Differential Involvement of BB Loops of Toll-IL-1 Resistance (TIR) Domain-Containing Adapter Proteins in TLR4- versus TLR2-Mediated Signal T
http://www.100md.com
免疫学杂志 2005年第13期
Abstract
transmit intracellular signals via the use of specific adapter proteins. We designed a set of "blocking peptides" (BPs) comprised of the 14 aa that correspond to the sequences of the BB loops of the four known Toll-IL-1 resistance (TIR) domain-containing adapter proteins (i.e., MyD88, TIR domain-containing adapter inducing IFN- (TRIF), TRIF-related adapter molecule (TRAM), and TIR-domain containing adapter protein (TIRAP)) linked to the cell-penetrating segment of the antennapedia homeodomain. LPS (TLR4)-mediated gene expression, as well as MAPK and transcription factor activation associated with both MyD88-dependent and -independent signaling pathways, were disrupted by all four BPs (TRAM MyD88 > TRIF > TIRAP), but not by a control peptide. In contrast, none of the BPs inhibited TLR2-mediated activation of MAPKs. Only the MyD88 BP significantly blocked Pam3Cys-induced IL-1 mRNA; however, the inhibitory effect was much less than observed for LPS. Our data suggest that the interactions required for a fully functional TLR4 signaling "platform" are disrupted by these BPs, and that the adapter BB loops may serve distinct roles in TLR4 and TLR2 signalosome assembly.
Introduction
Toll-like receptors are a family of signaling molecules that act as sensors for recognition of diverse pathogen-associated molecular patterns. All TLRs share a similar domain architecture: an extracellular region that contains multiple leucine-rich repeat regions (1) that function in recognition of pathogen-associated molecular patterns and coreceptors (2); a transmembrane domain comprised of one membrane-spanning -helix; and an intracellular "Toll-IL-1 resistance" (TIR) 4 domain located within the C terminus (1).
In addition to TLRs, a group of TIR domain-containing proteins that lack a transmembrane domain has been identified. MyD88 was the first TIR domain-containing protein that was shown to serve as an adapter for IL-1R, and later, for TLR signaling (3). With rapid progress in sequencing and gene annotation, this family grew quickly and now includes five members, four of which are involved in TLR4 signaling: MyD88; TIR-domain containing adapter protein (TIRAP) (4), also known as Mal (5); TIR domain-containing adapter inducing IFN- (TRIF) (6), also referred as TIR domain-containing adapter molecule-1 (7); and TRIF-related adapter molecule (TRAM) (8), also named as TIR-containing protein (9) or TIR domain-containing adapter molecule-2 (10).
MyD88 and TRIF have a more complex domain architecture than TIRAP and TRAM. In addition to the TIR domain, MyD88 has a mid-region and an N-terminal death domain that enable interaction with IL-1R-associated kinase 4 (IRAK-4) and IRAK-1, respectively, serine/threonine kinases that signal downstream of IL-1R (11) and TLRs (3), leading to NF-B activation (reviewed in Ref. 12). TRIF is the largest member of adapter family, with the TIR domain located in the middle of its sequence. The N terminus of TRIF is necessary for activation of NF-B and IFN regulatory factor (IRF)-3 (6), a transcription factor responsible for induction of IFN- and other genes. IB kinase (IKK)- and TRAF family member-associated NF-B binding kinase (TANK)-binding kinase-1 (TBK-1), act downstream of TRIF to phosphorylate IRF-3 (13). Recruitment of IKK- and TBK-1 to TRIF does not require TIR domain of the protein (8), and overexpression of a TRIF mutant that lacks the C terminus still activates NF-B and IFN- reporter constructs (6, 7).
TIRAP and TRAM are the smallest adapters involved in TLR signaling. The N terminus of TRAM has a myristoylation site, the mutation of which alters its normal membrane localization (14). Both TIRAP and TRAM have been found to associate constitutively with TLR4 (5, 8). Current models of TLR signaling ascribe different roles to the adapters (10, 14). TIRAP and TRAM have been suggested to serve as "platform-forming" components responsible for recruitment of the larger adapters, MyD88 and TRIF, respectively, that in turn, recruit downstream effector molecules (e.g., IRAK-1 and IRAK-4 to MyD88 and TBK-1, TRAF6, and IKK- to TRIF via non-TIR domains). Distinct combinations of adapters involved in a signaling platform differ among TLRs and have been postulated as the basis for the observed specificity of gene subsets induced by different TLR agonists (see Refs. 14 and 15 for review).
The structure of TLR TIR domains was first predicted by Rock et al. (1). Analysis of sequences of five human TLRs and their diverse homologues led the authors to conclude that TIR domains are composed of five -strands alternating with five -helices. Resolution of crystal structures of human TLR1 and TLR2 confirmed these theoretical findings and suggested that a functionally important proline-glycine combination that is highly conserved among TIR domains is located in the loop connecting the second -strand with the second helix, the "BB loop," using the terminology proposed by Xu et al. (16).
The importance of BB loops in TLR signaling derives from observations that the function of TIR domain-containing proteins is highly susceptible to mutations in this region. A naturally occurring point mutation (P712H) in murine TLR4, which leads to complete unresponsiveness of the C3H/HeJ mouse to LPS (17, 18), lies in this region. Similarly, replacement of Pro681 with His in human TLR2 was reported to disrupt signal transduction induced by Gram-positive bacteria and essentially abolishes MyD88 recruitment (16); however, this finding was not confirmed by Dunne et al. (19). Mutations of residues in the vicinity of the conserved BB loop proline also led to the decreased activity of TLR2 (16) and TLR4 (20). Many TIR-domain containing proteins that bear mutations homologous to the mutation in C3H/HeJ mice exert dominant-negative effects on TLR signaling (e.g., TLR4 (21), TLR2 (22), TIRAP (4), and TRAM (8)). Overexpression of the P200H mutation of MyD88, however, fails to prevent induction of NF-B (4) or its binding to human TLR4 (19). Similarly, in a yeast two-hybrid system, interaction of TLR4 with TRAM is not disrupted by mutation of the TRAM cysteine 117 in its BB loop (19).
Cell-penetrating cationic peptides are potent and efficient tools for delivery into the intracellular space of diverse substances that normally would not penetrate plasma membranes (23). Cargoes that can be carried into cells vary widely in size and chemical nature and include oligonucleotides and proteins. After the initial observation that aa 47–57 of HIV-1 TAT protein are crucial for the ability of this protein to enter cells (24, 25), functionally similar sequences were identified in the homeodomain of the Drosophila transcription factor, antennapedia (26), and the herpes-simplex-virus-1 DNA-binding protein, VP22 (27). The mechanism by which cell-penetrating cationic peptides enter cells is still controversial (28).
