key: cord-0023291-asteuhze
authors: Sun, Jiejie; Wang, Liyan; Yang, Wenwen; Li, Yinan; Jin, Yingnan; Wang, Lingling; Song, Linsheng
title: A novel C-type lectin activates the complement cascade in the primitive oyster Crassostrea gigas
date: 2021-10-27
journal: J Biol Chem
DOI: 10.1016/j.jbc.2021.101352
sha: 4ae2215ea50fbcc8f8677296c2137eac63b13cf2
doc_id: 23291
cord_uid: asteuhze

The ancient origin of the lectin pathway of the complement system can be traced back to protochordates (such as amphioxus and tunicates) by the presence of components such as ficolin, glucose-binding lectin, mannose-binding lectin-associated serine protease (MASP), and C3. Evidence for a more primitive origin is offered in the present study on the Pacific oyster Crassostrea gigas. C3 protein in C. gigas (CgC3) was found to be cleaved after stimulation with the bacteria Vibrio splendidus. In addition, we identified a novel C-type lectin (defined as CgCLec) with a complement control protein (CCP) domain, which recognized various pathogen-associated molecular patterns (PAMPs) and bacteria. This protein was involved in the activation of the complement system by binding CgMASPL-1 to promote cleavage of CgC3. The production of cytokines and antibacterial peptides, as well as the phagocytotic ratio of haemocytes in CgCLec-CCP-, CgMASPL-1-, or CgC3-knockdown oysters, decreased significantly after V. splendidus stimulation. Moreover, this activated CgC3 participated in perforation of bacterial envelopes and inhibiting survival of the infecting bacteria. These results collectively suggest that there existed an ancient lectin pathway in molluscs, which was activated by a complement cascade to regulate the production of immune effectors, phagocytosis, and bacterial lysis.

The complement system represents one of the most ancient cornerstones of immunity, which plays essential roles in both innate and adaptive immunity. The mammalian complement system consists of many serum proteins involved in a chain reaction of proteolysis and protein complex assembly that leads to the elimination of pathogens. The complement system can be activated by three distinct pathways, including classical, lectin and alternative pathways, and each of them is composed of immune recognition and a common lytic process to the destruction of invading pathogens (1) (2) (3) (4) . The lectin pathway is thought to be the most primitive complement pathway. It is activated by the binding between mannose-binding lectin (MBL)/ficolin and carbohydrate patterns from pathogens to assemble different MBL-associated serine proteases (MASPs). Complement proteins C2 and C4 are cleaved by MASPs to form C3 convertase complex (C4aC2b), which converts C3 into a variety of products (5, 6) to mediate various processes, including inflammation, phagocytosis, chemotaxis, and cell lysis (7) (8) (9) .

The origin and evolution of the complement lectin pathway can be traced back to early deuterostomes, such as lamprey and/or hagfish, amphioxus (lancelets) and ascidian (tunicates) (7) . The key components of the lectin pathway, including C3, MBL, and MASP, have been identified in lamprey and/or hagfish (7, 10) . In lamprey, the lectin pathway can be activated by MBL and MASP to generate C3 cleavage fragments (7, 11) and a subsequently cytolytic response, which is distinct from the mammalian lytic pathway (10, (12) (13) (14) . A ficolin-mediated complement pathway has been identified in amphioxus, in which MASP1/3 can form complexes with ficolin (FCN1) to facilitate the activation of C3 (11) . The more ancient origin of lectin pathway is further proved by the identification of glucose-binding lectin (GBL) (15) , MASP, (16) and C3 (17) from ascidian. The GBL from ascidian lacks a collagen-like domain presenting in mammalian MBLs, but it can associate with MASPs to activate C3 and mediate haemocyte phagocytosis (18) . These pieces of evidence indicate that the lectindependent complement system has been developed in Urochordates and Chordates (19) .

The identification of MASP or MASP-like molecule in protostomes suggests that the lectin pathway antedates the classical and alternative pathways of the complement activation. Actually, the existence of complement systems in insects has been controversial for a long time, even the thioestercontaining proteins (TEPs) are proved to execute similar functions as mammalian C3 (20) . In shrimp Litopenaeus vannamei, only one C3-like gene has been identified (21) . Although C3 and MASP have been identified in the sea anemone Nematostella vectensis, their functions are still not fully defined (22) (23) (24) . In molluscs, the homologues of C3 have been recently documented in oyster (25) , clam (26) (27) (28) , and mussel (29) , which contain all the conservative motifs identified in mammalian C3. Moreover, MASP-like molecule 1 (MASPL-1) lacking complement control protein (CCP) domains was also identified in oyster (30) , suggesting that a complement lectin pathway might exist in molluscs. Though MASP has been identified in protostomes, there are still no reports about protostome MBL/ficolin homologues with collagen-like regions. It is suspected that the ancient complement pathway is likely to undergo major changes during evolution (7, 15, 31) . At the same time, due to the lack of some key molecules, especially most of the proteins involved in the formation of membrane attack complex (MAC), such as C5, C6, C7, C8, and C9 (32) in early deuterostomes and protostomes, the lectin pathway is suspected to be assembled in some special ways to induce the cytolytic systems, which is different from mammalian ones (7) .

Molluscs are protostome animals with evolutionary significance, and they also represent a significant proportion of world aquaculture production, in which the Pacific oyster Crassostrea gigas is considered as one of the most important commercial marine species. Recently, a large number of gene models containing complement-related domains have been characterized in oyster C. gigas, and C3 has been identified in the cell-free haemolymph as a processed mature protein, suggesting that the complement system might be of significance for the innate immune response of molluscs (25) . However, the activation mechanism of the complement system and its functions in molluscs are still not clear. Although hundreds of C-type lectins (CTLs), fibrinogen-related proteins (FREPs), and MASP-like proteins have been found in C. gigas (25, 33) , there are no homologues sharing similar domain organization with mammalian MBL/ficolin/MASP. These findings encourage us to suspect that the lectin pathway in primitive invertebrates may assemble in flexible ways different from that in vertebrates, in which other paralogous molecules with complement-related domains may be dedicated to the activation cascade. In the present study, a novel C-type lectin was identified from C. gigas (designated CgCLec-CCP) with the objectives to confirm the existence of the ancient complement lectin pathway and understand its activation mechanism and roles in antibacterial immunity in molluscs.

