Site-directed mutagenesis of Cys-15 and Cys-20 of pulmonary surfactant protein D. Expression of a trimeric protein with altered anti-viral properties.

Surfactant protein D (SP-D) molecules are preferentially assembled as dodecamers consisting of trimeric subunits associated at their amino termini. The NH2-terminal sequence of each monomer contains two conserved cysteine residues, which participate in interchain disulfide bonds. In order to study the roles of these residues in SP-D assembly and function, we employed site-directed mutagenesis to substitute serine for cysteine 15 and 20 in recombinant rat SP-D (RrSP-D), and have expressed the mutant (RrSP-Dser15/20) in Chinese hamster ovary (CHO-K1) cells. The mutant, which was efficiently secreted, bound to maltosyl-agarose, but unlike RrSP-D, was assembled exclusively as trimers. The constituent monomers showed a decreased mobility on SDS-polyacrylamide gel electrophoresis resulting from an increase in the size and sialylation of the N-linked oligosaccharide at Asn-70. Although RrSP-Dser15/20 contained a pepsin-resistant triple helical domain, it showed a decreased Tm, and acquired susceptibility to proteolytic degradation. Like RrSP-D, RrSP-Dser15/20 bound to the hemagglutinin of influenza A. However, it showed no viral aggregation and did not enhance the binding of influenza A to neutrophils (PMN), augment PMN respiratory burst, or protect PMNs from deactivation. These studies indicate that amino-terminal disulfides are required to stabilize dodecamers, and support our hypothesis that the oligomerization of trimeric subunits contributes to the anti-microbial properties of SP-D.

example, SP-D has been shown to enhance the binding of influenza A to neutrophils, to inhibit the viral hemagglutinin and inhibit infectivity in vitro, to potentiate the neutrophil respiratory burst in response to bound virus, and to protect neutrophils from viral deactivation (4).
Cloning and sequence analysis of SP-D has shown that the 43-kDa monomer has four major structural domains (8 -11). The short, non-collagenous amino-terminal domain contains two absolutely conserved cysteine residues (Cys- [15][16][17][18][19][20] involved in interchain disulfide bonding. The uninterrupted collagenous domain contains 59 Gly-X-Y repeats (in rat and human), and is connected to the carboxyl-terminal carbohydrate recognition domain (CRD) by a short linking sequence (L). A single N-linked glycosylation site at Asn-70 is found within the collagenous domain, and studies of SP-D from natural and recombinant sources have shown that this site is utilized and contains a sialylated sugar moiety (12,13).
Other collagenous C-type lectins include: pulmonary surfactant protein A (SP-A), serum mannose-binding protein (MBP), bovine serum conglutinin, and bovine serum lectin 43 (CL-43) (14,15). SP-D is predominantly assembled as a dodecamer consisting of four homotrimeric subunits associated at their amino termini (12,16). Analysis of proteolytic fragments and limited reduction and alkylation of SP-D dodecamers suggests that the amino-terminal cysteine residues form intra-and intertrimeric disulfide bonds, which stabilize the fully assembled dodecamer (12).
Trimeric CRDs (but not monomers) show high affinity binding to various saccharide ligands in vitro (17). In addition, ultrastructural studies suggest that the rigid collagenous arms restrict the spatial distribution of the lectin domains and allow for bridging interactions between ligands separated by up to 100 nm (8,12). Accordingly, we hypothesized that amino-terminal cross-linking of trimeric subunits is critical for SP-Dmediated microbial agglutination. We further hypothesized that trimeric subunits of SP-D (i.e. single arms) will bind to microorganisms but show decreased agglutinating activity. Although SP-D "arms" can be generated by reduction and alkylation, the intrachain bonds required for lectin activity are also disrupted (12). We therefore sought to test these hypotheses by mutating the amino-terminal cysteine residues in rat SP-D and utilizing a CHO K1 expression system. Previous work in our laboratory has shown that wild type recombinant rat SP-D (RrSP-D) produced by this system is chemically, morphologically, and functionally indistinguishable from natural rat SP-D (13).