For this study, we designed a set of "blocking peptides" (BPs) comprised of the translocating sequence of the antennapedia homeodomain, in tandem with 14-aa sequences corresponding to the BB loops of MyD88, TRAM, TIRAP, and TRIF. This approach was first used by Horng et al. (4) in their original studies on the identification of TIRAP in TLR4 signaling. We tested these four constructs and a control peptide (CP) in primary murine macrophages for the ability to interfere with LPS-induced signal transduction. All four BPs blocked TLR4-mediated gene expression, MAPK, and transacting factor activation, but failed to block TLR2-mediated activation of MAPKs. Only the MyD88 BP inhibited TLR2-mediated induction of IL-1 mRNA; however, its inhibitory effect was considerably weaker than its effect on TLR4-mediated IL-1 gene expression. Collectively, our findings suggest that these BPs interfere with protein-protein interactions involved in the assembly and/or stabilization of the TLR4 "platform" and support the hypothesis that surfaces on MyD88 and TIRAP apart from the BB loop enable their interaction with TLR2.
Materials and Methods
Animals
C3H/OuJ mice were purchased from The Jackson Laboratory. Thioglycolate-elicited peritoneal macrophages were harvested and cultured in the presence of Escherichia coli K235 LPS or (S-[2,3-bis(palmitoyloxy)-(2-RS)-propyl]-N-palmitoyl-(R)-Cys-Ser-Lys4-OH) (Pam3Cys) as described previously (29).
Design of peptides
The peptides were composed of 16-aa carrier sequence from Drosophila homeodomain protein (aa 366–381, GenBank ID number (GI): 4389425), the C terminus of which was synthesized in tandem with the N terminus of the 14 aa that surround the conserved P(C)G sequence in BB loops of murine TIR domain-containing adapter proteins: TRAM (GI: 27734184), MyD88 (GI: 6754772), TRIF (GI: 33859797), and TIRAP (GI: 16905131) (sequences provided in Table I). A CP was a scrambled amino acid sequence shown by Basic Local Alignment Search Tool (BLAST) to not be homologous to known proteins. Peptides were synthesized and purified by HPLC by Biosynthesis. Purity was confirmed by mass spectrometry. Ten-millimolar stocks were dissolved in 25% DMSO and kept frozen at –80°C. Forty micromolar was the highest concentration used in these experiments due to limited solubility of BPs in the cell culture medium.
SDS-PAGE and Western blot analysis were performed as described previously (30) with minor modifications. Abs for detection of activated MAPKs (pERK, pJNK, and pP38) were purchased from Cell Signaling Technology. IRF-1 was detected using a rabbit polyclonal Ab from Santa Cruz Biotechnology. Ab for detection of tyrosine 701 phosphorylated STAT1 was purchased from Zymed Laboratories.
Native PAGE for detection of IRF-3 activation was performed with minor modifications (31). A total of 1 x 107 thioglycolate-elicited cells were plated into a 6-cm dish and washed with PBS to remove nonadherent cells. Primary macrophages were stimulated with 100 ng/ml LPS for 1.5 h and lysed in buffer containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 1 mM sodium orthovanadate, and 1x Complete protease inhibitor mixture (Roche). Lysates were mixed with native lysis buffer (Bio-Rad) supplemented with 1% deoxycholate. Fifteen to 20 μg of protein per lane was electrophoresed on a nonreducing gel and immunoblotted. IRF-3 dimers and monomers were detected using rabbit anti-IRF-3 Ab (Zymed Laboratories) by Western analysis.
Results
BB loop peptides block LPS-induced gene expression in primary macrophages, but do not distinguish between MyD88-dependent and -independent pathways
Based on the observation that macrophages from MyD88 knockout mice do not synthesize TNF- or IL-6 in response to LPS, while they are still capable of inducing IFN- and IFN--inducible genes (32, 33), two different signaling pathways propagating from TLR4 were defined (reviewed in Ref. 15). The "MyD88-independent" pathway of TLR4 signaling uses TRAM and TRIF to activate IRF-3 that, in turn, leads to induction of IFN- and IFN--inducible genes, while the "MyD88-dependent" pathway leads to induction of genes that encode IL-1, TNF-, and IL-6 that do not depend on IRF-3 for their expression. Mice with targeted mutations in MyD88 (32, 33) or TIRAP (34) retain the capacity to signal through the MyD88-independent pathway. In contrast, the effect of targeted mutations in TRIF or TRAM affects both arms of the TLR4 signaling pathway: macrophages derived from TRIF and TRAM knockouts not only exhibited a complete loss of MyD88-independent signaling, but also exhibited a severe loss of signaling associated with the MyD88-dependent pathway (e.g., TNF-, IL-6, and IL-12 p40 secretion is essentially eliminated, although there is some retention of early NF-B translocation) (35, 36). Therefore, the first question we sought to address experimentally is whether interactions of the BB loops of the TIR domains of these adapters affect protein interactions leading to gene expression. To this end, we designed cell permeable peptides whose sequences correspond to the BB loops of each adapter molecule and measured the ability of each BP to interfere with the induction of immediate response genes induced via MyD88-dependent and -independent pathways. Fig. 1A shows the effect of the BP treatment of macrophage cultures on the induction of genes that encode two MyD88-dependent cytokines, IL-1 and MIP-1 (37, 38), and two MyD88-independent cytokines, IFN- and RANTES (32, 33, 37), 1 h after LPS stimulation of primary macrophage cultures. All BB loop-containing peptides exerted an inhibitory effect on LPS-induced expression of MyD88-dependent IL-1 and MIP-1, as well as MyD88-independent IFN- and RANTES mRNA, as measured by real-time PCR (Fig. 1A). In general, the inhibitory effect of the TRAM BP was strongest, with MyD88 BP being comparable or only slightly less inhibitory; the TIRAP BP was consistently weakest among the four tested. Fig. 1B demonstrates the effect of BB loop peptides on IL-1 and IL-6 protein expression. The BPs, but not the CP, also blocked secretion of IL-1 and IL-6. The concentration of BP used in Fig. 1 and all other experiments was 40 μM based on optimal inhibitory activity of the two most active BPs, TRAM BP and MyD88 BP, against an MyD88-dependent gene (e.g., IL-1) and an MyD88-independent gene (e.g., IFN-) (Fig. 1C). Fig. 1C also shows that BP-mediated inhibition is seen over a very narrow concentration range (i.e., between 10 and 40 μM).
