Identification and Cloning of a Glucan- and Lipopolysaccharide-binding Protein from Eisenia foetidaEarthworm Involved in the Activation of Prophenoloxidase Cascade

Coelomic fluid of Eisenia foetidaearthworms contains a 42-kDa protein named coelomic cytolytic factor 1 (CCF-1) that was described previously to be involved in cytolytic, opsonizing, and hemolytic properties of the coelomic fluid. Cloning and sequencing of CCF-1 reveal significant homology with the putative catalytic region of β-1,3- and β-1,3–1,4-glucanases. CCF-1 also displays homology with coagulation factor G from Limulus polyphemus and with Gram-negative bacteria-binding protein of Bombyx mori silkworm, two proteins involved in invertebrate defense mechanisms. We show that CCF-1 efficiently binds both β-1,3-glucan and lipopolysaccharide. Moreover, CCF-1 participates in the activation of prophenoloxidase cascade via recognition of yeast and Gram-negative bacteria cell wall components. These results suggest that the 42-kDa CCF-1 protein of E. foetida coelomic fluid likely plays a role in the protection of earthworms against microbes. Coelomic fluid of Eisenia foetidaearthworms contains a 42-kDa protein named coelomic cytolytic factor 1 (CCF-1) that was described previously to be involved in cytolytic, opsonizing, and hemolytic properties of the coelomic fluid. Cloning and sequencing of CCF-1 reveal significant homology with the putative catalytic region of β-1,3- and β-1,3–1,4-glucanases. CCF-1 also displays homology with coagulation factor G from Limulus polyphemus and with Gram-negative bacteria-binding protein of Bombyx mori silkworm, two proteins involved in invertebrate defense mechanisms. We show that CCF-1 efficiently binds both β-1,3-glucan and lipopolysaccharide. Moreover, CCF-1 participates in the activation of prophenoloxidase cascade via recognition of yeast and Gram-negative bacteria cell wall components. These results suggest that the 42-kDa CCF-1 protein of E. foetida coelomic fluid likely plays a role in the protection of earthworms against microbes. prophenoloxidase lipopolysaccharide smooth LPS chemotype smooth phenoloxidase bovine serum albumin 3-(3,4-dihydroxylphenyl)-l-alanine coelomic cytolytic factor-1 recombinant coelomic cytolytic factor-1 monoclonal antibody polymerase chain reaction fluorescein isothiocyanate. The prophenoloxidase (pro-PO)1-activating system represents an important defense mechanism in a large variety of invertebrates (reviewed in Refs. 1Smith V.J. Adv. Comp. Environ. Physiol. 1996; 23: 75-114Crossref Google Scholar and 2Johansson M. Söderhäll K. Prog. Mol. Subcell. Biol. 1996; 15: 46-67Crossref PubMed Scopus (93) Google Scholar). This system is based on the recognition of bacterial antigens such as lipopolysaccharide (LPS) or peptidoglycan and β-1,3-glucan present as major components of the cell wall of yeast and fungi (3Ashida M. Ishizaki Y. Iwahana H. Biochem. Biophys. Res. Commun. 1983; 113: 562-568Crossref PubMed Scopus (148) Google Scholar, 4Silverman D.H. Kreuger J.M. Karnovsky M.L. J. Immunol. 1986; 136: 2195-2201PubMed Google Scholar). Generally, upon the recognition of such saccharides proteases cleave by limited proteolysis inactive pro-PO to its active state, phenoloxidase (PO). The active enzyme catalyzes the o-hydroxylation of monophenols as well as the oxidation of diphenols to quinones that are subsequently polymerized nonenzymatically to melanin. Melanin and its precursors involved in the pro-PO-activating system have cytotoxic and antimicrobial properties and participate in a wide range of other biological activities including phagocytosis/opsonization, encapsulation/nodule formation, degranulation, and wound healing (5Pawelek J.M. Lerner A.B. Nature. 