Elsevier

Immunology Letters

Volume 85, Issue 2, 22 January 2003, Pages 85-95
Immunology Letters

Review
Recognition of pathogen-associated molecular patterns by TLR family

https://doi.org/10.1016/S0165-2478(02)00228-6Get rights and content

Abstract

Toll-like receptors (TLRs) are type I transmembrane proteins involved in innate immunity by recognizing microbial conserved structures. Recent studies have shown that TLR3 recognizes dsRNA, a viral product, whereas TLR9 recognizes unmethylated CpG motifs frequently found in the genome of bacteria and viruses, but not vertebrates. TLR7 recognizes small synthetic immune modifiers including imiquimod, R-848, loxoribine, and bropirimine, all of which are already applied or promising for clinical use against viral infections and cancers. Plasmacytoid dendritic cells express TLR7 and TLR9, and respond to TLR7 and TLR9 ligands by producing a large amount of interferon (IFN-α). These results indicate that TLR3, TLR7 and TLR9 may play an important role in detecting and combating viral infections.

Introduction

The immune system in vertebrates comprises innate and acquired immunity, both of which work cooperatively to protect the host from microbial infections. Acquired immunity is characterized by specificity and memory, and exerted by T and B lymphocytes. Each lymphocyte expresses a cell surface receptor with a single specificity, which is generated by somatic recombination during maturation. However, since initiation of acquired immunity usually takes several days to occur, innate immunity is the first line of the host defense against bacterial growth and spread in the early phase of infection. The major players of innate immunity are phagocytes such as macrophages, neutrophils and dendritic cells (DCs). The primary role of these cells was formerly thought to be phagocytosis of pathogens in a non-specific manner, digestion of these, and presentation of pathogen-derived antigens to T cells. This response was also called the non-specific immune response. However, recent identification of Toll-like receptors (TLRs) has changed that notion. It has been revealed that innate immunity can recognize conserved pathogen-associated molecular patterns (PAMPs) through TLRs expressed on the cell surface of immune cells [1], [2], [3]. Recognition of invading pathogens then triggers cytokine production and upregulation of costimulatory molecules in phagocytes, leading to activation of T cells. Thus, innate immunity is closely linked to acquired immunity.

Toll is a type I transmembrane receptor in Drosophila. The extracellular domain contains leucine-rich repeat (LRR) whereas the cytoplasmic domain shows striking homology with that of IL-Rs, and is referred to as Toll/IL-1R (TIR) domain. The Toll gene was identified as a gene essential for the dorsal-ventral development in the Drosophila embryo. Toll receptor signaling is initiated by binding of an endogenous peptide ligand, termed Spatzle. Spatzle is synthesized as an inactive precursor protein that is cleaved by the protease Easter. Easter is also generated from the precursor protein in a series of protease activation cascade. Triggering of Toll receptor causes recruitment to the membrane of the adapter Tube and the protein kinase Pelle, and finally leads to activation of the NF-κB-like proteins, Dorsal. Subsequently, Toll signaling has been shown to play an essential role in the immune response to microbial infection [4], [5]. In adult Drosophila, microbial infection promptly induces production of a battery of antimicrobial peptides in the fat body, which is functionally equivalent to the liver in mammals. Antimicrobial peptides are categorized into antifungal and antibacterial peptides. Interestingly, the genes of these two groups of peptides are differentially regulated depending on the invading pathogen. Toll-mutant flies show compromised induction of antifungal peptide but not antibacterial peptides, and are highly sensitive to fungal infection, but not to bacterial infection. On the other hand, the immune-deficiency (Imd) pathway is involved in the induction of antibacterial peptides [6]. Imd mutants do not produce antibacterial peptides and are highly susceptible to Gram-negative bacterial infection. The Toll and Imd pathways do not share any signaling components. The imd gene has recently been identified and has been shown to encode an adaptor protein with homology to the mammalian RIP. At present the receptor involved in this pathway remains unclear.

Toll receptors are evolutionarily conserved and their homologs are found in insects, plants and mammals. To date ten TLRs have been identified in humans and mouse, respectively [1]. The essential role of Drosophila Toll in antifungal response suggested that mammalian TLRs also participate in innate immunity. In fact, TLRs have been found to be essential for microbial recognition (Table 1).

