A Structural Perspective on the Interaction between Lipopolysaccharide and Factor C, a Receptor Involved in Recognition of Gram-negative Bacteria*

The recognition of broadly conserved microorganism components known as pathogen-associated molecular patterns is an essential step in initiating the innate immune response. In the horseshoe crab, stimulation of hemocytes with lipopolysaccharide (LPS) causes the activation of its innate immune response, and Factor C, a serine protease zymogen, plays an important role in this event. Here, we report that Factor C associates with LPS on the hemocyte surface and directly recognizes Gram-negative bacteria. Structure-function analyses reveal that the LPS binding site is present in the N-terminal cysteine-rich (Cys-rich) region of the molecule and that it contains a tripeptide sequence consisting of an aromatic residue flanked by two basic residues that is conserved in other mammalian LPS-recognizing proteins. Moreover, we have demonstrated that the Cys-rich region specifically binds to LPS on Gram-negative bacteria and that mutations in the tripeptide motif abrogate its association with both LPS and Gram-negative bacteria, underscoring the importance of the tripeptide in LPS interaction. Although the innate immune response to LPS in the horseshoe crab is distinct from that of mammals, it appears to rely on structural features that are conserved among LPS-recognizing proteins from diverse species.

Innate immunity, which defends the host against infectious microorganisms, is an ancient and ubiquitous system in both vertebrates and invertebrates. Although different organisms employ a variety of environment-specific adaptations to ensure host defense, several generalized features underlie the innate immune response. First, germline-encoded pattern recognition receptors (PRRs) 2 of the host recognize broadly conserved microbial components known as pathogen-associated molecular patterns, including lipopolysaccharide (LPS) of Gram-negative bacteria, ␤-1,3-glucans of fungi, and peptidoglycans of Gram-positive bacteria (1)(2)(3). For example, the Drosophila peptidoglycan recognition protein family members and vertebrate Toll-like receptors have been well characterized as PRRs (4,5). Second, the stimulation of PRRs by pathogen-associated molecular patterns activates intracellular signaling cascades that result in transcriptional activation (6). Finally, phagocytosis and the production of antimicrobial peptides by the host are essential for the clearance and killing of infectious microbes (7,8).
Hemocytes play a fundamental role in the invertebrate innate immune system both as the initial determinants of non-selfrecognition and subsequently as the mediators of phagocytosis, encapsulation, and melanization (9). Granular hemocytes predominate in the hemolymph of the horseshoe crab (accounting for Ͼ99% of total hemocytes) and constitute a key component of the host defense against infectious microorganisms, as they contain an array of defense molecules that are deployed to protect against infectious microorganisms (10). Horseshoe crab granular hemocytes are especially sensitive to LPS, a major cell wall component of Gram-negative bacteria. Stimulation of the granular hemocytes with LPS causes the activation of hemolymph coagulation within a few seconds, leading to the immobilization and engulfment invading microorganisms. Hemocytes concurrently release defense molecules by LPS-induced exocytosis, thereby killing infectious microbes (10,11).
Because of their sensitivity to LPS, it has been thought that granular hemocytes express LPS receptors on their surfaces and that binding of LPS to these receptors could initiate a series of signaling processes important for the innate immunity, as has been observed in vertebrate. A close homolog of vertebrate Toll-like receptors, designated tToll, was identified in the horseshoe crab Tachypleus tridentatus. However, the apparent absence of a pathogen-associated molecular pattern binding domain in tToll as well as its relatively nonspecific tissue expression pattern (12) cast doubt upon the idea that tToll itself functions as an LPS receptor in horseshoe crab hemocytes. Another candidate for the putative LPS receptor is Factor C, a serine protease zymogen predominantly stored in hemocyte intracellular granules that initiates the hemolymph coagulation cascade following proteolytic activation. Features of Factor C that are consistent with its potential role as an LPS receptor include its high sensitivity toward LPS (13,14) and its restricted hemocyte-specific expression pattern (see supplemental Fig.  S1). We recently proposed that a portion of Factor C might localize to the hemocyte surface and undergo autocatalytic activation upon association with LPS, thereby initiating signal transduction via protease-activated receptors coupled to heterotrimeric GTP-binding proteins (15).
In this study, we have established the essential involvement of the N-terminal Cys-rich region of Factor C in LPS recognition and determined the importance of a tripeptide motif contained within this region of the molecule. Based on these findings, we propose a mode of LPS recognition by Factor C that employs this motif.
Immunofluorescence Microscopy-One milliliter of T. tridentatus hemolymph was collected into 50 ml of pyrogen-free 10 mM HEPES buffer (pH 7.0) containing 0.5 M NaCl, and the diluted hemolymph (200 l) was plated on coverslips. After a 20-min incubation to allow attachment, hemocytes were fixed with 3.7% formaldehyde for 10 min and washed twice with phosphate-buffered saline (PBS). For LPS treatment, the attached cells were incubated with 50 g/ml FITC-conjugated LPS for 30 min and then fixed with 3.7% formaldehyde. The cells were blocked with 5% bovine calf serum and incubated for 1 h with anti-Factor C monoclonal antibody 2C12 (16) diluted in blocking buffer. For detection, Cy3-conjugated anti-mouse secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) was used. Cells were imaged with an Olympus BX-FLA fluorescence microscope (Tokyo, Japan).