1978; 276: 627-628Crossref Scopus (162) Google Scholar, 6Graham D.G. Tiffany S.M. Vogel F.S. J. Invest. Dermatol. 1978; 70: 113-116Abstract Full Text PDF PubMed Scopus (118) Google Scholar, 7Söderhäll K. Ajaxon R. J. Invertebr. Pathol. 1982; 39: 105-109Crossref Scopus (105) Google Scholar, 8Nappi A.J. Vass E. Pigm. Cell Res. 1993; 6: 117-126Crossref PubMed Scopus (254) Google Scholar, 9Thörnqvist P.-O. Johansson M.W. Söderhäll K. Dev. Comp. Immunol. 1994; 18: 3-12Crossref PubMed Scopus (114) Google Scholar). The pro-PO-activating system has been detected both in protostomian and deuterostomian species. Although pro-PO-activating system is well documented in arthropods, data in other protostomian groups are more scarce. In annelids, melanization reactions and formation of "brown bodies" or nodules have been described in polychaetes and oligochaetes (10Poinar Jr., G.O. Hess R.T. Bulla L.A. Cheng T.C. Comparative Pathobiology. 3. Plenum Publishing Corp., New York1977: 69-84Crossref Google Scholar, 11Dales R.P. J. Invertebr. Pathol. 1983; 42: 288-291Crossref Scopus (11) Google Scholar, 12Valembois P. Lassègues M. Roch P. Dev. Comp. Immunol. 1992; 16: 95-101Crossref PubMed Scopus (52) Google Scholar, 13Porchet-Hennerè E. Vernet G. Cell Tissue Res. 1992; 269: 167-174Crossref PubMed Scopus (37) Google Scholar, 14Porchet-Hennerè E. M′Berri M. J. Invertebr. Pathol. 1987; 50: 58-66Crossref Scopus (12) Google Scholar). However, biochemical detection of PO activity was so far restricted to a few species with rather controversial results. Whereas Smith and Söderhäll (15Smith V.J. Söderhäll K. Dev. Comp. Immunol. 1991; 15: 251-261Crossref PubMed Scopus (217) Google Scholar) failed to detect pro-PO system in the polychaete Aphrodite aculeata and Arenicola marina, Fischer (16Fischer E. Acta Histochem. 1978; 63: 210-223Crossref Scopus (10) Google Scholar), Valembois et al.(17Valembois P. Seymour J. Roch P. J. Invertebr. Pathol. 1991; 57: 177-183Crossref Scopus (24) Google Scholar), and Porchet-Hennerè and Vernet (13Porchet-Hennerè E. Vernet G. Cell Tissue Res. 1992; 269: 167-174Crossref PubMed Scopus (37) Google Scholar) have documented PO activity in Lumbricus terrestris, Eisenia foetida andrei, and Nereis diversicolor, respectively. More recently using l-DOPA as substrate, a 38-kDa protein responsible for PO activity was identified in the coelomic fluid of E. foetida andrei (18Seymour J. Nappi A. Valembois P. Anim. Biol. 1993; 2: 1-6Google Scholar). A report showing that the oxidative activity of the coelomic fluid of earthworms toward l-DOPAin vitro is not affected by trypsin but completely blocked by subtilisin reflects the importance of a correct proteolytic digestion as an initial step for inactive pro-PO activation (17Valembois P. Seymour J. Roch P. J. Invertebr. Pathol. 1991; 57: 177-183Crossref Scopus (24) Google Scholar). Since the factor that recognizes microbial saccharides and triggers the pro-PO system has not yet been described in annelids (1Smith V.J. Adv. Comp. Environ. Physiol. 1996; 23: 75-114Crossref Google Scholar, 2Johansson M. Söderhäll K. Prog. Mol. Subcell. Biol. 1996; 15: 46-67Crossref PubMed Scopus (93) Google Scholar), investigations were initiated to identify such molecules in the coelomic fluid of E. foetida foetida. Surprisingly, we found that a previously described 42-kDa protein named CCF-1 (coelomic cytolytic factor 1; see Ref. 19Bilej M. Brys L. Beschin A. Lucas R. Vercauteren E. Hanusova R. De Baetselier P. Immunol. Lett. 1995; 45: 123-128Crossref PubMed Scopus (75) Google Scholar) binds specifically β-1,3-glucan and LPS from the smooth (S) chemotype. Cloning and sequence analysis show that CCF-1 displays significant amino acid homology with β-1,3- and