The amino-terminal region of toll-like receptor 4 is essential for binding to MD-2 and receptor translocation to the cell surface.

Toll-like receptor 4 (TLR4) and MD-2 are pivotal components that elicit inflammatory responses to lipopolysaccharide (LPS). They have been shown to form a physical complex on the cell surface that responds directly to LPS. However, the functional region of TLR4 required for association with MD-2 and LPS responsiveness is poorly understood. To identify the region of TLR4, we created a series of mutants with deletions in the extracellular domain and examined their activities in human embryonic kidney 293 cells. A mutant with a 317-amino acid deletion from the membrane proximal region of TLR4 was capable of associating with MD-2, while only a 9-amino acid truncation of the N terminus severely impaired the interaction. The association between the two molecules was well correlated with TLR4 maturation into an endoglycosidase H-resistant form and the cell surface expression. Mouse MD-2 bound to human TLR4, but its activity to facilitate the cell surface expression of TLR4 and confer LPS responsiveness was much weaker than that of human MD-2, indicating species specificity. A chimeric receptor composed of the N-terminal region of human TLR4 and the adjacent region of mouse TLR4 showed preference for human MD-2 in its transport to the cell surface and responsiveness to LPS. Taken together, the N-terminal region of TLR4 is essential for association with MD-2, which is required for the cell surface expression and hence the responsiveness to LPS.

Innate immunity is the first line of host defense in the mammalian immune system and is conserved among multicellular organisms from plants and insects to humans (1,2). While adaptive immunity responds to specific antigens, innate immunity recognizes certain molecular patterns specific to microorganisms, known as pathogen-associated molecular patterns (PAMPs), 1 via germ line-encoded pattern recognition receptors (PRRs). Therefore, the recognition of PAMPs by PRRs underlies self/nonself discrimination by multicellular organisms.
Despite the biological significance of PRRs, only limited information is currently available regarding the molecular mechanisms of PRRs for recognition of PAMPs and the subsequent activation of downstream mediators. In Drosophila, the crystal structure of peptidoglycan recognition protein, a PRR for peptidoglycan, has recently been determined (3). However, the mechanisms for activation of the protease cascade leading to induction of antimicrobial peptides and/or melanin formation remain to be determined. In another arthropod, horseshoe crabs, factor C and factor G, PRRs for the hemolymph coagulation cascade, are synthesized as serine protease zymogens and converted to active proteases upon binding to their target PAMPs, lipopolysaccharide (LPS) and ␤-glucan, respectively (4). While the tertiary structures of some LPS-binding proteins/ peptides have been determined (5,6), the crystal structures of factor C and factor G have not been solved.
A growing body of evidence has indicated that Toll-like receptors (TLRs) are central transducers for various PAMPs to invoke mammalian innate immunity (7,8). TLRs are type I transmembrane proteins with an extracellular domain composed of leucine-rich repeats (LRRs) and a cytoplasmic domain called the toll/interleukin-1 receptor (TIR) domain, which shares similarity with that of the interleukin-1 receptor. Upon stimulation, a homotypic interaction(s) between the TIR domain of TLR and a cytosolic adaptor molecule(s) harboring a TIR domain leads to the activation of downstream signaling (8). Activation of TLRs culminates in the production of proinflammatory cytokines, antimicrobial peptides, and co-stimulatory molecules that induce acute inflammation or subsequent activation of the adaptive immune system.
LPS, a major cell wall constituent of Gram-negative bacteria, is well known as one of the strongest elicitors of the innate immune system (9). LPS interacts with LPS-binding protein in serum and then with a glycosylphosphatidylinositol-anchored cell surface protein, CD14 (10 -12). Recent genetic studies and reconstitution analyses with LPS-unresponsive cells have revealed that TLR4 is indispensable for LPS-mediated intracellular signaling (13)(14)(15)(16)(17). TLR4 alone is not sufficient for the LPS signaling, and a secretory protein, MD-2, has also been identified as an essential component (18). MD-2 binds to the extracellular domain of TLR4 to form an active complex on the cell surface.
