serum albumin (BSA) conjugated mannan (100 ng/ml in 0.1 M NaHCO₃-buffer, pH 9.6) was coated on a 96-well microtiter plate (Maxisorp; Nunc, Roskilde, Denmark) and after blocking with 3% (wt/vol) nonfat dry milk (Protifar; Nutricia, Zoetermeer, The Netherlands) the wells were washed three times with washing buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.05% [vol/ vol] Tween-80, and 5 mM CaCl₂). Samples were diluted in 1% BSA containing washing buffer and 50 µl was applied to the wells. After incubation for 1 h, plates were washed six times with washing buffer. Mannan-bound SP-D was detected using anti-rat SP-D IgG diluted in washing buffer containing 1% (wt/vol) BSA (50 µl per well). After an additional incubation for 1 h, the wells were washed six times with washing buffer and a horseradish peroxidase-conjugated swine anti-rabbit IgG (Nordic) was applied to the wells for 1 h. The plates were stained using 400 µM tetramethylbenzidine reagent (Merck), 1 mM H₂O₂, 0.1 M citric acid buffer pH 4.0, and the reaction was stopped by adding 50 µl 2 M H₂SO₄. Absorption was measured at 450 nm.

Inhibition of Influenza A Virus Hemagglutination Activity

HA inhibition assays were performed as previously described (33). Influenza A virus strains tested were A/Philippines/82 (H3N2) and its variant A/Philippines/82/BS, both kindly provided by Dr. E. Margot Anders (University of Melbourne, Melbourne, Australia). For HA inhibition studies, nondenatured N-deglycosylated pSP-D was obtained by treating 50 µg purified pSP-D with 5 mU recombinant N-Glycanase (Glyko, Inc.) in a final volume of 200 µl incubation buffer containing 20 mM sodium phosphate pH 7.5 and 0.02% sodium azide. After incubation for 6 h at room temperature, Ca²⁺ was added to a final concentration of 5 mM, followed by addition of 100 µl bed volume of equilibrated mannan-sepharose beads. Washing and elution of the mannan-bound fraction was performed as described for the purification of pSP-D from BALF. Analysis by SDS-PAGE indicated that ~ 90% of the protein was deglycosylated after treatment with N-Glycanase. Sham-treated pSP-D was obtained by performing the same procedure in the absence of N-Glycanase. Recombinant rat SP-D produced in CHO-K1 cells (34), a gracious gift of Dr. E. C. Crouch, was used for comparison. Statistical analysis was performed using the paired Student's t test. See also the legend of Table 1 for experimental details.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of Porcine SP-D

For the purification of pSP-D from lung lavage fluid we used techniques as previously described (24), based upon Ca²⁺-dependent carbohydrate affinity chromatography followed by gel filtration. Immobilized mannan proved to be more efficient in binding pSP-D than maltose or mannose, and in terms of yield a batchwise overnight incubation procedure was preferred over column chromatography. The average yield of the purification procedure was 1.5-2.0 mg pSP-D from lung lavage fluid derived from one animal. Western analysis showed that the mannan-sepharose-extracted lavage supernatant still contained substantial amounts of soluble SP-D. This fraction could not be extracted by repetitive affinity treatment using either immobilized mannan or maltose, indicating the presence of an SP-D fraction that displays a low carbohydrate binding capacity. Elution of the mannan-bound proteins using an EDTA buffer released semipure SP-D; no SP-A contamination was present, although a 30-kD protein coeluted as visualized by SDS-PAGE analysis of the mannan-fraction (Figure 2A, lane 1). Application of the mannan-sepharose eluate to a Bioprep 1,000/17 gel filtration column resulted in SP-D eluting in the void volume of the column, indicating a molecular mass of more than 1,000 kD. Resolution of this column was insufficient to separate dodecameric pSP-D from higher order multimers. No smaller molecular weight forms of SP-D were detected in the other fractions. The void volume fraction contained pure SP-D as judged by SDS-PAGE analysis and silver staining (Figure 2A, lane 2). The minor band of ~ 100 kD is most likely nonreducable dimeric pSP-D and might result from the presence of a high proportion of "fuzzy ball" complexes present in the preparation (35). This was confirmed by Western blot analysis, because both the 50-kD and the 100-kD band of this preparation were specifically detected by the anti-rat SP-D antibody (Figure 2B, lane 1). N-terminal amino acid sequencing (10 amino acids) showed that the 50-kD band was identical to the predicted amino acid sequence from the pSP-D cDNA sequence (Figure 1, residues 21-31), confirming that the isolated protein was indeed mature pSP-D. SDS-PAGE analysis of unreduced pSP-D showed an increased intensity of the dimeric SP-D 100-kD band, whereas a second band of ~ 200 kD in size appeared, probably due to the presence of tetrameric SP-D (Figure 2B, lane 2). To illustrate the size difference between monomeric pSP-D (50 kD) and monomeric hSP-D (43 kD), hSP-D isolated from lung lavage obtained from patients with alveolar proteinosis was also analyzed by reducing SDS-PAGE and Western blotting using rabbit anti-human SP-D IgG (Figure 2B, lane 3).