Effect of BB loop BPs on activation of STAT-1 by LPS
STAT-1 propagates signal transduction emanating from both type I and type II IFN receptors (reviewed in Ref. 39). Two phosphate acceptor sites are involved in activation of STAT-1. Although phosphorylation of Tyr701 is critical for homo- or heterodimerization and nuclear translocation of STAT-1, phosphorylation of Ser727 modulates its transcriptional activity (reviewed in Ref. 39). We previously reported that tyrosine phosphorylation of STAT1 by LPS depends on MyD88-independent induction of IFN- and the subsequent autocrine activation of type I IFN receptors (33).
LPS-induced Tyr701 phosphorylation of STAT-1 was strongly inhibited by the TRAM and TRIF BPs (Fig. 2B). Consistent with its strong ability to block IFN- gene expression (Fig. 1), the MyD88 BP was also highly effective at inhibiting Tyr701 phosphorylation of STAT-1 (Fig. 2B). However, treatment of macrophages with the TIRAP BP did not completely block tyrosine phosphorylation of STAT-1 in response to LPS, although it was significantly diminished compared with the level seen in the presence of the CP (Fig. 2B). LPS-induced Ser727 phosphorylation of STAT-1 was similarly blocked by the BPs (data not shown). The BPs failed to inhibit IFN--induced Tyr701 phosphorylation of STAT-1 (data not shown), consistent with the observation that the type I IFN receptor does not depend on any of these adapters to signal (reviewed in Ref. 39).
Effect of BB loop BPs on activation of IRF-1 and IRF-3
IRFs were first identified as important transcriptional mediators of the antiviral immune response (reviewed in Ref. 40). Later, it was recognized that this family of transcription factors plays an essential role in signal transduction mediated by different TLRs. IRF-1 is regulated transcriptionally by signaling from different TLRs, as well as by cytokines induced by TLR engagement (e.g., IL-1, TNF, and IFN-) (41). In contrast, IRF-3 is constitutively expressed in many cell types (42). Upon stimulation through TLR4 or TLR3, IRF-3 is phosphorylated at multiple sites in its C terminus, and this modification enables it to dimerize and translocate to the nucleus where it targets IRF-3-responsive genes (32, 43). Using MyD88 knockout mice, Kawai et al. (32) demonstrated that LPS activates IRF-3 via the MyD88-independent pathway, and this, in turn, leads to the induction of IFN-, the signature cytokine gene for the MyD88-independent pathway (33). Pretreatment of cells with TRAM, TRIF, or MyD88 BPs significantly inhibited LPS-induced dimerization of IRF-3 (Fig. 3). The TIRAP BP also decreased the amount of detectable IRF-3 dimer; however, inhibition was not as complete as with the other BPs.
IRF-1 is an immediate response gene, quickly induced by signaling from different TLRs, which contributes to the induction of other important TLR-inducible genes (e.g., inducible NO synthase). Importantly, MyD88 knockout macrophages are capable of inducing IRF-1 mRNA (35). Stimulation of macrophages by LPS for 1.5 h after pretreatment of cells with medium only or CP resulted in a strong up-regulation of IRF-1 protein (Fig. 3). As was observed for IRF-3, TRAM, and MyD88 peptides were the strongest inhibitors with respect to both IRF-3 and IRF-1. The TRIF BP again showed intermediate efficacy, while the TIRAP peptide was the weakest of the BPs.
Limited capacity of adapter BB loop inhibitory peptides to block TLR2-induced signaling
Apart from the use of a CP, we sought to analyze further the specificity of the inhibitory effects of BB loop peptides on TLR signaling. Therefore, we also tested the effect of the peptides on signal transduction elicited from TLR2. TLR2 was chosen because it shares MyD88-dependent signaling pathway with TLR4, and TLR2-mediated responses are depreciated in both MyD88 and TIRAP-deficient cells (34, 44); however, TLR2-mediated signaling does not appear to use MyD88-independent arm for signal transduction (reviewed in Ref. 15). Stimulation of MyD88 knockout macrophages with the Pam3Cys, a TLR2 agonist, does not activate MAPKs (Ref. 44 and our unpublished observation), whereas LPS-stimulated MyD88 knockout macrophages exhibit MAPKs and NF-B activation, albeit with delayed kinetics (38). Fig. 4A demonstrates that the BPs differentially affect phosphorylation of ERK and JNK in response to LPS vs Pam3Cys. Surprisingly, under conditions where LPS-induced MAPK phosphorylation was strongly blocked by all BPs, none inhibited Pam3Cys-induced phosphorylation of MAPKs.
The effect of BPs on Pam3Cys-induced expression of IL-1 mRNA is shown in Fig. 4B. LPS and Pam3Cys induced comparable levels of IL-1 mRNA, which were not significantly affected by treatment of cells with CP (data not shown). Similar to their effect on activation of MAPKs, BPs were much weaker inhibitors with respect to Pam3Cys- vs LPS-induced expression of IL-1 mRNA. Only the MyD88 BP exerted a statistically significant inhibitory effect (Fig. 4B) and this effect was considerably less than that seen in the case of LPS stimulation; in LPS-stimulated cells, MyD88 BP reduced IL-1 mRNA expression to <5% of the level detected in the presence of the CP, while in Pam3Cys-stimulated cells, IL-1 gene expression was reduced to 25% of control levels. Taken collectively, these data demonstrate that TLR4-induced signaling is much more susceptible to inhibition by the adapter BB loop BPs than is TLR2-mediated signaling.
Discussion
This study represents an attempt to use penetrating oligopeptides containing homologous sequences that presumably interfere with the protein-protein interaction surfaces formed between the BB loops of four adapter proteins with their target protein(s) to investigate mechanisms of signal transduction emanating from TLR4 and TLR2. The impetus for generating these four adapter BPs stemmed from contradictory results obtained using the TIRAP BP and a TIRAP dominant-negative mutant construct vs macrophages from TIRAP knockout mice. The TIRAP BP was first used by Horng et al. (4) as a complementary tool to confirm the involvement of this adapter in LPS-signaling and was reported to inhibit TLR4-, but not TLR9-mediated signaling (4, 33). However, using this same TIRAP BP, we found that it also inhibited LPS-induced IFN- gene expression and IFN--luciferase reporter activity (33), while later studies using macrophages from TIRAP knockout mice found that induction of IRF-3 and IRF-3/STAT1-dependent genes by LPS was intact (34). Therefore, we sought to carry out a systematic comparison of the effects of BPs based on the homologous sequences of the other known TLR4 adapters to gain insights into the role of adapter BB loops for protein interactions required for TLR4 signaling. Although the mechanism by which cell-penetrating, BB loop peptides inhibit TLR signaling has not been formally demonstrated, it has been postulated that such peptides, once inside the cell, can occupy the docking site of the cognate adapter on its target and prevent binding of the native adapter. This, in turn, would be predicted to disrupt the formation of a functional signaling platform (14, 45).