β-1,3–1,4-glucanases and furthermore shares similarity with clotting factor G from horseshoe crab and a Gram-negative binding protein from silkworm. Finally, we report that CCF-1 participates in the pro-PO cascade in the coelomic fluid of E. foetida. Cell-free coelomic fluid was isolated from adult specimens of E. foetida foetida (Oligochaeta; Annelida) in the absence of serine protease inhibitor as described (19Bilej M. Brys L. Beschin A. Lucas R. Vercauteren E. Hanusova R. De Baetselier P. Immunol. Lett. 1995; 45: 123-128Crossref PubMed Scopus (75) Google Scholar). Gram-negativeSalmonella typhimurium LT2 (S), Salmonella minnesota (Re595), and Gram-positive Bacillus firmus2212 were kind gifts of Drs. J. Hofman and L. Prokesova (Institute of Microbiology, Academy of Sciences of the Czech Republic, Prague, and Faculty of Medicine, Charles University, Prague). YeastSaccharomyces cerevisiae was kindly provided by Dr. M. Novak (Institute of Microbiology, Academy of Sciences of the Czech Republic, Prague). LPS from Escherichia coli 055:B5 (S, delipidized, FITC-labeled), E. coli 0111:B4 (S), E. coli H100 (Ra), E. coli J5 (Rc), E. coli F583 (Rd),S. typhimurium (S), S. minnesota (Re595), and diphosphoryl lipid A (from E. coli F583 Rd) were purchased from Sigma. All LPS fractions were suspended in 50 mmTris-HCl, pH 8, 50 mm NaCl (TN buffer) at 1 mg/ml and sonicated until clarification. Insoluble LPS from S. typhimurium LT2, a gift of Dr. J. Hofman (Institute of Microbiology, Academy of Sciences of the Czech Republic, Prague), was suspended in TN buffer without sonication. Commercial S-LPS preparations were contaminated with less than 1% protein and less than 1% RNA as stated in the product data sheet. Insoluble S-LPS fromS. typhimurium LT2 contains 1.6% protein (CBQCA protein quantitation kit, Molecular Probes) and less than 1% nucleic acid (PicoGreen and RiboGreen quantitation kit, Molecular Probes). Insoluble glucans, curdlan, zymosan, and lichenan (Sigma), were suspended in water (10 mg/ml) and sonicated. Further dilutions were performed in TN buffer. Soluble saccharides, laminarin, cellobiose, glucose, mannitol, and gentiobiose (Sigma), were solubilized directly in TN buffer. Laminarin coupled to FITC or to biotin via a two-carbon residue spacer according to the procedure described previously (20Novotna V. Mikes L. Horak P. Jonakova V. Ticha M. Int. J. Biochromatogr. 1996; 2: 37-47Google Scholar) was kindly provided by Drs. L. Mikes and P. Horak (Faculty of Science, Charles University, Prague). LPS contamination of buffers or samples other than LPS fractions was below the detection limit of E-TOXATE test (Sigma). Coelomic fluid (500 μl) was incubated 1 h at 4 °C with curdlan aggregates or LPS particles from S. typhimurium LT2 (1 mg/100 μl). After washings in PBS (1 ml, 13,000 ×g, 10× for 2 min) bound material was eluted in 50 μl of 0.0625 m Tris-HCl buffer, pH 6.8, containing 2% SDS, 20% glycerol, and 5% β-mercaptoethanol (when mentioned), boiled (6 min), and subjected to SDS-polyacrylamide gel electrophoresis (12%). To investigate the specificity of the interaction with β-1,3-glucan or LPS, coelomic fluid (500 μl) or recombinant CCF-1 (rCCF-1; 10 μg/200 μl) was preincubated with soluble glucans or soluble LPS (300 μg/300 μl, 30 min, 4 °C). The mixture was then examined for its ability to bind curdlan or LPS particles from S. typhimurium as described above. Free sites on nitrocellulose sheet were blocked with 1% BSA. Biotinylated laminarin, anti-CCF-1 monoclonal antibody (mAb) 12C9 (19Bilej M. Brys L. Beschin A. Lucas R. Vercauteren E. Hanusova R. De Baetselier P. Immunol. Lett. 1995; 45: 123-128Crossref PubMed Scopus (75) Google Scholar), or an irrelevant IgG1 were added (10 μg/ml), and