Lipopolysaccharide is in close proximity to each of the proteins in its membrane receptor complex. transfer from CD14 to TLR4 and MD-2.

The structural features of some proteins of the innate immune system involved in mediating responses to microbial pathogens are highly conserved throughout evolution. Examples include members of the Drosophila Toll (dToll) and the mammalian Toll-like receptor (TLR) protein families. Activation of Drosophila Toll is believed to occur via an endogenous peptide rather than through direct binding of microbial products to the Toll protein. In mammals there is a growing consensus that lipopolysaccharide (LPS) initiates its biological activities through a heteromeric receptor complex containing CD14, TLR4, and at least one other protein, MD-2. LPS binds directly to CD14 but whether LPS then binds to TLR4 and/or MD-2 is not known. We have used transient transfection to express human TLRs, MD-2, or CD14 alone or in different combinations in HEK 293 cells. Interactions between LPS and these proteins were studied using a chemically modified, radioiodinated LPS containing a covalently linked, UV light-activated cross-linking group ((125)I-ASD-Re595 LPS). Here we show that LPS is cross-linked specifically to TLR4 and MD-2 only when co-expressed with CD14. These data support the contention that LPS is in close proximity to the three known proteins of its membrane receptor complex. Thus, LPS binds directly to each of the members of the tripartite LPS receptor complex.

The structural features of some proteins of the innate immune system involved in mediating responses to microbial pathogens are highly conserved throughout evolution. Examples include members of the Drosophila Toll (dToll) and the mammalian Toll-like receptor (TLR) protein families. Activation of Drosophila Toll is believed to occur via an endogenous peptide rather than through direct binding of microbial products to the Toll protein. In mammals there is a growing consensus that lipopolysaccharide (LPS) initiates its biological activities through a heteromeric receptor complex containing CD14, TLR4, and at least one other protein, MD-2. LPS binds directly to CD14 but whether LPS then binds to TLR4 and/or MD-2 is not known. We have used transient transfection to express human TLRs, MD-2, or CD14 alone or in different combinations in HEK 293 cells. Interactions between LPS and these proteins were studied using a chemically modified, radioiodinated LPS containing a covalently linked, UV light-activated crosslinking group ( 125 I-ASD-Re595 LPS). Here we show that LPS is cross-linked specifically to TLR4 and MD-2 only when co-expressed with CD14. These data support the contention that LPS is in close proximity to the three known proteins of its membrane receptor complex. Thus, LPS binds directly to each of the members of the tripartite LPS receptor complex.
Bacterial endotoxin (lipopolysaccharide, LPS) 1 okonp61 is a complex glycolipid composed of a hydrophilic polysaccharide moiety and a hydrophobic domain known as lipid A (1). LPS is an outer membrane constituent of all Gram-negative bacteria where it has indispensable barrier functions. LPS is also a potent activator of innate immune responses that result in the production of pro-and anti-inflammatory mediators from my-eloid lineage and other cell types (2). LPS-induced cell activation depends on the presence of three proteins comprising a multiprotein cell surface receptor complex here termed the LPS receptor complex. One essential protein of the LPS receptor complex is CD14 (3), a 55-kDa glycoprotein present in soluble form (sCD14) in blood or as a membrane-bound form (mCD14) in myeloid lineage cells. This latter form is attached to the outer leaflet of the cell membrane via a glycosylphosphatidylinositol anchor. Multiple lines of biochemical and genetic evidence support the contention that CD14 principally acts to bind LPS and does not participate in signaling directly. Thus, others and we postulated that there must be at least one transmembrane protein that acts in concert with CD14 (2). This putative transmembrane protein is now identified as a member of the mammalian Toll-like receptor (TLR) family and is TLR4 (4,5). Genetic and biochemical studies suggest that TLR4 plays an important role in LPS signaling under physiological conditions. Positional cloning and sequencing of the lps d locus localized the defect to the tlr4 gene (6). The importance of TLR4 in LPS signaling is further supported by the fact that TLR4-deficient mice are LPS hyporesponsive but respond normally to products of Gram-positive organisms (7). MD-2 is another protein that appears to be important in LPS signaling (8). However, its function in the LPS receptor complex is currently unknown.