Analysis of the effects of BPs on the expression of early inducible genes revealed an interesting and completely unexpected pattern; the BPs showed little specificity with respect to genes induced by LPS via MyD88-dependent or -independent pathways. For example, the MyD88 peptide did not preferentially block IL-1 and MIP-1 genes, which, according to data derived from knockout mice, require MyD88 for induction (37, 38). Conversely, TRAM BP was equally effective in inhibiting both MyD88-dependent and -independent genes, in contrast to the conclusion drawn from experiments with TRAM knockout mice (35). The fact that all four BPs block TLR4 signaling through either the MyD88-dependent or -independent pathways suggests strongly that the stability of the TLR4 signaling platform is indeed disrupted or fails to assemble correctly as a consequence of BP-target protein interactions.
In this regard, the disruption of TLR4 signaling by BPs is more reminiscent of a "receptor-knockout" phenotype, where both branches of the TLR4 signaling pathway are inhibited, rather than an "adapter-knockout" phenotype characterized by expression of distinct subsets of genes. The observation that BPs are capable of inhibiting various manifestations of LPS signaling suggests that they act before the signal bifurcates into separate pathways, most likely at the level of the receptor.
By examining events upstream of gene expression and cytokine secretion, it was possible to identify differences in the relative inhibitory capacities of BPs on TLR4 signaling. Thus, Figs. 2 and 3 show that MyD88 TRAM > TRIF > TIRAP BPs with respect to inhibition of MAPK, STAT-1, or IRF activation. Various reasons can be invoked to explain the observed differences in the relative efficiency of individual BPs to inhibit LPS signaling. Differences in peptide permeability, intracellular peptide turnover, existence of stable, "inefficient" conformations that do not bind as strongly to the intended molecular target, or, simply, nonspecific binding to unrelated sites could all potentially contribute to our observations. Importantly, the efficiency of the peptides failed to correlate with their general physical-chemical parameters that are summarized in Table I.
In addition to the surprising result that all four BPs inhibited both MyD88-dependent and -independent branches of TLR4-mediated signaling, the observations that the BPs were much weaker inhibitors of TLR2-mediated signaling was equally unexpected. The ability of BB loop peptides to disrupt TLR4 signaling again supports the hypothesis that these regions of the adapters are critical for interaction with TLR4. The fact that the BPs did not affect TLR2-induced activation of JNK and ERK and exerted a much weaker effect on TLR2-induced IL-1 gene expression implies that the adapter BB loops do not belong to the adapter surfaces that interact with TLR2. It is possible that the MyD88 BP partially interferes with the MyD88-TLR2 interaction surface, thus resulting in a lesser degree of inhibition. By this same logic, it would be predicted that TIRAP does not use its BB loop in the formation of an active TLR2 signaling complex. These findings suggest that the architecture of receptor-adapter complexes differs between TLR2 and TLR4. The idea that adapters shared by different TLRs can interact with individual receptors in fundamentally different ways supports the existence of asymmetrical models of TIR-TIR complex formation. Interestingly, recent computer modeling performed by Dunne et al. (19) predicted that TLR4 and TLR2 bind TIRAP or MyD88 through different nonoverlapping sites further support such a model.
Disclosures
The authors have no financial conflict of interest.
Footnotes
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.
1 This work was supported by National Institutes of Health Grants AI-18797 and AI-44936 (to S.N.V.) and AI-57490 (to M.J.F.).
2 M.J.F. and S.N.V. contributed equally to this work.
3 Address correspondence and reprint requests to Dr. Stefanie N. Vogel, Department of Microbiology and Immunology, University of Maryland, Baltimore, 655 West Baltimore Street, 13-009, Baltimore, MD 21201. E-mail address: svogel@som.umaryland.edu
4 Abbreviations used in this paper: TIR, Toll-IL-1 resistance; BP, blocking peptide; CP, control peptide; TIRAP, TIR-domain containing adapter protein; TRIF, TIR domain-containing adapter inducing IFN-; TRAM, TRIF-related adapter molecule; TBK-1, TRAF family member-associated NF-B binding kinase (TANK)-binding kinase-1; IKK-, IB kinase ; IRAK-1/4, IL-1R-associated kinase 1/4; Pam3Cys, (S-[2,3-bis(palmitoyloxy)-(2-RS)-propyl]-N-palmitoyl-(R)-Cys-Ser-Lys4-OH); HPRT, hypoxanthine phosphoribosyltransferase; IRF, IFN regulatory facto; GI, GenBank ID number.
Received for publication January 10, 2005. Accepted for publication April 16, 2005.
References
ock, F. L., G. Hardiman, J. C. Timans, R. A. Kastelein, J. F. Bazan. 1998. A family of human receptors structurally related to Drosophila Toll. Proc. Natl. Acad. Sci. USA 95: 588-593.
Kobe, B., A. V. Kajava. 2001. The leucine-rich repeat as a protein recognition motif. Curr. Opin. Struct. Biol. 11: 725-732.
Medzhitov, R., P. Preston-Hurlburt, E. Kopp, A. Stadlen, C. Chen, S. Ghosh, C. A. Janeway. 1998. MyD88 is an adaptor protein in the hToll/IL-1 receptor family signaling pathways. Mol. Cell. 2: 253-258.
Horng, T., G. M. Barton, R. Medzhidov. 2001. TIRAP: an adapter molecule in the Toll signaling pathway. Nat. Immunol. 2: 825-826.
Fitzgerald, K. A., E. M. Palsson-McDermott, A. G. Bowie, C. A. Jefferies, A. S. Mansell, G. Brady, E. Brint, A. Dunne, P. Gray, M. T. Harle, et al 2001. Mal (MyD88-adapter-like) is required for Toll-like receptor-4 signal transduction. Nature 413: 78-83.
Yamamoto, M., S. Sato, K. Mori, K. Hoshino, O. Takeuchi, K. Takeda, S. Akira. 2002. Cutting edge: a novel Toll/IL-1 receptor domain-containing adapter that preferentially activates the IFN- promoter in the Toll-like receptor signaling. J. Immunol. 169: 6668-6672.