Co-immunoprecipitation-3Myc-tagged Factor C variants and anti-LPS factor (ALF) proteins were transiently expressed in HEK293 cells and secreted into the culture medium. HEK293 cell lines were maintained in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 1% glutamine, 1% penicillin-streptomycin, and 10% bovine calf serum, respectively, at 5% CO 2 and 37°C. Four micrograms of the 3Myctagged Factor C variants and ALF expression plasmids were used to transfect HEK293 cells at 50% confluence in a six-well plate using the calcium phosphate method. Conditioned media were collected 2 days after transfection, and the protein expres-

RESULTS AND DISCUSSION
Factor C Is a Membrane-bound LPS Receptor on the Hemocyte Surface-At least two proteins, ALF and Factor C, have been determined to associate with LPS in the horseshoe crab. We recently reported that Factor C, an LPS-sensitive serine protease zymogen, exists on the surface of hemocytes and initiates signaling through heterotrimeric GTP-binding proteins (15). We reasoned that hemocyte surface-bound Factor C could serve as an LPS receptor, and if so, it would be a likely candidate for the LPS-responsive PRR that induces exocytosis of granular components.
To address this issue, we morphologically investigated whether Factor C and LPS co-localize on the hemocyte surface (Fig. 1). Immunostaining of hemocytes with an antibody to Factor C (red) demonstrated that endogenous Factor C was pres-  immunoprecipitation with Factor C-(1-116)/3Myc and its mutants. The R36E/ R38E mutant completely lacked the ability to bind bacteria. C, flow cytometric analysis of Factor C-(1-116)/3Myc (blue line) and its R36E/R38E mutant (green line) binding to E. coli K12. For detection of surface-bound protein, FITC-labeled anti-myc monoclonal antibody 9E10 was used. As a negative control, untreated bacteria were detected with FITC antibody (black line). w.t., wild type. FEBRUARY 9, 2007 • VOLUME 282 • NUMBER 6 ent in a punctate distribution on the hemocyte surface (Fig. 1B,  arrows). We next tested whether the surface-localized Factor C could bind to LPS. Hemocytes were incubated with FITC-labeled LPS (E. coli O111:B4). FITC-LPS accumulated on the cell surface of hemocyte (Fig. 1G) and co-localized with Factor C (Fig. 1, H and I, arrowheads). Interestingly, FITC-LPS accumulation was not observed in fixed hemocytes (supplemental Fig.  S2), which would be expected to undergo chemical modification and inactivation of surface proteins. These results suggest that Factor C is a membrane-bound LPS receptor in horseshoe crab hemocytes. Factor C can bind to cholesterol and acidic phospholipids with high affinity (14,15) likely through a different site (19), suggesting that it might associate with plasma membrane lipid rafts enriched with cholesterol and glycosphingolipids (20) or microdomains and subsequently initiate signaling processes through protease-activated receptors following LPS-induced autocatalytic activation.

Lipopolysaccharide Recognition Mode by Factor C
Identification of the Factor C Domain Essential for LPS Recognition-We therefore attempted to identify the region of Factor C responsible for LPS recognition. Factor C is a multidomain glycoprotein with an estimated molecular mass of ϳ120 kDa. In addition to the typical serine protease domain at its C terminus, it also contains a Cys-rich region, an epidermal growth factor (EGF)-like domain, five complement control protein (CCP) modules, an LCCL module (derived from a conserved domain of Limulus Factor C, Coch-5b2, and Lg11 (21)), and lectin-like domains (supplemental Fig. S3) (22).
We also investigated the specificity of the Factor C-(1-296)/3Myc-LPS complex (Fig. 2C). For both ALF/3Myc and Factor C-(1-296)/ 3Myc, complex formation with biotinylated LPS was diminished with increasing amounts of unlabeled Re595 LPS, with complex formation completely disrupted by the addition of a 10-fold excess of unlabeled competitor (Fig. 2C, lanes 3  and 6). Taken together, these results indicate that the N-terminal region of Factor C is sufficient for LPS recognition.