Lipopolysaccharide (LPS) is an integral component of the outer membranes of Gram-negative bacteria and a potent activator of macrophages as a causative agent of endotoxic shock. Structurally, LPS is a complex glycolipid composed of a hydrophilic polysaccharide portion and a hydrophobic domain known as lipid A, which is responsible for its biological activity. Stimulation of macrophages with LPS results in the production of various cytokines such as TNF-α, IL-1, IL-6, IL-10, macrophage inflammatory protein-1α/β (MIP-1α/β), and inflammatory effector substances such as prostanoids, leukotrienes, and nitric oxide. CD14 is a glycosylphosphatidyl inositol (GPI) anchored cell surface glycoprotein that functions as a binding receptor for LPS. When LPS is present in the blood stream, it is immediately captured by lipopolysaccharide-binding protein (LBP), which is one of the acute phase proteins produced in the liver. In CD14-negative cells such as endothelial cells and fibroblasts, the soluble form of CD14 present in serum can functionally replace membrane-bound CD14. Occupation of the LPS/LBP/CD14 ternary complex on the cell surface triggers activation of several members of the mitogen-activated protein kinase (MAP kinase) family, and the transcriptional factor NF-κB. Although LPS is bound primarily by LBP and CD14, the fact that CD14 lacks a transmembrane domain suggests the existence of an additional coreceptor that initiates signal transduction. Indeed, LPS-mediated cellular activation does occur in CD14-deficient mice.

C3H/HeJ is a mutant mouse strain hyporesponsive to LPS. Positional cloning of the lps locus demonstrated that C3H/HeJ mice harbor a missense point mutation within the cytoplasmic portion of TLR4 that results in the substitution of histidine for a proline that is highly conserved among TLR family members [7], [8]. The C57BL/ScCr and C57BL/ScN strains lack TLR4 mRNA expression altogether due to a chromosomal deletion of the gene. Mice deficient in the TLR4 gene were also generated by gene disruption techniques [9]. Macrophages and B cells from these TLR4 knockout (KO) mice were hyporesponsive to LPS to a similar extent to C3H/HeJ mice. Furthermore, over-expression of wild-type TLR4 but not C3/HeJ TLR4 activated NF-κB, confirming that TLR4 is required for LPS signaling [9].

TLR2 has also been shown to be involved in LPS-mediated signaling. Of the human TLR 1–5, only over-expression of human TLR2 solely conferred LPS responsiveness on the human embryonic kidney cell line 293 [10], [11]. In addition, evidence suggested that human TLR2 interacts with CD14 to form an LPS receptor complex [12]. LPS treatment leads to the oligomerization of this receptor and to the subsequent recruitment of IRAK. However, results from TLR4 mutant mice demonstrated more definitely that LPS specifically activates via the TLR4 receptor. Although transfection of cell lines with human TLR4 alone did not confer the ability to respond to LPS, dose-dependent LPS responsiveness occurred following cotransfection of the novel molecule, MD2, which physically associates with TLR4 on the cell surface [13]. Finally, TLR2 KO mice as well as Chinese hamster ovary cells with a defective TLR2 gene exhibit normal LPS responsiveness [14], [15]. Thus, the initial proposal that TLR2 is the LPS receptor is strongly challenged by these subsequently published data. Indeed repurification of commercially available LPS preparations showed that neither human nor murine TLR2 play a role in LPS signaling [16]. However, some LPSs (from Leprospira interrogans and Porphyromonas gingivalis) that are structurally different from those from enterobacteria have been shown to activate cells through TLR2 [17], [18]. Gram-positive bacteria such as Staphylococcus aureus do not produce LPS but trigger a toxic shock syndrome similar to that induced by LPS. The major immunostimulatory components are peptidoglycan (PGN), lipoteichoic acid (LTA) and lipoproteins. Several studies with TLR2-expressing cell lines and TLR2 KO mice showed that PGN and lipoproteins are recognized by TLR2 [19], [20]. In fact, TLR2 KO mice are highly susceptible to Staphylococcus aureus infection [21]. It has also been shown that TLR2 is involved in the responses to lipoproteins/lipopeptides from mycobacterium tuberculosis [22], Gram-negative bacteria [23], Borrelia burgdoreri [24], and Mycoplasma fermentans [25], GPI anchors from Trypanosoma cruzi [26], lipoarabinomannan from mycobacterium tuberculosis [27], [28] and zymosan, a component of yeast cell wall [29].