The distribution, expression, activation, and antibacterial activity of CgC3 after V. splendidus stimulation A C3 homologue was found by bioinformatic means from oyster (CgC3) with an A2M_N_2 (Alpha-2-macroglobulin) domain, an A2M (Alpha-2-macroglobulin family) domain, a TBR (Alpha-macro-globulin thiol-ester bond-forming region) domain, an A2Mc (Complement component region of A2M), an A2MR (A-macroglobulin receptor) domain, and an NTR (Netrin) module (Figs. 1A and S1). The mRNA transcripts of CgC3 were examined by qRT-PCR, and they were found to be expressed in all the tested tissues with relatively higher expression level in haemocytes, mantle, and adductor muscle (Fig. 1B) . Haemocytes are the main invertebrate immune cells circulating in a semi open circulatory system containing haemolymph (the circulatory fluid). The relatively expression level of CgC3 mRNA in haemocytes increased significantly at 6, 12, 24, and 48 h after Vibrio splendidus stimulation, which were 3.5-fold, 16.6-fold, 8.2-fold, and 6.9-fold than that at 0 h, respectively (Fig. 1C) . The polyclonal antibody against CgC3-NTR (anti-CgC3) was prepared from mouse, and its specificity was examined by western blotting. There was a distinct band of recombinant His-CgC3-NTR, which was consistent with the predicted molecular weight of His-CgC3-NTR (Fig.  1D) . A band of CgC3 from cell-free haemolymph of untreated oysters was shown, which was identical to the prediction of molecular mass of CgC3 (Fig. 1E) . Usually, C3 molecule consists of an αand a β-chain, which can be cleaved into different fragments. Coimmunoprecipitation with anti-CgC3 revealed that there were four bands (the fulllength CgC3, alpha chain of CgC3, alpha chain fragment of iCgC3b, and C3γ fragment) for the proteins from cell-free haemolymph and haemocytes from the oysters stimulated by V. splendidus for 30 min (Figs. 1E and S1). The mass spectrometric analysis was conducted using the cell-free haemolymph of oysters. Anti-CgC3 antibody was used to immunoprecipitate CgC3 fragments in the cell-free haemolymph from V. splendidus-infected oysters, and the proteins were analyzed by SDS-PAGE. The immunoprecipitated CgC3 fragment (alpha chain fragment of iC3b) was identified by mass spectrometric analysis (Beijing Protein Innovation, BPI) (Fig. S2, A and B) . In the immunocytochemistry assay with anti-CgC3, CgC3 was marked with Alexa Fluor 488-labeled Goat Anti-Mouse IgG in green, and nucleus was stained by 4 0 ,6-diamidino-2-phenylindole (DAPI) in blue. The positive signals of CgC3 were mainly distributed in haemocyte cytoplasm, which were transferred to the membrane after V. splendidus stimulation, compared with those in phosphate buffered saline (PBS) group (Fig. 1F ). After V. splendidus were incubated with the native complex proteins from oysters coimmunoprecipitated by anti-CgC3 after V. splendidus stimulation, the bacteria growth was inhibited, and the OD value decreased significantly from 2 to 5 h, which was 0.7-to 0.8-fold of that in PBS group (Fig. 1G) . Meanwhile, the envelope of V. splendidus was observed to be perforated (Fig. 1H) .

The molecular features, mRNA expression patterns, and ligand-binding activities of CgCLec-CCP A C-type lectin (defined as CgCLec-CCP) with a CCP domain and a carbohydrate recognition domain (CRD) domain was identified in oyster C. gigas ( Fig. 2A) . A short stretch of the N-terminal nonlectin region in CgCLec-CCP shared limited similarity with the collagen region in MBLs from other species, and six residues (CPD and WPD) in CRD domain of CgCLec-CCP were predicted to be involved in carbohydrate binding (Fig. S3) . The mRNA transcripts of CgCLec-CCP were distributed in all tested tissues, with relatively higher abundance in haemocytes and gonad

The complement cascade in the primitive oyster Ha, haemolymph; Gi, gills; Ma, mantle; Go, gonad; Ad, adductor muscle; He, hepatopancreas. C, the temporal and spatial mRNA expression patterns of CgC3 in haemocytes. D, the purification of rCgC3-NTR protein. E, detection of CgC3 cleavage in cell-free haemolymph and haemocytes by using anti-CgC3 antibody. F, the subcellular location of CgC3 after V. splendidus stimulation. G, the direct bacteriostatic/bacteriocidal activity of the native complex proteins coimmunoprecipitated by CgC3 antibody after V. splendidus stimulation. H, SEM image of bacterial cells after incubation with the native complex proteins coimmunoprecipitated by CgC3 antibody after V. splendidus stimulation. The sphere next to the bacterium was Protein G Agarose Bead. Data shown were the mean ± SD (n = 3). *p < 0.05, **p < 0.01 (Student's t test). Different letters: p < 0.05 (one-way ANOVA).

(21.4-fold and 11.7-fold of that in hepatopancreas, p < 0.05, respectively) (Fig. 2B) . The mRNA expression level of CgCLec-CCP in haemocytes increased significantly at 3, 6, 12, 24, 48, and 72 h after V. splendidus stimulation (Fig. 2C) . The recombinant forms of CgCLec-CCP (rHis-CgCLec-CCP) and Trx-His (rTrx-His) were expressed in Escherichia coli and purified by affinity chromatography (Fig. 2, D and E) . rCgCLec-CCP exhibited binding activity toward G − bacteria (E. coli and V. splendidus) and G + bacteria (Staphylococcus aureus and Micrococcus luteus) with the relatively higher binding activity to V. splendidus and M. luteus (Fig. 2F) . It bound LPS, PGN, MAN, and GLU in a dose-dependent manner and displayed relatively higher binding affinity toward LPS and PGN (Fig. 2G ). Western blot assay with anti-CgCLec-CCP antibody revealed that there was a distinct band of CgCLec-CCP in both haemocyte and cell-free haemolymph samples (Fig. 2, H and I) .