after repeated washings, peroxidase-labeled avidin (Amersham Pharmacia Biotech) or swine anti-mouse IgG antibody (SEVAC, Prague, Czech Republic) was used for visualization (4-chloro-1-naphthol as substrate). CCF-1 was purified by immunoaffinity from the coelomic fluid as described earlier (19Bilej M. Brys L. Beschin A. Lucas R. Vercauteren E. Hanusova R. De Baetselier P. Immunol. Lett. 1995; 45: 123-128Crossref PubMed Scopus (75) Google Scholar), separated on SDS-polyacrylamide gel electrophoresis, blotted on polyvinylidene difluoride problot membrane, and stained with Amido Black. One part of the immobilized protein was NH2-terminally sequenced by Edman degradation. The remaining part of the CCF-1 band was digested with trypsin for internal sequence determination according to the method of Fernandez et al. (21Fernandez J. Andrews L. Mischee S.M. Anal. Biochem. 1994; 218: 112-118Crossref PubMed Scopus (143) Google Scholar). The released peptides were separated on a reverse phase column (Vydac C4, 2.1 × 250 mm), eluted with a linear gradient (0% to 70%) of acetonitrile in 0.1% trifluoroacetic acid, and the most prominent peaks were sequenced. Sequencing was performed using an Applied Biosystems 477A sequenator. E. foetida foetidaearthworms were injected with S-LPS from E. coli 055:B5 (10 μg/worm) which was shown to increase CCF-1 levels (22Bilej M. Rossmann P. Sinkora M. Hanusova R. Beschin A. Raes G. De Baetselier P. Immunol. Lett. 1998; 60: 23-29Crossref PubMed Scopus (44) Google Scholar) and put on filter paper soaked with PBS containing penicillin/streptomycin (100 units/100 μg/ml) for 1 day. Three worms were frozen in liquid nitrogen and ground to powder, and total RNA was prepared as described by Chomczynski and Sacchi (23Chomczynski P. Sacchi N. Anal. Biochem. 1987; 162: 156-159Crossref PubMed Scopus (63191) Google Scholar). Poly(A)+ RNA was isolated using the poly(A) tract system (Promega). cDNA was synthesized using a cDNA synthesis module (Amersham Pharmacia Biotech). For the cDNA library construction the lambda zap II vector of Stratagene was used. Before inserting the cDNA in the lambda vector,EcoRI adapters were ligated onto the ends of the cDNA. The adapted cDNA was size-fractionated, phosphorylated, and ligated in the lambda arms. The recombinant lambda DNA was packaged in vitro with Gigapack II Gold Packaging extracts (Stratagene). Finally E. coli (XL1-Blue MRF′ Stratagene) was infected with the phage suspension for amplification and determination of the phage titer. Degenerate PCR primers were deduced from a number of native earthworm CCF-1 peptide sequences in order to identify the CCF-1 cDNA from the E. foetidalibrary. Combination of the following primers (5′TIACIGAITGGGAICA(A/G)TA(T/C)ATIGTITGGCA3′ and 5′AAIGTITGIAAITT(G/A)TCICC(G/A)TA(G/A)TTCCA3′) resulted in a specific PCR fragment. A digoxigenin-labeled PCR product (digoxigenin labeling mix from Boehringer Mannheim) was subsequently used as CCF-1-specific probe in plaque hybridization. Positive plaques were isolated, and the presence of CCF-1 cDNA was confirmed by PCR. In order to obtain pBluescript phagemids in vivo excision was performed according to recommendations of Stratagene. To sequence the entireEcoRI insert, deletions were made with the Erase-a-base system of Promega. The sequencing reactions were performed at Innogenetics (Zwijnaarde, Belgium). To confirm the sequence information, the second strand was sequenced making use of first strand-specific primers. Sequence analysis was carried out on the Intelligenetics program PC-gene. The cDNA sequence encoding for mature CCF-1 (mCCF-1) was amplified by PCR using PWO polymerase (Boehringer Mannheim) and the pBluescript phagemid as template. The