Soluble Collectin-12 (CL-12) Is a Pattern Recognition Molecule Initiating Complement Activation via the Alternative Pathway

The complement system is an important humoral and cellular innate immune defense system that functions as a rapid and efficient primary host defense barrier against invading pathogens and altered host cells. Its initiation and amplification is mediated by three defined pathways: the classical pathway (CP), the lectin pathway (LP), and the alternative pathway (AP) (1). The CP is mainly activated in an Ab-dependent fashion by IgM or IgG clusters via interaction with C1q, but it is also initiated by C1q directly through recognition of distinct structural moieties on microbial and apoptotic cells or through versatile soluble pattern-recognition molecules (sPRMs), such as PTX3 and C-reactive protein (2–4). However, LP activation is primarily induced by collagen-like sPRMs, such as mannose-binding lectin (MBL) and the ficolins (5). It has also been shown that the CP and LP could be cross-activated through formation of heterocomplexes among PTX3, C1q, and the ficolins, as well as MBL (6, 7). In contrast, the AP does not require Abs or sPRMs for its initiation, and it is often referred to as spontaneous and nondiscriminatory because it does not differentiate between self and nonself. Rather, its initiation occurs by spontaneous low-rate hydrolysis of the thioester in C3, which is quickly amplified on nearby surfaces (8).

Recently, three novel human defense collagens: collectin-10 (collectin liver 1, CL-L1, or CL-10), collectin-11 (collectin kidney 1, CL-K1, or CL-11), and collectin-12 (collectin placenta 1, CL-P1, or CL-12) derived from the COLEC10, COLEC11, and COLEC12 genes, respectively, were identified and characterized (9–11). CL-10 was initially demonstrated as a cytoplasmic protein, whereas recent work has shown its potential presence in human blood circulation (12). CL-11 is a serum protein with a low concentration of ∼0.3 μg/ml (13, 14), and it has been shown to bind certain microorganisms and apoptotic cells through its carbohydrate-recognition domains (CRDs) mainly to l-fucose and d-mannose (11, 15). Furthermore, CL-11 has recently been shown to be an LP-activating molecule associated with the LP-associated serine proteases (MASPs) in addition to MBL and the ficolins (16). Different from the other novel collectins, CL-12 was originally defined as scavenger receptor C-type lectin because it shares structural and functional similarities with type A scavenger receptor and collectins (10, 17). As a transmembrane scavenger receptor, CL-12 is mainly localized in vascular endothelial cells in general, but in particular in umbilical cord endothelial cells (18). CL-12 revealed pattern-recognition characteristics toward yeast, as well as Escherichia coli and Staphylococcus aureus, by mediating phagocytosis by CL-12 transfected cells (17, 18). Furthermore, it has been shown that CL-12 selectively bind glycans bearing both terminal galactose and fucose moieties and presents a striking high-affinity selectivity for Lewis^X trisaccharide, which is commonly displayed on adhesion molecules of various leukocytes, suggesting that CL-12 could modulate leukocyte recruitment (19, 20).

CL-12 has also been shown to bind fibrillar β-amyloid protein, thus indicating a potential role in scavenging of amyloid (21). Intriguingly, recent findings in zebrafish embryo models have shown that knockdown of zebrafish COLEC12 expression causes severe defects in vasculogenesis and development during the early embryogenic stages (22). These results suggest a pivotal role for CL-12 not only in innate immune defense, but also in fundamental developmental processes. Based on its structure and functional similarities with soluble collectins, we hypothesized the existence of a fluid-phase analog of CL-12 released from cells, which may function as a sPRM. Based on our data, we report a novel mechanism of complement activation where the AP may be directly initiated by sPRMs.

Recombinant CL-12 full length (rCL-12FL) was produced using the Flp-In System (Invitrogen) according to the manufacturer’s instructions. In brief, cDNA encoding CL-12 full length (Met¹∼Leu⁷⁴²) was designed into the Flp-In expression vector (pcDNA5/FRT) containing the Hygromycin B resistance gene and ordered from Invitrogen. Stably growing Flp-In-CHO cells containing the lacZ-Zeocin fusion gene were cotransfected with a mixture of Lipofectamine, the Flp recombinase expression plasmid pOG44, and the pcDNA5/FRT vector containing COLEC12 gene. As negative control, the pcDNA5/FRT vector lacking the COLEC12 gene was applied in parallel. The cells were cultivated in Ham’s F-12 medium (Sigma) for 2 d. Stable transfectants were then selected by cultivation in the medium containing hygromycin (500 μg/ml; Invitrogen). In addition, recombinant CL-12 extracellular domain (rCL-12ED) (Lys⁶⁰∼Leu⁷⁴²) tagged with N-terminal 6× histidine was also produced as described earlier.