Using this approach, we have produced a full-length mutant SP-D molecule, RrSP-Dser15/20, that is secreted as a trimer. These studies indicate that amino-terminal disulfide bonds are required to form or stabilize dodecamers, and provide support for our hypothesis that the oligomerization of trimeric subunits to form dodecamers is required for some of the previously observed interactions of SP-D with influenza A. Our results also show that the cysteine substitutions decrease the thermal stability of the collagen helix and increase the protein's susceptibility to proteolytic degradation.

MATERIALS AND METHODS
Site-directed Mutagenesis of rSP-D-Site-directed mutagenesis was performed on a full-length rat SP-D cDNA clone provided by Drs. J. H. Fisher and D. R. Voelker, Denver, CO (10), using the PCR overlap extension method (18). Primers for substituting serine for Cys-15 and Cys-20 of the mature protein were: CAATAACCAACACGAGCACCC-TAGTCTTGAGTAGTCCAACAG (forward) and CTGTTGGACTACTC-AAGACTAGGGTGCTCGTGTTGGTTATTG (reverse). Sp6 and T7 were used as outside primers. PCR reactions were performed using 400 ng of ScaI linearized rSP-DpGEM-3Z template. The reaction buffer contained 2 mM MgCl 2 , 2 mM dithiothreitol, and 5 units of Taq polymerase (Promega). Separate reactions containing the forward and T7 primers or reverse and Sp6 primers were performed for 15 cycles with a 45°C annealing temperature. The resulting products were gel-purified and combined in a second PCR reaction for 10 cycles at 52°C annealing, followed by 20 cycles at 42°C annealing using Sp6 and T7 as primers. The ϳ1.3-kilobase pair DNA product was gel-purified, digested with EcoRI, and subcloned into pGEM-3Z. Clones were sequenced up to the ApaI site within rSP-D to verify the presence of the desired mutations and the absence of any additional point mutations. An EcoRI/ApaI fragment containing the mutated region was then subcloned into the corresponding site in wild type rSP-DpGEM-3Z. Subcloning of the mutated region was performed in order to avoid introduction of extraneous Taq generated mutations.
Expression of Rat SP-D cDNA Mutants in CHO K1-Mutated rSP-D cDNAs were excised from pGEM-3Z with EcoRI and subcloned into the corresponding site within the multiple cloning site of pEE14 (13,19,20). Orientation of the subclones was determined by restriction mapping and DNA sequencing. Transfection of pEE14 constructs into CHO K1 cells and selection of stably expressing cell lines was performed as described previously (13).
Metabolic Labeling and Purification of Recombinant SP-D-Confluent cultures were washed in serum-free DMEM and then incubated for 16 h in DMEM containing 1% (v/v) dialyzed fetal calf serum, 50 g/ml ascorbic acid and 5 Ci/ml L-[ 14 C]proline (283 mCi/mmol; DuPont NEN). Secreted recombinant protein was isolated from the culture medium by sequential maltosyl-agarose and gel filtration chromatography as described previously (13). The elution of radiolabeled protein was monitored by liquid scintillation counting and/or SDS-PAGE and autoradiography.
Immunoprecipitation-Cells were plated in six-well tissue culture plates and grown to confluence. Metabolic labeling was performed with L-[ 14 C]proline as described above or with Tran 35 S-label (ICN, 1255 Ci/mmol). For labeling with Tran 35 S-label, cells were washed and preincubated for 30 min with DMEM deficient in cysteine and methionine, DMEM(ϪCys/Met), supplemented with 50 g/ml ascorbate. Cells were then incubated in 1 ml of DMEM(ϪCys/Met) plus 10 Ci/ml Tran 35 Slabel for up to 8 h. Conditioned medium was harvested and centrifuged at 1000 ϫ g to pellet any cells or debris, and the supernatant was used in subsequent steps. For analysis of intracellular protein, cell layers were washed twice with cold DMEM and lysed with 1 ml of immunoprecipitation buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 0.5% (w/v) deoxycholate, 1% (v/v) Triton X-100, 0.1% (w/v) SDS, 2.5 mM EDTA, 0.1 mM PMSF) supplemented with 50 mM iodoacetamide to trap any free sulfhydryls. Cell lysates were incubated on ice in the dark for 30 min and centrifuged at 12,000 ϫ g to pellet cell debris. Immunoprecipitations were performed by mixing 200 -800 l of cell lysate or conditioned medium with 500 l of immunoprecipitation buffer and 5 l of rabbit anti-rat SP-D antibody for 2 h at room temperature or overnight at 4°C. Samples were then mixed with 300 l of a 10% suspension of IgGsorb (The Enzyme Center, Malden, MA) for 1 h at room temperature. After centrifugation at 12,000 ϫ g for 30 s, the supernatant was discarded and the pellets were washed three times with immunoprecipitation buffer. Immunoprecipitates were resolved by 5/10% or 5/12% SDS-PAGE and visualized by fluorography.