CgC3 cleavage, haemocyte phagocytosis, and cytokine production after CgCLec-CCP was blocked by the corresponding antibody or knocked down by RNAi

As no definite MBL ortholog had been identified in oysters, the CgCLec-CCP with additional CCP domain was suspected to display the same function as the mammalian MBL. The cleavage of CgC3 was examined after CgCLec-CCP was blocked by its antibody. In the pre-serum-blocked oysters, there were two bands (the full-length CgC3 and alpha chain fragment of iCgC3b) in cell-free haemolymph after V. splendidus stimulation, while the band of the full-length CgC3 became thicker and the band of alpha chain fragment of iCgC3b became thinner in anti-CgCLec-CCP antibody-blocked oysters after V. splendidus stimulation (Fig. 3A) . After CgCLec-CCP was knocked down by RNA interference (RNAi) (Fig. 3B) , the phagocytotic rate toward V. splendidus and mRNA transcripts of CgIL17-1 and CgTNF-1 all decreased compared with The complement cascade in the primitive oyster those in EGFP-RNAi oysters, which was 0.8-fold, 0.4-fold, and 0.1-fold (p < 0.05) of that in EGFP-RNAi oysters, respectively ( Fig. 3 , C-E). However, the number of V. splendidus in the haemolymph increased significantly (16.2-fold of that in EGFP-RNAi oysters, p < 0.01) in CgCLec-CCP-RNAi oysters (Fig. 3F ).

MASP is a member of the lectin pathway of complement, which can be activated by MBL/ficolin to induce the cleavage of C3 in mammals. A MASP-like protein (CgMASPL-1) was identified in oyster with four CUB domains, three LDLs domains, and a Tryp_SPc domain (Figs. 4A and S4A). Multiple alignment and phylogeny analysis revealed relatively lower conservative property between CgMASPL-1 and MASPs from different species (Fig. S4 , B and C). There were two additional cysteines (Cys683 and Cys808) in CgMASPL-1 to form the "histidine loop" disulfide bridge in its protease domain with the same manner as that in HsMASP1 and MsMASP1 (Fig. S4 , A and C). The mRNA transcripts of CgMASPL-1 were expressed in all tested tissues, with higher level in hepatopancreas (Fig.  4B ). The expression level of CgMASPL-1 mRNA in haemocytes increased significantly at 6, 12, and 24 h after V. splendidus stimulation, which were 10.2-fold, 4.4-fold, and 6.0-fold (p < 0.05) of that in Blank group, respectively (Fig. 4C ). The purified recombinant protein of CgMASPL-1-CUB1 (Fig. 4D ) was used for pull-down assay. There was a band corresponding to rHis-CgCLec-CCP, while no band was observed in rTrx-His control group (Fig. 4 , E and F). CgMASPL-1 was coimmunoprecipitated by anti-CgCLec-CCP antibodies, and the band of CgMASPL-1 became thicker after V. splendidus stimulation, compared with that in PBS group (Fig. 4G ). The purified rGST-CgMASPL-1-CUB1 was used to bind the native CgC3 from haemolymph. Pull-down assay with anti-CgC3 antibody revealed that there was a band of alpha chain fragment of CgiC3b (Fig. 4H) . When the haemolymph sample from the oyster stimulated by V. splendidus was coimmunoprecipitated with the CgMASPL-1 antibodies, there were four bands observed, which were correspondent to the complex of the full-length CgC3 and CgMASPL-1, the full-length CgC3, alpha chain of CgC3, and alpha chain fragment of iCgC3b, respectively (Fig. 4I ).

The phagocytotic rate of haemocytes, productions of CgIL17-1, CgIL17-2, CgTNF-1, and CgBigDef1 in haemocytes, and number of V. splendidus in haemolymph after CgMASPL-1 was knocked down

RNAi was used to specifically inhibit the expression of CgMASPL-1 in haemocytes at 24 h after the injection of CgMASPL-1 double-stranded RNA (dsRNA). The mRNA The complement cascade in the primitive oyster expression level of CgMASPL-1 decreased significantly, which was 0.2-fold lower than that in EGFP-RNAi oysters (Fig. 5A ). In CgMASPL-1-RNAi oysters, the phagocytotic rate of haemocytes toward V. splendidus was 0.5-fold (p < 0.01) of that in EGFP-RNAi oysters (Fig. 5, B and C) . Meanwhile, the mRNA expression levels of CgIL17-1, CgIL17-2, CgTNF-1, and CgBigDef1 also decreased significantly, which were 0.1-fold, 0.4-fold, 0.3-fold, and 0.1-fold (p < 0.05) of that in EGFP-RNAi oysters, respectively (Fig. 5, D-G) . However, the number of V. splendidus in the haemolymph increased significantly, which was 18.4-fold of that in EGFP-RNAi oysters (p < 0.01) (Fig.  5H) .

The phagocytotic rate of haemocytes, productions of CgIL17-1, CgIL17-2, CgTNF-1, and CgBigDef1 in haemocytes, and number of V. splendidus in haemolymph after CgC3 was knocked down

The complement lectin pathway is able to induce proinflammatory factor secretion and cytolysis. The phagocytosis and production of some cytokines were determined after CgC3 was knocked down by RNAi. The immunocytochemistry assay by using the anti-CgC3 antibody and FITC-V. splendidus was conducted to detect the location of activated CgC3 on the cell surface of V. splendidus. CgC3 was labeled with Alexa Fluor 488-labeled Goat Anti-Mouse IgG in red, and nucleus was stained by DAPI in blue. The positive red signals of CgC3 distributed on the cell surface of V. splendidus, while no positive signals were observed in the negative control with Preserum (Fig. 6A) . The expression level of CgC3 mRNA decreased significantly at 24 h after the injection of CgC3 dsRNA, which was 0.48-fold lower than that in EGFP-RNAi oysters (Fig. 6B ). The number of phagocytized V. splendidus in haemocytes was counted by flow cytometry analysis (Fig.  6C ) to calculate the phagocytotic rate. The phagocytotic rate decreased significantly in CgC3-RNAi oysters (Fig. 6D) , which was 0.6-fold of that in EGFP-RNAi oysters (Fig. 6, E and F) . Meanwhile, the mRNA expression levels of CgIL17-1, CgIL17-2, and CgBigDef1 all decreased significantly in CgC3-RNAi oysters, which were 0.2-fold, 0.2-fold, and 0.1-fold (p < 0.05) of that in EGFP-RNAi oysters, respectively (Fig. 6 , G-J). However, the number of V. splendidus in the haemolymph increased significantly (4.1-fold of that in EGFP-RNAi oysters, p < 0.01) (Fig. 6K ).

