Regions of the Mouse CD14 Molecule Required for Toll-like Receptor 2- and 4-mediated Activation of NF-κB*

Regions of mouse CD14 required for Toll-like receptor 2 (TLR2)- and TLR4-mediated activation of NF-κB were studied in transiently transfected 293 cells. Wild-type CD14 enhanced lipopolysaccharide (LPS)-induced NF-κB-dependent reporter activity in cells expressing TLR4/MD-2, and deletion of amino acid regions 35–44, 144–153, 235–243, and 270–275 impaired the TLR4-mediated activation. Unlike human CD14, mouse CD14 truncated at amino acid 151 lost the activity. Deletion of amino acids 35–44 or 235–243 also abrogated TLR2-mediated activation of NF-κB, whereas mutants lacking 144–153 and 270–275 retained the activity. Deletion and alanine substitution experiments revealed that amino acids 151–153 and 273–275 were required for the TLR4-mediated activation. Both deletion mutants lacking amino acids 35–44 and 235–243 and alanine substitution mutants in regions 151–153 and 273–275 were expressed on the cell surface and retained the ability to associate with TLR4. A cross-linking study with photoreactive LPS showed that the labeling intensities to CD14 mutants/TLR4/MD-2 were paralleled by the ability of CD14 mutants to increase TLR4-mediated activation. These results indicate that different regions of mouse CD14 are required for TLR4- and TLR2-mediated activation of NF-κB and suggest that amino acids 35–44, 151–153, 235–243, and 273–275 of mouse CD14 play an important role in LPS binding and its transfer to TLR4/MD-2.

Bacterial lipopolysaccharide (LPS) 1 is a constituent of the outer membrane of the cell wall of Gram-negative bacteria and plays a major role in septic shock in humans (1,2). Exposure of macrophages to nanogram quantities of LPS results in rapid activation of a number of transcription factors, including NF-B, which leads to the synthesis of inflammatory cytokines (3). The cell surface molecules that bind to LPS have been extensively studied, and CD14 has been found to be the major receptor (4 -6). Recently, Toll-like receptor 4 (TLR4) has been identified as another molecule that transmits LPS signaling into intracellular components (7,8).
TLR4 is a mammalian homologue of the Drosophila Toll protein, and it was initially recognized as a molecule that increases constitutive NF-B activity but not LPS-inducible NF-B activity (9). However, the finding of a novel accessory molecule, MD-2 (7), which confers LPS responsiveness on TLR4, and analyses of TLR4-deficient (10 -14) mice have provided strong evidence for involvement of TLR4 in LPS signaling. Although TLR2 was initially recognized as a signaling molecule for LPS (15,16), analyses of TLR2-deficient mice (14,17) and modified phenol extraction studies (18,19) showed that an endotoxic substance(s) other than LPS is responsible for TLR2-mediated signaling. CD14 is a glycosylphosphatidylinositol-anchored glycoprotein expressed on leukocytes and is the major receptor responsible for the effects of LPS on macrophages, monocytes, and neutrophils (for a review, see Ref. 2). Since CD14 lacks transmembrane and intracellular domains, it is postulated that CD14 presents LPS to its signaling molecule, TLR4 (20). CD14 has also been reported to be a membrane receptor for various bacterial products, such as peptidoglycan and lipoarabinomannan (for a review, see Ref. 21), and TLR2 has been reported to transmit their signalings (14,22,(22)(23)(24)(25). Thus, CD14 plays a critical role in discriminating bacterial products and dividing their signals into TLR4 or TLR2 depending on the nature of the product. We previously reported that lipid A preparations from various Salmonella strains and synthetic Salmonella-type lipid A (compound 516) possess very little stimulatory activity in human macrophages despite being potently active in murine macrophages (26). On the other hand, lipid A preparations from Escherichia coli and synthetic E. coli-type lipid A (compound 506) were equally active in both human and murine macrophages (26). Thus, Salmonella lipid A shows animal species-specific actions. The mechanism of the animal speciesspecific actions of Salmonella lipid A is still unknown, but it may be attributable to species differences in the CD14 molecules in recognizing these lipid A molecules. Several studies using monoclonal antibodies to CD14 or CD14 deletion and point mutants to identify the structural requirements for human CD14 as a membrane-bound (27)(28)(29)(30)(31) or soluble (4,(32)(33)(34)(35)(36) receptor for LPS have been performed, and results have shown that the N-terminal region of human CD14 is required for LPS recognition and signal transduction. However, since no structural studies of mouse CD14, especially the regions of the mouse CD14 molecule required for TLR4/MD-2-or TLR2-mediated activation of NF-B, had ever been performed, in the present study we investigated these regions by transiently transfecting 293 cells as a first step in clarifying the speciesspecific actions of Salmonella lipid A.