View larger version (62K): [in this window] [in a new window]   Figure 2.   Purification of pSP-D. (A) Silver-stained reducing SDS-PAGE (9%) of the EDTA-eluted fraction of mannan-bound pSP-D before (lane 1) and after subsequent purification by gel filtration (void volume fraction; lane 2). (B) Western blot analysis of pSP-D (reduced, lane 1; unreduced, lane 2) and hSP-D (reduced, lane 3) using rabbit anti-rat SP-D and rabbit anti-human SP-D antibodies, respectively. pSP-D migrates under reducing conditions as a 50-kD monomer which is larger than hSP-D (43 kD); unreduced pSP-D migrates as two bands of ~ 100 and 200 kD. (C) Ultrastructural analysis of pSP-D. pSP-D isolated from BALF using mannan affinity chromatography and gel filtration was prepared for rotary shadowing electron microscopy as described in MATERIALS AND METHODS. Preparations contain both dodecameric "four-armed" structures and higher order oligomers resulting in "fuzzy ball" complexes. Bar: 200 nm.

Ultrastructural studies on the isolated pSP-D by rotary shadowing showed that most of the pSP-D was present as dodecameric structures and "fuzzy ball" complexes which consisted of up to 10 or more dodecameric SP-D molecules (Figure 2C).

Collagenase and NGF Digestion of SP-D

To investigate whether the size difference between porcine and human SP-D is caused by dissimilarities in the CRD or in the collagen domain, both SP-D species were subjected to a collagenase treatment producing a pSP-D-CRF and a hSP-D-CRF. pSP-D-CRF was analyzed by N-terminal amino acid sequencing and shown to contain only the last two collagen triplet sequences in addition to the neck/CRD (Figure 1, residues 217-222). Collagenase digestion on SP-D from both species also produced minor amounts of an unknown immunoreactive fragment with an apparent molecular mass ~ 3 kD lower than the major digestion product, which was N-terminally sequenced (Figure 3A, lanes 2 and 5). Analysis by SDS-PAGE and immunoblotting (Figure 3A) indicated that pSP-D-CRF (ranging from 20-25 kD; Figure 3A, lane 2) is ~ 5 kD larger in size than hSP-D-CRF (range: 15-20 kD; Figure 3A, lane 5). A subsequent treatment of pSP-D-CRF with NGF resulted in a size shift toward the position of human CRF (Figure 3A, lanes 3 and 5, respectively), indicating that the difference in apparent molecular mass between porcine and human monomeric SP-D is largely due to N-glycosylation of the CRD of pSP-D. Detection of carbohydrates on the same digestion products by glycan staining further supports this finding, because the strong carbohydrate staining of pSP-D-CRF disappears after specific removal of Asn-linked carbohydrate moieties accomplished by NGF digestion (Figure 3B, lanes 2 and 3). In contrast, hSP-D only stained positive before treatment with collagenase (Figure 3B, lane 4), confirming that hSP-D is exclusively glycosylated in the collagen region and not in the neck/ CRD domain (36).