Proteins were separated by 4–12% SDS-PAGE gel under reducing or nonreducing conditions according to the method of Laemmli (23) and visualized by Coomassie Brilliant blue or Western blot, as described previously (6, 7).

To determine soluble CL-12, we cultivated rCL-12FL–expressing CHO cells (CHO/CL-12FL) in EGM-2 BulletKit medium (Lonza). Expression level of rCL-12FL on CHO/CL-12FL was determined with goat anti-human CL-12 polyclonal Ab (CL-12 pAb; R&D Systems) and Alexa Fluor 488–conjugated donkey anti-goat IgG (Life Technologies), and finally analyzed with a BD FACSCalibur (Becton Dickinson) as previously described (6). When the culture flasks were 80–90% confluent, the supernatants were harvested and analyzed for the presence of CL-12 by Western blot as described earlier. In brief, the supernatants were subjected to SDS-PAGE, and CL-12 was detected with rat anti-human CL-12 mAb (R&D Systems) and HRP-conjugated rabbit anti-rat IgG (Dako). As a negative control, CHO cells that were transfected with the blank vectors lacking the COLEC12 gene (CHO/Control) were applied in parallel. Alternatively, transmembrane CL-12 was isolated from CHO/CL-12FL using Mem-PER Plus Membrane Protein Extraction Kit (Thermo Scientific) according to the manufacturer’s instructions. The presence of CL-12 was then determined by Western blot as described earlier. As a negative control, CHO/Control was applied in parallel.

Human umbilical cord tissue was obtained freshly from 10 healthy puerperal women with informed consent. After the umbilical cords had been cut from the newborn babies, the bloods that remained in the umbilical veins after parturition were collected in hirudin blood tubes (Roche Diagnostics). We also collected venous blood in hirudin tubes from 10 healthy staff members with informed consent. Preparation of total protein extracts from human umbilical cords was achieved by Chemicon’s total protein extraction kit (Millipore A/S) according to the manufacturer’s instructions. Furthermore, platelet‐poor plasma was isolated by centrifugation (2000 × g, 10 min at room temperature) of human umbilical vein blood, and plasma samples were stored in small aliquots at −80°C until analysis.

To determine CL-12 in human umbilical cord, we subjected samples of the total protein extracts (50 μg) from human umbilical cords to SDS-PAGE, and the presence of CL-12 with biotinylated CL-12 mAb and streptavidin-conjugated HRP (Amersham Bioscience) was visualized in Western blot as described earlier. In some experiments, the samples were immunoprecipitated with Dynabead sheep anti-rat IgG (Invitrogen) in combination with CL-12 mAb or rat IgG1 (R&D Systems), and analyzed with CL-12 pAb by Western blot.

To determine the level of soluble CL-12 in human umbilical cord plasma, we coated ELISA plates with CL-12 pAb (1 μg/ml) in PBS buffer (10 mM Na₂HPO₄, 1.5 mM KH₂PO₄, 137 mM NaCl, 2.7 mM KCl, pH 7.4). All reaction volumes were 100 μl, and plates were washed after each step in PBS containing 0.05% Tween 20. The plates were then incubated with purified recombinant CL-12 from R&D (rCL-12R&D; 100 ng/ml; N terminally 9× histidine tagged extracellular domain of human CL-12, Leu⁵⁷∼Leu⁷⁴²) or human umbilical cord plasma (20%) that was diluted with PBS containing 0.05% Tween 20 containing EDTA (10 mM) plus normal goat IgG (50 μg/ml; R&D system) at room temperature for 3 h with agitation in triplicate. Bound proteins were detected with CL-12 mAb or rat IgG1 as isotype control, followed by incubation of HRP-conjugated rabbit anti-rat IgG (Dako). Final peroxidase reaction was performed as previously described (24). CL-12 level in the plasma was calculated based on the standard curve fitted with rCL-12R&D. In some experiments, the plasma samples were incubated in ELISA plates coated with CL-12 pAb (1 μg/ml) or normal goat IgG (1 μg/ml). Bound proteins were stripped and visualized with biotinylated CL-12 mAb and streptavidin-conjugated HRP in Western blot as described earlier.