The lectin pathway is one of the well-studied complement activation pathways in vertebrates, which plays significant roles in opsonization of pathogens, chemotaxis, activation of leukocytes, directly killing of pathogens, and modulation of inflammatory (34, 35) . It can be triggered by the specific activation cascade, in which MBLs and ficolins firstly recognize PAMPs and subsequently associate with MASPs. The activation of the lectin pathway promotes the cleavage of C3, and the resulting products can bind the invading bacteria and destruct them by lytic pathway (36) . The increasing evidence demonstrates that the complement system also exists in invertebrates (31) . A number of genes encoding the components in the lectin pathway have been annotated in the genome of ascidian Ciona intestinalis and amphioxus B. floridae (37, 38), which The complement cascade in the primitive oyster undoubtedly indicates that the lectin pathway can be evolutionary traced back to Urochordata (7, 39) . Recently, C3 was characterized to play crucial roles in antibacterial immunity of molluscs (25) . However, as the homologues of MASP or MBL/ ficolin have not been identified in molluscs, the existence of molluscan lectin pathway is more controversial. It is suspected that the lectin pathway in C. gigas is likely to have undergone major changes during its evolution, and its activation cascade needs further confirmation.

In vertebrates, C3 is the central component of the complement system. It is mainly synthesized in liver and also locally macrophages, and rich in serum. C3 can be cleaved by the C3 convertases (C4bC2a) to generate C3b and C3a (40) . C3b can be further cleaved by H factor and I factor into iC3b, then C3dg and C3c (41) . In the present study, CgC3 mRNA was expressed in all examined tissues with higher abundance in haemocytes. CgC3 protein was distributed in haemolymph and haemocytes. It was cleaved to generate the fragments (CgC3α, CgiC3bα, and CgC3γ) after V. splendidus stimulation.

In mammals, MBL has been identified as the main recognition molecule of the lectin pathway, which contains a conserved collagen-like domain and a CRD (36, 39) . After the binding to carbohydrates via its CRD, MBL forms oligomer through the collagen-like region, which then activates lectin pathway (42) . Glucose-binding lectin (GBL) with an N-terminus coiled-coil region instead of the collagen region is the only MBL homologue reported in tunicate Halocynthia roretzi, which participates in the lectin pathway (15) . In the present study, CgCLec-CCP identified from oyster C. gigas harbored an extra CCP domain, but lacked a collagen region, compared with mammalian MBLs. There were six conserved residues (CPD and WPD) in the CRD of CgCLec-CCP, which was suspected to bind mannose, GlcNAc, and glucose directly in the presence of Ca 2+ similar to that of other reported MBLs (43) , indicating that CgCLec-CCP was a member of the mannose-binding lectin family. The in vitro assay confirmed that rCgCLec-CCP was able to recognize different bacteria and PAMPs with a broad recognition spectrum. There is a CCP domain in CgCLec-CCP, which is an evolutionarily conserved module essential for complement-mediated immune functions (44) . CCP domain has been identified in several proteins of the complement system, such as B factor, C1, C2, and MASPs. In mammals, the CCP domains in MASP-2 are likely to house an exosite required for efficient cleavage of C4 by binding to the C345 C domain of C4 (45) , and the CCP domain in C1r is responsible for its dimer formation (46) . It is speculated that the CCP domain in CgCLec-CCP may have the same function as the collagen region. As CgCLec-CCP was upregulated after bacterial stimulation, our data suggest that this protein functions as a PRR to recognize bacteria and activate the complement-mediated immune response through its CCP domain.

MASP is the key component in the complement system, which mainly performs the cleavage of the complement components, such as C2, C4, and C3 (47) . The homologues of mammalian MASP-1 have been identified in invertebrates, including lamprey (5) and amphioxus (13) and ascidian (16) . But those identified MASP homologues from these primitive invertebrates share low similarity with mammalian MASPs. In this present study, CgMASPL-1 had N-terminal CUB domains and LDLa domains, while it lacked CCP domains, compared with mammalian MASPs. The lack of the CCP in MASPL-1 might be compensated by CgCLec-CCP. There were two additional cysteines (Cys683 and Cys808) in CgMASPL-1 to form the "histidine loop" disulfide bridge in the protease domain, which shared the same manner as HsMASP1 and MsMASP1 (14) . In mammals, the N-terminal CUB1-EGF-CUB2 segment of MASPs plays a key role in binding MBL or ficolins, while its C-terminal CCP1-CCP2-SP segment contributes to the catalysis of substrate cleavage (47) . In the present study, rCgMASPL-1-CUB1 interacted directly with CgCLec-CCP in vitro, suggesting that the Nterminal of oyster CgMASPL-1 played a similar role as mammalian MASPs in binding CTL via its CUB1 domain. In amphioxus, FCN1 interacts with the N-terminal CUB1-EGF-CUB2 segment of MASP1/3, indicating that there exists the classical MASPs and similar binding mechanism as that in mammals (11) . The MASP family is divided into two phylogenetic lineages, TCN-type and AGY-type lineages (48) . CgMASPL-1 is a typical AGY-type MASP. These structural features suggested that CgMASPL-1 was the orthologue of mammalian MASP-1/3 (48) , and more interestingly, it lacked the conserved CCP domain, which was different from the classic complement serine proteases in other species. All these results indicated that CgCLec-CCP might have evolved early as a prototype of MBL to activate CgMASPL-1 in molluscs. C3 can be cleaved by MASPs to trigger the activation of lectin pathway. Mammalian MASP1 is able to mediate the proteolytic activation of C4 and C2 to eventually induce the cleavage of C3 (47) . The CCP1-CCP2-SP segment in human MASP-1 and the CCP-SP segment in amphioxus BjMASP1/3 can directly activate C3 (11, 49) . The MBL-MASP complex in lamprey as well as GBL-MASPa complex in ascidian is also able to activate C3 in the same manner as the human MBL-MASP1 complex (10) . Even the main components of lectin pathway, such as MASP and C3, have also been found in protostomes, the existence of activable lectin pathway is still in controversy. In the present study, CgCLec-CCP was found to associate CgMASPL-1 in response to bacterial infection and then to cleave CgC3 directly. The cleaved fragments of CgC3 were similar to those of C3 in other species, suggesting that CgMASPL-1 was able to directly cleave oyster CgC3 (15, 39, 50) . Similarly, ascidian MASPa associated with GBL was also demonstrated to cleave ascidian C3 directly (15) . As there is no CCP domain in CgMASPL-1, which is different from human MASP-1, amphioxus BjMASP1/3, and ascidian MASPa, the cleavage mechanism of CgMASPL-1 should be unique and needs further investigation. These results collectively indicated that CgMASPL-1 was able to interact with C3 to promote the cleavage of C3, suggesting a new activation mechanism of C3 in molluscs and the presence of the primordial lectin pathway in primitive protostomes (18) .