View larger version (30K): [in this window] [in a new window]   Figure 3.   Enzymatic degradation of pSP-D and hSP-D using collagenase and NGF. (A) Control samples and digests were, after separation by reducing SDS-PAGE (11%), transferred to nitrocellulose membrane and immunostained using rabbit anti-porcine SP-D (lanes 1-3) or rabbit anti-human SP-D (lanes 4 and 5). (B) Corresponding samples were also analyzed for the presence of carbohydrates using a DIG glycan detection kit. Samples are: lane 1, pSP-D; lane 2, pSP-D-CRF; lane 3, pSP-D-CRF, NGF treated; lane 4, hSP-D; lane 5, hSP-D-CRF.

Isolation of the Glycosylated Peptide from the CRD of pSP-D

Previous studies on the primary structure of pSP-D by cDNA cloning (23) had revealed that the CRD contains an N-glycosylation sequence motif Asn³²³/Phe³²⁴/Thr³²⁵ (Figure 1) in the middle of a 5-kD trypsin fragment of the CRD. To investigate linkage of a carbohydrate moiety to Asn³²³, purified SP-D was trypsinized and the digest was separated by reverse-phase chromatography. The obtained UV-chromatogram indicated the presence of two overlapping peaks eluting at 40% acetonitrile; SDS-PAGE analysis showed that a peptide fragment of ~ 8 kD was present in one of the peak eluting fractions (data not shown). Dot blotting on nitrocellulose followed by glycan staining indicated that the fraction contained carbohydrate. Protein sequencing showed that the peptide was derived from the CRD of SP-D (Figure 1; residues 310-330) whereas no signal was observed during analysis of residue Asn³²³, which strongly suggested that the residue was modified.

Structural Characterization of the N-glycosylation Linked to Asn³²³

The structure of the glycan linked to residue Asn³²³ located in the CRD of pSP-D was studied by digestion of the fluorescently labeled carbohydrate with different glycosidases followed by separation using a PAGE Oligo sequencing gel (Figure 4). The profile of the intact carbohydrate released after N-glycanase treatment of pSP-D-CRF (Figure 4, lane 3) showed a complex mixture of oligosaccharides migrating very closely at the position between glucose polymers G6 and G8, which indicated that the glycan is highly heterogeneous. Digestion with endoglycosidase H (Figure 4, lane 2) had little or no effect on the profile, which implies that there is a lack of, or very limited amounts of, oligomannose-type oligosaccharides present. After treatment with sialidase (Figure 4, lane 4), all bands shifted toward higher degree of polymerization values (G9-G13) due to the loss of negatively-charged sialic acid monomers from the oligosaccharide causing a decreased electrophoretic mobility. Digestion with both sialidase and -galactosidase resulted in a downward shift of the entire profile compared with digestion in the presence of sialidase only (Figure 4, lane 5). Treatment with -galactosidase alone did not affect the profile compared with that of the control (Figure 4, lane 6). These results indicate that all the released oligosaccharides are terminally sialylated whereas galactose residues are only accessible after sialic acids have been removed. The overall size of the oligosaccharide can be predicted based upon established mobility shifts (in units of degree of polymerization) that can be measured after release of specific monosaccharides from the oligosaccharide. Thus, it was calculated that the molecular weight of the carbohydrate moiety present in the CRD of pSP-D ranges from 3,000 to 3,500. Elucidating the monosaccharide sequence of this glycan using FACE analysis was not possible because the heterogeneity of the carbohydrate makes it impossible to identify where individual bands originate or shift.

View larger version (102K): [in this window] [in a new window]   Figure 4.   Profiling of the Asn323-linked glycan of pSP-D using various glycosidase digestions followed by FACE analysis. pSP-D-CRF was treated with N-Glycanase and the released carbohydrate, after purification and labeling with ANTS, was subjected to different glycosidase digestions. The resulting products were separated using a PAGE Oligo Profiling Gel and visualized by imaging. The released intact carbohydrate (lane 3) was digested with endoglycosidase H (lane 2), sialidase (lane 4), both sialidase and -galactosidase (lane 5), and -galactosidase only (lane 6). The degree of polymerization of the oligosaccharide can be roughly estimated using an ANTS-labeled glucose polymer (lane 1).