TLR2 is a signaling receptor for a variety of microbial products that in some cases require CD14 for maximum activation (9). Although TLR2 was initially thought to function as an LPS receptor current evidence suggests that it does not play a major role in the physiological response to LPS (9 -11). Nonetheless a recent publication provides data showing that when TLR2 is overexpressed in 293 cells it mediates cell activation induced by a variety of purified LPS isolates (12). It thus appears that TLR2 is more promiscuous with respect to ligand recognition and that LPS must be added to the list of TLR2 ligands.
Numerous reports describe the direct binding of LPS to CD14 (3,(13)(14)(15). The question of whether LPS binds to, or even is in close proximity to the other proteins of the LPS receptor remains unanswered. Here we have used radioiodinated Re595 LPS with a covalently attached UV-activated phenylazide that is capable of cross-linking to nearby proteins in order to characterize interactions between LPS and the LPS receptor complex at the cell surface. Our results show that LPS is brought into close proximity to TLR4 only when it is present as an LPS⅐CD14 complex and when CD14 and TLR4 are co-expressed with MD-2.

MATERIALS AND METHODS
Cell Culture and Transfection-Human embryonic kidney 293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 units/ml penicillin, * This work was supported in part by National Institutes of Health Grants AI15136 (to R. J. U.), GM28485 (to R. J. U.), GM37696 (to R. J. U.), AI32021 (to P. S. T.), and HL23584 (to P. S. T.). This is Publication 136-IMM from The Scripps Research Institute, Department of Immunology. 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.
Reagents-Murine anti-human TLR4 monoclonal antibodies were from Dr. K. Miyake. The following reagents were prepared as described in the references: Re595 LPS (16), rabbit LBP (17), and recombinant human soluble CD14 (18). Mouse monoclonal antibodies against CD14 (28C5, 63D3, and 18E12) were a gift from A. Moriarty and D. Leturcq (R. W. Johnson Pharmaceutical Research Institute, La Jolla, CA). M2 anti-FLAG monoclonal antibody was from Sigma. Escherichia coli LCD25 LPS and 0111:B4 LPS were from List Biological Labs (Campbell, CA). E. coli 055:B5 LPS was from Sigma. Protein A-Sepharose was from Amersham Pharmacia Biotech (Piscataway, NJ). In order to mitigate concerns about protein contaminants the Re595 LPS was also repurified according to the methods of Hirschfeld et al. (10). This material was also used to prepare 125 I-ASD-Re595 LPS; results obtained with the 125 I-ASD derivative of the repurified Re595 LPS were not distinguishable from the parent LPS.
Mammalian Expression Constructs-Human CD14 cDNA was cloned in pRc/RSV vector as described (18). The cDNAs for human TLR2 and TLR4 were gifts from Drs. P. Godowski and C. Janeway, respectively. The cDNAs for Drosophila Toll and human TLR1 were gifts from Dr. Alan Aderem. FLAG-dToll was cloned in pFLAG-CMV1 using HindIII/ BamHI restriction sites. Expression vectors for FLAG-MD-2 was engineered by introducing a NotI/SmaI DNA fragment into pFLAG-CMV1 plasmid, in which the prepro trypsin leader sequence precedes an NH 2 -terminal FLAG epitope.