MATERIALS AND METHODS
Cell Culture and Reagents-The human embryonic 293 cell line (obtained from the Human Science Research Resources Bank, Tokyo, Japan) was grown in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% (v/v) heat-inactivated fetal calf serum (Invitro-* This work was supported in part by grants from the Japan Health Sciences Foundation and the Ministry of the Environment. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ To whom correspondence should be addressed: Division of Microbiology, National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya, Tokyo 158-8501, Japan. Tel.: 81-3-3700-1141 (ext. 272); Fax: 81-3-3707-6950; E-mail: tanamoto@nihs.go.jp. 1 The abbreviations used are: LPS, lipopolysaccharide; TLR, Toll-like receptor; PEX, phenol extract; SBED, sulfosuccinimidyl-2-[6-(biotinamido)-2-(p-azidobenzamido) hexanoamido]ethyl-1,3Ј-dithiopropionyl; PBS, phosphate-buffered saline. gen), penicillin (100 units/ml), and streptomycin (100 g/ml). LPS was prepared from E. coli 03K2a2b:H2 or R3 F653. R3 F653 LPS and E. coli O111:B4 LPS (Sigma) preparations were repurified according to Hirschfeld et al. (18). The phenol extract (PEX) used to stimulate the TLR2-dependent pathway was obtained as follows. E. coli O111:B4 LPS (Sigma) that had not been subjected to repurification was dissolved in a 0.2% triethylamine aqueous solution containing 0.5% deoxycholate and mixed with an equal volume of phenol. The phenol phase was then extracted twice with a 0.2% triethylamine aqueous solution containing 0.5% deoxycholate. Next, the phenol phase was then extensively dialyzed against purified water, and after vacuum drying, the dried residue was dissolved in purified water and used as the PEX. Antibody against the EIAV tag epitope (antiserum no. 1060) was a kind gift of Dr. Nancy Rice (NCI-Frederick). Anti-mouse CD14 antibody (M-20) was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Biotinylated LPS was prepared by dissolving 1 mg of repurified E. coli R3 F653 LPS and 3 mg of PFP-biotin (Pierce) in a mixture of 0.3 ml of Me 2 SO and 0.05 ml of purified water, and after incubating the mixture for 16 h at 4°C, it was dialyzed against methanol, dried, and dissolved in purified water to obtain the final product. Sulfosuccinimidyl-2-[6-(biotinamido)-2-(p-azidobenzamido) hexanoamido]ethyl-1,3Ј-dithiopropionyl LPS (SBED-LPS) was prepared as follows. Repurified E. coli R3 F653 LPS (2.7 mg) was dissolved in 0.8 ml of purified water plus 0.1 ml of PBS and mixed with 5 mg of sulfo-SBED (Pierce) dissolved in 0.1 ml of Me 2 SO. The mixture was incubated overnight at 4°C and then dialyzed against methanol in a dark place. The dialysate was dried and dissolved in purified water to obtain the final product.
Expression Plasmids-Plasmids containing human CD14 and mouse CD14 cDNAs were provided by Dr. Shunsuke Yamamoto (Medical College of Oita, Oita, Japan). The coding regions of mouse TLR2 and mouse TLR4 were amplified by reverse transcriptase-PCR from total RNA prepared from murine fibroblast L929 cells. The coding region of mouse MD-2 was amplified from a mouse embryo cDNA library (CLONTECH, Palo Alto, CA). The coding regions of mouse CD14, mouse TLR2, mouse TLR4, and mouse MD-2, minus their respective signal peptide sequences, were subcloned downstream of a mammalian expression vector in which the preprotrypsin signal peptide sequence precedes the NH 2 -terminal EIAV tag epitope (amino acid sequence ADRRIPG-TAEE). A similar LPS response was observed in 293 cells transfected either with tagged or untagged versions of plasmids described above. A plasmid for hexahistidine-tagged TLR4 was constructed by inserting a coding sequence of hexahistidine into the EIAV-TLR4 plasmid described above at the position just downstream of the EIAV tag epitope sequence. A luciferase reporter plasmid pELAM-L was constructed by inserting the PCR fragment (Ϫ730 to ϩ52) of the E-selectin promoter (ELAM-1) (22) into the SacI-HindIII site of pGL3-Basic vector (Promega, Madison, WI). Plasmids for all CD14 mutants were constructed by PCR-mediated mutagenesis. Mouse (mCD14-DAF) or human (hCD14-DAF) CD14 truncated at amino acid 151 attached to the GPI anchor sequence of decay-accelerating factor was constructed according to Lee et al. (27).