Thermal Stability Analysis-Thermal denaturation temperatures were determined essentially as described (23). L-[ 3 H]Proline-labeled wild type RrSP-D and L-[ 14 C]proline-labeled RrSP-Dcys3ser15/20 were diluted in 100 mM Tris, 400 mM NaCl (pH 7.4), and 45 l was added to 200 l microcentrifuge tubes. Two tubes of each sample were kept at room temperature. The remaining tubes were placed in a Perkin-Elmer Cetus thermal cycler at 30°C and programmed to increase 2°C every 10 min between 30°and 44°C. At the end of each 10-min plateau, one tube of each sample was removed, quenched at 20°C for 30 s, and digested with 5 l of 1 mg/ml L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin (Sigma, 12,000 units/mg) for 2 min at 20°C. Digestion was stopped by addition of reducing SDS-PAGE sample buffer containing PMSF, followed by boiling.
To test susceptibility to proteolytic degradation, wild type SP-D and the Cys mutant (1.5 g each) were incubated with 1 ϫ 10 Ϫ6 M of each enzyme in a total mixture volume of 50 l (buffer ϭ 0.05 M Tris, pH 7.5, 0.01 M CaCl 2 , 0.15 M NaCl, 0.02% Brij) at 37°C overnight. The reactions were terminated by the addition of 1 mM o-phenanthroline for the metalloproteinases and 1 mM PMSF for leukocyte elastase. The mixtures were then boiled in reducing sample buffer. Proteins were resolved by SDS-PAGE and visualized by silver staining.
Preparation of Virus-Virus stocks were grown in the chorioallantoic fluid of 10-day-old embryonated hen's eggs, purified on a discontinuous sucrose density gradient, stored, and titered as described previously (28). The A/Texas 77/H3N2 (Texas 77) and A/PR/8/34/H1N1 (PR8) strains of IAV were gracious gifts of Dr. Jon Abramson (Bowman-Gray School of Medicine, Winston-Salem, NC). The A/Mem71 H -Bel N (H3N1) and its glycosylation mutant derivative strain Mem71 H -Bel N /BS were kindly provided by Dr. E. Margot Anders (University of Melbourne, Melbourne, Australia) (29). Incubation of SP-Ds with IAV stocks was carried out at 37°C in TBS with 2 mM Ca 2ϩ except in indicated experiments in which maltose was added, or in which the buffer contained 10 mM EDTA and no added Ca 2ϩ (28).
Assessment of Viral Aggregation-Aggregation was assessed by measuring changes in light transmission through suspensions of IAV after addition of various concentrations of SP-D (28). Addition of SP-D alone in the presence of 2 mM Ca 2ϩ did not cause any decrease in light transmission.
Measurement of Neutrophil Activation and Protection-Neutrophils (PMN) from healthy volunteer donors were isolated to Ͼ95% purity using dextran sedimentation, followed by a Ficoll-Hypaque gradient centrifugation for removal of mononuclear cells, and hypotonic lysis to remove contaminating erythrocytes (28). H 2 O 2 production was measured by the oxidation of scopolectin, and O 2 . assessed by monitoring the superoxide dismutase inhibitable reduction of cytochrome c (30). PMN protection assays measured the superoxide response to FMLP (5 ϫ 10 Ϫ7 M) after exposure to IAV preincubated in the absence or presence of SP-D (4, 31, 32).