The possibility that TLR2 ligands are recognized by heterodimers between TLR2 and other TLR was first indicated by Aderem et al. [30]. Both dominant negative (DN)-TLR2 or DN-TLR6 mutants completely inhibited TNF-α production elicited by PGN or zymosan. In contrast, TLR2 recognizes bacterial lipopeptide without TLR6. They also showed that dimerization of the cytoplasmic domain of TLR2 does not induce TNF-α production in macrophages, whereas the cytoplasmic domain of TLR2 can form functional pairs with TLR6, leading to cytokine production. These results suggested that TLR2/6 heterodimers recognize PGN and zymosan whereas bacterial lipoprotein is recognized by a heterodimer between TLR2 and another, as yet unidentified TLR. We have generated TLR6 KO mice and examined the response to PGN, zymosan, and lipoproteins [20]. In contrast with the Aderem group's findings, the response to PGN was not impaired, indicating that TLR6 does not participate in PGN recognition. Nevertheless, TLR6-deficient macrophages did not produce any detectable levels of inflammatory cytokines in response to synthetic mycoplasmal lipopeptide, MALP-2 as is the case for TLR2-deficient macrophages. The requirement for both TLR2 and TLR6 in MALP-2 recognition was further confirmed by reconstitution experiments in which TLR2 or TLR6 expression vectors were transfected into TLR2−/−TLR6−/− deficient embryonic fibroblasts and then IL-6 production was measured. Only when both expression vectors were transfected, the mutant MEF restored the ability to respond to MALP-2. On the other hand, in the case of synthetic bacterial lipopeptide, PAM3CSK4, TLR6-deficient macrophages responded normally to this. Fig. 1 shows structural differences between MALP-2 and PAM3CSK4. All membrane-anchored lipoproteins contain a lipolyated amino-terminal residue. The immunostimulatory activity of lipoproteins is attributed to the lipid portion. The lipolylation of the amino-terminal cysteinyl residue by a diacylglyceryl moiety via a thioether bond is a feature common to all known bacterial membrane lipoproteins [31]. Some lipoproteins undergo further acylation at the cysteinyl residue via an amide bond. For triacylated lipoproteins such as the Braun's lipoprotein in Escherichia coli, the maturation pathway comprises three steps, that is, diacylglyceryl modification of the pro-lipoprotein by diacylglyceryl transferase, cleavage of the leader peptide by signal peptidase II, and N-acylation of the amino-terminal diacylglyceryl cysteinyl residue. PAM3CSK4 contains triacylated cysteinyl residue at the N-terminus, whereas the cysteinyl residue in MALP-2 is only diacylated. Therefore, it was expected that this difference might explain the TLR6 requirement in MALP-2 recognition. In fact, the replacement of the lipid portion of MALP-2 with that of PAM3CSK4 activated TLR6-deficient macrophages, demonstrates that TLR6 discriminates between the subtle differences in the lipid portion of these lipopeptides.

We have recently generated TLR1 KO mice and examined the response to a variety of bacterial components [32]. TLR1 KO mice responded to PGN, LPS and mycoplasmal lipopeptide to the same extent as wild-type mice. Nevertheless, TLR1-deficient macrophages showed partial impairment of the response to mycoplasmal 19 kD lipoprotein and bacterial lipopeptide, PAM3CSK4. Furthermore, the response to N-PAM-S-Lau2CSK4 was more profoundly impaired in TLR1-deficient cells, indicating that TLR1 is involved in the recognition of lipoproteins and the configuration of the lipid moiety of lipoproteins is critical for the TLR1 requirement.

As shown in Fig. 2, mycoplasmal lipopeptide (diacylated form) is recognized by a heterodimer between TLR2 and TLR6 whereas bacterial lipopeptides PAM3CSK4 and N-PAM-S-Lau2CSK4 (triacylated forms) are preferentially recognized by a heterodimer between TLR2 and TLR1. The responses to zymosan and Trypanosoma cruzi GPI are also recognized by a TLR2/TLR6 heterodimer (our unpublished data). It is possible that a TLR other than TLR1 and TLR6 pairs with TLR2 to recognize PGN, or TLR2 alone may suffice to recognize it.