NF-B Reporter Assay-The NF-B-dependent luciferase reporter assay was performed as described previously (37). Briefly, human embryonic kidney 293 cells (3-5 ϫ 10 5 /well) were plated in six-well dishes and on the following day transfected by the calcium phosphate precipitation method with the indicated amounts of the expression plasmids, 0.2 g of pELAM-L, and 0.05 g of pRL-TK (Promega, Madison, WI) for normalization. After 24 h, cells were stimulated for 6 h, and the reporter gene activity was measured according to the manufacturer's recommendations (Promega).
Western Blotting-Western blotting was carried out in the following manner as previously described (38). Cell extracts were prepared by incubating cells on ice for 10 min with a lysis buffer (10 mM HEPES-KOH, 5 mM EDTA, 0.5% Nonidet P-40, 1 mM dithiothreitol, and 10 mM KCl, pH 7.9) containing a protease inhibitor mix (Roche Molecular Biochemicals GmbH). Following centrifugation at 1000 ϫ g for 5 min, the supernatants obtained were used as cell extracts, and the cell extracts were subjected to SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore Corp., Bedford, MA). After blocking with 5% nonfat dry milk, the membranes were probed with the antibody indicated and then with a peroxidase-labeled second antibody. After washing the membranes, the signals were visualized with the enhanced chemiluminescence system (Amersham Biosciences).
LPS Binding-LPS binding was studied according to the method described by da Silva et al. (39) with modification using SBED-LPS. SBED-LPS is a UV-activable cross-linking LPS carrying biotin covalently attached to a photoactivable aryl azide moiety in its molecule. After plating 293 cells in 10-cm dishes precoated with pig collagen type I (Asahi Techno Glass, Tokyo, Japan), they were transfected with the plasmids indicated (18 g for TLR4 and MD-2, 3 g for CD14) by the calcium phosphate precipitation method, and 24 h later they were incubated at 15°C for 30 min with 200 ng/ml of SBED-LPS in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. After UV irradiation at 25°C for 4 min, the cells were washed with PBS, and cell extracts were prepared with PBS containing 1% Nonidet P-40, 2 mM EDTA and a protease inhibitor mix (Roche Molecular Biochemicals). To the cell extracts, sodium deoxycholate and SDS were added to final concentrations of 1 and 0.25%, respectively. After the addition of immobilized streptavidin-agarose, the mixture was incubated at 4°C for 1 h. The agarose was washed three times with radioimmune precipitation buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 2 mM EDTA, 1% sodium deoxycholate, 1% Triton X-100, 0.25% SDS) and boiled in an SDSsample buffer. The supernatant obtained was subjected to SDS-PAGE followed by Western blot analysis.
Detection of CD14 Proteins Expressed on the Cell Surface-Detection of cell surface CD14 was performed as described previously (37) with a slight modification. Briefly, 293 cells were plated in 6-cm dishes and transfected with the plasmids indicated (5 g each) by the calcium phosphate precipitation method. After 24 h, the cells were washed twice with PBS and exposed at 15°C for 30 min to 0.5 mg/ml of a membraneimpermeable biotinylation reagent (sulfo-N-hydroxysuccinimide-LC-LC-biotin; Pierce) dissolved in PBS. After stopping the biotinylation by adding 3 ml of the culture medium, cell extracts were prepared with 200 l of the lysis buffer, as described above. The cell extracts were diluted to 500 l with PBS containing 0.1% Nonidet P-40 and incubated with immobilized streptavidin-agarose at 4°C for 1 h. The agarose was washed three times with PBS containing 0.1% Nonidet P-40 and boiled in an SDS-sample buffer. The supernatant obtained was subjected to SDS-PAGE followed by Western blot analysis.