Analysis of Differently Charged SP-D Isoforms

Analysis of reduced pSP-D by 2D IEF/SDS-PAGE showed the presence of at least six differently charged isoforms (pI range: 6-9; Figure 5, panel A). After NGF treatment of pSP-D the different isoforms migrated as a single, more basic SP-D mono-isomer (Figure 5, panel B). Two-dimensional analysis of pSP-D-CRF also revealed a charge train although the isoforms were more acidic and the pI range was smaller (pI 4-5.5) compared with that of the isoforms of intact pSP-D (Figure 5, panel C). Again, treatment with NGF led to the disappearance of the pSP-D-CRF charge train and only one, less acidic, isoform was detected (Figure 5, panel D).

View larger version (49K): [in this window] [in a new window]   Figure 5.   Effect of NGF treatment on the differently charged isoforms of pSP-D and pSP-D-CRF. Two-dimensional IEF/SDS-PAGE was performed on pSP-D (A and B) and pSP-D-CRF (C and D) before (A and C) and after release of N-linked carbohydrates by NGF digestion (B and D). After two-dimensional analysis the proteins were Western blotted and immunostained using rabbit anti-porcine SP-D antibodies. A charge train can be observed when analyzing both pSP-D and pSP-D-CRF, although their pI ranges differ. Removal of N-linked glycosylations results in the disappearance of the charge train, leaving a single, increasingly basic isomer to be detected.

In an attempt to isolate differently charged isoforms, we performed anion exchange chromatography on native pSP-D. Results showed that distinct protein peaks could be detected, containing isoforms of oligomeric SP-D that eluted at increasing salt concentration (Figure 6). Analysis of the SP-D isoforms eluting in the first and the fifth peak by 2D IEF/SDS-PAGE showed that these peaks were assembled from a charge train of monomeric isoforms (Figure 6, panel inserts). The first peak contained isoforms with higher pI values (7) compared with those derived from the fifth peak (pI 5-7).

View larger version (26K): [in this window] [in a new window]   Figure 6.   Anion exchange chromatography of oligomeric pSP-D. Mannan affinity purified pSP-D (approx. 0.5 mg) was applied to a mono-Q anion exchange column and eluted with a nonlinear NaCl gradient (broken line). Protein was monitored by measuring the absorbance at 280 nm (solid line) and collected fractions were screened for Ca2+-dependent carbohydrate binding using an enzyme-linked mannan-binding assay (dotted lines). Two of the eluted peaks (peaks I and V) were analyzed by reducing two-dimensional IEF/SDS-PAGE and were, after Western blotting, immuno- stained using a rabbit anti-porcine SP-D antibody. A charge train of SP-D mono-isomers was found in the pooled fractions of both peak I and peak V; the average pI value of the monomeric isoforms decreases as the oligomers assembled from these monomers elute at increasing salt concentration.

Inhibition of Influenza A Virus HA Activity by SP-D

The finding of an N-linked oligosaccharide moiety in the CRD of pSP-D suggested that this moiety might be important for the function of SP-D. Therefore, the functional significance of the N-linked oligosaccharide in the CRD of pSP-D was studied by IAV HA inhibition assays using both the Phil82 and the mutant Phil82/BS strain. This BS variant is resistant to growth inhibition by bovine serum -inhibitor and conglutinin, which results from loss of a single high-mannose oligosaccharide chain, overlying the sialic acid receptor of hemagglutinin (37). The pSP-D preparation shows an increased activity against the Phil82/BS strain as compared with recombinant rat SP-D (Table 1). The inhibitory effect of deglycosylated pSP-D on the HA-acitivity of both Phil82 and Phil82/BS IAV strains was significantly decreased compared with that of sham-treated pSP-D.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this report, we have described the isolation and characterization of pSP-D from porcine lung lavage. The procedure used to purify native pSP-D was similar to the method reported by Lu and coworkers (24) for the isolation of human and bovine SP-D from amniotic fluid and lung lavage, respectively. Ultrastructural analysis showed that the purified SP-D was assembled as dodecamers or multimers (> 8 dodecamers) as seen in SP-D preparations from other species (10, 35). In addition, the molecular dimensions of the collagenous "arms" and the globular "heads" of the assembled molecule matched those calculated from human and rat SP-D preparations.