The Cys-rich/EGF-like Domain Encompasses the LPS-recognizing Domain-We performed structure-function analyses to further define the LPS-recognizing sequence within the N-terminal region of Factor C. Two fragments of the previously identified LPS-recognizing domain, Factor C-(1-116)/3Myc and Factor C-(117-296)/3Myc, were generated and tested for their ability to bind to LPS (Fig. 3A). Of these two, only the most N-terminal fragment (Factor C-(1-116)/3Myc), which contains the tandem Cys-rich region and EGF-like domain, was capable of forming a complex with biotinylated S. minnesota Re595 LPS (Fig. 3A, lane 4). The interaction between Factor C-(1-116)/3Myc and LPS was determined to be specific, because the complex was disrupted by the addition of excess competing LPS (unlabeled Re595 LPS), as was previously observed for Factor C-(1-296)/3Myc (Fig. 3B). The specific interaction between Factor C-(1-116)/3Myc and LPS was further demonstrated by the inability of either cholesterol or acidic phospholipids to disrupt the complex (Fig. 3C, lanes 3-5). Our results contrast with those presented in a previous study of the Factor C ortholog from the Southeast Asian horseshoe crab Carcinoscorpius rotundicauda, which defined a region within the tandem CCP domains as important for LPS binding (24). However, these authors did not specifically investigate the role of the N-terminal region identified in the present study, nor did they include negative controls similar to those described in the following sections.
Mutations within the Cys-rich Region Abolish LPS Binding-From the present study, it is clear that the potential LPS recognition domain in Factor C is located in its N-terminal region, which consists of a Cys-rich region and an EGF-like domain. An interesting feature of the Cys-rich region is the presence of a unique conserved tripeptide motif, namely an aromatic residue immediately flanked by two basic residues, which appears in five other known LPS-recognizing proteins (Fig. 4A, bottom). This sequence motif is present in two copies within the LPSrecognizing domain of Factor C from Arg 36 to Arg 38 and from Lys 55 to Lys 57 (Fig. 4A, top).
To test whether these tripeptide motifs are essential for the LPS binding ability, we introduced paired glutamate substitutions into Factor C-(1-116)/3Myc at locations corresponding to basic residue pairs (Arg 36 /Arg 38 and Lys 55 /Lys 57 ). The results of co-immunoprecipitation experiments indicate that the R36E/ R38E mutant is incapable of binding to LPS (Fig. 4, B and C, lanes  2), whereas the K55E/K57E mutant retains the ability to bind LPS (Fig. 4, B and C, lanes 3). We also substituted an alanine residue for Trp 37 to assess the importance of an aromatic residue at this position and found that the resulting W37A mutant also lacked the ability to bind LPS (supplemental Fig. S4). These results establish the essential nature of the Cys-rich region for LPS recognition by Factor C and further highlight the importance of the Arg 36 -Trp 37 -Arg 38 motif.
The Cys-rich/EGF-like Domain of Factor C Binds to Gramnegative Bacteria-The preservation of LPS binding activity in Factor C-(1-116)/3Myc allowed us to explore the direct LPSdependent recognition of live bacteria. Using bacterial immunoprecipitation, we investigated whether Factor C-(1-116)/ 3Myc could selectively recognize Gram-negative bacteria, which are rich in LPS on their surface. Four types of live bacteria (two Gram-positive bacteria, S. aureus and E. hirae, and two Gram-negative bacteria, E. coli B and K12) were incubated with Factor C-(1-116)/3Myc, and the association of the protein with the bacteria was assessed by Western blotting with the anti-Myc monoclonal antibody 9E10. Factor C-(1-116)/3Myc was specific for Gram-negative bacteria, showing significant binding to E. coli B and K12 but no binding to Gram-positive bacteria (Fig. 5A). In addition, the binding of Factor C-(1-116)/3Myc to Gram-negative bacteria (E. coli) was confirmed by flow cytometry (Fig. 5C). Finally, the R36E/R38E mutant was incapable of binding to E. coli, as judged by both bacterial immunoprecipitation (Fig. 5B) and flow cytometry (Fig. 5C), indicating that the association between Factor C-(1-116)/3Myc and bacteria is LPS-dependent.
Implications for Innate Immunity-Factor C associates with LPS on the hemocyte surface, and the N-terminal Cys-rich region can selectively bind Gram-negative bacteria in an LPSdependent manner. Taken together, these findings suggest that Factor C may serve as a membrane-bound LPS receptor capable of tethering invading bacteria to the hemocyte surface (Fig. 6), ultimately leading to hemolymph coagulation by Factor C secreted from granules. We propose that the first conserved tripeptide motif (Arg 36 -Trp 37 -Arg 38 ) in the Cys-rich region of Factor C plays an important role in the recognition of LPS, an assumption that is consistent with a modeling study of ALF that predicts direct docking of the corresponding tripeptide motif with LPS (25). In this model, two lysine residues that correspond to Arg 36 and Arg 38 of Factor C interact with glucosamine (GlcN) I-1-phosphate in lipid A, and a tryptophan residue, which corresponds to Trp 37 of Factor C, interacts with a hydrophobic portion of lipid A (Fig. 6, right panel). Our results are consistent with such a model and underscore the importance of these elements within the Cys-rich region of Factor C for mediating its association with LPS, an essential structural characteristic that appears to be conserved among other LPS-recognizing proteins. Finally, these findings imply that Factor C acts as a PRR for Gram-negative bacteria in the horseshoe crab and lay the groundwork for future studies to elucidate the mechanism whereby hemocyte-bound Factor C potentiates LPS-induced exocytosis degranulation.