Immunoprecipitation and Membrane Surface Protein-Protein Association-Immunoprecipitation from total cell extract was performed as follows. After plating 293 cells in 6-cm dishes, they were transfected with the plasmids indicated (4 g each) by the calcium phosphate precipitation method, and 24 h later, cell extracts were prepared by incubating the cells on ice for 10 min with 0.2 ml of a buffer (10 mM HEPES-KOH, 0.5% Nonidet P-40 and 10 mM KCl, pH 7.9) containing a protease inhibitor mix (Roche Molecular Biochemicals). The cell extracts were diluted to 500 l with PBS containing 0.1% Nonidet P-40, and after adding an anti-mouse CD14 antibody (M-20) and Protein LA-Sepharose (Sigma), the diluted cell extracts were incubated for 1 h at 4°C with rocking. The Sepharose was washed three times with PBS containing 0.1% Nonidet P-40 and boiled in an SDS-sample buffer. The supernatant obtained was subjected to SDS-PAGE followed by Western blot analysis. Membrane surface protein-protein association was studied as follows. After plating 293 cells in 75-cm 2 culture flasks, they were transfected with the plasmids indicated (4 g for CD14, 24 g for TLR4 and MD-2) by the calcium phosphate precipitation method, and 24 h later, the cells were washed once with PBS and collected into 1.5-ml microcentrifuge tubes. The cells were exposed at 15°C for 30 min to 1 mg/ml membrane-impermeable bifunctional cross-linking reagent, sulfosuccinimidyl-2-[p-azidosalicylamido]ethyl-1,3Ј-dithiopropionate (Pierce) in PBS with rocking in a dark place. After UV irradiation at 25°C for 4 min, the cells were washed with PBS, and cell extracts were prepared with PBS containing 0.1% Nonidet P-40 and a protease inhibitor mix (Roche Molecular Biochemicals). To the extracts, an equal volume of 2ϫ binding buffer (40 mM Tris, pH 7.9, 1 M NaCl, 10 mM imidazole, 12 M urea) was added, and hexahistidine-tagged TLR4 was collected by precipitation with nickel-agarose (Novagen, Madison, WI). The agarose was washed twice with 1ϫ binding buffer followed twice by a wash buffer (20 mM Tris, pH 7.9, 500 mM NaCl, 60 mM imidazole, 6 M urea). Then the agarose was suspended in an elution buffer (20 mM Tris, pH 7.9, 500 mM NaCl, 1 M imidazole) and incubated for 30 min. The supernatant was subjected to trichloroacetic acid precipitation, and the precipitated proteins were dissolved in an SDS-sample buffer containing 3% 2-mercaptoethanol. Finally, cross-linked CD14 protein was detected by Western blotting. Immunoprecipitation from membrane surface proteins were performed as follows. After plating 293 cells in 150-cm 2 culture flasks, they were transfected with the plasmids indicated (54 g for TLR4 and MD-2, 9 g for CD14) by the calcium phosphate precipitation method, and 24 h later, the cells were washed twice with PBS and collected into 1.5-ml microcentrifuge tubes. The cells were exposed at 15°C for 60 min to 2 mg/ml membraneimpermeable biotinylation reagent (PEO-iodoacetyl biotin; Pierce) dissolved in PBS. After washing the cells with culture medium followed by PBS, cell extracts were prepared with PBS containing 1% Nonidet P-40, 2 mM EDTA and a protease inhibitor mix (Roche Molecular Biochemi-cals). After adding immobilized streptavidin-agarose, the cell extracts were incubated for 1 h at 4°C with rocking. The agarose was washed three times with PBS containing 1% Nonidet P-40, 2 mM EDTA, and biotinylated proteins were eluted from the agarose by incubating with 5 mg/ml water-soluble biotin derivative (sulfo-N-hydroxysuccinimidebiotin; Pierce) dissolved in a buffer (50 mM Tris, pH 8, 150 mM NaCl, 5 mM EDTA, 0.5% Nonidet P-40). After the addition of an anti-mouse CD14 antibody (M-20) and Protein G-Sepharose (Amersham Biosciences), the supernatants were incubated for 1 h at 4°C with rocking. The Sepharose was washed three times with PBS containing 1% Nonidet P-40, 2 mM EDTA, and immunoprecipitated proteins were eluted from the antibody by incubating with excess amounts of the antigen peptide dissolved in PBS containing 0.1% Nonidet P-40 for 18 h at 4°C with rocking. Finally, the obtained supernatants were boiled in an SDS-sample buffer and subjected to SDS-PAGE followed by Western blot analysis.

RESULTS
Effects of CD14 Deletion Mutants on TLR4-and TLR2-mediated Activation of NF-B-To explore the structure-activity relationship of mouse CD14, we first examined the effect of wild-type CD14 on TLR4-and TLR2-mediated activation of NF-B. Human embryonic 293 cells were transfected with an NF-B-dependent E-selectin (ELAM-1) promoter luciferase reporter gene together with expression plasmids for mouse TLR4 and MD-2 or an expression plasmid for mouse TLR2. The transfected 293 cells were stimulated either with 0.1 ng/ml LPS and 1 g/ml PEX to activate the TLR4-dependent pathway and TLR2-dependent pathway, respectively. Stimulation with LPS did not increase reporter activity in cells expressing TLR4 and MD-2, whereas LPS markedly increased reporter activity in cells co-expressing mouse CD14 (Fig. 1A). PEX did not increase reporter activity in cells expressing TLR2 either, whereas it markedly increased reporter activity in cells co-expressing mouse CD14 (Fig. 1B).