CpG DNA also stimulates immune cells [33], [34], [35]. CpG DNA is largely equivalent to bacterial DNA. Unmethylated CpG-dinucleotide-containing sequences (CpG ODNs) are found much more frequently in bacterial genomes than in vertebrate genomes, whereas the frequency of CpG dinucleotides are suppressed and usually methylated. Methylated CpG ODNs lack immunostimulatory activities. Bacterial DNA and synthetic ODN containing unmethylated CpG-dinucleotide (CpG DNA) stimulate B cell proliferation and activate macrophages and DCs. Genomic DNA from viruses, yeast and insects stimulate mammalian immune cells as well. CpG DNA is known to be an excellent immune adjuvant in various murine disease models and to drive Th1 immune responses. CpG DNA activates the intracytoplasmic signaling molecules such as IRAK, TRAF6, NF-κB and MAP kinase like other pathogen-derived immunostimulatory components. However, unlike LPS, which can activate TLR4 at the cell surface, uptake of CpG DNA as well as endosomal maturation is likely to be required for its immunostimulatory activity. Indeed, chloroquine and related compounds that prevent acidification of endosome are shown to inhibit CpG DNA response [36]. The immunostimulatory activity of bacterial DNA was initially reported by Tokunaga et al. who showed that DNA purified from bacilli Calmette-Guerin inhibited the growth of various syngeneic animal tumors, augmented NK cell activity and induced IFNα/β and IFN-γ from mouse spleen cells and human PBL [37]. Krieg et al. subsequently reported that bacterial DNA as well as synthetic ODN containing a central CpG induces B cell proliferation [38]. They conducted extensive studies to define the DNA sequences with immunostimulatory activities. The most immunostimulatory motif usually has the structure of 5′-purine-purine-CpG-pyrimidine-pyrimidine-3′.

Inversion to GpC or methylation completely abrogates its stimulatory potential, and the regions adjacent to the CpG dinucleotide also affect the immunostimulatory activity. The optimal sequence differs significantly between human and mouse. Mouse cells respond maximally to GACGTT, while for humans the optimal sequence is GTCGTT. Recent studies indicate that immunostimulatory DNAs may exert different immune responses depending on the nucleotide sequence and backbone. There are two major types of immunostimulatory CpG DNAs [39]. One type has an entirely phosphorothioate backbone with CpG dinucleotides. This CpG DNA stimulates B cell proliferation and induces production of IL-6 and IL-12 by monocytes. The other type contains phosphorothioate G-rich sequences at the ends and a phosphodiester palindromic sequences with a CpG dinucleotide in the middle. This CpG DNA preferentially stimulates IFN-γ production by NK cells.

We have generated TLR9 KO mice, and found that the response to CpG DNA is mediated by TLR9 [40]. TLR9-deficient mice did not show any inflammatory cytokine production of macrophages, proliferation of B cells, or maturation of DCs in response to unmethylated CpG DNA. CpG DNA-induced activation of several signaling molecules such as NF-kB, JNK, and IRAK was severely impaired in TLR9-deficient mice. These results demonstrate that cellular activation in response to CpG DNA is mediated by TLR9. In spite of the existence of various types of CpG DNA with different biological effects, TLR9 KO mice were unresponsive to all CpG DNA we used, demonstrating that TLR9 is essential for the response to a variety of CpG DNA. Recent complementation experiment using CpG-unresponsive 293 cells transfected with human or mouse TLR9 unambiguously demonstrated that human TLR9 recognizes the human CpG motif most efficiently, whereas mouse TLR9 has a high preference for the mouse CpG motif, suggesting direct interaction of the TLR9 receptor with its CpG DNA ligand [41].