It is well known that C3 is cleaved by the C3 convertases to generate C3b and C3a, and C3b can bind the microbe surface to enhance cell phagocytosis (40) . The phagocytic promotion mediated by C3b is currently suspected as one of the most important functions of the complement system in invertebrates (25) . In ascidian and sea urchin, C3 has been reported to promote haemocyte phagocytosis toward yeast (15, 51) . In the present study, the activated CgC3 induced phagocytosis of oyster haemocytes, suggesting that the complement components probably functioned as opsonins to assist the haemocytes to phagocytose the opsonized pathogenic organisms (25) .

It has been documented that the activation of complement lectin pathway can induce proinflammatory factor secretions and cytolysis (52) . For example, C3 could mediate the IL17A secretions in mice T cells (53) . In the present study, the productions of CgIL17-1, CgIL17-2, and CgBigDef-1 in CgCLec-CCP-, CgMASPL-1-, or CgC3-RNAi oysters were reduced after V. splendidus stimulation, suggesting that the ancient lectin pathway in oyster exhibited similar functions as that of mammals. For the vertebrate complement system, the most important immune reaction is to induce the cytolysis of invading pathogen. Mammalian C3b combines with C4bC2a in a covalent bond to form C5 convertases and initiate the assembly of C6 to C9 to form MACs to disturb the integrity of cell membranes of microbes (40) . Because the components of C6 to C9 have not been identified in agnathans and invertebrates (7) , the lytic pathway is suspected to be different from that in mammals. It has been reported that C3 in human saliva, lamprey and amphioxus humoral fluids, and haemolymph from shrimp L. vannamei and clam Sinonovacula

The complement cascade in the primitive oyster constricta are all essential to kill bacteria (21, 27, 54, 55) . In lampreys, complement activation causes the polymerization of pore-forming protein (LPFP) and finally leads to the cytolysis of pathogens, which is distinct from the mammalian lytic pathway (7) . In the present study, the activated CgC3 in oyster was found to directly inhibit the bacteria growth and perforated the envelope of bacteria in haemolymph, which was different from that in mammals and lampreys (56, 57) , suggesting that molluscs might have evolved an ancient cytolytic system. The results indicated that the complement system in oyster was activated by lectin pathway to induce phagocytosis, production of IL17s and defensin, and cytolysis, which played important roles in the immune defense against pathogen invasion.

In conclusion, the complete complement lectin pathway with unique activation mechanism was characterized in oyster. Upon recognizing extracellular ligands, CgCLec-CCP with a CCP domain interacted with the CUB domain of CgMASPL-1 to cleave CgC3 directly. The activated CgC3 successively mediated multiple immune responses, such as haemocyte phagocytosis, secretions of cytokines and AMPs, perforating the bacterial envelopes, and inhibition of bacterial growth and survival (Fig. 7) . Defining the ancient lectin pathway, as we have done, is important to understand the activation mechanism and immune functions of complement system in invertebrates as well as the origin and evolution of mammalian complement system.

Pacific oysters, C. gigas (shell length 12-16 cm) were collected from a local breeding farm in Dalian, Liaoning, China. They were cultured in aerated seawater at 15 C ± 2 deg. C for 14 days before processing.

The oysters individually received an injection with 100 μl of V. splendidus at 2 × 10 8 CFU/ml dissolved in PBS. Nine oysters were randomly sampled from each group at 0, 3, 6, 12, 24, 48, 72, and 96 h after V. splendidus stimulation. The haemolymphs collected from three oysters were pooled together as one sample, and there were three replicates for each time point. The haemocytes, as the immune cells of oysters, were collected by centrifugation at 800g, 4 C for 10 min, and the total RNA was extracted using Trizol reagent (Invitrogen) to detect the temporal expression patterns of CgCLec-CCP, CgMASPL-1, and CgC3 as the previous description (58) . Tissues including muscle, gill, mantle, hepatopancreas (the important gland of certain invertebrates that combines the functions of liver and pancreas) (33) , and haemolymph (the circulatory fluid of invertebrates) (34) from other nine untreated oysters were pooled as three replicates for RNA extraction to examine the mRNA distribution of the target genes in different tissues.

The full-length cDNA fragments of CgCLec-CCP, CgMASPL-1, and CgC3 were obtained by PCR with the primers (Table S1 ) designed according to the sequence information of CGI_10008644, LOC105342011, and LOC105333243 deposited in the NCBI database (https://www. ncbi.nlm.nih.gov/), respectively. A translation tool (http://web. expasy.org/translate/) and SMART (http://smart.emblheidelberg.de/) were used to predict the amino acid sequences and the domains in CgCLec-CCP, CgMASPL-1, and CgC3.

qRT-PCR assays with primers of CgCLec-CCP-RT, CgMASPL-1-RT, and CgC3-RT (Table S1 ) were performed to examine their expression in tissues and their temporal expression profiles in haemocytes after V. splendidus stimulation. The fragment of elongation factor (CgEF) (NP_001292242.2) amplified with the primers of CgEF-RT (Table S1 ) was used as an internal reference. The mRNA transcripts of interleukin 17-1 (CgIL17-1), tumor necrosis factor 1 (CgTNF-1), and big defensing 1 (CgBigDef1) after the interference of CgCLec-CCP, CgMASPL-1, and CgC3 were determined by using the primers of CgIL17-1-RT, CgIL17-1-RT, CgTNF-1-RT, and CgBigDef1-RT (Table S1) , respectively. qRT-PCR was programmed at 95 C for 10 min, followed by 40 cycles at 95 C for 10 s, and 60 C for 45 s. The final product was analyzed via melting analysis from 65 to 95 C. Their relative mRNA expression levels were calculated using the 2 −ΔΔCt method (59) . The data were statistically analyzed with t test, and the significant differences were accepted at p < 0.05. Recombinant expression, purification, and antiserum production of CgCLec-CCP, CgMASPL-1, and CgC3