Next, we examined the effects of CD14 deletion mutants on TLR4-and TLR2-mediated activation of NF-B. When mutants d35-44, d144 -153, d235-243, and d270 -270, in which the corresponding amino acid regions were deleted, were coexpressed in 293 cells together with TLR4 and MD-2, little or no increase in reporter activity was observed in response to LPS, but LPS-induced activation was partially retained with d244 -247 and fully retained with d312-320 (Fig. 1A). Both d35-d44 and d235-243 lost the ability to increase PEX-induced reporter activity in 293 cells expressing TLR2. Despite the fact that the d144 -153 and d270 -275 mutants had lost the ability for the TLR4-mediated activation, they were fully able to increase TLR2-mediated activation (Fig. 1B). Both the d244 -247 and d312-320 mutants retained the ability for the TLR2mediated activation (Fig. 1B).
Effect of CD14 Deletion and Substitution Mutants on TLR4and TLR2-mediated Activation of NF-B-The 144 -153 and 270 -275 amino acid regions of CD14 were further analyzed to identify the structural requirements, because they were selectively required for TLR4-mediated activation. Amino acids 144 -153 were analyzed first ( Fig. 2A). Two deletion mutants, d144 -147 and d151-153, were created, and the LPS-induced reporter activity of 293 cells was examined after expressing these mutants in them together with TLR4 and MD-2. LPS stimulation of cells co-expressing d144 -147 showed an increase in reporter activity that was comparable with the activity of cells co-expressing wild-type CD14, whereas no increase in reporter activity was observed in cells coexpressing d151-153. LPS-induced activation was also severely impaired in cells co-expressing a CD14 mutant (151-153A) in which amino acids 151-153 were replaced by alanine, whereas replacement of one of these amino acids with alanine (P151A, G152A, and L153A) or arginine (G152R) had no effect on LPS-induced activation. These re-sults indicate that amino acids 151-153 are indispensable to the ability of CD14 to increase TLR4-mediated activation.
Next, we examined the structural requirement of amino acids 270 -275 (Fig. 2B). Co-expression of 270 -272A instead of wild-type CD14 in 293 cells expressing TLR4 and MD-2 had no effect on LPS-induced activation of NF-B, but LPS-induced activation was severely impaired in cells co-expressing 273-275A. Replacement of each amino acid from 271 to 275 with alanine (P271A, K272A, G273A, L274A, and P275A) or arginine (G273R), however, had no effect on LPS-induced activation. These results indicate that amino acids 273-275 are indispensable to the ability of CD14 to increase TLR4-mediated activation.
To know the property of CD14 mutants further, we next examined the concentration-response relationship of LPS in 293 cells expressing CD14 mutants, TLR4 and MD-2 (Fig. 3). When wild-type CD14 was expressed with TLR4 and MD-2, the activation of NF-B was observed at 0.1 ng/ml and reached to the maximum at 10 ng/ml of LPS. When each of CD14 mutants was expressed instead of wild-type CD14, ϳ10-fold higher con- centrations of LPS in 151-153A and 273-275A, and 100-fold higher concentrations of LPS in d35-44 were required to achieve equivalent responses to the wild type, and their max-imal responses were found to be 70 -80% of the maximal response of the wild-type. In cells expressing d235-243, the LPSresponse was severely impaired, and even 100 ng/ml LPS produced only 40 -50% of the activity.
The substitution mutants that lost the ability to increase TLR4-mediated activation were also examined for PEX-induced activation of NF-B (Fig. 4). The increase in reporter activity in response to PEX observed in 293 cells expressing TLR2 and wild-type CD14 was also observed in cells expressing 151-153A or 273-275A instead of wild-type CD14 (Fig. 4).
Cell Surface Expression of CD14 Mutants-Next, we attempted to determine whether the CD14 mutants examined above are expressed on the cell surface. The cell surface proteins of 293 cells expressing each of the CD14 mutants were labeled by exposing them to a membrane-impermeable biotinylation reagent, and the biotinylated proteins were collected by precipitation with a streptavidin gel and assayed for CD14 mutants by Western blotting (Fig. 5). In the absence of biotinylation, no CD14 was detected in the precipitate (lane 1, upper panel), although CD14 was expressed (lane 1, lower panel). When CD14 was not expressed, no signals were detected in either the precipitate or the lysate even when biotinylation was performed (lane 2). When the CD14 mutants indicated in Fig. 5 were expressed, all of them were detected in both the precipitates and the lysates (lanes 3-8). These results indicate that these CD14 mutants were expressed on the membrane surface.