Despite the established physiological roles for TLR4 and MD-2, biochemical characterization of LPS recognition by the TLR4⅐MD-2 complex remains obscure. While LPS binding to LPS-binding protein or CD14 was clearly detected (10,11), efficient LPS binding to the TLR4⅐MD-2 complex expressed on cells has not been observed despite its capacity for LPS respon-siveness. Nevertheless, several groups have attempted to reveal the molecular mechanisms for ligand recognition by the TLR4⅐MD-2 complex, although the results are controversial. Several LPS derivatives such as lipid IV A and Salmonella lipid A and a plant-derived antitumor agent, Taxol, act as TLR4 agonists on mouse cells, although they are inactive or antagonistic on human cells. In reconstitution experiments with TLR4 and MD-2 derived from the two species, the species-specific responses to these compounds are attributable to the origin of MD-2, indicating that MD-2 is directly involved in the recognition and discrimination of the TLR4 ligands (19 -21). On the other hand, other groups have reported that the species-specific responses to lipid IV A or Rhodobacter sphaeroides lipid A are ascribed to the species of TLR4 (22,23). More recent studies have provided evidence for direct binding of LPS to the cell surface TLR4⅐MD-2 complex (24,25). Furthermore other studies have indicated that MD-2 binds to LPS and that the complex of MD-2 and LPS behaves as an active ligand for TLR4 (26 -28).
While the biological significance of the TLR4⅐MD-2 complex in the LPS responses is becoming clear, little is known about the structure-function relationships of the ectodomain of TLR4.
In the present study, as a first step toward understanding the mechanisms of TLR4 for recognizing LPS and transferring an activation signal downstream, we carried out molecular dissection of the extracellular domain of TLR4. We provide evidence that the amino-terminal region of TLR4 is a crucial determinant of its association with MD-2, which is essential for the cell surface expression of the receptor and hence the recognition of LPS. This is the first report to determine the functional region of TLR4 that associates with MD-2.

EXPERIMENTAL PROCEDURES
Cells and Reagents-HEK293 cells were cultured in Dulbecco's modified Eagle's medium containing 10% heat-inactivated fetal calf serum supplemented with 100 units/ml penicillin and 100 g/ml streptomycin. LPS from Escherichia coli 0111:B4 was purchased from List Biological Laboratories Inc. (Campbell, CA).
Plasmids-The expression plasmids for human TLR4 have been described previously (17). A cDNA for mouse TLR4 was obtained by reverse transcriptase-polymerase chain reaction from RAW264.7 cells, and a fragment encoding its mature protein portion was subcloned into pFLAG-CMV-1 (Sigma). cDNAs for human and mouse MD-2 were obtained from THP-1 cells and RAW264.7 cells, respectively, and subcloned into pcDNA3 (Invitrogen) with a carboxyl-terminal Myc tag. Truncated and chimeric mutants were created by polymerase chain reactions. pELAM1-Luc (29) and pRL-TK (Promega Corp., Madison, WI) were used as a nuclear factor-B (NF-B) reporter and an internal control reporter, respectively.
Immunoprecipitation and Immunoblotting-Cells were transfected with FLAG-tagged TLR4 or its mutant (5 g) with or without Myctagged MD-2 (5 g) by the calcium phosphate method (30). The total amount of transfected DNA was kept constant with an empty vector. Two days after transfection, the cells were lysed in lysis buffer containing 50 mM Tris-HCl (pH 7.5), 0.15 M NaCl, 1% Nonidet P-40, 2 mM EDTA, 50 mM NaF, and 0.2 mM Na 3 VO 4 supplemented with Complete TM protease inhibitor mixture (Roche Diagnostics) on ice for 15 min. The cell lysates were clarified by centrifugation and then subjected to immunoprecipitation with anti-FLAG monoclonal antibody (M2)agarose (Sigma) or protein G-Sepharose 4 Fast Flow (Amersham Biosciences) coupled with an anti-c-Myc monoclonal antibody (9E10). Immunoprecipitates were washed three times in lysis buffer and eluted with FLAG peptide (Sigma) or released by boiling in SDS-PAGE sample buffer under reducing conditions. In some experiments, the eluted samples were treated with endoglycosidase H (Endo H, Roche Diagnostics) according to the manufacturer's instructions. Samples were resolved by 10 or 12.5% SDS-PAGE, transferred to a polyvinylidene fluoride membrane (Millipore Corp.), and probed with peroxidase-conjugated anti-FLAG (M2, Sigma) or anti-c-Myc (9E10, Roche Diagnostics) monoclonal antibodies.