The recombinant proteins of CgCLec-CCP (rCgCLec-CCP), the N-terminal CUB domain of CgMASPL-1 (rCgMASPL-1-CUB1), and the NTR domain of CgC3 (CgC3 -NTR) were expressed in E. coli according to the previous report (25) . The coding sequences of CgCLec-CCP, CgMASPL-1-CUB1, and CgC3-NTR were amplified by using the primer pairs of CgCLec-CCP-ExF and -ExR, CgMASPL-1-CUB1-ExF and -ExR, CgC3 -NTR-ExF and -ExR (Table S1 ), respectively. The PCR products were inserted into the pGEX-4T-1 or pET-30a expression vector and expressed in E. coli transetta (DE3). His-tag (rHis) from pET-32a expression vector was used as control. The recombinant proteins were purified by affinity chromatography using GST-Bind or His-Bind resin following the manufacturer's instructions. The antisera of CgCLec-CCP, CgMASPL-1, and CgC3 were prepared as the previous report (58) , which was validated by using western blotting.

The determination of CgCLec-CCP, CgMASPL-1, and CgC3 proteins by western blotting analysis

The deduced amino acid sequence of CgTubulin (NP_001292292.1) was found to be highly conserved with 81% identity to that of human beta-Tubulin (NP_001184110.1). The anti-human beta-Tubulin (AF1216, Beyotime Biotechnology) antibodies were therefore employed for western blotting assay. The proteins from cell free haemolymph and haemocytes were extracted from oysters at 30 min after V. splendidus stimulation, separated by 10% or 15% SDS-polyacrylamide gel electrophoresis, and then transferred onto the nitrocellulose membrane by mini transfer tank for electrophoresis. After blocked with 3% nonfat milk in TBST (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.2% Tween-20) for 2 h, the membranes were incubated with 1/1000 diluted antisera against CgCLec-CCP, CgMASPL-1, and CgC3 in TBST with 3% nonfat milk at room temperature for 2 h, respectively. The subsequent steps for determining CgCLec-CCP, CgMASPL-1, and CgC3 proteins were performed as the previous report (58) .

The synthesis of dsRNA and RNA interference of CgCLec-CCP, CgMASPL-1, and CgC3

The 3 0 -terminal cDNA sequences (about 500 bp) of CgCLec-CCP, CgMASPL-1, CgC3, and EGFP were amplified by the primers Fi and Ri linked to the T7 promoter (Table S1 ) as templates to synthesize dsRNA using T7 polymerase (Takara) according to the instruction, respectively. The dsRNAs (50 μg) for CgCLec-CCP, CgMASPL-1, CgC3, and EGFP were individually injected into each oyster. To enhance the RNAi effect, a second injection was conducted at 12 h after the first injection. Haemolymph was collected from nine oysters at 24 h after the second injection and were pooled together as three replicates (each replicate was a mixture from three individuals). At 24 h after the first injection of dsRNA, the oysters received an injection of 100 μl V splendidus (2 × 10 8 CFU/ml). The haemocytes were harvested by centrifugation, and the total RNA was extracted and assessed by qRT-PCR with specific primers RT-F and RT-R (Table S1) to evaluate the RNAi efficacy. The qRT-PCR reactions were carried out on Quan Studio 6 Flex (Thermo Fisher) using SYBR premix ExTaq (RR420, Takara, Dalian). The mRNA transcripts of CgIL17-1, CgIL17-1, CgTNF-1, and CgBigDef1 in CgCLec-CCP-, CgMASPL-1-, or CgC3-RNAi oysters were detected at 12 h after V. splendidus stimulation. The relative expression levels were calculated using the 2 −ΔΔCt method (59) and statistically analyzed with t test. The significant differences were accepted at p < 0.05.

Haemolymph was obtained from three oysters at 24 h after the first injection of dsRNA and centrifuged at 400g, 4 C for 10 min to collect the haemocytes, which were then incubated with V. splendidus for 1 h to determine the phagocytic rate of haemocytes by flow cytometry assay with Amnis ImageStream mkII and IDEAS analysis software in CgCLec-CCP-, CgMASPL-1-, and CgC3-RNAi oysters, respectively. The phagocytic rate was the percentage of the haemocytes that phagocytosed bacteria in total haemocytes. Ten thousand haemocytes were analyzed for each sample.

At 24 h after the first injection of dsRNA, the oysters received an injection of 100 μl V splendidus (2 × 10 8 CFU/ml). The haemolymph was collected from three oysters at 1 h after V. splendidus stimulation and diluted 1000-fold with PBS. Each aliquot of 100 μl diluted haemolymph from CgCLec-CCP-, CgMASPL-1-, or CgC3-RNAi oysters was cultured in a plate containing 2216E medium at 25 C for 3 days to count the colony-forming units in the plate. The data were statistically analyzed with t test, and significant differences were accepted at p < 0.05.

Gram-negative bacteria (E. coli and V. splendidus) and Gram-positive bacteria (S. aureus and M. luteus) were used to test the bacterial-binding activities of rHis-CgCLec-CCP with rTrx-His tag as control. Bacteria were cultured in 3 ml of Luria-Bertani (LB) medium (1% tryptone, 0.5% yeast extract, and 1% NaCl) at 37 C overnight and then collected by centrifugation at 1000g for 5 min. After washed three times with TBS, the collected bacteria were resuspended in TBS and adjusted to an OD 600 of 1.0. The bacterial suspensions (400 μl) in TBS were separately incubated with 400 μl purified rHis-CgCLec-CCP (4 mM) and rTrx-His tag at room temperature with rotation for 1 h and then collected by centrifugation. The subsequent steps were carried out as the previous report (58) .