CD14 Mutants Are Still Capable of Associating with TLR4 -Next, we used a co-immunoprecipitation method to investigate whether the CD14 mutants that had lost TLR4-mediated activity were capable of associating with TLR4. After transfecting 293 cells with wild type or a mutant CD14 plasmid together with a control vector or the plasmid(s) for TLR4 and/or MD-2, CD14 protein was immunoprecipitated from cell extracts prepared from the cells, and co-precipitated TLR4 and MD-2 proteins were detected by Western blotting (Fig. 6A, top panel). No signals corresponding to MD-2 or TLR4 were observed when a control vector or plasmid for MD-2 was co-transfected with the wild-type CD14 plasmid (lanes 1 and 3), but co-precipitation of TLR4 was detected when TLR4 was co-expressed with wildtype CD14 (lane 2). When both TLR4 and MD-2 were coexpressed with wild-type CD14, both TLR4 and MD-2 were co-precipitated (lane 5). When each of the CD14 mutants indi- cated in Fig. 6 was co-expressed with TLR4 and MD-2, both TLR4 and MD-2 were co-precipitated with every CD14 mutant tested (lanes [5][6][7][8]. Western blotting of the cell extract showed that all proteins transfected were expressed normally (Fig. 6A,  lower panel). These results indicate that these wild-type and mutant CD14 proteins are capable of associating with TLR4 but not MD-2. Thus, we further examined whether these CD14 mutant proteins are capable of associating with TLR4 on the membrane surface. After transfecting 293 cells with wild type or a mutant CD14 plasmid together with a control vector or the plasmids for TLR4 and MD-2, cell surface proteins were crosslinked by a membrane-impermeable bifunctional cross-linking reagent, sulfosuccinimidyl-2-[p-azido-salicylamido]ethyl-1,3Јdithiopropionate. This cross-linking reagent possesses aminereactive N-hydroxysuccinimide in one end and photoreactive hydroxyphenyl azide in the other end with a cleavable 18.9-Å spacer arm. Thus, proteins present in close proximity on the membrane surface are cross-linked to each other with this reagent. After it was cross-linked, hexahistidine-tagged TLR4 was collected by precipitation with nickel-agarose in the presence of 6 M urea to disrupt noncovalent protein-protein association. Then CD14 protein cross-linked with TLR4 was detected by Western blotting after the cross-linking reagent was cleaved with a reducing agent (Fig 6B). When cells were not treated with this cross-linking reagent, CD14 protein was not detected (lane 8), indicating that noncovalently associated CD14 was not co-precipitated with TLR4 in this experimental condition. However, upon cross-linking, all of wild-type and mutant CD14 proteins examined were clearly co-precipitated with TLR4 (lanes 3-7). This result indicates that these wild-type and mutant CD14 proteins are present on the membrane surface in close proximity to TLR4 with a distance of less than 18.9 Å, which is close enough for these proteins to associate with each other. A co-immunoprecipitation experiment was also carried out. After membrane surface proteins were labeled with a membrane-impermeable biotinylation reagent, the biotinylated proteins were collected by precipitation with streptavidin-agarose. Collected proteins were eluted from the agarose by incubating with an excess amount of a watersoluble biotin derivative. Then CD14 was immunoprecipitated from eluted proteins, and co-precipitated TLR4 was detected by Western blotting. The result showed that TLR4 was co-precipitated with all of wild-type and mutant CD14 proteins used in Fig. 6B (data not shown). This result supported the idea that these mutant CD14 proteins are capable After 24 h, cell surface proteins were biotinylated with a membraneimpermeable biotinylating reagent, and cell extracts were prepared. The cell extracts were divided into two portions, and biotinylated proteins were collected from one portion of the extracts with streptavidinagarose. After washing, the agarose beads were boiled in an SDSsample buffer, and the supernatants (ppt.; top) and cell extracts set aside (cell ext.; bottom) were analyzed for CD14 by Western blotting (WB) with an anti-EIAV tag antibody.
FIG. 6. Association of CD14 mutants with TLR4. A, 293 cells were transiently transfected with the expression plasmids indicated (1 g for CD14 and its mutants; 6 g for TLR4 and MD-2). After 24 h, cell extracts were prepared and divided into two portions. CD14 and its mutants were immunoprecipitated with an anti-CD14 antibody from one portion of the extracts, and co-precipitated TLR4 and MD-2 were detected by Western blotting with an anti-EIAV tag antibody (IP; top). Other portions of the cell extracts were detected for CD14, TLR4, and MD-2 in the same way (cell ext.; bottom). B, 293 cells were transiently transfected with the expression plasmids indicated (4 g for CD14 and its mutants; 24 g for TLR4 and MD-2). After 24 h, cells were either untreated (lane 8) or treated (lanes 1-7) with a membrane-impermeable cross-linking agent, and then cell extracts were prepared. From the extracts, hexahistidine-tagged TLR4 was collected by precipitation with nickel-agarose in the presence of 6 M urea to disrupt noncovalent protein-protein association. Then CD14 protein cross-linked with TLR4 was detected by Western blotting after the cross-linking reagent was cleaved with a reducing agent. of associating with TLR4 on the membrane surface.