Flow Cytometric Analyses-Cells were transfected with FLAGtagged TLR4 or its mutants (2 g) with or without Myc-tagged MD-2 (8 g) by the calcium phosphate method (30). The total amount of trans-fected DNA was kept constant with an empty vector. Two days after transfection, the cells were harvested, washed in phosphate-buffered saline, and incubated with an anti-FLAG monoclonal antibody (M2) in phosphate-buffered saline containing 0.1% bovine serum albumin on ice for 30 min. After washing with phosphate-buffered saline, the cells were stained with goat F(abЈ) 2 anti-mouse IgG-Alexa Fluor® 488 (Invitrogen) on ice for 30 min. The cells were then washed with phosphate-buffered saline and analyzed using a flow cytometer (FACSCalibur TM , BD Biosciences).
Luciferase Assay-HEK293 cells (1.5ϫ10 5 cells/well) were plated in a 24-well plate and transfected with TLR4 (0.1 g) with or without Myc-tagged human or mouse MD-2 (0.1 g) together with the pELAM1-Luc reporter plasmid (0.125 g) and pRL-TK control plasmid (0.05 g) by the calcium phosphate method (30). The total amount of transfected DNA was kept constant with an empty vector. Two days after transfection, the cells were stimulated with or without 1 g/ml LPS for 6 h, and their luciferase activities were measured using a dual luciferase reporter assay system (Promega Corp.). The NF-B reporter activity was divided by the activity of the Renilla control reporter to normalize the transfection efficiency. Data are shown as the mean Ϯ S.E. of duplicate samples and are representative of two independent experiments.

RESULTS
The Amino-terminal Region of TLR4 Is Essential for Association with MD-2-Since MD-2 stably associates with the extracellular domain of TLR4 (18), we first analyzed the association between the two molecules. For efficient and specific immunoprecipitation and immunoblotting, we utilized aminoterminally FLAG-tagged human TLR4 (17) in which the mature polypeptide of TLR4 (predicted to begin with amino acid 25) was preceded by the signal sequence of preprotrypsin and a FLAG epitope (Fig. 1A, Full-length). The FLAG-tagged TLR4 responded to LPS resulting in NF-B activation, indicating that it was functionally active (data not shown). We cotransfected FLAG-tagged TLR4 or TLR2 and Myc-tagged MD-2 into HEK293 cells and immunoprecipitated FLAG-TLR4 or FLAG-TLR2 with an anti-FLAG antibody. In contrast to TLR2, the transfected TLR4 appeared as two bands on the SDS-polyacrylamide gel, and the upper band was slightly increased by coexpression of MD-2. The MD-2 coprecipitated with TLR4 was visualized by immunoblotting with an anti-Myc antibody. Consistent with previous reports (18,31), MD-2 specifically coprecipitated with TLR4 but not with TLR2, indicating a specific association (Fig. 1B). The amount of MD-2 in the cell lysate was not altered by transfection of TLR4 or TLR2.
To explore the functional region of TLR4, we constructed a series of human TLR4 mutants with deletions from the amino or carboxyl terminus of the extracellular domain (Fig. 1A). A small deletion of the first 3 amino-terminal amino acids (⌬25-27) did not affect TLR4 expression but significantly impaired its association with MD-2 (Fig. 1C). A further deletion (⌬25-33) completely abolished the association. On the other hand, mutant TLR4 molecules with large deletions in the carboxyl-terminal region of the extracellular domain (⌬314 -637 and ⌬321-637) were still capable of associating with MD-2. Truncations from the carboxyl-terminal region gradually diminished the association with MD-2. We also analyzed an internal deletion mutant of TLR4 (⌬151-175), and no association with MD-2 was observed (data not shown). Thus, the amino-terminal region of the extracellular region of TLR4 up to amino acid 320 is indispensable for the association with MD-2.