Construction and Expression of RrSP-D Site-directed
Mutants-Sequence analysis of the constructs showed that we had generated the expected Cys-15/Cys-20 double substitution mutant, designated RrSP-Dser15/20. RrSP-Dser15/20 was efficiently secreted into the culture medium, and bound quantitatively to maltosyl-agarose ( Figs. 1 and 2). The recovery was comparable to that obtained for wild type SP-D under compa-rable conditions of culture (ϳ5 g/ml) (13). Non-reducing SDS-PAGE indicated the absence of disulfide-bonded dimers, trimers, or other multimers, consistent with the absence of amino-terminal cysteine residues (Fig. 1). However, migration of RrSP-Dser15/20 on SDS-PAGE was slightly slower following reduction, indicating the presence of intramolecular disulfide bonds in the globular CRD domain. Comparison of the migration of reduced wild type RrSP-D and RrSP-Dser15/20 on SDS-PAGE showed that the double cysteine mutant SP-D had a higher apparent molecular mass (ϳ55 kDa versus 43 kDa) attributable to differences in post-translational modification (see below). The lower mobility of the secreted wild type and mutant proteins relative to the major intracellular species is attributed to maturation and terminal sialylation of the Nlinked oligosaccharide at Asn-70. The secreted and intracellular components migrated with approximately the same apparent molecular mass following incubation with endoglycosidase F (see below). As observed for wild type SP-D, incubation of RrSP-Dser15/20 with purified bacterial collagenase liberated an ϳ18-kDa fragment (reduced) corresponding to the carboxylterminal domain (data not shown).
Assembly of RrSP-Dser15/20 -Oligomerization of affinitypurified RrSP-Dser15/20 was assessed by gel filtration chromatography (Fig. 2B). RrSP-Dser15/20 eluted at approximately the same position as natural hSP-D trimers (single arms) or components liberated by reduction and alkylation of rSP-D under non-denaturing conditions (12). The presence of a collagenous triple helix of approximately normal length was verified by overnight incubation with 100 g/ml porcine pepsin A (Worthington, 3352 units/mg) at 27°C. Pepsin liberated two fragments, a minor component (ϳ29 kDa, globular standards, reduced) that co-migrated with the major pepsin-resistant fragment of RrSP-D, and a major fragment of ϳ25 kDa.
To identify the sites of peptic cleavage, fragments of RrSP-Dser15/20 were resolved by SDS-PAGE, transferred to Immobilon membranes, and submitted for gas-phase amino-terminal microsequencing as described previously (22). The 29-kDa fragment gave sequence beginning within the amino-terminal peptide domain (Val 18 -Gly-Ser-Pro-Thr-Glu-Asn-Gly-Leu) and in-cluding the substituted serine at position 20 (10). However, the major fragment gave sequence beginning within the 10th Gly-X-Y triplet of the collagen domain (Ser 55 -Gly-Leu-Hyp-Gly-Pro-Arg-Gly) (9).
Post-translational Processing of RrSP-Dser15/20 -To determine if the size differential of RrSP-Dser15/20 versus reduced wild type RrSP-D was due to overhydroxylation of proline or lysine residues in the collagen domain and/or overglycosylation of hydroxylysine, transfected cells were metabolically labeled in the presence or absence of 2,2Ј-dipyridyl, an inhibitor of prolyl and lysyl hydroxylase. Comparison of mutant and wild type proteins synthesized under these conditions suggested that there was little if any overmodification of the helical domain of RrSP-Dser15/20 (data not shown).