Binding of soluble CL-12 to A. fumigatus (clinical isolate 6871; University of Perugia, Perugia, Italy) was performed as described previously (6, 24). rCL-12R&D was applied as a soluble form of CL-12 in the following experiments. In brief, A. fumigatus were incubated with rCL-12R&D (∼0.03–0.25 μg/ml) at 37°C for 1 h. Bound proteins were detected with CL-12 pAb and Alexa Fluor 488–conjugated donkey anti-goat IgG and then finally analyzed with a BD FACSCalibur. In some experiments, A. fumigates conidia were incubated with rCL-12R&D in the presence of N-acetyl-d-glucosamine (GlcNAc; 0.3 M), curdlan (500 μg/ml), or a mixture of both (all from Sigma). Bound CL-12 was assessed as described earlier. Alternatively, rCL-12R&D binding to A. fumigatus was also determined in the presence of EDTA (10 mM) or EGTA (10 mM)-Mg²⁺ (5 mM) instead of Ca²⁺ (2.5 mM).

Serum and plasma samples were obtained from healthy volunteer donors with informed consent. MBL-defect serum (serum^MBL−) and properdin-defect serum (serum^fP−) were obtained from an MBL-defect individual and a properdin-defect individual with normal complement activities, respectively. The sera were diluted with TBS/Ca²⁺ (10 mM Tris/150 mM NaCl/2.5 mM CaCl₂, pH 7.4) or TBS/EGTA-Mg²⁺ (10 mM Tris/150 mM NaCl/10 mM EGTA/5 mM Mg²⁺, pH 7.4) before use based on experimental intentions.

Complement C4b deposition was determined as previously described (7). In brief, A. fumigatus were incubated with rCL-12R&D (5 μg/ml) before addition of serum^MBL− (10%) as a complement source in the presence of TBS/Ca²⁺. After washing, C4b deposition was detected by FITC-linked rabbit anti-human C4c pAb (Dako) and analyzed by flow cytometry. Complement C3b and terminal complement complex (TCC) deposition were determined in the presence of TBS/EGTA-Mg²⁺ as follows: A. fumigatus were preincubated with rCL-12R&D (5 μg/ml) or elevated concentrations of rCL-12R&D (∼0–4 μg/ml) before incubation of serum^MBL− (10%) in the presence of EGTA-Mg²⁺. After washing, C3b and TCC deposition were detected by rabbit anti-human C3c pAb (Dako)/FITC-linked swine anti-rabbit IgG (Dako) or mouse anti-human C5b-9 mAb (BioPorto)/FITC-linked goat anti-mouse IgG (Dako), respectively, and analyzed as described earlier.

In some experiments, complement C3b and TCC deposition were also induced with serum^fP− (10%) instead and assessed as described earlier. Alternatively, serum^fP− was spiked with exogenous properdin (5 μg/ml; CompTech) under physiological concentration before induction of complement activation. Properdin that was thawed once from initial preparation was analyzed by SDS-PAGE before use to determine formation of properdin aggregates during storage. Preparations that formed aggregates were excluded from testing. During complement activation, properdin that bound A. fumigatus was also detected using mouse anti-human properdin mAb (Hyb 39-6, characterized in our laboratory) and FITC-linked goat anti-mouse IgG. Bound proteins were analyzed by flow cytometry as described earlier. In some experiments, to determine whether properdin binds the transmembrane form of CL-12, we incubated properdin (5 μg/ml) with either CHO/CL-12FL or CHO/Control and analyzed it as described earlier.

A. fumigatus were preincubated with rCL-12R&D (5 μg/ml) and then blocked residual exposed sites with serum^fP− (10%). After washing, properdin binding was determined by incubation of exogenous properdin (5 μg/ml) using goat anti-human properdin polyclonal Ab (CompTech) and Alexa Fluor 488–conjugated donkey anti-goat IgG. Bound proteins were analyzed by flow cytometry as described earlier.