125
I-ASD-Re595 LPS Cross-linking Assays-293 cells cultured in 10-cm Petri dishes were transfected and harvested 2 days after transfection. Cells were washed 2 times with 50 mM HEPES, pH 7.4, 140 mM NaCl, 5 mM EDTA, pH 7.4, and 2 times with DMEM containing 50 mM HEPES, pH 7.4. Cells were pelleted and resuspended in Dulbecco's modified Eagle's medium containing 50 mM HEPES, pH 7.4, and 2.5% endotoxin-free human bovine serum albumin (Serologicals Proteins Inc., Kankakee, IL). LBP⅐LPS complexes were formed and utilized as previously described (19). Briefly, 25 g/ml rLBP and 2.5 g/ml 125 I-ASD-Re595 LPS were incubated in a solution containing 25 mM HEPES, pH 7.4, and 2.5 mM EDTA, pH 8, 10 min at 37°C. LBP⅐LPS complexes were added to cells, incubated 10 min at 37°C, and photolyzed with a 253-nm UV light for 4 min on ice. A typical incubation reaction contained a final concentration of 250 ng/ml 125 I-ASD Re595 LPS and 2.5 g/ml rLBP. When metabolic inhibition was specified, cells were incubated for 30 min at 37°C in binding buffer containing 10 mM sodium azide, 2 mM sodium fluoride, and 5 mM deoxyglucose prior to addition of LBP⅐LPS complexes.
Immunoprecipitation and Western Blot Analysis-Cell lysates were pre-cleared 3 times for 20 min at 4°C with 60 l of protein A-Sepharose beads, and mixed with 30 g of M2 monoclonal antibodies for 3 h at 4°C under constant agitation. Immune complexes were allowed to bind to 60 l of protein A-Sepharose beads overnight, beads were washed 3 times with lysis buffer and the washed beads resuspended in 100 l of Laemmli buffer and boiled for 10 min. Aliquots from 20 l of the original mixtures were separated on 12% SDS-PAGE and transferred to nitrocellulose membranes. Filters were blocked with 5% nonfat milk in blocking buffer (TBS-T, 50 mM Tris-Cl, pH 7.5, 150 mM NaCl, 0.1% Tween 20), and incubated with anti-FLAG antibody for 2 h and with peroxidase-conjugated secondary antibody for 1 h at ambient temperature. Specific bands were revealed using the ECL Plus system (Amersham Pharmacia Biotech). Aliquots of 80 l of immunoprecipitates were separated on 12% SDS-PAGE. The gels were dried and the radioactivity associated with protein was quantitated with a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Each experiment described here was performed at least three times. Here we show individual representative experimental data.
Flow Cytometry Analysis-293 cells were plated at a density of 5 ϫ 10 5 cells/well in 6-well plates, and transfected with the indicated plasmids together with pEGFP-N3 vector (CLONTECH Laboratories, Inc., Palo Alto, CA) for 2 days. Subsequently, cells were harvested, washed twice in phosphate-buffered saline containing 1% fetal calf serum and 0.1% NaN 3 and incubated with anti-FLAG or HTA125 (10 g/ml each) for 45 min at 4°C. After 2 washes, cells were labeled for 45 min with phycoerythrin-conjugated goat anti-mouse antibody (BD PharMingen, San Diego, CA). Cells were then washed twice and analyzed with a FACSCalibur flow cytometer (Becton Dickinson, Mountain View, CA).

LPS Is Cross-linked to TLR4 and MD-2 in a CD14-dependent
Manner-HEK 293 cells were transfected with various combinations of DNA encoding wild-type human CD14 and epitopetagged (FLAG) versions of TLR4 and MD-2. Two days after transfection the cells were washed and exposed to a preformed complex of LBP and 125 I-ASD-Re595 LPS as described (19). After photolysis, cells were lysed and the lysates treated with anti-FLAG monoclonal antibody. The resultant immunoprecipitates were subjected to SDS-PAGE first to reveal radioactive bands by phosphorimaging (top panels) and then to determine TLR4 and MD-2 protein expression (bottom panels) by immunoblotting with anti-FLAG antibody. Evidence for LPS cross-linking to TLR4 and MD-2 was obtained by the radiolabeled protein bands noted only when CD14 was expressed (Fig.  1A, lane 7). We failed to observe binding of LPS to TLR4 or MD-2 when expressed alone (lanes 2 and 4, respectively) or in combination with CD14 (lanes 3 and 5). When TLR4 and MD-2 were expressed together a barely detectable band corresponding to MD-2 was found (lane 6). In dose-response studies, not shown, identical labeling patterns were observed even when we reduced the input LPS concentration to as low as 10 ng/ml. Unless otherwise noted all experiments were done with 250 ng/ml 125 I-ASD-Re595 LPS. Western blotting revealed similar levels of expressed proteins. MD-2 appears as two species characterized by faster and slower migrating forms due to differential glycosylation patterns. 2 Although CD14 does not contain an epitope tag we were unable to fully eliminate it from anti-FLAG immunoprecipitates despite a pre-clear step and extensive washing. Several lines of evidence suggest this reflects a minor amount of nonspecific carryover in the immunoprecipitates. First, quantitation of the amount of bound LPS showed that LPS bound to TLR4 and MD-2 represented less than 0.1% of total LPS bound to the CD14⅐TLR4⅐MD-2 complex. The remainder is bound to CD14 or other proteins. This is not surprising given the major function of CD14 to bind and internalize LPS in a high capacity manner. Second, analysis of whole cell lysates only revealed the presence of LPS bound to CD14 whereas immunoprecipitation was required to reveal LPS bound to TLR4 and MD-2 (data not shown).