At the same time as we published the TLR9 paper, a report showed that CpG DNA activates DNA protein kinase (DNA-PK), which phosphorylates IκB-kinases, leading to the activation of NF-κB, which further leads to the production of cytokines [42]. Furthermore, it was shown that DNA-PK is also essential for the response to CpG DNA based on the lack of the response to CpG DNA in DNA-PK KO mice. Therefore, we reevaluated the role of DNA-PK in CpG DNA responsiveness with DNA-PK KO mice that were kindly provided by Dr Fred Alt. DNA-PK-KO DCs showed upregulation of costimulatory molecules and cytokine production in response to CpG DNA, indicating that DNA-PK is not essential for the response to CpG DNA (our unpublished data).

Flagellin is a 55-kDa monomer obtained from bacterial flagella, a polymeric rod-like appendage extending from the outer membrane of Gram-negative bacteria that propels the organism through its aqueous environment. Flagellin is also a potent proinflammatory factor. Flagellated bacteria, purified flagellin, and medium conditioned by flagellated bacteria all induce NF-κB activation and iNOS expression in transformed human epithelial cells and murine macrophages. These actions of flagellin have been shown to be mediated through TLR5 [43]. TLR5-stimulating activity was purified from L. monocytogenes culture supernatants and identified as flagellin by tandem mass spectrometry. Expression of L. monocytogenes flagellin in non-flagellated E. coli conferred on the bacterium the ability to activate TLR5, while deletion of the flagellin genes from S. typhimurium abrogated TLR5-stimulating activity. TLR5 is expressed on the basolateral, but not apical, surface of intestinal epithelia [44]. Therefore, flagellin activates proinflammatory gene expression in the epithelia only if it crosses intestinal epithelia and contacts their basolateral membranes, which explains why commensal microbes can secrete flagellin into the intestinal lumen yet not induce inflammation. Flagellin acts as a PAMP in plants as well. FLS2, the gene affected in several independent mutants exhibiting flagellin insensitivity, encodes a LRR transmembrane receptor-like kinase with similarities to TLRs in mammals [45]. Evidence indicates that the LRR domain of FLS physically interacts with flagellin [46]. There is no significant homology between the LRR domains of TLR5 and FLS2.

Viral replication within infected cells results in generation of dsRNA which can stimulate immune cells. It is well known that dsRNA binds and activates the dsRNA activated protein kinase, PKR, initiating downstream signaling that inhibits viral replication and upregulation of inflammatory cytokine production. However, cells derived from PKR KO mice still respond to the viral RNA mimic, poyinosine-polycytidylic acid, poly(I:C), suggesting the existence of another receptor, which recognizes dsRNA. Recently, TLR3 KO mice have been generated and showed reduced responses to poly(I:C), suggesting that TLR3 is involved in recognition of dsRNA [47]. This may indicate the role of TLR3 in viral recognition.

Imiquimod (Aldara, R-837, S-26308) and R-848 (resiquimod, S-28463) are low molecular weight compounds of the imidazoquinoline family that have potent antiviral and antitumor properties in animal models [48](Fig. 3). Imiquimod's antiviral activity was first discovered in guinea-pigs infected with herpes simplex virus. Imiquimod is also shown to be effective against arbovirus and cytomegalovirus. The activity of imiquimod is mediated predominantly through the induction of cytokines including IFN-α and IL-12. Topical imiquimod therapy is now approved for the treatment of external genital and perianal warts caused by Papilloma virus infection. R-848 is a more potent analogue of imiquimod, and trials are under way to assess its use in treatment of genital herpes and hepatitis C virus.

We found that the response to these imidazoquinolines is completely abolished in MyD88 KO mice, indicating the involvement of TLRs in the response to imidazoquinolines. Further studies showed that TLR7 is essential for the response to imidazoquinolines [49]. Peritoneal macrophages from wild-type mice produced TNF-α, IL-6 and IL-12 in response to imiquimod or R-848, whereas TLR7 KO macrophages produced no detectable amounts of these cytokines in response to either compound. Wild-type splenocytes proliferated in response to imiquimod or R-848 in a dose-dependent manner. In contrast, TLR7 KO splenocytes did not proliferate in response to either imiquimod or R-848. Furthermore, TLR7-deficient bone marrow-derived DCs did not show upregulation of the costimulatory molecules, CD40, CD80 and CD86 in response to R-848. Activation of intracytoplasmic molecules involved in TLR signaling such as IRAK, NF-κB and JNK was completely abolished in TLR7 KO macrophages. Intraperitoneal injection of wild-type mice with R-848 induced increased serum concentrations of IFN-α, TNF-α and IL-12. In contrast, increase in serum levels of these cytokine was not observed in TLR7 KO mice.