Microtiter plates were coated with rCL-12R&D or BSA in 2-fold serial dilutions starting at 5 μg/ml in PBS. All reaction volumes were 100 μl, and plates were washed after each step in TBS/Ca²⁺ containing 0.05% Tween 20. After blocking with BSA (1%), the plates were incubated with properdin (5 μg/ml) at room temperature for 3 h. Bound protein was detected using goat anti-human properdin polyclonal Ab/HRP-conjugated donkey anti-goat IgG (H&L; GenScript) and visualized by enzymatic reactions as described earlier. In a reverse order, rCL-12R&D binding to the plates immobilized with properdin (or BSA; 5 μg/ml) was analyzed by incubation in 2-fold serial dilutions starting at 5 μg/ml, followed by detection with CL-12 mAb/HRP-conjugated rabbit anti-rat IgG as described earlier.

Comparisons between groups were made by one-way ANOVA and GraphPad Prism, version 5.0 (GraphPad Software). The p values <0.01 were considered significant.

To determine a soluble form of CL-12, we established stably transfected Flp-In-CHO cells expressing rCL-12FL and rCL-12ED, respectively. The supernatants derived from in vitro–cultured CHO/CL-12FL were used to analyze soluble CL-12 expression. Under nonreducing conditions, rCL-12FL presented multimeric ladder bands with monomers, dimers, and trimers, whereas under reducing conditions it migrated as a single band of ∼140 kDa. Consistently, rCL-12 from R&D (rCL-12R&D) and rCL-12ED displayed electrophoretic mobility similar to that of rCL-12FL, showing slight difference in gel mobility presumably because both rCL-12R&D and rCL-12ED lack in domain structure integrity compared with rCL-12FL (Fig. 1A, 1B). When the supernatants from in vitro–cultured CHO/CL-12FL were analyzed, as shown in Fig. 1A and 1B, we could show distinct bands of soluble form of CL-12 in the culture supernatants, but not in the supernatants harvested from mock-transfected CHO cells (CHO/Control). No changes in the growth and morphology of the cells were observed during the cultivation. However, the soluble form of recombinant CL-12 in the culture supernatants indicated three bands with gel mobility corresponding to ∼150, ∼250, and ∼400 kDa under nonreducing conditions. The ∼150-kDa band represented bovine IgG that cross-reacted with the detection Ab. Nevertheless, under reducing conditions it restored to ∼120 kDa (Fig. 1A, 1B).

To further substantiate the potential existence of a soluble form of CL-12 under physiological conditions, we collected fresh human umbilical cord tissue and plasma from healthy puerperal women and used them to analyze CL-12 expression. Western blot showed that native CL-12 derived from human umbilical cord tissue had an electrophoretic mobility equal to that of rCL-12R&D (∼130 kDa) under reducing conditions, which was further confirmed by the subsequent immunoprecipitation of the total protein extract (Supplemental Fig. 1). ELISA for human umbilical cord plasma also revealed that the median concentration of soluble CL-12 from 10 healthy subjects was measured to be 92.8 ng/ml (range ∼55–121.4 ng/ml; Fig. 1C), whereas reactivity in venous plasma from healthy adults (n = 10) was shown to be negligible. Furthermore, Western blot analysis showed that soluble CL-12 in the plasma migrated mainly as a doublet with molecular mass of ∼150 and ∼230 kDa, whereas under reducing conditions it migrated as single band with a molecular mass of ∼110 kDa (Fig. 1D, 1E). No apparent bands emerged with similar gel mobility in the negative control (Fig. 1D, 1E).

To characterize the potential physiological relevance of soluble CL-12 as recognition molecule in the complement activation, we used A. fumigatus as a model of infection and determined the binding of soluble CL-12 by purified rCL-12R&D. As shown in Fig. 2A and 2B, rCL-12R&D bound A. fumigatus in a dose-dependent manner and induced substantial agglutination of A. fumigatus conidia. However, addition of GlcNAc or curdlan inhibited the binding of rCL-12R&D, and when combined together, the inhibitory effects were even increased (Fig. 2C). Of interest, removal of Ca²⁺ by addition of EDTA or EGTA-Mg²⁺ did not reduce, but apparently enhanced CL-12 binding to the conidia (Fig. 2D).