Both mCD14 and sCD14 function to enhance cellular responses to LPS (3, 20 -22). We next examined whether sCD14 could substitute for mCD14 in facilitating LPS binding to TLR4 and MD-2. 293 cells were transfected with TLR4 and MD-2 either alone or in combination. The cells were then harvested, extensively washed, and exposed to the preformed LBP⅐LPS complex plus sCD14; sCD14 mediated LPS binding to TLR4 and MD-2 only when both proteins were co-expressed (Fig. 1B,  lane 4). No labeling of TLR4 or MD-2 was noted when these proteins are expressed alone (lanes 2 and 3). These data suggest that LPS was delivered from an LBP⅐LPS complex to a TLR4⅐MD-2 complex in a CD14-dependent manner.
To determine whether any other member of the TLR family would bind LPS in a CD14 and MD-2 dependent fashion, experiments were carried out with cells expressing TLR4, TLR1, or dToll. 293 cells were transfected with MD-2 and CD14 either in the presence of TLR4, TLR1, or dToll. Fig. 2A shows a specific binding of LPS only to TLR4. No binding to TLR1 or dToll could be observed under identical experimental conditions. Expression of each of the proteins was evaluated by immunoblotting ( Fig. 2A, bottom panel) and flow cytometry (Fig. 2B). TLR4 and MD-2 were clearly detectable at cell surface of 293 cells when expressed alone using HTA125, a monoclonal anti-TLR4 antibody, or with anti-FLAG antibody. The co-expression of TLR4 with MD-2 dramatically enhanced the fluorescence intensity of TLR4. This effect was observed with either anti-FLAG or HTA125 antibody. Moreover it was dependent on MD-2, since the introduction of CD14 in the absence of MD-2 had no effect on the fluorescence staining of TLR4 (data not shown). Both TLR1 and dToll were expressed at cell surface as shown in Fig. 2B (bottom panels). Unlike our observations with TLR4, co-expression with MD-2 did not modify cell surface expression of TLR1 or dToll (data not shown).
TLR2 had been reported to be a receptor for Gram-negative LPS (23,24). However, these observations resulted in part from trace contaminants present in the LPS isolates (10,25). On the other hand recent work of Dziarski et al. (12) strongly suggests that TLR2 can recognize some forms of purified, protein-free LPS including that of Re595 LPS. As part as our exploration of the specificity of the labeling reaction we asked whether Re595 LPS might also be brought into close proximity with TLR2 and if so whether CD14 and MD-2 are involved. We observed crosslinking of Re595 LPS to TLR2 in the absence of CD14 and MD-2 (Fig. 3, lane 2). We noted that co-expression of CD14 somewhat enhanced the amount of LPS bound to TLR2 (lane 3). In contrast to our findings with TLR4, co-expression of MD-2 had no detectable influence on Re595 LPS cross-linking to TLR2 whether or not CD14 was present (lanes 4 and 5). Moreover, we failed to observe any detectable binding of Re595 LPS to MD-2 regardless of whether TLR2 and/or CD14 were expressed. TLR2 protein expression level was constant in each of the conditions considered in this experiment (Fig. 3A, bottom  panel) and by flow cytometry analysis (Fig. 3C). Interestingly, and in contrast to findings with TLR4, expression of MD-2

FIG. 1. LPS is cross-linked to TLR4 and MD-2 in a CD14-dependent manner.