The Amino-terminal Region of TLR4 Is Required for Its Cell Surface Expression-Nagai et al. (32) reported that transport of TLR4 to the cell surface is impaired in MD-2-deficient cells. A reconstitution experiment also showed that surface labeling of TLR4 is enhanced by expression of MD-2 (24). Therefore, we next analyzed the effect of the association between TLR4 and MD-2 on the cell surface expression of the receptor.
FLAG-tagged TLR4 was transfected with or without MD-2, and the cell surface TLR4 was stained with an anti-FLAG antibody followed by flow cytometric analysis ( Fig. 2A). When FLAG-TLR4 alone was transfected, a histogram with the anti-FLAG antibody staining produced a right shoulder compared with the histogram for mock-transfected cells, indicating that a certain amount of TLR4 reached the cell surface. Coexpression of MD-2 resulted in a further right shift of the histogram, demonstrating enhancement of the surface expression of TLR4. On the other hand, mutants of TLR4, ⌬25-33 and ⌬300 -637, that were unable to associate with MD-2 did not appear on the cell surface even when cotransfected with MD-2. In contrast to these mutants, the other mutants with the 3-amino acid deletion at the amino terminus (⌬25-27) and with the large deletion in the carboxyl-terminal region (⌬321-637) that were capable of associating with MD-2 were expressed on the surface, and this expression was further enhanced by cotransfection of MD-2.
The lack of cell surface expression of the ⌬25-33 and ⌬300 -637 mutants was not due to their degradation in the cells (Fig.  1C). A previous report showed that TLR4 accumulates in the Golgi apparatus in the absence of MD-2 (32). To analyze the intracellular transport of TLR4 from the Golgi apparatus to the plasma membrane, we examined the sensitivity of the transfected TLR4 to Endo H treatment (Fig. 2B). Endo H treatment of the full-length TLR4 without MD-2 resulted in a shift of the main band to a lower molecular weight band, indicating that the full-length TLR4 contained Endo H-sensitive carbohydrate chains. Thus, it is strongly suggested that this molecular species enters the secretory pathway and is glycosylated in the endoplasmic reticulum but does not reach the medial Golgi compartment where the carbohydrate chains on glycoproteins are modified to Endo H-resistant forms. Coexpression of MD-2 increased the amount of the upper minor band, which was Endo H-resistant. Taken together with the results of the flow cytometric analyses, the band is likely to be the mature TLR4 on the cell surface. On the other hand, the ⌬25-33 and ⌬300 -637 mutants, which were unable to bind to MD-2, were entirely sensitive to Endo H in the presence or absence of MD-2. The other mutants, ⌬25-27 and ⌬321-637, also exhibited an Endo H-sensitive band in the absence of MD-2. However, cotransfection of MD-2 with these mutants that could associate with MD-2 resulted in the generation of the Endo H-resistant molecular species. These results strongly suggest that the association of MD-2 with the amino-terminal region of TLR4 is required for the cell surface expression of the receptor.
Both the Amino-terminal and the Carboxyl-terminal Regions of TLR4 Are Required for the LPS Responsiveness-We next measured the signaling capability of these mutants to examine whether the capacity of the TLR4 mutants to bind to MD-2 and to be delivered to the cell surface correlates with their LPS responsiveness. We transfected HEK293 cells with wild type or the mutants of TLR4 together with an NF-B reporter plasmid. The transfected cells were stimulated with LPS, and the resulting reporter activity was measured (Fig. 3). The cells transfected with the full-length TLR4 responded to LPS as shown by elevated NF-B reporter activity. As expected, the mutants that did not bind to MD-2 and therefore did not reach the cell surface (⌬25-33, ⌬300 -637, ⌬307-637, and ⌬314 -637) barely responded to LPS. On the other hand, the amino-terminal deletion mutant (⌬25-27) with MD-2 binding capacity showed the LPS responsiveness. However, the other mutant (⌬321-637) acted as a constitutively active mutant that activated the signaling in the absence of the stimulation and did not exhibit the LPS responsiveness, although it retained the ability to bind to MD-2 (Fig. 1C).