Oshiumi, H., M. Matsumoto, K. Funami, T. Akazawa, T. Seya. 2003. TICAM-1, an adaptor molecule that participates in Toll-like receptor 3-mediated interferon- induction. Nat. Immunol. 4: 161-167.(Vladimir U. Toshchakov*, )
transmit intracellular signals via the use of specific adapter proteins. We designed a set of "blocking peptides" (BPs) comprised of the 14 aa that correspond to the sequences of the BB loops of the four known Toll-IL-1 resistance (TIR) domain-containing adapter proteins (i.e., MyD88, TIR domain-containing adapter inducing IFN- (TRIF), TRIF-related adapter molecule (TRAM), and TIR-domain containing adapter protein (TIRAP)) linked to the cell-penetrating segment of the antennapedia homeodomain. LPS (TLR4)-mediated gene expression, as well as MAPK and transcription factor activation associated with both MyD88-dependent and -independent signaling pathways, were disrupted by all four BPs (TRAM MyD88 > TRIF > TIRAP), but not by a control peptide. In contrast, none of the BPs inhibited TLR2-mediated activation of MAPKs. Only the MyD88 BP significantly blocked Pam3Cys-induced IL-1 mRNA; however, the inhibitory effect was much less than observed for LPS. Our data suggest that the interactions required for a fully functional TLR4 signaling "platform" are disrupted by these BPs, and that the adapter BB loops may serve distinct roles in TLR4 and TLR2 signalosome assembly.
Introduction
Toll-like receptors are a family of signaling molecules that act as sensors for recognition of diverse pathogen-associated molecular patterns. All TLRs share a similar domain architecture: an extracellular region that contains multiple leucine-rich repeat regions (1) that function in recognition of pathogen-associated molecular patterns and coreceptors (2); a transmembrane domain comprised of one membrane-spanning -helix; and an intracellular "Toll-IL-1 resistance" (TIR) 4 domain located within the C terminus (1).
In addition to TLRs, a group of TIR domain-containing proteins that lack a transmembrane domain has been identified. MyD88 was the first TIR domain-containing protein that was shown to serve as an adapter for IL-1R, and later, for TLR signaling (3). With rapid progress in sequencing and gene annotation, this family grew quickly and now includes five members, four of which are involved in TLR4 signaling: MyD88; TIR-domain containing adapter protein (TIRAP) (4), also known as Mal (5); TIR domain-containing adapter inducing IFN- (TRIF) (6), also referred as TIR domain-containing adapter molecule-1 (7); and TRIF-related adapter molecule (TRAM) (8), also named as TIR-containing protein (9) or TIR domain-containing adapter molecule-2 (10).
MyD88 and TRIF have a more complex domain architecture than TIRAP and TRAM. In addition to the TIR domain, MyD88 has a mid-region and an N-terminal death domain that enable interaction with IL-1R-associated kinase 4 (IRAK-4) and IRAK-1, respectively, serine/threonine kinases that signal downstream of IL-1R (11) and TLRs (3), leading to NF-B activation (reviewed in Ref. 12). TRIF is the largest member of adapter family, with the TIR domain located in the middle of its sequence. The N terminus of TRIF is necessary for activation of NF-B and IFN regulatory factor (IRF)-3 (6), a transcription factor responsible for induction of IFN- and other genes. IB kinase (IKK)- and TRAF family member-associated NF-B binding kinase (TANK)-binding kinase-1 (TBK-1), act downstream of TRIF to phosphorylate IRF-3 (13). Recruitment of IKK- and TBK-1 to TRIF does not require TIR domain of the protein (8), and overexpression of a TRIF mutant that lacks the C terminus still activates NF-B and IFN- reporter constructs (6, 7).
TIRAP and TRAM are the smallest adapters involved in TLR signaling. The N terminus of TRAM has a myristoylation site, the mutation of which alters its normal membrane localization (14). Both TIRAP and TRAM have been found to associate constitutively with TLR4 (5, 8). Current models of TLR signaling ascribe different roles to the adapters (10, 14). TIRAP and TRAM have been suggested to serve as "platform-forming" components responsible for recruitment of the larger adapters, MyD88 and TRIF, respectively, that in turn, recruit downstream effector molecules (e.g., IRAK-1 and IRAK-4 to MyD88 and TBK-1, TRAF6, and IKK- to TRIF via non-TIR domains). Distinct combinations of adapters involved in a signaling platform differ among TLRs and have been postulated as the basis for the observed specificity of gene subsets induced by different TLR agonists (see Refs. 14 and 15 for review).
The structure of TLR TIR domains was first predicted by Rock et al. (1). Analysis of sequences of five human TLRs and their diverse homologues led the authors to conclude that TIR domains are composed of five -strands alternating with five -helices. Resolution of crystal structures of human TLR1 and TLR2 confirmed these theoretical findings and suggested that a functionally important proline-glycine combination that is highly conserved among TIR domains is located in the loop connecting the second -strand with the second helix, the "BB loop," using the terminology proposed by Xu et al. (16).
The importance of BB loops in TLR signaling derives from observations that the function of TIR domain-containing proteins is highly susceptible to mutations in this region. A naturally occurring point mutation (P712H) in murine TLR4, which leads to complete unresponsiveness of the C3H/HeJ mouse to LPS (17, 18), lies in this region. Similarly, replacement of Pro681 with His in human TLR2 was reported to disrupt signal transduction induced by Gram-positive bacteria and essentially abolishes MyD88 recruitment (16); however, this finding was not confirmed by Dunne et al. (19). Mutations of residues in the vicinity of the conserved BB loop proline also led to the decreased activity of TLR2 (16) and TLR4 (20). Many TIR-domain containing proteins that bear mutations homologous to the mutation in C3H/HeJ mice exert dominant-negative effects on TLR signaling (e.g., TLR4 (21), TLR2 (22), TIRAP (4), and TRAM (8)). Overexpression of the P200H mutation of MyD88, however, fails to prevent induction of NF-B (4) or its binding to human TLR4 (19). Similarly, in a yeast two-hybrid system, interaction of TLR4 with TRAM is not disrupted by mutation of the TRAM cysteine 117 in its BB loop (19).
Cell-penetrating cationic peptides are potent and efficient tools for delivery into the intracellular space of diverse substances that normally would not penetrate plasma membranes (23). Cargoes that can be carried into cells vary widely in size and chemical nature and include oligonucleotides and proteins. After the initial observation that aa 47–57 of HIV-1 TAT protein are crucial for the ability of this protein to enter cells (24, 25), functionally similar sequences were identified in the homeodomain of the Drosophila transcription factor, antennapedia (26), and the herpes-simplex-virus-1 DNA-binding protein, VP22 (27). The mechanism by which cell-penetrating cationic peptides enter cells is still controversial (28).