Panel A, LPS requires the presence of the tripartite complex composed of CD14⅐TLR4⅐MD-2 to bind TLR4 and MD-2. 293 cells were co-transfected with indicated plasmids (2.5 g of FLAG-TLR4, 2.5 g of FLAG-MD-2, and 1 g of CD14) and cultured for 2 days. Cells were incubated with 250 ng/ml 125 I-ASD-Re595 LPS and 2.5 g/ml LBP at 37°C for 10 min, and photolyzed for 4 min on ice. Standard M2 precipitates of lysates were separated by SDS-PAGE, and analyzed by autoradiography (top panel) or by immunoblotting (bottom panel) using anti-FLAG antibody, indicating expression of FLAG-TLR4 and FLAG-MD-2 proteins. Positions of molecular mass standards (kDa) are shown on the left. Panel B, sCD14 can substitute mCD14 in transferring LPS to TLR4 and MD-2. Cells were transfected with the indicated vectors and cultured for 2 days. Cells were incubated with LBP and 125 I-ASD-LPS plus 5 g/ml sCD14. Cells were photolyzed and immunoprecipitates analyzed as described above.

FIG. 2. LPS specifically binds to TLR4 but not to TLR1 and dToll.
Panel A, LPS specifically binds to TLR4. 293 cells were transfected with vectors encoding for CD14, FLAG-MD-2 either in the presence of FLAG-TLR4, FLAG-TLR1, or FLAG-dToll. Cells were harvested after 2 days and incubated with 2.5 g/ml LBP and 250 ng/ml 125 I-ASD-Re595 LPS at 37°C for 10 min. Cells were photolyzed and anti-FLAG immunoprecipitates analyzed as described above. Panel B, cell surface expression of TLR4, MD-2, TLR1, and dToll. 293 cells were transfected with pEGFP-N3 vector and with specified plasmids for 2 days. Cells were stained with anti-FLAG or HTA125 antibodies and analyzed with FACSCalibur. Analysis gate was set on GFP positive cells. Experiments were performed 3 times with similar results. failed to enhance cell surface expression of TLR2 (not shown). Others recently showed that co-expression of TLR2 and MD-2 resulted in increased total expression of each protein in cell lysates when compared with the levels detected with each expressed individually (12). However, the latter study did not determine the surface expression levels so we cannot directly compare our findings.
In order to be certain of the purity of the LPS used in our studies we re-extracted Re595 LPS according to the methods described in Hirschfeld et al. (10). We used the repurified LPS to prepare the radioiodinated derivative for cross-linking studies. The derivatized, repurified Re595 LPS behaved identically in all experiments including the findings of cross-linking to TLR2 as noted above (data not shown). Thus it is virtually certain that the results we describe here with the derivatized Re595 LPS solely reflect interactions of LPS and not protein contaminants with cell surface proteins.
Binding of LPS to CD14 has been shown to be independent of cellular energy metabolism (26). We next investigated cellular energy requirements for LPS binding to MD-2 and TLR4. To do this we compared CD14-dependent binding of Re595 LPS to TLR4 and MD-2 expressed in 293 cells; LPS was added to the cells in the presence and absence of metabolic inhibitors (Fig.  4). Addition of sodium azide, sodium fluoride, and 2-deoxyglucose failed to prevent Re595 LPS cross-linking to TLR4⅐MD-2. Thus transfer of LPS from CD14 to the TLR4⅐MD-2 complex does not require cellular energy. Similar results were observed with TLR2.