To determine if the increased mass of double cysteine mutant reflected differences in the size or structure of the N-linked oligosaccharide at Asn-70, we compared the mobility of pro- Equal volumes (300 l) of cell lysate (C) or medium (M) were immunoprecipitated with P9, a polyclonal antibody to rat SP-D. Immune complexes were resolved on 5-10% discontinuous SDS-PAGE, fixed, treated with En 3 Hance, dried, and exposed to x-ray film for fluorography. The secreted mutant migrates significantly slower than natural or wild type, which migrates slightly faster than ovalbumin (43 kDa, reduced) (13). The higher mobility of the cellular protein as compared to components in the medium is attributed to incomplete maturation of the oligosaccharide at Asn-70 (see below). The lower mobility of both proteins in the absence of dithiothreitol reflects unfolding of intrachain disulfide bonds in the CRD. teins synthesized in the absence or presence of up to 6 g/ml tunicamycin (TM), and after glycosidase digestion. Consistent with the possibility, wild type and mutant proteins synthesized in the presence of tunicamycin showed approximately the same molecular mass (data not shown). Furthermore, incubation of the wild type and mutant proteins with Endoglycosidase F generated cleavage products with the same mobility by SDS-PAGE (Fig. 3). Neuraminidase digestion of RrSP-Dser15/20 resulted in a greater size shift than digestion of RrSP-D. However, differences in sialylation did not completely account for the size differential (Fig. 3).
Thermal Stability Analysis-Although intra-trimer disulfide bonds were not required for trimerization and formation of a pepsin-resistant triple helix, we hypothesized that they might contribute to stabilization of the collagenous triple helix at physiologic temperatures. To address this possibility, we examined the temperature-dependent acquisition of sensitivity to degradation by trypsin, a technique commonly used to characterize the effects of mutations on collagen helix stability in inherited connective tissue disorders (23). As expected, the wild type protein was relatively insensitive to trypsin digestion following incubation at temperatures through 38°C, but showed increasing degradation after incubation between 40 and 44°C (Fig. 4). Even after heating to 44°C, some intact RrSP-D remained. RrSP-Dser15/20 was similarly resistant to trypsin degradation following incubation at temperatures up to 34°C. However, in contrast to wild type, RrSP-Dser15/20 exhibited an abrupt increase in susceptibility to degradation after heating to 36°C and was completely degraded by trypsin after heating to 40°C (Fig. 4). Based on densitometric measurements of bands corresponding to undigested protein, RrSP-D has a T m of ap-proximately 40°C, while the mutant RrSP-Dser15/20 has a T m of 37°C (Fig. 4).

Susceptibility to Degradation by Mammalian Proteases-Un
published experiments demonstrated that natural SP-Ds are highly resistant to degradation by a wide range of proteolytic enzymes that might be released at sites of pulmonary inflammation including: human and rat interstitial collagenases (MMP-1 and -13), 72-and 92-kDa gelatinases (MMP-2 and -9, respectively), stromelysin (MMP-3), matrilysin (MMP-7), human and mouse macrophage metalloelastase (MMP-12), and human leukocyte elastase. Given the differences in thermal stability, we explored the possibility that the mutant SP-D might be abnormally susceptible to proteolytic attack. Consistent with our previous results, wild type RrSP-D was not degraded during prolonged incubation with high concentrations (1 ϫ 10 Ϫ6 M each) of these enzymes at 37°C (Fig. 5, top panel). By contrast, RrSP-Dser15/20 was degraded to varying degrees by all of the enzymes (Fig. 5, bottom panel) and appeared particularly susceptible to degradation by the gelatinases and human leukocyte elastase.
Viral Interactions-Interactions of the wild type and mutant proteins with the hemagglutinin of selected strains of influenza A virus (IAV) were initially compared in hemagglutination inhibition assays. For example, RrSP-Dser15/20 showed an identical hemagglutination inhibition titer for Mem71 H -Bel N /BS (Table I), a strain that is resistant to the serum collectin, conglutinin, and MBP. On the other hand, the mutant was significantly less effective at inhibiting the hemagglutinin of the Bangkok strain. The hemagglutination inhibition activity of RrSP-Dser15/20 was inhibited by maltose or EDTA, as documented previously for the wild type protein. Surprisingly, the mutant also inhibited the hemagglutination activity of PR-8, a viral mutant lacking oligosaccharides on the hemagglutinin head which does not bind to wild type rSP-D. However, the hemagglutination inhibiting activity was not significantly inhibited by EDTA.