To determine the potential effects of soluble CL-12 on complement activation, we used serum^MBL− as a source of complement and determined the effect on A. fumigatus in TBS/Ca²⁺ or TBS/EGTA-Mg²⁺. In TBS/Ca²⁺, no clear effect of rCL-12R&D was detected on C4b deposition (Supplemental Fig. 2). In TBS/EGTA-Mg²⁺, which inhibits activation of the Ca²⁺-dependent CP and LP and permits selective activation of the Mg²⁺-dependent AP, C4b deposition was minimal, whereas both C3b and TCC deposition were easily detected (Fig. 3). However, addition of rCL-12R&D enhanced both C3b and TCC deposition induced by serum^MBL− alone in the presence of EGTA-Mg²⁺ (Fig. 3B, 3C). Furthermore, a shift in concentration of rCL-12R&D from 0 to 4 μg/ml was accompanied with a significant amplification in both C3b and TCC deposition on A. fumigatus (Fig. 4).

Properdin is the only known positive regulator of complement activation so far. Therefore, to determine the involvement of properdin in CL-12–amplified AP activation, we used serum^fP− as a complement source instead. To allow the AP to work at serum concentration with minimal functional activity, we treated serum^fP− with TBS/EGTA-Mg²⁺ as a final concentration of 10% in the reaction mixtures. As shown in Fig. 5A and 5B, the AP activation was drastically reduced when EGTA-Mg²⁺–treated serum^fP− was applied and only minimal autoactivation of the AP was observed. However, reconstitution of exogenous rCL-12R&D or properdin alone under physiological concentration did not restore AP activation. However, when reconstituting both rCL-12R&D and properdin, we were able to restore AP activation validated by detection of C3b and TCC deposition (Fig. 5A, 5B). Furthermore, we found that properdin binding was substantially increased on the conidia when rCL-12R&D was present (Fig. 5C). In some experiments, the results were also confirmed by the rCL-12ED instead of purified rCL-12R&D as an independent example of soluble CL-12 (Supplemental Fig. 3). However, when incubating properdin with CHO/CL-12FL, we could not detect any apparent binding of properdin on the cell surfaces where rCL-12FL was abundantly expressed (data not shown).

Properdin binding to immobilized CL-12 was determined in both cell-based context and microtiter plate. In FACS analysis, in contrast with the case without serum, properdin binding was drastically inhibited when A. fumigatus was preincubated with serum^fP− (Supplemental Fig. 4A), indicating that some serum factors block access of properdin to the conidia by competing the surface binding sites with properdin. Therefore, to assess solely CL-12–mediated properdin binding, we minimized direct properdin binding to the cell surface glycosaminoglycans by blocking with serum^fP− after preopsonization of CL-12. As shown in Fig. 6A, opsonization of rCL-12R&D apparently enhanced the binding of properdin in contrast with the controls where rCL-12R&D was not present. Moreover, the results were also substantiated by the rCL-12ED that was evaluated by quantitative ELISA instead of purified rCL-12R&D (Supplemental Fig. 4B). In some experiments, this was also verified in ELISA where rCL-12R&D–immobilized wells bound properdin in a dose-dependent manner (Fig. 6B) and vice versa (Supplemental Fig. 4C).

CL-12 was originally identified and characterized from placenta and liver as a membrane-anchored molecule functioning as an endocytic receptor present on vascular endothelial cells (10, 17). It has previously been shown that macrophage-specific scavenger receptor/collagen-like domain 163, which mediates the internalization of hemoglobin–haptoglobin complexes by macrophages, can be shed as a soluble form from the cell surface by activation of TLR, cross-linking of the Fcγ receptor or oxidative stress, and so on during inflammation and macrophage activation (25). The level of soluble collagen-like domain 163 has been found to be elevated in patients with different types of infections and inflammatory conditions (26), implying potential generation of soluble scavenger receptor under certain pathophysiological conditions. Interestingly, scavenger receptor CL-12 also shows homology with soluble collectins like MBL, CL-10, and CL-11, all of which activate the LP of complement and are characterized by structural features of collagen-like domain and CRD (27). Thus, we hypothesized the potential generation of a soluble CL-12 from the cell membrane–anchored receptor and a putative role in the complement system.