Specificity of Re595 LPS Binding to TLR4 and MD-2-The lipid A moiety has been identified as the LPS component re-sponsible for LPS-induced biological effects through direct interaction with CD14 (1). To determine whether LPS binding to TLR4 and MD-2 is a lipid A-dependent phenomenon, competition experiments were carried out in 293 cells transfected with CD14, TLR4, and MD-2. Cells were preincubated with an excess of various LPS isolates, a partial lipid A structure known as lipid IVa (27) and lipoteichoic acid for 3 min prior to the addition of 125 I-ASD-Re595 LPS (Fig. 5). All LPS from E. coli serotypes tested, 0111:B4, 055:B5, and LCD25, inhibited binding of Re595 LPS. The different LPS isolates as well as lipid IVa blocked LPS binding to TLR4 and MD-2 (Fig. 5, B and C). In contrast, lipoteichoic acid did not alter Re595 LPS binding to TLR4 and MD-2. Others have suggested that lipoteichoic acid is a ligand for TLR2 (28 -30). These data support the contention that the interactions between Re595 LPS and CD14⅐TLR4⅐MD-2 are lipid A-dependent.
To further confirm the importance of CD14 in transfer of LPS to the TLR4⅐MD-2 complex, we evaluated 125 I-ASD-Re595 LPS binding to TLR4 and MD-2 in the absence or presence of various anti-CD14 monoclonal antibodies (Fig. 6). Three monoclonal antibodies were used; 28C5 known to prevent LPS binding to both mCD14 and sCD14 (31), 18E12 an antibody which prevents LPS-induced cell activation but does not block LPS binding to CD14 (26); and 63D3 an antibody that has minimal effects of cell activation and on LPS binding to CD14 (32). Pretreatment with 28C5 markedly inhibited LPS binding to TLR4 and MD-2. Treatment with 18E12 did not prevent LPS binding to TLR4 and slightly enhanced binding to MD-2. The antibody 63D3 slightly decreased cross-linking to TLR4 and MD-2. In totality these results suggest that binding of LPS to CD14 is essential for the subsequent step(s) that facilitate interactions of LPS with TLR4 and MD-2. Interestingly 18E12, an anti-human CD14 monoclonal antibody that prevents cell activation without blocking LPS binding to CD14 did not prevent LPS interactions with other members of the receptor complex. This suggests that additional steps subsequent to LPS binding to TLR4 and MD-2 are required for cell activation. Further support for this concept derives from studies with anti-TLR4 monoclonal antibodies. Of the four monoclonal antibodies available to us, HTA405, HTA414, and HTA1216 all strongly inhibited LPS-induced TNF and IL-8 release from blood cells (25). One, HTA125, failed to show the inhibitory activity in the same assay system (data not shown). Thus we used these antibodies to see whether they could modulate LPS binding to TLR4 and MD-2. Surprisingly none of the inhibitory antibodies reduced radiolabeling of TLR4 or MD-2. One of the three blocking antibodies, HTA405, reduced the radiolabeling by about 30% (data not shown) suggesting it may react at or near an LPS-binding site of TLR4.

DISCUSSION
Here we have used a radioiodinated derivative of Re595 LPS substituted with a UV-activated phenylazide to determine whether, after binding to CD14, we could detect cross-linking of Re595 LPS to other proteins in the putative LPS receptor complex. Using this approach we showed that after binding to CD14, Re595 LPS is in close enough proximity to both TLR4 and MD-2 that cross-linking occurred. These interactions do not require cellular energy, but are strikingly facilitated by CD14, as in cellular activation. Anti-CD14 monoclonal antibodies that block LPS binding to CD14 prevent the subsequent interactions with TLR4 and MD-2. Interestingly anti-TLR4 monoclonal antibodies that inhibit LPS-induced cell activation do not prevent LPS binding to CD14, TLR4, or MD-2. Altogether these data suggest that LPS binds directly to proteins in the LPS membrane receptor complex but that binding to TLR4 and/or MD-2 is not sufficient to induce cell activation. Other steps such as protein oligomerization or involvement of another member of the LPS receptor complex may be required to induce transmembrane signaling.