An enzyme-linked immunosorbent assay (ELISA) was used to examine the activities of rHis-CgCLec-CCP to directly bind pathogen-associated molecular patterns (PAMPs) (60) . The wells of the microplate were coated with 50 μl (2 μg) of LPS from E. coli, peptidoglycan (PGN) from S. aureus, mannose (MAN) from yeast and glucose (GLU), and incubated at 37 C overnight, respectively. The subsequent steps were performed as the previous report (58) . Samples with P (sample)-B (blank)/N (negative)-B (blank) > 2.1 were considered as positive. Three replications were performed for each sample, and the data were presented as mean ± SD (n = 3).

The immunocytochemical assay to determining the localization of CgC3 in haemocytes One milliliter of oyster haemolymph was fixed with 1 ml of a mixture containing anticoagulant (pH 7.4) and 4% paraformaldehyde and then centrifuged at 600g, 4 C for 10 min to collect haemocytes. The immunocytochemical assay of haemocytes was conducted as the previous description (58) with the polyclonal anti-CgC3 as the primary antibody and Alexa Fluor 488-labeled Goat Anti-Mouse IgG (Beyotime Biotechnology) as the secondary antibody. The 4 0 -6-diamidino-2phenylindole dihydrochloride (Beyotime Biotechnology) was used to label haemocyte nuclei. Fluorescence was observed under inversion fluorescence microscope (Axio Imager A2; ZEISS).

The confocal laser scanning microscope observation of CgC3 localization on the cell surface of V. splendidus

The haemolymph was collected from nine oysters and mixed with an anticoagulant solution (pH 7.4). The haemocytes were harvested by centrifugation at 600g for 10 min and resuspended in M-L15 medium at a concentration of 10 6 cell/ ml. The bacteria V. splendidus were grown in 2216E media at 28 C with shaking at 160 rpm for 12 h to the mid-log phase and collected via centrifugation at 6000g for 10 min. The bacteria were treated with absolute formaldehyde for 10 min, washed with NaHCO3 (0.1 M, pH 9.0) four times, and incubated with FITC (1 mg/ml; Sigma) with continuously gentle stirring at 4 C overnight. After four times of washing with PBS, the FITC-labeled V. splendidus were resuspended in PBS at a concentration of 2 × 10 8 CFU/ml. Twenty microliters of FITC labeled V. splendidus was incubated with 1 ml of haemocytes in dark at room temperature with slight rotation for 0.5 h. The immunocytochemical assay of haemocytes was conducted as the previous description (58) with anti-CgC3 as the primary antibody and Alexa Fluor 647-labeled Goat Anti-Mouse IgG (Beyotime Biotechnology) as the secondary antibody. Fluorescence was observed by using confocal laser scanning microscope (LSM 800, ZEISS).

The proteins extracted from haemolymph and haemocytes were incubated with protein G agarose (Beyotime Biotechnology) for 10 min. The supernatants were collected by centrifugation at 400g, 4 C for 10 min and then incubated with antibodies specific for CgC3 at room temperature for 3 h. The mixture (bound protein and antibody) was incubated with protein G agarose at room temperature for 3 h. The resulting pellet (bound protein, antibody, and protein G agarose) was washed with TBS five times and analyzed by western blotting with the anti-CgC3. Meanwhile, the pellet was also incubated with V. splendidus for 1 h and then the V. splendidus was collected for scanning electron microscope (SEM) analysis to observe the perforation of bacterial envelopes. Moreover, the antibacterial activity of the pellet was determined following the previous report (61) .

The blockage assay with CgCLec-CCP and CgC3 antibodies

The blockage assay was conducted by treating the oysters with an injection of 50 μl anti-CgCLec-CCP (15 mg/ml) or anti-CgC3 (15 mg/ml). The oysters that received an injection with the same volume of preserum were used as the control. Another injection with 100 μl V splendidus (2 × 10 8 CFU/ml) was conducted at 1 h after the first injection with anti-CgCLec-CCP. The haemolymph was collected at 2 h after V. splendidus stimulation, and the proteins were extracted for western blotting with anti-CgC3 as mentioned above to detect the cleavage of CgC3.

The anti-CgC3 antibody-blocked assay was conducted to determine the phagocytic rate of haemocytes toward V. splendidus. Thirty minutes after the injection of anti-CgC3, the oysters received another injection of 100 μl FITC-V. splendidus (2 × 10 8 CFU/ml) with the same volume of preserum as control. The haemolymph was collected from three oysters at 1 h after the injection of FITC-V. splendidus and centrifuged at 400g, 4 C for 10 min to collect the haemocytes. The phagocytic rate of haemocytes toward FITC-V. splendidus was determined by flow cytometry assay.

Pull-down assay to determine the interaction between CgCLec-CCP and CgMASPL-1

The interaction between rHis-CgCLec-CCP and rGST-CgMASPL-1-CUB1 was determined by GST Pull-down analysis according to the previously reported method with some modification (62) . Briefly, 400 μl of rGST-CgMASPL-1-CUB1 (100 μg/ml) was incubated with 40 μl of Glutathione Agarose resin at room temperature for 1 h, and then 400 μl of rHis-CgCLec-CCP (100 μg/ml) was added into the mixture and incubated with gentle rocking for 2 h. The mixture containing rGST-CgMASPL-1-CUB1, resin, and rHis-CgCLec-CCP was washed five times with TBS and then analyzed by SDS-PAGE.

The collected oyster haemolymph was incubated with protein G for 10 min to remove nonspecific binding proteins. After centrifuged at 400g for 3 min, the haemolymph was incubated with anti-CgCLec-CCP and anti-CgMASPL-1 at room temperature for 3 h, which was then incubated with protein A + G at room temperature for 3 h, respectively. After washed with TBS for five times, the resulting pellet (bound protein, Ab, and protein G) was analyzed by western blotting using anti-CgMASPL-1 and anti-CgC3 antibodies as described above.

All data generated for this study are included within this manuscript.

Supporting information-This article contains supporting information.