As predicted, RrSP-Dser/15/20 failed to give detectable viral aggregation in light transmission assays (Fig. 6) and was similarly defective in its capacity to generate aggregates that could be sedimented by low speed centrifugation (data not shown). The mutant also failed to enhance binding of IAV to neutrophils (Fig. 7, A and B), failed to enhance the H 2 O 2 response of PMNs to virus (Fig. 8), and was unable to protect these cells against the deactivating effects of IAV on FMLP-stimulated respiratory burst response (Fig. 9). DISCUSSION The combined substitution of serine for cysteine at positions 15 and 20 of rat SP-D resulted in the production of a mutant SP-D consisting exclusively of trimers. The identification of pepsin-resistant fragments comparable in size to the major pepsin-resistant fragments of wild type SP-D establishes that the monomers within each trimer are aligned in parallel and can fold to form a triple helical domain of approximately normal length. Thus, the mutant assembles as a single "arm" of the "four-armed" rSP-D molecule (dodecamer). In this respect, RrSP-Dser15/20 is structurally similar to the bovine serum collectin CL-43 (14,16) and to subpopulations of hSP-D trimers isolated from human amniotic fluid (33) and bovine bronchoalveolar lavage (8,16). Because we did not detect dodecamers or higher order oligomers by gel filtration chromatography, we conclude that interchain disulfide bonds are necessary for stable oligomerization of the trimeric subunits.
The assembly of triple helical arms in the absence of the amino-terminal interchain disulfide bonds is consistent with studies of other collagenous proteins. In the fibrillar matrix collagens, helix formation proceeds toward the amino terminus by a "zipper-like" mechanism following association of the carboxyl-terminal propeptides and disulfide cross-linking of the COOH-terminal domains (34,35). Although the collectins lack carboxyl-terminal interchain disulfide bonds, structural analysis of a peptide corresponding to the linking domain of bovine SP-D suggests that this region forms a stable ␣-helical coiledcoil, thereby associating collectin monomers and driving collagenous helix formation in vitro (36). Finally, deletion analysis of recombinant SP-A, MBP, and rat SP-D have shown that the linking and/or CRD domains are necessary and sufficient for the assembly of trimers (37)(38)(39).
Wild type SP-D displays a remarkable insensitivity to proteases, 2 which we hypothesized is due to tight folding of the CRD and the collagenous domain, and steric hindrance of protease binding near the hub of SP-D dodecamer. The capacity of pepsin to cleave RrSP-Dser15/20 within the 10th Gly-X-Y triplet of the collagen domain (at 27°C), and the decreased thermal stability and increased protease sensitivity of the mutant protein relative to wild type, together suggest that a major function of the conserved Cys residues is to stabilize the aminoterminal end of the triple helix. It is possible that the formation of dodecamers further enhances the stability of the collagen domain or limits accessibility of proteases to cleavage sites near the amino terminus in RrSP-D. Although the measured T m for the wild type protein is consistent with the formation of a stable triple helix in vivo, the melting temperature of the mutant (T m 37°C) suggests that its helical domain is incompletely folded in the cell. To our knowledge, these studies provide the first demonstration of stabilization of a contiguous collagenous helix by interchain disulfide bonds. We speculate that the amino-terminal and intrahelical cysteine residues found in other collectins play a similar role.