In this study we used HEK 293 cells since they do not express endogenous TLR4, MD-2, or CD14. Through transient transfection we were able to control expression levels of each of the three components of LPS receptor, particularly of TLR4 that was reported to be expressed at very low levels in most cells (11,33). Furthermore, this cell system allowed us to detect TLR4 and MD-2 as epitope-tagged proteins thereby enriching their concentrations in immunoprecipitates and facilitating detection by Western blotting. Thus the epitope-tagged proteins are essential for detection of cross-linking. Previous studies failed to demonstrate binding to proteins other than CD14 at the cell surface presumably as a result of the low levels of surface expression of TLR4 and MD-2 (19,34).
LPS binding to TLR4 and MD-2 readily occurred only in the presence of the ternary complex composed of CD14, TLR4, and MD-2. As expected, CD14 represents the major binding site on the cell surface since it contained more than 98% of total radioactive LPS bound to the cell. Our data suggest that only a small fraction of the LPS bound to CD14 is transferred to TLR4. This observation is reminiscent of the results of Gegner et al. (31) who showed that most of the cellular CD14-associated LPS appeared not to be involved in cellular activation. Unfortunately the present system only provides qualitative binding data and detailed analyses of affinities will require the development of more precise analytical approaches.
The mechanism by which TLR4 transduces LPS signaling across the plasma membrane is not known. Our observations suggest a complex mechanism for activation of cells by LPS. LPS is transferred from a CD14⅐LBP complex to a TLR4⅐MD-2 complex at cell surface. Several lines of evidence support this model. First, LPS binds efficiently to TLR4 and MD-2 only in the presence of CD14. Although MD-2 and TLR4 were both strongly detected at cell surface of cells when co-expressed as shown by FACS analysis, only faint binding of LPS to either protein could be observed in the absence of CD14. CD14 can be present either on the cell surface as a membrane-bound protein, or added to the cells as a soluble form. Second, 28C5, an anti-CD14 antibody that has been shown to mask the LPSbinding site on CD14, strongly abolishes LPS binding not only to CD14, but also to TLR4 and MD-2. Thus there is a critical role of CD14 in delivering LPS to TLR4 and MD-2. Finally, LPS binding to TLR4 and MD-2 was not affected in the presence of a mixture of metabolic inhibitors suggesting that transfer of LPS takes place at the plasma membrane. The exact role of LPS binding to TLR4 and MD-2 with respect to signal transduction is not yet clear. Surprisingly the anti-CD14 monoclonal 18E12 known to block LPS-induced cell activation without preventing binding of LPS to CD14 (26), failed to prevent LPS binding to TLR4 and MD-2. Moreover inhibitory anti-TLR4 monoclonal antibodies had a limited or no effect of LPS binding to TLR4 or MD-2. One interpretation of these data is that TLR4 has different domains in the ectodomain; an LPS-binding site, an activation domain, and regions involved in protein-protein interaction with TLR4 itself, MD-2 and/or CD14. It is likely that oligomerization of TLR4 occurs following interactions with FIG. 4. LPS binding to TLR4 and MD-2 occurs at cell membrane. 293 cells were transfected with 2.5 g of FLAG-TLR4 in the presence of CD14 and FLAG-MD-2. Two days after transfection cells were washed and preincubated in a buffer containing sodium azide, sodium fluoride, and deoxyglucose for 10 min on ice. Cells were incubated with 2.5 g/ml LBP and 250 ng/ml 125 I-ASD-Re595 LPS at 37°C for 10 min. Cells were photolyzed and immunoprecipitates analyzed as described above.
LPS. Similarly, Poltorak et al. (33) have also argued that LPS binds to TLR4 and initiates subsequent oligomerization required for cell activation.