Species-specific Effect of MD-2 on the Surface Expression of TLR4 -Several lines of evidence have suggested that speciesspecific cellular responses against LPS derivatives or Taxol can be attributed to the species of MD-2 (19 -21, 33) or TLR4 (22,23). These observations prompted us to examine whether the activity of MD-2 to enhance the surface expression of TLR4 was also species-specific. We transfected human or mouse TLR4 together with MD-2 from either of the two species and analyzed the cell surface expression of TLR4 by flow cytometric analysis and the Endo H sensitivity assay. As shown in Fig. 4A, the   FIG. 1. The amino-terminal region of TLR4 is essential for the association with MD-2. A, structures of the TLR4 mutants used in this study. Mature TLR4 protein is predicted to begin with amino acid 25 (the amino acid numbering begins with the initiation methionine). SS, signal sequence derived from preprotrypsin; TM, transmembrane region; Cyt, cytoplasmic region. B and C, HEK293 cells were transfected with FLAG-tagged TLR4 or TLR2 (B) or FLAG-tagged TLR4 mutants (C) with or without Myc-tagged MD-2. The cell lysates were subjected to immunoprecipitation followed by immunoblotting using anti-FLAG or anti-Myc antibodies. IP, immunoprecipitation; IB, immunoblotting.
surface expression of the full-length human TLR4 was specifically enhanced by human MD-2 but not by mouse MD-2. The Endo H sensitivity of the human TLR4 was not detectably affected by mouse MD-2, supporting the species specificity (Fig.  4B). In contrast to human TLR4, the cell surface expression of mouse TLR4 was similarly enhanced by either human or mouse MD-2. Consistent with these observations, both human and mouse MD-2 increased the Endo H-resistant form of mouse TLR4. The inability of mouse MD-2 to enhance the cell surface expression of human TLR4 was not attributed to a defect in its association with human TLR4 since mouse MD-2 efficiently associated with both human and mouse TLR4 (data not shown).

A Chimeric Receptor with the Amino-terminal Region of Human TLR4 Collaborates with MD-2 in a Species-specific Manner-
The different behaviors of human and mouse TLR4 stimulated us to analyze the sequences of the TLR4 ectodomain. Careful comparison of the amino acid sequences of human and mouse TLR4 revealed a less conserved region between the two molecules (amino acids 293-343 of human TLR4). To determine the region of TLR4 responsible for the species-specific collaboration with MD-2 in the transport of TLR4, we constructed a chimeric TLR4 molecule consisting of the aminoterminal portion of human TLR4 (amino acids 25-343) and the adjacent region of mouse TLR4 (amino acids 342-835) (Fig. 5A,  NhCm). The chimera contained the region of human TLR4 required for association with MD-2. The cell surface expression of the chimeric TLR4 was preferentially enhanced by human MD-2 rather than mouse MD-2 in flow cytometric analyses (Fig. 5B). As expected, human MD-2, but not mouse MD-2, generated the Endo H-resistant form of TLR4, indicating facilitation of the transport to the cell surface (Fig. 5C). These results strongly suggest that the surface expression of TLR4 on HEK293 cells requires the complex formation between the amino-terminal region of TLR4 and MD-2.