For this study, we designed a set of "blocking peptides" (BPs) comprised of the translocating sequence of the antennapedia homeodomain, in tandem with 14-aa sequences corresponding to the BB loops of MyD88, TRAM, TIRAP, and TRIF. This approach was first used by Horng et al. (4) in their original studies on the identification of TIRAP in TLR4 signaling. We tested these four constructs and a control peptide (CP) in primary murine macrophages for the ability to interfere with LPS-induced signal transduction. All four BPs blocked TLR4-mediated gene expression, MAPK, and transacting factor activation, but failed to block TLR2-mediated activation of MAPKs. Only the MyD88 BP inhibited TLR2-mediated induction of IL-1 mRNA; however, its inhibitory effect was considerably weaker than its effect on TLR4-mediated IL-1 gene expression. Collectively, our findings suggest that these BPs interfere with protein-protein interactions involved in the assembly and/or stabilization of the TLR4 "platform" and support the hypothesis that surfaces on MyD88 and TIRAP apart from the BB loop enable their interaction with TLR2.
Materials and Methods
Animals
C3H/OuJ mice were purchased from The Jackson Laboratory. Thioglycolate-elicited peritoneal macrophages were harvested and cultured in the presence of Escherichia coli K235 LPS or (S-[2,3-bis(palmitoyloxy)-(2-RS)-propyl]-N-palmitoyl-(R)-Cys-Ser-Lys4-OH) (Pam3Cys) as described previously (29).
Design of peptides
The peptides were composed of 16-aa carrier sequence from Drosophila homeodomain protein (aa 366–381, GenBank ID number (GI): 4389425), the C terminus of which was synthesized in tandem with the N terminus of the 14 aa that surround the conserved P(C)G sequence in BB loops of murine TIR domain-containing adapter proteins: TRAM (GI: 27734184), MyD88 (GI: 6754772), TRIF (GI: 33859797), and TIRAP (GI: 16905131) (sequences provided in Table I). A CP was a scrambled amino acid sequence shown by Basic Local Alignment Search Tool (BLAST) to not be homologous to known proteins. Peptides were synthesized and purified by HPLC by Biosynthesis. Purity was confirmed by mass spectrometry. Ten-millimolar stocks were dissolved in 25% DMSO and kept frozen at –80°C. Forty micromolar was the highest concentration used in these experiments due to limited solubility of BPs in the cell culture medium.
SDS-PAGE and Western blot analysis were performed as described previously (30) with minor modifications. Abs for detection of activated MAPKs (pERK, pJNK, and pP38) were purchased from Cell Signaling Technology. IRF-1 was detected using a rabbit polyclonal Ab from Santa Cruz Biotechnology. Ab for detection of tyrosine 701 phosphorylated STAT1 was purchased from Zymed Laboratories.
Native PAGE for detection of IRF-3 activation was performed with minor modifications (31). A total of 1 x 107 thioglycolate-elicited cells were plated into a 6-cm dish and washed with PBS to remove nonadherent cells. Primary macrophages were stimulated with 100 ng/ml LPS for 1.5 h and lysed in buffer containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 1 mM sodium orthovanadate, and 1x Complete protease inhibitor mixture (Roche). Lysates were mixed with native lysis buffer (Bio-Rad) supplemented with 1% deoxycholate. Fifteen to 20 μg of protein per lane was electrophoresed on a nonreducing gel and immunoblotted. IRF-3 dimers and monomers were detected using rabbit anti-IRF-3 Ab (Zymed Laboratories) by Western analysis.
Results
BB loop peptides block LPS-induced gene expression in primary macrophages, but do not distinguish between MyD88-dependent and -independent pathways
Based on the observation that macrophages from MyD88 knockout mice do not synthesize TNF- or IL-6 in response to LPS, while they are still capable of inducing IFN- and IFN--inducible genes (32, 33), two different signaling pathways propagating from TLR4 were defined (reviewed in Ref. 15). The "MyD88-independent" pathway of TLR4 signaling uses TRAM and TRIF to activate IRF-3 that, in turn, leads to induction of IFN- and IFN--inducible genes, while the "MyD88-dependent" pathway leads to induction of genes that encode IL-1, TNF-, and IL-6 that do not depend on IRF-3 for their expression. Mice with targeted mutations in MyD88 (32, 33) or TIRAP (34) retain the capacity to signal through the MyD88-independent pathway. In contrast, the effect of targeted mutations in TRIF or TRAM affects both arms of the TLR4 signaling pathway: macrophages derived from TRIF and TRAM knockouts not only exhibited a complete loss of MyD88-independent signaling, but also exhibited a severe loss of signaling associated with the MyD88-dependent pathway (e.g., TNF-, IL-6, and IL-12 p40 secretion is essentially eliminated, although there is some retention of early NF-B translocation) (35, 36). Therefore, the first question we sought to address experimentally is whether interactions of the BB loops of the TIR domains of these adapters affect protein interactions leading to gene expression. To this end, we designed cell permeable peptides whose sequences correspond to the BB loops of each adapter molecule and measured the ability of each BP to interfere with the induction of immediate response genes induced via MyD88-dependent and -independent pathways. Fig. 1A shows the effect of the BP treatment of macrophage cultures on the induction of genes that encode two MyD88-dependent cytokines, IL-1 and MIP-1 (37, 38), and two MyD88-independent cytokines, IFN- and RANTES (32, 33, 37), 1 h after LPS stimulation of primary macrophage cultures. All BB loop-containing peptides exerted an inhibitory effect on LPS-induced expression of MyD88-dependent IL-1 and MIP-1, as well as MyD88-independent IFN- and RANTES mRNA, as measured by real-time PCR (Fig. 1A). In general, the inhibitory effect of the TRAM BP was strongest, with MyD88 BP being comparable or only slightly less inhibitory; the TIRAP BP was consistently weakest among the four tested. Fig. 1B demonstrates the effect of BB loop peptides on IL-1 and IL-6 protein expression. The BPs, but not the CP, also blocked secretion of IL-1 and IL-6. The concentration of BP used in Fig. 1 and all other experiments was 40 μM based on optimal inhibitory activity of the two most active BPs, TRAM BP and MyD88 BP, against an MyD88-dependent gene (e.g., IL-1) and an MyD88-independent gene (e.g., IFN-) (Fig. 1C). Fig. 1C also shows that BP-mediated inhibition is seen over a very narrow concentration range (i.e., between 10 and 40 μM).