While biochemical and genetic studies establish a role of TLR2 as a receptor for lipoproteins and peptidoglycan from diverse Gram-positive bacteria (28, 30, 35, 36) its role as a receptor for most Gram-negative LPS under physiological conditions is in doubt (7,11). It does appear, however, that purified Cells were harvested 48 h post-transfection and either left untreated or preincubated with 100 g/ml 0111:B4 LPS, 100 g/ml 055:B5 LPS, 500 g/ml LCD25 LPS, 5 g/ml Lipid IVa, or 100 g/ml lipoteichoic acid for 3 min at 37°C prior to incubation with 250 ng/ml 125 I-ASD-Re595 LPS and 2.5 g/ml LBP. Panels B and C, quantification of incorporated radioactivity to TLR4 and MD-2, respectively, by PhosphorImager analysis.
FIG. 6. 28C5 blocks LPS binding to TLR4 and MD-2. Panel A, 293 cells were transfected with vectors encoding for CD14, FLAG-TLR4, and FLAG-MD-2. After 2 days, cells were harvested and either left untreated or preincubated with 10 g/ml of each antibody 28C5, 63D3, or 18E12 for 1 h at 37°C prior to incubation with 250 ng/ml 125 I-ASD-Re595 LPS and 2.5 g/ml LBP. Panels B and C, quantification of incorporated radioactivity to TLR4 and MD-2, respectively, by Phos-phorImager analysis.
LPS is capable of activating cells overexpressing TLR2 in the presence of MD-2 (12). Furthermore, it has been shown that ␥␦ T cells as well as macrophages derived from C3H/HeJ mice respond to LPS from E. coli and Porphyromonas gingivalis and lipid A specifically through TLR2 (37,38). Here we show that LPS and TLR2 are capable of interacting as evidenced by the finding of radiolabeled TLR2 following exposure to the derivatized Re595 LPS. However, in contrast to our findings with TLR4, co-expression of MD-2 failed to influence these interactions. It is important to emphasize that we could not detect any cross-linked products when other members of the Toll family, dToll or TLR1, were expressed at the surface of 293 cells even when co-expressed with CD14 and MD-2. Thus we are confident that the cross-linked products detected in cells expressing TLR4 or TLR2 reflect specific LPS interactions with proteins functioning as receptors.
Previous studies suggested an essential role for MD-2 in determining cellular responses to LPS (39). No data were provided to suggest a mechanism for MD-2 function. Here we show that MD-2 is essential to detect LPS binding to TLR4, a step that can be assumed to be essential in the activation pathway. Information about a related protein MD-1 may provide some insight into how MD-2 functions. MD-1 is a secreted protein that binds to a protein known as RP105, a member of Toll receptor family that is specifically expressed on B-lymphocytes (40,41). MD-1 is thought to control expression of RP105 at the cell surface. Shimazu and co-workers (8) have reported that MD-2 is not expressed on the cell surface of the murine pro-B cell line Ba/F3 in the absence of TLR4. In contrast, our flow cytometry analyses showed that MD-2 could be detected at the cell surface of 293 cells in the absence of TLR4. TLR4 was also present, although weakly, at cell surface when expressed alone. However, the presence of MD-2 dramatically enhanced cell surface-exposed TLR4. Alternatively MD-2 may complex with TLR4 and enhance reactivity with the detecting antibodies. However, since an identical enhancement was observed with anti-FLAG and HTA125, the anti-TLR4 monoclonal antibody, we feel the latter speculation is less likely. Preliminary studies in our laboratory have shown that MD-2 may interact with selected regions of the ectodomain of TLR4; studies are underway to define these interactions through mutagenesis of TLR4. Preliminary experiments demonstrate that the first N-terminal hundred amino acids of TLR4, containing 5 LRRs, play a central role in binding LPS, since its deletion abolished LPS binding but not association with MD-2 (data not shown).
The biochemical data we have presented together with results of others (33) strongly supports a model where LPS binds directly to the known proteins of the putative LPS receptor complex. This contrasts with what is believed to occur in Drosophila where products of microbial pathogens most likely do not interact directly with Toll-like proteins. An essential question that remains unanswered is how binding to TLR4 and MD-2 is linked to the generation of a transmembrane signal.