Determining the content of CL-12 in the culture supernatant of CHO/CL-12FL, we found that CL-12 was shed from the cell surfaces during cultivation. However, the resulting soluble CL-12 (∼120 kDa) displayed a slight difference in electrophoretic mobility from that of rCL-12FL (∼140 kDa) and rCL-12ED (∼130 kDa) under reducing conditions, whereas under nonreducing conditions it represented dimer and trimer, suggesting that membrane-anchored CL-12 might be cleaved in close vicinity of the plasma membrane at its N terminus by a hitherto unknown mechanism. This is a subject of our future investigation. Because CL-12 has previously been shown to be highly expressed in HUVEC and placenta, we reasoned that membrane CL-12 might be shed constitutively into human umbilical cord blood circulation. Analysis of human umbilical cord tissues and plasma further identified the expression of native soluble CL-12. Using our ELISA setup, we demonstrated that CL-12 circulates in normal human umbilical cord plasma as a soluble protein at ∼55–121.4 ng/ml, but we were not able to clearly detect CL-12 in normal adult venous plasma. Taken together, shedding of a soluble form of CL-12 from in vitro–cultured CHO/CL-12FL and detection of CL-12 immunoreactivity in human umbilical cord plasma show the existence of a soluble form of CL-12. Western blot analysis demonstrated that like soluble recombinant CL-12, the soluble plasma CL-12 was able to assemble into multimeric structures under nonreducing conditions. However, under reducing conditions it had electrophoretic mobility corresponding to ∼110 kDa, which differs from that of soluble recombinant CL-12 (∼120 kDa). Nevertheless, immunoprecipitation and Western blot analysis of human umbilical cord revealed that CL-12 migrated with gel mobility corresponding to ∼130 kDa. CL-12 possesses 13 putative N-glycosylation sites in the entire extracellular domain, thus making it possible to be modified posttranslationally via N-glycosylation. Previous studies have shown that HUVEC (or placenta) cell lysates revealed two major bands of CL-12 corresponding to 140 and 110 kDa under reducing conditions in immunoblotting due to differential glycosylation status in vivo, which could be depleted by transfection of small interfering RNAs specific to the COLEC12 gene (17, 18). These results suggest that there is considerable heterogeneity in posttranslational modification. Therefore, the remaining size difference between the two resultant soluble CL-12 forms from in vitro–cultured CHO/CL-12FL and human umbilical cord plasma could be because of the difference between the glycan produced by human and animal cells and/or the type of posttranslational modification.

Previous studies have pointed to scavenger receptor CL-12 as the phagocytic mediator responsible for recognition and phagocytosis of yeast, as well as E. coli and S. aureus (17, 18), indicating that CL-12 could recognize some constituent of cell-wall component on fungi and bacteria and may play a role in innate immunity. Intriguingly, we found that soluble CL-12 could bind A. fumigatus and subsequently induce massive agglutination via recognition of specific sugar moieties on the conidia, for instance, N-acetylglucosamine and β-1,3-glucan, which comprises the core of the cell wall through complex structure with repeating sugar arrays (28). Previously, Coombs et al. (20) have shown that recombinant CL-12 ED has affinities with some monosaccharides, such as galactose, fucose, N-acetylgalactosamine, and α-methyl galactoside, and selectively recognizes glycans containing Lewis^X trisaccharide Galβ (1-4)[Fucα (1-3)]GlcNAc or the closely related Lewis^a trisaccharide Galβ (1-3)[Fucα (1-3)]GlcNAc. Different from these reports, we revealed that soluble CL-12 might have affinity for glycans bearing N-acetylglucosamine and β-1,3-glucan, suggesting that soluble CL-12 could bind to certain fungal pathogens by interaction with repeating carbohydrate arrays on the cell wall.