LPS Binding to Wild-type or Mutant CD14 and TLR4/MD-2-Next we investigated the binding of LPS to CD14 mutants and the capability of these mutants to transfer LPS to TLR4/ MD2 using photoactivable LPS (SBED-LPS). SBED-LPS is a UV-activable cross-linking LPS carrying biotin covalently attached to a photoactivable aryl azide moiety. After transfecting 293 cells with wild type or a mutant CD14 plasmid together with a control vector or the plasmid(s) for TLR4 and/or MD-2, cells were incubated with SBED-LPS. After UV irradiation, proteins cross-linked with SBED-LPS were collected by precipitation with a streptavidin gel. Then collected CD14, TLR4, and MD-2 proteins were detected by Western blotting (Fig. 7). When all of the wild type CD14, TLR4, and MD-2 were expressed, all of these proteins were labeled with SBED-LPS (lane 4). When either TLR4 or MD-2 was omitted, labeling to CD14 was reduced, and no labeling to TLR4 and MD-2 was observed (lanes 2-4), indicating that labeling to TLR4 or MD-2 requires the presence of both TLR4 and MD-2. When CD14 was omitted, no labeling to TLR4 and MD-2 was observed (lane 1), indicating that the labeling to these proteins requires CD14. These results agreed with those described by da Silva et al. (39). When mutant CD14 proteins instead of wild-type CD14 Species Difference in CD14 Molecules-Since N-terminal 151 amino acids of human CD14 were reported to be sufficient to transmit LPS signaling (27,28), we compared the effects of human (hCD14-DAF) and mouse (mCD14-DAF) CD14 truncated at amino acid 151 attached to the GPI anchor sequence of decay-accelerating factor (Fig. 8). When hCD14-DAF was coexpressed in 293 cells together with human TLR4 and MD-2, the increase in NF-B-dependent reporter activity comparable with wild type CD14 was observed in response to LPS. On the other hand, mCD14-DAF lost the ability to increase LPS-in- After 24 h, cell surface proteins were biotinylated with a membrane-impermeable biotinylating reagent, and cell extracts were prepared. Then biotinylated proteins were collected with streptavidin-agarose. After washing, the agarose beads were boiled in an SDS-sample buffer, and the supernatants were analyzed for CD14 by Western blotting with an anti-EIAV tag antibody. duced reporter activity in 293 cells expressing mouse TLR4 and MD-2 (Fig. 8A), although both hCD14-DAF and mCD14-DAF were expressed on the membrane surface (Fig. 8B). DISCUSSION We created three types of deletion mutants to investigate the structural requirement of mouse CD14 molecules for TLR4and TLR2-mediated activations of NF-B. The first type of deletion mutant (d35-44) lacked the region (amino acids [35][36][37][38][39][40][41][42][43][44]) that corresponds to the region of human CD14 reported to be important for LPS recognition and signal transduction. The second type of deletion mutant (d144 -153, d235-243, and d270 -275) lacked regions conserved among species, and the third type of deletion mutant (d244 -247 and d312-320) lacked regions that vary among species. We avoided the deletion of possible glycosylation sites of CD14 because they might release the CD14 molecule from the cell membrane by disrupting its GPI anchoring. Deletion of variable regions (d244 -247 and d312-320) did not significantly affect either TLR2-or TLR4mediated activation (Fig. 1, A and B). Deletion of conserved regions (d144 -153, d235-243, and d270 -275) always caused loss of either TLR2-or TLR4-mediated activation, indicating an important role of these conserved regions in mouse CD14 function. Interestingly, deletion of amino acids 35-44 caused loss of both LPS-induced TLR4-mediated activation and PEX-induced TLR2-mediated activation (Fig. 1, A and B). We used a phenol extract of E. coli O111:B4 LPS (PEX) to stimulate the TLR2dependent pathway, because the LPS preparation repurified by the method described by Hirschfeld et al. (18) did not stimulate this pathway (data not shown) as shown by Hirschfeld et al. (18). We were also unable to use peptidoglycan and bacterial lipopeptides, which are known to be TLR2 ligands, because peptidoglycan yielded ambiguous results in our experimental system (data not shown), and the TLR2-mediated activity of a synthetic lipopeptide derived from E. coli murein lipoprotein, tripalmitoyl-Cys-Ser-Ser-Asn-Ala, was not dependent on CD14 (data not shown). The biochemical nature of the substance(s) contained in PEX is currently unknown. Hirschfeld et al. (18) and Dziarski et al. (40) reported that bacterial lipoproteins or endotoxin proteins may be responsible for the activation of TLR2-dependent pathway. However, PEX-induced activation of NF-B was clearly different in terms of CD14 dependence from that induced by bacterial lipoproteins. It is therefore likely that a substance(s) other than bacterial lipopeptides is responsible for the effect of PEX. Our finding that the d35-44 mutant had lost activity in response to PEX suggests involvement of an LPS-like substance(s) in the PEX-induced activation. An effort to identify the substance(s) contained in PEX is currently in progress in our laboratory.