primers were designed so that after PCR, the mCCF-1 cDNA contained BamHI/NsiI sites at the 5′ end (GGGGATCCATGCATTCACCGACTGGGATCAATATCAC) and aSalI site at the 3′ end (CCGTCGACTCAGTTGCGCTTGTAGACTCG). Hence, after cutting the PCR product with NsiI and blunting the sticky ends, the first codon of the mCCF-1 was blunt-end available for ligation. TheBamHI-SalI fragment was subcloned in pBluescript (pBSmCCF-1) and checked by sequencing. A NsiI-bluntedSalI fragment containing the mCCF-1 cDNA from pBSmCCF-1 was cloned into pIGRHISA vector (provided by Innogenetics, Zwijnaarde, Belgium). In this vector an amino-terminal His tag and an enterokinase cleavage site precede the mCCF-1 cDNA sequence. After transformation in the E. coli strain MC1061 pAcI the clones were ready for induction. Since the expression of pIGRHISA-CCF-1 is under the control of the PL promoter, cultures were grown at 32 °C and induced at 42 °C at an A 600 of 0.7. One liter of induced bacterial culture was resuspended in PBS and sonicated. After centrifugation the pellet was solubilized in 50 ml of a urea solution (8 m urea in TN buffer), applied on 2.5 ml of Ni-NTA agarose resin (Qiagen), and renatured by a linear decreasing gradient of urea solution while rCCF-1 was bound to the column. Three successive elutions were performed by imidazole (300 mm imidazole in TN buffer). Imidazole was removed by extensive dialysis against TN buffer. Microtiter plates were coated with rCCF-1 (100 μl/well of 10 μg/ml in TN buffer, overnight, 4 °C). Free sites were blocked with 1% BSA in TN buffer (200 μl/well, 3 h, 37 °C). After washings (TN buffer) different doses of FITC-labeled laminarin or FITC-labeled S-LPS were added (in 100 μl of TN buffer/0.1% BSA). After 3 h at 37 °C, plates were washed; 100 μl/well TN buffer were added, and the bound fluorescence was measured using a fluorescence multi-well plate reader CytoFluor II (PerSeptive Biosystems; emission/excitation 485 ± 20/530 ± 30 nm). Mean bound fluorescence ± S.D. of triplicates was then analyzed. Control experiments were performed on BSA-coated plates. For inhibition experiments, rCCF-1-coated plates were incubated with different doses of inhibitors (in 100 μl TN buffer/0.1% BSA, 3 h, 37 °C). After washings, FITC-labeled laminarin or FITC-labeled S-LPS was added (in 100 μl TN buffer/0.1% BSA, 3 h, 37 °C), and CCF-1-bound material was quantified as described above. The dose of inhibitor reducing by 50% the binding of 50 μg/ml laminarin or 10 μg/ml LPS was estimated (ED50). The level of the pro-PO system activation was assessed according to Valembois et al. (17Valembois P. Seymour J. Roch P. J. Invertebr. Pathol. 1991; 57: 177-183Crossref Scopus (24) Google Scholar). Briefly, in a total volume of 100 μl, 10 μl of coelomic fluid (without or with 1 mmPefabloc (serine protease inhibitor, Boehringer Mannheim)), 65 μl 0.1m Tris, pH 8, containing 50 mmCa2+, and 10 μl l-DOPA (3-(3,4-dihydroxylphenyl)-l-alanine; Fluka; final concentration 1.5 mm) were incubated at room temperature for 6 h in the absence or presence of different doses of lyophilized microorganisms, saccharides, or LPS fractions. The oxidation of l-DOPA was measured at 492 nm and expressed as the difference between the values without and with Pefabloc. To confirm the role of glucan- or LPS-binding protein in pro-PO activation, the coelomic fluid was incubated with anti-CCF-1 mAb (12C9) coupled to Affi-Gel (Bio-Rad; 1 h, 4 °C), centrifuged, and the depleted coelomic fluid was used in l-DOPA oxidation test as described above. To reconstitute the pro-PO activating cascade rCCF-1 (0.5–2 μg/ml) was added to CCF-1-depleted coelomic fluid before testing l-DOPA oxidation. A putative β-1,3-glucan-binding protein was detected