Novel mechanisms and functions of complement

COVID-19, microangiopathy, hemostatic activation, and complement

C3 glomerulopathy -understanding a rare complement-driven renal disease

Therapeutic inhibition of the complement system in diseases of the central nervous system

Cleavage of the third component of complement (C3) by mannose-binding protein-associated serine protease (MASP) with subsequent complement activation

Murine red blood cells lack ligands for B cell siglecs, allowing strong activation by erythrocyte surface antigens

The complement system of agnathans

Complement: A key system for immune surveillance and homeostasis

Revised mechanism of complement lectin-pathway activation revealing the role of serine protease MASP-1 as the exclusive activator of MASP-2

Cloning and characterization of mannosebinding lectin from lamprey (Agnathans)

Functional characterization of a ficolin-mediated complement pathway in amphioxus

Purification of a lamprey complement protein homologous to the third component of the mammalian complement system

Origin of mannose-binding lectin-associated serine protease (MASP)-1 and MASP-3 involved in the lectin complement pathway traced back to the invertebrate, amphioxus

Two lineages of mannose-binding lectin-associated serine protease (MASP) in vertebrates

An ancient lectin-dependent complement system in an ascidian: Novel lectin isolated from the plasma of the solitary ascidian, Halocynthia roretzi

Ancient origin of the complement lectin pathway revealed by molecular cloning of mannan binding protein-associated serine protease from a urochordate, the Japanese ascidian, Halocynthia roretzi

Opsonic complement component C3 in the solitary ascidian, Halocynthia roretzi

Two mannosebinding lectin homologues and an MBL-associated serine protease are expressed in the gut epithelia of the urochordate species Ciona intestinalis

Which came first, the lectin/classical pathway or the alternative pathway of complement?

The roles of programmed death ligand 1 in virus-associated cancers

A novel complement C3 like gene (Lv-C3L) from Litopenaeus vannamei with bacteriolytic and hemolytic activities and its role in antiviral immune response

Genomic view of the evolution of the complement system

Multi-component complement system of Cnidaria: C3, Bf, and MASP genes expressed in the endodermal tissues of a sea anemone, Nematostella vectensis

Characterization of PsTX-60B, a new membrane-attack complex/perforin (MACPF) family toxin, from the venomous sea anemone Phyllodiscus semoni

The RNA-seq analysis suggests a potential multi-component complement system in oyster Crassostrea gigas

Characterization of a C3 and a factor B-like in the carpet-shell clam, Ruditapes decussatus

Complement C3 gene: Expression characterization and innate immune response in razor clam Sinonovacula constricta

Expression of a novel complement C3 gene in the razor clam Sinonovacula constricta and its role in innate immune response and hemolysis

An updated molecular basis for mussel immunity

A single-CRD C-type lectin from oyster Crassostrea gigas mediates immune recognition and pathogen elimination with a potential role in the activation of complement system

Derivatives of the lectin complement pathway in Lophotrochozoa

Proteolytic cascades and their involvement in invertebrate immunity

The oyster genome reveals stress adaptation and complexity of shell formation

Complement. First of two parts

Biology of complement: The overture

The lectin-complement pathway-its role in innate immunity and evolution

Genomic analysis of immunity in a urochordate and the emergence of the vertebrate immune system

Genomic analysis of the immune gene repertoire of amphioxus reveals extraordinary innate complexity and diversity

Origin of the classical complement pathway: Lamprey orthologue of mammalian C1q acts as a lectin

Evolution of the complement system

HMGB1: LPS delivery vehicle for caspase-11-mediated pyroptosis

Ancient origin of the complement system: Emerging invertebrate models

Siglec-7 engagement by GBS beta-protein suppresses pyroptotic cell death of natural killer cells

C3d of complement as a molecular adjuvant: Bridging innate and acquired immunity

Structural basis for activation of the complement system by component C4 cleavage

Hepatic stellate cell autophagy inhibits extracellular vesicle release to attenuate liver fibrosis

Mannose-binding lectin serine proteases and associated proteins of the lectin pathway of complement: Two genes, five proteins and many functions?

Evolution of the lectin-complement pathway and its role in innate immunity

Proteolytic activities of two types of mannose-binding lectin-associated serine protease

The B cell inhibitory Fc receptor triggers apoptosis by a novel c-Abl family kinase-dependent pathway

Complement in disease: A defence system turning offensive

The complement system

Role of the complement system in antibody-dependent enhancement of flavivirus infections

Siglec-9 transduces apoptotic and nonapoptotic death signals into neutrophils depending on the proinflammatory cytokine environment

A novel chimeric PD1-NKG2D-41BB receptor enhances antitumor activity of NK92 cells against human lung cancer H1299 cells by triggering pyroptosis

Complement driven by conformational changes

Structure and function of complement C5 convertase enzymes

Molecular mechanism of SHP2 activation by PD-1 stimulation

Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method

CgCLec-HTM-mediated signaling pathway regulates lipopolysaccharide-induced CgIL-17 and CgTNF production in oyster

Binding of a C-type lectin's coiled-coil domain to the Domeless receptor directly activates the JAK/STAT pathway in the shrimp immune response to bacterial infection

Beta-arrestins negatively regulate the toll pathway in shrimp by preventing dorsal translocation and inhibiting dorsal transcriptional activity

Acknowledgments-We are grateful to all the laboratory members for their technical advice and helpful discussions. 

Abbreviations-The abbreviations used are: A2M, alpha-2macroglobulin family; A2MR, a-macroglobulin receptor; AMP, antimicrobial peptide; BSA, bovine serum albumin; C4bC2a, C3 convertases; CFU, colony-forming unit; Co-IP, Coimmunoprecipitation; CRD, carbohydrate recognition domain; CTL, C-type lectin; DAPI, 4 0 ,6-diamidino-2-phenylindole; dsRNA, double-stranded RNA; FITC, fluorescein isothiocyanate; FREP, fibrinogen-related protein; GBL, glucose-binding lectin; GLU, glucose; LPS, lipopolysaccharide; MAC, membrane attack complex; MAN, mannose; MASPs, MBL-associated serine proteases; MBL, mannose-binding lectin; PAMP, pathogen-associated molecular pattern; PBS, phosphate buffered saline; PGN, peptidoglycan; PRR, pattern recognition receptor; qRT-PCR, quantitative real-time PCR; RNAi, RNA interference; SEM, scanning electron microscope; TBR, alpha-macroglobulin thiol-ester bond-forming region; TEP, thioester-containing protein; V. splendidus, Vibrio splendidus.