To identify the molecular mechanism responsible for the loss of TLR4-mediated activity by the CD14 mutants, we tested them for membrane surface expression, and the results showed that they were expressed on the cell surface (Fig. 5). Since close interaction between CD14 and TLR4 has been suggested to participate in LPS signaling (20), we investigated the association between CD14 mutants and TLR4, and, as shown in Fig. 6, the results showed that TLR4 was co-precipitated with all mutants (d35-44, 151-153A, d235-243, and 273-275A). Therefore, these regions of mouse CD14 are unlikely to play a significant role in the interaction with TLR4. Our result also suggested that deletion or substitution of these regions does not cause any destructive change in the structure of CD14. In the present study, an association between CD14 and TLR4 was detected by the co-immunoprecipitation technique regardless of MD-2 expression, although our data do not necessarily mean a direct interaction between CD14 and TLR4. Co-precipitation of MD-2 with CD14 was also detected when MD-2 was co-expressed with CD14 and TLR4, but no co-precipitation of MD-2 was detected when CD14 and MD-2 were expressed without TLR4. It is therefore likely that MD-2 was co-precipitated with CD14 through the interaction between MD-2 and TLR4.
The amino acid regions of CD14 between 144 -153 and 270 -275 were investigated in detail in this study, because these regions were found to be selectively involved in TLR4-mediated activity (Fig. 1, A and B). We were able to narrow these required regions down to 151-153 and 273-275, respectively (Fig.  2, A and B), and since point mutants in which each of the amino acids in these regions was changed to alanine retained TLR4mediated activity (Fig. 2, A and B), the secondary structure created by these regions is likely to play an important role in the TLR4-mediated activity. By contrast, the amino acid regions between 35-44 and 235-243 were required for both TLR4-and TLR2-mediated activity (Fig. 1, A and B). As stated above, deletion of these regions seems to cause no destructive change in CD14 structure because CD14 mutants (d35-44 and d235-243) in which these regions were deleted retained their ability to associate with TLR4. These regions may play a fundamental role in CD14 function.
In this study, photoactivable LPS (SBED-LPS) was used to assess LPS binding to CD14 or mutant CD14 and the capability of these mutants to transfer LPS to TLR4/MD2 (Fig. 7). The labeling to TLR4/MD-2 was completely CD14-dependent, because no labeling was observed when CD14 was not expressed. These results agreed with those described by da Silva et al. (39). In this experimental condition, we found that labeling intensities to CD14 mutants, TLR4 and MD-2, were roughly paralleled by the ability of CD14 mutants to increase TLR4mediated activation (see Fig. 3). It is, therefore, likely that the impairment of LPS binding and its transfer to TLR4/MD-2 is involved in the impaired TLR4-mediated activity.
Viriyakosol et al. (28), using chimeric CD14 in which amino acids 1-151 of human CD14 were fused to the C-terminal region of decay-accelerating factor, demonstrated that the Cterminal portion of human CD14 beyond amino acid 152 is not required for its role as a membrane receptor for the LPS response. In our study, the amino acid region 273-275 of mouse CD14 was required for TLR4-mediated activation of NF-B (Fig. 2B). In addition, mouse CD14 truncated at amino acid 151 attached to the GPI anchor sequence of decay-accelerating factor (mCD14-DAF) lost the ability to increase LPS-induced reporter activity in 293 cells expressing mouse TLR4 and MD-2. Therefore, our results demonstrate that human and mouse CD14 are clearly different in terms of the requirement of their C-terminal portions in TLR4-mediated activation of NF-B. We previously reported that Salmonella lipid A exerts very little stimulatory activity on human macrophages despite exerting strong activity on murine macrophages (26). Although similar species-specific actions have been observed with lipid IV A (compound 406) and CD14 has been reported (41) not to be involved in its species-specific actions, the structure of lipid IV A is not found as a component of any LPS of bacterial origins, and emphasis should be put on the fact that Salmonella lipid A is a naturally occurring lipid A. Thus, it will be interesting to investigate the involvement of the C-terminal part of mouse CD14 in species-specific actions of Salmonella lipid A.