Reduced pSP-D migrates as a 50-kD monomer on SDS-PAGE, which is 7 kD larger than human and rat SP-D (43 kD). There is a report of a 50-kD variant isolated from human lung lavage which is due to extensive O-glycosylation of the N-terminal region. This anomalous form is only present as SP-D monomers or trimers (38). By contrast, the porcine 50-kD monomer is exclusively assembled into higher ordered multimers. Digestion studies using collagenase and NGF revealed that the higher molecular mass of monomeric pSP-D was predominantly caused by the presence of an N-linked carbohydrate moiety in the CRD. Amino acid sequence analysis of a trypsin fragment from the CRD indicated that the glycan was attached to asparagine 323, as predicted by the cDNA sequence (23). So far, no SP-D of any other animal species was found to be glycosylated in its lectin domain; both O- and N-linked carbohydrates were only found in the collagen region as recently demonstrated by Leth-Larsen and colleagues (36) for human and bovine SP-D. Studies on the structure of the Asn³²³-linked glycan by FACE analysis showed that it is a highly heterogeneous, complex-type oligosaccharide which is highly sialylated. The estimated size range of the glycan (3-3.5 kD) does not completely account for the size difference observed between pSP-D (50 kD) and SP-D from other species (43 kD). This is likely due to inaccuracies in molecular mass determination on SDS-PAGE in which globular proteins are used as molecular mass standards for analyzing elongated collagenous proteins.

Variations in the number of sialic acid residues present in the CRD-glycosylation result in the existence of differently charged isoforms of monomeric pSP-D, which, upon collagenase digestion, yield differently charged pSP-D-CRF isoforms (Figure 5). Calculations (based on the amino acid composition) of the changes in the pI value caused by introduction of a fixed number of additional negative charges (e.g., sialic acids) into full-length pSP-D and pSP-D-CRF, respectively, show a larger change for the former than for the latter. Therefore, the observation that the range of pI values of the various monomeric pSP-D isoforms (Figure 5, panel A) is larger than that of the pSP-D-CRF isoforms (Figure 5, panel C) gives no reason to assume that charge heterogeneity outside the neck/CRD region contributes to the heterogeneity observed for the full-length pSP-D. This is in line with a report that in hSP-D the N-linked glycan present in the collagen domain is not sialylated (36).

Analysis by anion exchange chromatography showed there are over five distinctly charged oligomeric isoforms of pSP-D. Early studies by Persson and associates (32) demonstrated that natural nondenatured rat SP-D could be separated by diethylaminoethyl-chromatography into variants of multimeric SP-D. These were assembled from SP-D monomers that were different in size, although within one multimeric variant only one monomeric isoform could be detected. Hence, for rat SP-D there appears to be a preferential assembly into multimers of similarly modified monomers. However, depending on the overall negative charge of the oligomer, pSP-D appears to be assembled from distinct sets of heterogeneously charged mono-isomers (Figure 6) that are sensitive to NGF treatment (Figure 5). Furthermore, it was shown that the charge differences of monomeric pSP-D are largely due to variations in sialic acid content of the N-linked oligosaccharide present in the CRD (Figures 4 and 5). Taken together, this implies that native pSP-D exists as discrete oligomeric isoforms that are assembled from subsets of monomers that contain differently charged lectin domains due to variations in the sialic acid composition of the glycan present in the CRD.