To our surprise, the mutant showed a slower mobility on SDS-PAGE than wild type RrSP-D or the predominant forms of natural SP-D. Although some degree of overhydroxylation of the collagen domain cannot be excluded, our data indicate that lower mobility of mutant monomers primarily results from an increase in the apparent size and sialylation of the N-linked sugar at Asn-70. We speculate that the mutant offers a less restricted access of Golgi enzymes to the glycosylation site, which is close to the hub of cross-linked dodecamers (12). In   7. Effect of wild type RrSP-D and RrSP-Dser15/20 on IAV binding to neutrophils. A, neutrophils were allowed to adhere to plastic coverslips followed by incubation of the cells with either FITClabeled Bangkok79 IAV alone (top panels), or the same amount of IAV, which had been preincubated with either 1.6 g/ml RrSP-D (middle panels) or RrSP-Dser15/20 (bottom panels). Phase contrast images are shown on the left and fluorescence images of the same fields on the right (magnification, ϫ100). In the absence of SP-D, faint fluorescence is seen distributed over the surface indicating bound virus; negative controls were devoid of surface immunostaining. In the presence of RrSP-D, coarse aggregates of viral particles (e.g. white arrows) were observed in association with the cell surface. The identification of viral aggregates in association with the cell surface was confirmed by electron microscopy (40). In the presence of the mutant, there was a marked decrease in the binding of virus and no coarse viral aggregates were observed. B, FITC-labeled Bangkok79 IAV was preincubated with the indicated concentrations of either wild type RrSP-D (dark squares) or RrSP-Dser15/20 (open squares), followed by assessment of binding of these samples to human neutrophils by flow cytometry. Results shown are mean Ϯ S.E. of four experiments. Viral binding was significantly enhanced by wild type RrSP-D (* indicates p Յ 0.05 compared to binding of untreated IAV). In contrast, RrSP-Dser15/20 did not cause any enhancement of viral binding. In fact, significantly reduced IAV binding was seen using samples preincubated with the highest concentration of RrSP-Dser15/20 (indicated by **). this regard, other studies indicate that wild type dodecamers assemble in the rough endoplasmic reticulum, and that conversion of the N-linked sugar from a high mannose to mature sialylated form occurs shortly prior to secretion. 3 The failure of RrSP-Dser15/20 to aggregate IAV provides strong support for our hypothesis that the multimers of trimers are necessary for SP-D to agglutinate viral particles and other microorganisms. While RrSP-Dser15/20 bound to IAV virions and showed significant hemagglutination inhibition activity, there was no detectable agglutinating activity. The protein was also deficient in other activities believed to be dependent on the state of viral aggregation including: IAV binding to neutrophils, augmentation of the neutrophil respiratory burst response to bound virus, and protection of neutrophils from viral deactivation (4).
Although distant conformational effects transmitted from the mutated amino terminus through the helical arms to the CRD cannot be entirely excluded, the available data strongly suggest that the trimeric CRD of SP-D is functionally univalent with regard to IAV binding, i.e. that each trimeric CRD interacts with one or more hemagglutinin molecules on a single viral particle. In this regard, crystallographic studies of trimeric carboxyl-terminal fragments of MBP suggest that the three sugar binding sites are located near the tip of the carboxylterminal globule (i.e. oriented at ϳ180°relative to the axis of the coiled-coil) and form a relatively planar binding surface which could favor cooperative interactions with up to three independent glycoconjugates displayed on a microbial surface (17).
We speculate that the enhanced anti-IAV activities of SP-D (and conglutinin) relative to SP-A, MBP, and monoclonal IgG antibodies to the hemagglutinin (40), primarily result from differences in the valency and spatial distribution of the hemagglutinin binding domains (i.e. CRD or Fab). The long arm length of SP-D may favor bridging interactions between viral particles, leading to the formation of large viral aggregates, thereby modifying the interactions of IAV with the leukocyte surface. The failure of the mutant SP-D to promote viral aggregation suggests that single-arm forms of natural SP-D are similarly "deficient" in aggregation-dependent anti-viral activity.
An unexpected finding relates to the capacity of RrSP-Dser15/20 to bind to the PR-8 strain of IAV in a calciumindependent manner (Table I). Previous studies have shown negligible binding of rSP-D, RrSP-D, and hSP-D to PR-8, a strain that lacks N-linked oligosaccharides on the head of the hemagglutinin. We speculate that RrSP-Dser15/20 binding to this strain reflects increased accessibility of the sialylated oligosaccharide at Asn-70 to sialic acid binding sites on the hemagglutinin or neuraminidase.
In conclusion, we have expressed a mutated SP-D molecule that is secreted as a trimer corresponding to a single "arm" of the SP-D dodecamer. Characterization of this mutant has shown that the amino-terminal disulfide bonds of SP-D play an important role in stabilizing the collagen triple helix and protecting the molecule from degradation by several proteases that may be released at sites of pulmonary inflammation. These studies have also confirmed that important anti-viral properties of SP-D are entirely dependent on the oligomerization of trimeric subunits.