Species-specific Enhancement of TLR4 Surface Expression by MD-2 Is Essential for the LPS Signaling-Next we examined the effect of MD-2 on the activation of TLR4. Since TLR4mediated cellular signaling culminates in activation of the transcription factor NF-B, we measured NF-B activity by reporter analysis. While the cells expressing human TLR4 alone responded weakly to LPS, cotransfection of human MD-2 dramatically enhanced the response as previously reported (Fig. 6) (18). The effect of MD-2 on human TLR4 was speciesspecific since mouse MD-2 did not enhance the LPS response at all. As expected from the transport analyses (Fig. 4), the LPS responsiveness of mouse TLR4 did not show species selectivity since both human and mouse MD-2 acted equally on mouse TLR4 to augment the LPS response (Fig. 6). The LPS signaling by the chimeric TLR4 (NhCm) was augmented by human MD-2 but not by mouse MD-2, similar to the results for its cell surface expression shown in Fig. 5. These results are consistent with the conclusion that the amino-terminal region of TLR4 forms a complex with MD-2, which is required for the intracellular transport of TLR4 to the cell surface, a prerequisite for the LPS signaling.

DISCUSSION
While recent intensive studies have established the functional importance of MD-2 in TLR4-mediated signaling, the interaction between the two proteins has not been fully understood. In the present study, we focused on MD-2 binding to the extracellular domain of TLR4, which is essential for both the cell surface expression of the receptor complex and the LPS responsiveness. Our current analyses demonstrated specific association between TLR4 and MD-2. A previous report described the interaction of MD-2 with TLR2, in addition to TLR4, and its importance in the TLR2-mediated cellular response (34). However, we could not detect any interaction between TLR2 and MD-2. A substantial amount of TLR2 was expressed on the cell surface without MD-2, and the surface expression of TLR2 was not affected by the expression of MD-2 (data not shown). Whether MD-2-like molecules are necessary for the surface expression or the receptor function of other TLRs is currently unknown. If present, HEK293 cells should express a sufficient amount of such a factor(s) for TLR2.
The extracellular domain of TLR4 consists of two subdomains of conserved LRRs connected by an intervening sequence of about 100 amino acid residues (35). The MD-2 bind- ing capacity of the mutant TLR4 with a deletion in the carboxyl-terminal 317 residues of the extracellular domain strongly suggests that the amino-terminal subdomain with eight LRRs is the binding site for MD-2. The intact structure of this region appears to be required for the MD-2 binding since amino-terminal or internal deletions within this region abolished the interaction with MD-2 and the LPS signaling capacity ( Fig. 1C and data not shown). Critical roles for the aminoterminal region of LRR-containing receptors in association with their accessory molecules can be found in the literature. In a report by Mitsuzawa et al. (36), a deletion in the aminoterminal region of TLR2 (amino acids 40 -64) abolished the receptor function for peptidoglycan. Furthermore Miyake et al. (37) reported that RP-105 associated with MD-1 via LRRs in the amino-terminal subdomain of RP-105.
In contrast to the essential requirement for the amino terminus of TLR4, a large deletion in the carboxyl-terminal region of the extracellular domain (⌬321-637) did not severely affect the association with MD-2 ( Figs. 1C and 2). A previous study identified the importance of glycosylation of Asn-526 or Asn-575 of TLR4 for its cell surface expression (38), but the ⌬321-637 mutant lacking this region was expressed on the cell surface. Although the ⌬321-637 mutant bound to MD-2 and was expressed on the cell surface, it was spontaneously activated and did not respond to LPS even in the presence of MD-2 (Fig.  3). Thus, the carboxyl-terminal subdomain of TLR4 is also indispensable for exerting the receptor functions. Several gainof-function mutants of Drosophila Toll have been isolated (39,40). The mutants harbored a mutation of 1 of 4 conserved cysteine residues of the membrane-proximal region or a deletion of this region and were constitutively active in the absence of the ligand. Therefore, the carboxyl-terminal region of the extracellular domain of TLR4 may be necessary to prevent the spontaneous activation that guarantees responsiveness to LPS.