in the coelomic fluid of E. foetida transferred onto nitrocellulose after SDS-polyacrylamide gel electrophoresis using soluble biotin-labeled laminarin (β-1,3-glucan). Out of all coelomic fluid proteins, laminarin reacts under reducing and nonreducing conditions with only one band with apparent molecular mass of 42 kDa (Fig. 1 A). Controls with avidin conjugate were negative. In enzyme-linked immunosorbent assay, biotinylated laminarin also revealed glucan-binding protein in the coelomic fluid of E. foetida diluted up to 1/800 (not shown). To isolate putative glucan-binding protein(s) coelomic fluid of E. foetida was incubated with curdlan particles (β-1,3-glucan). Glucan-bound material was eluted with electrophoresis SDS sample buffer, subjected to electrophoresis, and silver-stained (Fig. 1 A). Bound material consists of a single component with an apparent molecular mass of 42 kDa. By using LPS from S. typhimurium LT2 (S) which is insoluble at neutral pH the same protein could apparently be isolated (Fig. 1 A). To exclude the possibility that the binding of the 42-kDa protein to insoluble β-1,3-glucan or LPS resulted from aspecific interactions, the coelomic fluid was preincubated with soluble laminarin or soluble S-LPS from S. typhimurium or E. coli O55:B5, before incubation with curdlan or insoluble LPS. This treatment abolishes completely the binding of the 42-kDa protein to the glucan or LPS particles (Fig. 1, B and C). In contrast, neither cellobiose (Glcβ-1,4Glc) nor LPS from S. minnesota(Re) inhibits the binding of the 42-kDa protein to curdlan or LPS particles (Fig. 1, B and C). Thus, the binding of the 42-kDa protein from the coelomic fluid of E. foetida to β-1,3-glucan or S-LPS seems to be specific. Since the previously described cytolytic, hemolytic, and opsonizing protein CCF-1 from the coelomic fluid of E. foetida (19Bilej M. Brys L. Beschin A. Lucas R. Vercauteren E. Hanusova R. De Baetselier P. Immunol. Lett. 1995; 45: 123-128Crossref PubMed Scopus (75) Google Scholar) had the same molecular weight as glucan- and LPS-bound material, this material was probed with the anti-CCF-1 mAb 12C9 (19Bilej M. Brys L. Beschin A. Lucas R. Vercauteren E. Hanusova R. De Baetselier P. Immunol. Lett. 1995; 45: 123-128Crossref PubMed Scopus (75) Google Scholar). As shown on Fig. 1 D, the 12C9 mAb recognizes the 42-kDa band corresponding to the glucan- or LPS-binding protein. Moreover, biotinylated laminarin binds to CCF-1 purified on anti-CCF-1 immunoaffinity column (not shown). These data suggest that the glucan- and LPS-binding protein and CCF-1 are identical proteins. Peptides resulting from the NH2-terminal and internal sequencing of affinity purified native CCF-1 are underlined in Fig. 2. Based on the amino acid information of the NH2-terminal part of CCF-1 and one internal peptide (indicated in boldface in Fig. 2), a sense and an antisense degenerate primer, respectively, were deduced to synthesize a digoxigenin-labeled probe that was used to pick up the CCF-1 cDNA from the E. foetida library. The complete nucleotide sequence of CCF-1 cDNA and the deduced amino acid sequence are shown in Fig. 2. Based on the NH2-terminal sequence of the native purified protein, it was clear that the CCF-1 cDNA encodes for a 384-amino acid protein, synthesized with a 17-amino acid signal peptide. Consequently, the mature CCF-1 is a 367-amino acid protein of 42 kDa with a predicted pI of 4.2. This calculated molecular mass that is similar to that of native CCF-1 suggests the absence of a carbohydrate moiety. This is corroborated by the fact that the CCF-1 cDNA contains no potentialN-glycosylation site. Searching for sequence similarities of CCF-1 with known protein revealed close matches with highly conserved amino acid residues commonly shared by β-1,3 and β-1,3–1,4 bacterial glucanases (24Yamamoto M. Aono R. Horikoshi K. Biosci. Biotechnol. Biochem. 