The interactions of IAV with collectins have been well studied and it can be concluded from both in vitro and in vivo evidence that they play an important role in the pulmonary host defense against IAV (39). Neutralization of IAV infectivity by collectins involves binding to and aggregation of the virus, which interferes with the binding of the virus to receptors on the cell surface, and also contributes to enhanced clearance by phagocytes. Binding of SP-D to IAV involves the Ca²⁺-dependent interaction of the CRD with glycosylated proteins present in the viral envelope: hemagglutinin and neuraminidase (40). Of these viral spike proteins, hemagglutinin is involved in binding of the virus to the host cell, whereas neuraminidase is involved in release of the virus from the infected cell after proliferation. However, recognition of virus by the other lung collectin, SP-A, is dependent on the presence of a conserved sialylated N-linked oligosaccharide moiety present in the CRD (41). Specific removal of the sialic acids from SP-A by neuraminidase treatment prevented the binding of SP-A to IAV-infected cells and inhibited the viral neutralization by SP-A. This suggests that the sialylated oligosaccharide present on SP-A can interact with IAV via the sialic acid receptor present on the hemagglutinin of the virus (43). Our studies on pSP-D revealed the presence of a sialylated carbohydrate which is, analogous to SP-A, located in one of the loop regions that comprises the top half of the CRD. Therefore, experiments were performed to test the antiviral activity of pSP-D and, more specifically, to investigate the role of the N-linked oligosaccharide present in the CRD of pSP-D in IAV neutralization.

As a first step, HA inhibition assays were performed using IAV Phil82 (H3N2) and the mutant strain Phil82/BS. This mutant was shown earlier to be resistant to growth inhibition by bovine serum -inhibitor and conglutinin; sequence analysis showed that Phil82/BS lacks a high-mannose oligosaccharide moiety which is attached at residue Asn-165 of HA in all H3-subtype viruses. Previous studies with various recombinant SP-Ds showed that this strain is also relatively resistant to HA inhibition by SP-D. Absence of the Asn-165 glycosylation from Phil82/BS accounts for the decreased ability of this strain to interact with the CRD of SP-D (40). The HA inhibition data presented in this paper show that pSP-D has a similar activity against the Phil82 wild-type strain as compared with recombinant rat SP-D (Table 1). However, pSP-D proved to be significantly more effective in HA inhibition of the Phil82/BS mutant than recombinant rat SP-D. Furthermore, there was a substantial drop in HA-inhibiting activity of pSP-D against both Phil82 and Phil82/BS strains after treatment of pSP-D with N-Glycanase. From these data it can be concluded that the N-linked carbohydrate contributes to the activity of pSP-D, probably by providing an additional mechanism of attachment to the virus in addition to the usual mode of attachment of the CRD to virus-associated carbohydrates. These findings are different from results obtained with recombinant rat SP-D in which the conserved Asn⁹⁰-linked oligosaccharide located within the collagen domain (Figure 1) was deleted by site-directed mutagenesis (44). HA inhibition assays using this SP-D mutant resulted in a slightly increased anti-influenza activity, in contrast to our findings after N-glycanase treatment of pSP-D. Although it is not clear whether the Asn⁹⁰-linked oligosaccharide in the collagen domain of pSP-D is also removed by our deglycosylation procedure, it is highly probable that such is the case. As a consequence, the contribution of the CRD-linked oligosaccharide in pSP-D to viral neutralization might be even more profound compared with what we observed in our experiments.

In conclusion, the isolation and characterization of natural pSP-D has revealed that this species exhibits unique features that have not been found in SP-D from other species examined to date. Porcine SP-D has been shown to be 7 kD larger than hSP-D due to the presence of a N-linked carbohydrate in the CRD of pSP-D. The N-glycosylation, which displays heterogeneity in the degree of sialylation, results in the existence of differently charged monomeric units which are assembled into variously charged oligomeric isoforms. The presence of a carbohydrate moiety in the CRD of pSP-D affects the interaction with influenza A virus, which is demonstrated by HA inhibition experiments. Because pigs are thought to be involved in the interspecies transmittance of IAV strains with birds and humans (45, 46), one may speculate that the distinct structural and functional properties of pSP-D are related to the presence of certain virus strains that emerge in pigs after genetic reassortment of different influenza A viruses and that may result in the appearance of new pandemic strains in humans.