Effect of BB loop BPs on activation of STAT-1 by LPS
STAT-1 propagates signal transduction emanating from both type I and type II IFN receptors (reviewed in Ref. 39). Two phosphate acceptor sites are involved in activation of STAT-1. Although phosphorylation of Tyr701 is critical for homo- or heterodimerization and nuclear translocation of STAT-1, phosphorylation of Ser727 modulates its transcriptional activity (reviewed in Ref. 39). We previously reported that tyrosine phosphorylation of STAT1 by LPS depends on MyD88-independent induction of IFN- and the subsequent autocrine activation of type I IFN receptors (33).
LPS-induced Tyr701 phosphorylation of STAT-1 was strongly inhibited by the TRAM and TRIF BPs (Fig. 2B). Consistent with its strong ability to block IFN- gene expression (Fig. 1), the MyD88 BP was also highly effective at inhibiting Tyr701 phosphorylation of STAT-1 (Fig. 2B). However, treatment of macrophages with the TIRAP BP did not completely block tyrosine phosphorylation of STAT-1 in response to LPS, although it was significantly diminished compared with the level seen in the presence of the CP (Fig. 2B). LPS-induced Ser727 phosphorylation of STAT-1 was similarly blocked by the BPs (data not shown). The BPs failed to inhibit IFN--induced Tyr701 phosphorylation of STAT-1 (data not shown), consistent with the observation that the type I IFN receptor does not depend on any of these adapters to signal (reviewed in Ref. 39).
Effect of BB loop BPs on activation of IRF-1 and IRF-3
IRFs were first identified as important transcriptional mediators of the antiviral immune response (reviewed in Ref. 40). Later, it was recognized that this family of transcription factors plays an essential role in signal transduction mediated by different TLRs. IRF-1 is regulated transcriptionally by signaling from different TLRs, as well as by cytokines induced by TLR engagement (e.g., IL-1, TNF, and IFN-) (41). In contrast, IRF-3 is constitutively expressed in many cell types (42). Upon stimulation through TLR4 or TLR3, IRF-3 is phosphorylated at multiple sites in its C terminus, and this modification enables it to dimerize and translocate to the nucleus where it targets IRF-3-responsive genes (32, 43). Using MyD88 knockout mice, Kawai et al. (32) demonstrated that LPS activates IRF-3 via the MyD88-independent pathway, and this, in turn, leads to the induction of IFN-, the signature cytokine gene for the MyD88-independent pathway (33). Pretreatment of cells with TRAM, TRIF, or MyD88 BPs significantly inhibited LPS-induced dimerization of IRF-3 (Fig. 3). The TIRAP BP also decreased the amount of detectable IRF-3 dimer; however, inhibition was not as complete as with the other BPs.
IRF-1 is an immediate response gene, quickly induced by signaling from different TLRs, which contributes to the induction of other important TLR-inducible genes (e.g., inducible NO synthase). Importantly, MyD88 knockout macrophages are capable of inducing IRF-1 mRNA (35). Stimulation of macrophages by LPS for 1.5 h after pretreatment of cells with medium only or CP resulted in a strong up-regulation of IRF-1 protein (Fig. 3). As was observed for IRF-3, TRAM, and MyD88 peptides were the strongest inhibitors with respect to both IRF-3 and IRF-1. The TRIF BP again showed intermediate efficacy, while the TIRAP peptide was the weakest of the BPs.
Limited capacity of adapter BB loop inhibitory peptides to block TLR2-induced signaling
Apart from the use of a CP, we sought to analyze further the specificity of the inhibitory effects of BB loop peptides on TLR signaling. Therefore, we also tested the effect of the peptides on signal transduction elicited from TLR2. TLR2 was chosen because it shares MyD88-dependent signaling pathway with TLR4, and TLR2-mediated responses are depreciated in both MyD88 and TIRAP-deficient cells (34, 44); however, TLR2-mediated signaling does not appear to use MyD88-independent arm for signal transduction (reviewed in Ref. 15). Stimulation of MyD88 knockout macrophages with the Pam3Cys, a TLR2 agonist, does not activate MAPKs (Ref. 44 and our unpublished observation), whereas LPS-stimulated MyD88 knockout macrophages exhibit MAPKs and NF-B activation, albeit with delayed kinetics (38). Fig. 4A demonstrates that the BPs differentially affect phosphorylation of ERK and JNK in response to LPS vs Pam3Cys. Surprisingly, under conditions where LPS-induced MAPK phosphorylation was strongly blocked by all BPs, none inhibited Pam3Cys-induced phosphorylation of MAPKs.
The effect of BPs on Pam3Cys-induced expression of IL-1 mRNA is shown in Fig. 4B. LPS and Pam3Cys induced comparable levels of IL-1 mRNA, which were not significantly affected by treatment of cells with CP (data not shown). Similar to their effect on activation of MAPKs, BPs were much weaker inhibitors with respect to Pam3Cys- vs LPS-induced expression of IL-1 mRNA. Only the MyD88 BP exerted a statistically significant inhibitory effect (Fig. 4B) and this effect was considerably less than that seen in the case of LPS stimulation; in LPS-stimulated cells, MyD88 BP reduced IL-1 mRNA expression to <5% of the level detected in the presence of the CP, while in Pam3Cys-stimulated cells, IL-1 gene expression was reduced to 25% of control levels. Taken collectively, these data demonstrate that TLR4-induced signaling is much more susceptible to inhibition by the adapter BB loop BPs than is TLR2-mediated signaling.
Discussion
This study represents an attempt to use penetrating oligopeptides containing homologous sequences that presumably interfere with the protein-protein interaction surfaces formed between the BB loops of four adapter proteins with their target protein(s) to investigate mechanisms of signal transduction emanating from TLR4 and TLR2. The impetus for generating these four adapter BPs stemmed from contradictory results obtained using the TIRAP BP and a TIRAP dominant-negative mutant construct vs macrophages from TIRAP knockout mice. The TIRAP BP was first used by Horng et al. (4) as a complementary tool to confirm the involvement of this adapter in LPS-signaling and was reported to inhibit TLR4-, but not TLR9-mediated signaling (4, 33). However, using this same TIRAP BP, we found that it also inhibited LPS-induced IFN- gene expression and IFN--luciferase reporter activity (33), while later studies using macrophages from TIRAP knockout mice found that induction of IRF-3 and IRF-3/STAT1-dependent genes by LPS was intact (34). Therefore, we sought to carry out a systematic comparison of the effects of BPs based on the homologous sequences of the other known TLR4 adapters to gain insights into the role of adapter BB loops for protein interactions required for TLR4 signaling. Although the mechanism by which cell-penetrating, BB loop peptides inhibit TLR signaling has not been formally demonstrated, it has been postulated that such peptides, once inside the cell, can occupy the docking site of the cognate adapter on its target and prevent binding of the native adapter. This, in turn, would be predicted to disrupt the formation of a functional signaling platform (14, 45).