1993; 57: 1518-1525Crossref PubMed Scopus (25) Google Scholar) and β-1,3-glucanase of sea urchinStrongylocentrotus purpuratus (25Bachman E. McClay D. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 6808-6813Crossref PubMed Scopus (78) Google Scholar). This region is also present in the α-subunit of the coagulation factor G of the horseshoe crab Limulus polyphemus (26Seki N. Muta T. Oda T. Iwaki D. Kuma K. Miyata T. Iwanaga S. J. Biol. Chem. 1994; 269: 1370-1374Abstract Full Text PDF PubMed Google Scholar) and the Gram-negative bacteria-binding protein of the silkworm Bombyx mori (27Lee W. Lee J. Kravchenko V. Ulevitch R. Brey P. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7888-7893Crossref PubMed Scopus (191) Google Scholar) (Fig. 3). Alignment of the complete CCF-1 sequence reveals the closest homology with the sea urchin β-1,3-glucanase. By using the pIGRHISA vector high expression of CCF-1, up to 8 mg/liter, was obtained. Western blot using mAb against native CCF-1 confirms that this protein of approximately 43 kDa is CCF-1 (Fig. 4). Moreover rCCF-1 binds to curdlan or S-LPS particles from S. typhimurium (not shown) and shows a clear reaction with biotinylated laminarin in ligand blot (Fig. 4). The binding of laminarin or S-LPS from E. coli O55:B5 to microtiter plate-immobilized rCCF-1 is a saturable process indicating the specificity of these interactions (Fig. 5). A qualitative analysis of inhibition experiments summarized in Table Iindicates that saccharides bearing β-1,3-glucosidic link inhibit efficiently the binding of laminarin or S-LPS to rCCF-1. In contrast glucose and saccharides bearing β-1,4- or β-1,6-glucosidic links are ineffective to block the interaction of laminarin or S-LPS with rCCF-1. Rough LPS mutants, particularly Rc-, Rd-, and Re-LPS, do not inhibit the binding of laminarin or S-LPS to rCCF-1. Moreover, lipid A moiety of LPS from E. coli F583 Rd does not influence the binding of β-1,3-glucan or S-LPS to rCCF-1. Collectively, these results indicate that rCCF-1 binds both Glcβ-1,3Glc link and the O-antigen of LPS.Table IInhibition of laminarin or S-LPS binding to rCCF-1 (ED50)InhibitorLaminarin bindingS-LPS bindingS-LPS (E. coli 055:B5)306.5Boiled S-LPS (E. coli055:B5)357.8S-LPS (S. typhimurium)429.8Boiled S-LPS (S. typhimurium)4510.8Delipidized LPS601.0Ra-LPSNI>80Rc-LPSNINIRd-LPSNINIRc-LPSNINILipid ANINILaminarin (β-1,3-Glc link)451.8Curdlan (β-1,3-Glc link)1004.5Zymosan (β-1,3-Glc link)1157.2Cellobiose (β-1,4-Glc link)NINILichenan (β-1,4-Glc link)NINIGentiobiose (β-1,6-Glc link)NINIGlcNINIMannitol (linear Glc)NINIDose of inhibitor required to inhibit by 50% (ED50 in μg/ml) the binding of laminarin (50 μg/ml) to immobilized rCCF-1.Dose of inhibitor required to inhibit by 50% (ED50in μg/ml) the binding of S-LPS from E. coli 055:B5 (10 μg/ml) to immobilized rCCF-1.Inhibitors were boiled 1 h to denature most protein and nucleic acid contaminants.NI, not inhibiting at 1 mg/ml.Contains α-mannan.Possible contaminants of laminarin. Open table in a new tab Dose of inhibitor required to inhibit by 50% (ED50 in μg/ml) the binding of laminarin (50 μg/ml) to immobilized rCCF-1. Dose of inhibitor required to inhibit by 50% (ED50in μg/ml) the binding of S-LPS from E. coli 055:B5 (10 μg/ml) to immobilized rCCF-1. Inhibitors were boiled 1 h to denature most protein and nucleic acid contaminants. NI, not inhibiting at 1 mg/ml. Contains α-mannan. Possible contaminants of laminarin. The activation of pro-PO system is based on the recognition of microbial glucans or LPS. It was thus interesting to test whether the glucan- and LPS-binding protein from the coelomic fluid of E. foetida trigger this defense mechanism. Since the activation of pro-PO is dependent on proteolytic enzymes (28Aspan A. Sturtevant J.E. Smith V.J. Söderhäll K. Insect Biochem. 