It has been shown that a variant form of CL-12 lacking the CRD still shows binding activity for bacteria (10). Recently, studies using CL-12 domain deletion mutants have shown that the collagen-like domain in CL-12 mainly recognizes microbial and fungal ligands and the CRD specifically binds some sugar substances, whereas the coiled-coil domain additionally interacts with modified low density lipoprotein (29). These findings suggest that CL-12 differs from the classical collectins both in the arrangement and the functions of these domains. Combining with these reports, our observations that soluble CL-12 bound A. fumigatus in a divalent cation-independent fashion also imply that other functional domains could be involved in the opsonization in addition to the CRD. Previously, we have shown that recombinant CL-11 could bind A. fumigatus in a Ca²⁺-independent fashion, which could not be inhibited by its defined sugar ligands, such as l-fucose or d-mannose, and this did not lead to the subsequent serum-induced complement activation (16). However, binding of CL-11 to C. albicans occurring via its CRD led to MASP binding and complement activation (16). This suggests that if the binding site involves the collagen-like domain, it might hamper complex formation with MASPs and thus preclude LP activation. In analogy with this finding, soluble CL-12 did not show a clear influence in complement C4b deposition upon support of only Ca²⁺. These results suggest that soluble CL-12 might be incapable of activating the complement system via the CP or LP. Because of this, as soluble CL-12 binding to A. fumigatus was Ca²⁺ independent, we investigated whether soluble CL-12 could activate complement in the presence of EGTA-Mg²⁺, which selectively allows activation of the AP. Our results demonstrated that soluble CL-12 drastically augmented the AP-mediated C3b and TCC deposition in a dose-dependent fashion. However, it is noteworthy that properdin is capable of binding only to the soluble form of CL-12, but not to the CHO cell membrane-anchored form, emphasizing that appropriate orientation of CL-12 may be critical during interaction with properdin. These findings suggest that the CL-12/properdin–directed AP complement amplification might be highly restricted to the soluble form of CL-12 perhaps to prevent constitutive AP complement activation.

Properdin is a plasma glycoprotein of the complement system and plays a crucial role in host defenses against bacteria and fungi (30). However, isolated properdin and properdin in serum have shown striking differences in target-binding activities (31, 32). In agreement with another previous recent report, we observed that isolated properdin binding was drastically interfered with in the context of serum on the microbial surface as on A. fumigatus (33). However, our results revealed that soluble CL-12 might also provoke and direct target-binding activity of properdin from serum under certain physiological and pathophysiological conditions.

Decades of research have solidified the role of properdin as a stabilizer of the AP C3 convertase C3bBb, but the properdin protein itself was viewed upon having a relatively nonspecific role (34). Previously, properdin has been shown to bind AP activator surfaces strictly in a C3-dependent manner (33, 35). However, several lines of recent evidence suggest that properdin by itself may provide a platform for the in situ assembly of C3bBb complex on target surfaces, therefore providing new insights into properdin function (31, 36). Nevertheless, it must be emphasized that this is a disputed area of research (33). In this study, we demonstrated that soluble CL-12 recruits properdin to target surfaces, thereby allowing AP activation. However, our results also indicated that unlike the soluble form of CL-12, the membrane-anchored CL-12 is incapable of mediating properdin-directed complement activation, implying that nonspecific and uncontrolled self-activation may not occur under normal conditions. Taken together, our results revealed that given a platform via a cooperative immune response integrating signals from another “sensory” input, for instance, sPRM as CL-12, properdin is able to mediate de novo C3 convertase assembly. Therefore, properdin may initiate complement activation on certain pathogens more selectively via recognition of specific target surfaces by its “sensory” input, compared with standard AP of complement activation via a highly reactive but nondiscriminating nascent C3b, which binds covalently to a target. Thus, the evidence presented in this article supports the idea that properdin as a PRM may direct AP of complement activation (30, 31). Nevertheless, we envision that properdin plays a role in AP of complement activation not only by stabilizing preformed C3bBb, but also by providing a platform for de novo C3 convertase assembly once surface-bound through either surface glycosaminoglycans or cooperative recognition by the humoral innate immune system. However, both mechanisms require tick-over–generated nascent C3b.

In summary, like the CP and LP of complement relying on sPRMs for initiation, the AP of complement may also be initiated by specific recognition molecules via collaboration with properdin. This soluble CL-12/properdin collaboration–directed AP of complement activation could be particularly effective against A. fumigatus, an opportunistic fungal pathogen, which causes mortality in critically ill and immunocompromised patients. Therefore, we provide insight into understanding a novel mechanism of the complement amplification via the AP, which might contribute adjunct to antibiotic therapy in the treatment of certain fungal infections or possibly other pathogens.

The authors have no financial conflicts of interest.