Analysis of the effects of BPs on the expression of early inducible genes revealed an interesting and completely unexpected pattern; the BPs showed little specificity with respect to genes induced by LPS via MyD88-dependent or -independent pathways. For example, the MyD88 peptide did not preferentially block IL-1 and MIP-1 genes, which, according to data derived from knockout mice, require MyD88 for induction (37, 38). Conversely, TRAM BP was equally effective in inhibiting both MyD88-dependent and -independent genes, in contrast to the conclusion drawn from experiments with TRAM knockout mice (35). The fact that all four BPs block TLR4 signaling through either the MyD88-dependent or -independent pathways suggests strongly that the stability of the TLR4 signaling platform is indeed disrupted or fails to assemble correctly as a consequence of BP-target protein interactions.
In this regard, the disruption of TLR4 signaling by BPs is more reminiscent of a "receptor-knockout" phenotype, where both branches of the TLR4 signaling pathway are inhibited, rather than an "adapter-knockout" phenotype characterized by expression of distinct subsets of genes. The observation that BPs are capable of inhibiting various manifestations of LPS signaling suggests that they act before the signal bifurcates into separate pathways, most likely at the level of the receptor.
By examining events upstream of gene expression and cytokine secretion, it was possible to identify differences in the relative inhibitory capacities of BPs on TLR4 signaling. Thus, Figs. 2 and 3 show that MyD88 TRAM > TRIF > TIRAP BPs with respect to inhibition of MAPK, STAT-1, or IRF activation. Various reasons can be invoked to explain the observed differences in the relative efficiency of individual BPs to inhibit LPS signaling. Differences in peptide permeability, intracellular peptide turnover, existence of stable, "inefficient" conformations that do not bind as strongly to the intended molecular target, or, simply, nonspecific binding to unrelated sites could all potentially contribute to our observations. Importantly, the efficiency of the peptides failed to correlate with their general physical-chemical parameters that are summarized in Table I.
In addition to the surprising result that all four BPs inhibited both MyD88-dependent and -independent branches of TLR4-mediated signaling, the observations that the BPs were much weaker inhibitors of TLR2-mediated signaling was equally unexpected. The ability of BB loop peptides to disrupt TLR4 signaling again supports the hypothesis that these regions of the adapters are critical for interaction with TLR4. The fact that the BPs did not affect TLR2-induced activation of JNK and ERK and exerted a much weaker effect on TLR2-induced IL-1 gene expression implies that the adapter BB loops do not belong to the adapter surfaces that interact with TLR2. It is possible that the MyD88 BP partially interferes with the MyD88-TLR2 interaction surface, thus resulting in a lesser degree of inhibition. By this same logic, it would be predicted that TIRAP does not use its BB loop in the formation of an active TLR2 signaling complex. These findings suggest that the architecture of receptor-adapter complexes differs between TLR2 and TLR4. The idea that adapters shared by different TLRs can interact with individual receptors in fundamentally different ways supports the existence of asymmetrical models of TIR-TIR complex formation. Interestingly, recent computer modeling performed by Dunne et al. (19) predicted that TLR4 and TLR2 bind TIRAP or MyD88 through different nonoverlapping sites further support such a model.
Disclosures
The authors have no financial conflict of interest.
Footnotes
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.
1 This work was supported by National Institutes of Health Grants AI-18797 and AI-44936 (to S.N.V.) and AI-57490 (to M.J.F.).
2 M.J.F. and S.N.V. contributed equally to this work.
3 Address correspondence and reprint requests to Dr. Stefanie N. Vogel, Department of Microbiology and Immunology, University of Maryland, Baltimore, 655 West Baltimore Street, 13-009, Baltimore, MD 21201. E-mail address: svogel@som.umaryland.edu
4 Abbreviations used in this paper: TIR, Toll-IL-1 resistance; BP, blocking peptide; CP, control peptide; TIRAP, TIR-domain containing adapter protein; TRIF, TIR domain-containing adapter inducing IFN-; TRAM, TRIF-related adapter molecule; TBK-1, TRAF family member-associated NF-B binding kinase (TANK)-binding kinase-1; IKK-, IB kinase ; IRAK-1/4, IL-1R-associated kinase 1/4; Pam3Cys, (S-[2,3-bis(palmitoyloxy)-(2-RS)-propyl]-N-palmitoyl-(R)-Cys-Ser-Lys4-OH); HPRT, hypoxanthine phosphoribosyltransferase; IRF, IFN regulatory facto; GI, GenBank ID number.
Received for publication January 10, 2005. Accepted for publication April 16, 2005.
References
ock, F. L., G. Hardiman, J. C. Timans, R. A. Kastelein, J. F. Bazan. 1998. A family of human receptors structurally related to Drosophila Toll. Proc. Natl. Acad. Sci. USA 95: 588-593.
Kobe, B., A. V. Kajava. 2001. The leucine-rich repeat as a protein recognition motif. Curr. Opin. Struct. Biol. 11: 725-732.
Medzhitov, R., P. Preston-Hurlburt, E. Kopp, A. Stadlen, C. Chen, S. Ghosh, C. A. Janeway. 1998. MyD88 is an adaptor protein in the hToll/IL-1 receptor family signaling pathways. Mol. Cell. 2: 253-258.
Horng, T., G. M. Barton, R. Medzhidov. 2001. TIRAP: an adapter molecule in the Toll signaling pathway. Nat. Immunol. 2: 825-826.
Fitzgerald, K. A., E. M. Palsson-McDermott, A. G. Bowie, C. A. Jefferies, A. S. Mansell, G. Brady, E. Brint, A. Dunne, P. Gray, M. T. Harle, et al 2001. Mal (MyD88-adapter-like) is required for Toll-like receptor-4 signal transduction. Nature 413: 78-83.
Yamamoto, M., S. Sato, K. Mori, K. Hoshino, O. Takeuchi, K. Takeda, S. Akira. 2002. Cutting edge: a novel Toll/IL-1 receptor domain-containing adapter that preferentially activates the IFN- promoter in the Toll-like receptor signaling. J. Immunol. 169: 6668-6672.
Oshiumi, H., M. Matsumoto, K. Funami, T. Akazawa, T. Seya. 2003. TICAM-1, an adaptor molecule that participates in Toll-like receptor 3-mediated interferon- induction. Nat. Immunol. 4: 161-167.(Vladimir U. Toshchakov*, )