1990; 20: 709-718Crossref Scopus (77) Google Scholar, 29Aspan A. Söderhäll K. Insect Biochem. 1991; 21: 363-373Crossref Scopus (134) Google Scholar), the level of the activation was estimated as the difference ofl-DOPA oxidation in the absence and presence of serine protease inhibitor. Coelomic fluid of E. foetida by itself causes only low level of oxidation of the substrate. However, in the presence of a triggering stimulus such as laminarin or LPS, the oxidation of l-DOPA by coelomic fluid proteins increases reaching a maximum between 6 and 8 h of incubation (not shown). As shown in Fig. 6, the yeast S. cerevisiae and the Gram-negative bacteria S. typhimurium LT2 (S) induce a dose-dependent activation of the pro-PO cascade. E. coli 055:B5 (S) and E. coli 0111:B4 (S) also trigger pro-PO activation in the coelomic fluid (not shown). In contrast, the induction of pro-PO activation byS. minnesota (Re) or the Gram-positive B. firmusis close to background. As few as 100 pg/ml laminarin or S-LPS fromS. typhimurium are sufficient to triggerl-DOPA oxidation in the coelomic fluid. Other glucans bearing β-1,3-glucosidic link (curdlan, zymosan) could also trigger pro-PO activation, whereas glucose or saccharides with β-1,4-(cellobiose, lichenan) or β-1,6-glucosidic link (gentiobiose) are inactive (not shown). Finally, we observed that LPS fails to trigger l-DOPA oxidation in E. foetida coelomic fluid following removal of the O-antigen (Fig. 7). Lipid A is also unable to activate the pro-PO cascade.Figure 7Activation of pro-PO in the coelomic fluid by LPS fractions. Coelomic fluid levels of l-DOPA oxidation are expressed as the mean ± S.D. of theA 492 values difference of the sample without and with protease inhibitor. For activation of pro-PO, Gram-negative bacteria or LPS fractions were given at 50 ng/ml. l-DOPA oxidation in the presence of nonactivated coelomic fluid was considered as background level.View Large Image Figure ViewerDownload Hi-res image Download (PPT) To confirm that glucan- and LPS-binding protein CCF-1 is involved in the activation of pro-PO cascade, CCF-1 was removed from the entire coelomic fluid by preincubation with anti-CCF-1 mAb coupled to Affi-Gel (anti-CCF-1 mAb does not induce PO activity, not shown). CCF-1 removal results in a significant decrease of the oxidative activity of the coelomic fluid even in presence of yeast or Gram-negative bacteria (S) cell wall compounds (Fig. 8). The PO activity of the CCF-1-depleted coelomic fluid can be recovered by addition of rCCF-1. However rCCF-1 by itself does not trigger the oxidation of l-DOPA. Altogether, these results indicate that, upon interaction with yeast β-1,3-glucans or with the bacterial external O-specific chain of LPS, CCF-1 may play a role in the pro-PO-activating system of E. foetida earthworms. Extensive studies were carried out over the past 3 decades on the biochemical aspects of the defense system in invertebrates. It is now clear that invertebrates lack specific immunoglobulins, lymphocytes, or other features of the vertebrate adaptive immune system but possess innate defense components (30Roch P. Adv. Comp. Environ. Physiol. 1996; 23: 115-150Crossref Google Scholar). One candidate for such immediate recognition and defense mechanisms is the prophenoloxidase (pro-PO) system that is often put in analogy with the alternate pathway of complement activation (1Smith V.J. Adv. Comp. Environ. Physiol. 1996; 23: 75-114Crossref Google Scholar, 15Smith V.J.