The flow cytometric analyses revealed a defect in the cell surface expression of the mutant TLR4 that did not bind to MD-2 (Fig. 2B). However, the mutant was present in the cells as a form with Endo H-sensitive carbohydrate chains, indicating that it does not reach the medial Golgi cisternae. When wild-type TLR4 was transfected, a small fraction of TLR4 was expressed on the cell surface and responded to LPS in the absence of cotransfected MD-2. This is most likely due to the presence of minute amounts of endogenous MD-2 in HEK293 cells (17). Nevertheless, the majority of the transfected TLR4 was present in an Endo H-sensitive form inside the cells, and coexpression of MD-2 conferred the Endo H resistance on TLR4. Thus, the MD-2 binding is a prerequisite for the exit of TLR4, probably from the endoplasmic reticulum to the Golgi apparatus, which allows the cell surface expression. The function of MD-2 in the transport of TLR4 is reminiscent of the role for ␤ 2 -microglobulin in the exit of the major histocompatibility complex class I polymorphic ␣ chain from the endoplasmic reticulum (41,42). RP-105, another LRR-containing cell surface receptor, also requires association with soluble MD-1 protein for its surface expression (37,43,44).
Schromm et al. (45) reported that addition of soluble MD-2 to the culture medium conferred LPS responsiveness on TLR4expressing cells, suggesting that the role of MD-2 is mainly to form an active receptor complex with TLR4 on the cell surface. The absolute requirement of MD-2 for the cell surface expression of TLR4 suggests that the transfected TLR4 associated with endogenous MD-2 for delivery to the cell surface and that excess amounts of MD-2 augmented the LPS responsiveness. Our studies did not determine the stoichiometry of MD-2 and TLR4 in the complex. Since MD-2 has a tendency to form multimers (18,45), the numbers of MD-2 molecules required for efficient transfer of TLR4 and LPS responsiveness could be different.
It is noteworthy that the activity of MD-2 to transport TLR4 was species-specific (Fig. 4). The inability of mouse MD-2 to transport human TLR4 was not simply due to its failure in binding to TLR4 or its malfunction in human cells since it efficiently bound to human TLR4 and was fully active in the transport of mouse TLR4 in the same cells. Interestingly, mouse TLR4 did not show any preferences for human or mouse MD-2 in the transport. The experiments with the chimeric TLR4 revealed that the amino-terminal region of human TLR4 determines the species-specific effect of MD-2. Thus, the complex of MD-2 and the amino-terminal region of TLR4 should interact with a molecule(s) that allows the transport of the complex to the Golgi apparatus. One such candidate is the endoplasmic reticulum chaperone gp96 whose defect results in intracellular retention of TLRs (46).
Several previous studies have attempted to identify the molecules that directly recognize TLR4 ligands by evaluating the contribution of each molecule in the heterologous expression system to the species-specific responses to LPS derivatives or Taxol (19 -21, 33, 47-50). Some studies have suggested that MD-2 determines the species-specific responses (19 -21), and others have pointed out the importance of TLR4 (22,23). Our observations of the species-specific effects of MD-2 on the cell surface expression of TLR4 suggest that strategies using the heterologous expression system require careful interpretation of the obtained results. Whereas one group reported significant LPS-mediated activation of HEK293 cells transfected with human TLR4 and mouse MD-2 (21), another group did not detect any activation of the same cells under similar conditions (19). The apparent discrepancies might simply reflect differences in the expression level of endogenous MD-2 but not the transfected MD-2.
Since recognition of PAMPs by PRRs initiates the activation of innate immunity, studies on the structure-function relation-FIG. 6. Species-specific requirement of MD-2 for the LPS responsiveness of TLR4. HEK293 cells were transfected with TLR4 with or without Myc-tagged human or mouse MD-2 together with the pELAM1-Luc reporter plasmid and pRL-TK control plasmid. Two days after transfection, the cells were stimulated with LPS (1 g/ml) for 6 h, and the luciferase activity was measured. hTLR4, human TLR4; mTLR4, mouse TLR4; hMD-2, human MD-2; mMD-2, mouse MD-2.
ships of the ectodomain of TLRs should provide answers for the molecular basis for self/nonself discrimination by multicellular organisms. Although the recently determined structure of a complex of glycoprotein Ib␣ and von Willebrand factor (51) suggests that extracellular LRRs provide a surface for proteinprotein interactions, no three-dimensional structures of LRRs of TLRs are currently available. Our current study will help to elucidate the molecular mechanisms of LPS recognition on the cell surface.