Recently it has been shown that two other immunomodulators, loxoribine and bropirimine are also TLR7 ligands (Fig. 3) (our unpublished data). Loxoribine (7-allyl-8-oxoguanosine) enhances NK cell activity, B lymphocyte proliferation, and induces production of interferons and cytokines [50]. Bropirimine (2-amin-5-bromo-6-phenyl-4(3)-pyrimidinone, ABPP) is an orally active immunomodulator that increases endogenous IFN-α and other cytokines, and is used clinically against carcinoma of the bladder [51]. Further studies will be needed to disclose the structure-activity relationship of TLR7 and its ligands.

Triggering of IL-1R or TLR recruits IRAK to the receptor complex via adaptor MyD88, then TRAF6 is activated, which finally results in activation of NF-κB and MAP kinases (Fig. 4). MyD88-deficient mice were generated and found to be completely unresponsive to IL-1 as well as IL-18, an IL-1-related cytokine, demonstrating that MyD88 is a critical component in the signaling cascades mediated by IL-1R and IL-18R. Furthermore, MyD88 KO mice did not respond to LPS and other immunostimulatory bacterial components including PGN, lipoproteins, CpG DNA, and imidazoquinolines, demonstrating essential roles of MyD88 in the response to all pathogen-derived immunostimulatory molecules.

However there is a difference in the signaling pathways triggered by LPS and by these latter stimuli. Mycoplasmal lipopeptide-dependent activation of NF-kB and MAP kinases, which is mediated by TLR2, is completely abolished in TLR2 KO or MyD88 KO macrophages [25]. However, LPS activation of MAP kinases and NF-κB remains intact in MyD88 KO macrophages, although activation was delayed compared with wild-type mice [52]. This indicates that the LPS response may be mediated by both MyD88-dependent and -independent pathways, each of which leads to the activation of MAP kinases and NF-κB. Nevertheless, the MyD88-dependent pathway is essential for the inflammatory response mediated by LPS. Further studies demonstrated that the MyD88-independent pathway is involved in IRF-3 activation, and induction of type I IFN and IFN inducible genes including IP-10 [53]. Consistent with the lack of MyD88-independent pathway in TLR2 signaling, wild-type macrophages do not activate IRF-3 or induce IP-10 mRNA in response to mycoplasmal lipopeptide. The MyD88-independent pathway has also been shown to regulate LPS-mediated maturation of DCs [54]. Recently, a novel protein containing a TIR domain and designated TIRAP/Mal has been identified [55], [56]. TIRAP/Mal over-expression led to activation of NF-κB in 293 cells and is associated with TLR-4 but not other TLR members. It is most likely that TIRAP/Mal participates in MyD88-independent NF-κB activation. A further interesting point is whether TIRAP is also involved in MyD88-independent activation of IRF-3 and subsequent induction of IFN-β and IFN-inducible genes.

Section snippets

Dendritic cells and TLRs

DCs reside in most tissues and organs in the ‘immature state’, and play a sentinel role in detecting invading pathogens [57]. The immature DCs are characterized by high endocytotic capacity, low expression of MHC and costimulatory molecules on their surface. The function of DCs is not primarily to destroy pathogens but to present pathogen-associated antigens to T cells, leading to activation of acquired immunity. Upon encountering a pathogen, immature DCs can be activated and undergo

Myeloid DC and plasmatoid DC

There are two DC subsets identified in human [60]. One is CD11c+myeloid dendritic cell (MDC) and the other CD11c plasmacytoid dendritic cell (PDC). These two subsets differ in TLR expression as well as cytokine production profile; MDC express TLR4 but not TLR9, and thereby respond to LPS but not CpG DNA. On the other hand, PDC express TLR9 but not TLR4 and respond to CpG DNA. A characteristic of PDC is the production of a large amount of type I IFNs in response to viral infection. We studied

Conclusion

Evidence is now accumulating that the recognition mechanism and subsequent signaling pathway are quite different between insects and mammals. Toll in insects is not a receptor that recognizes bacteria itself, but is activated by an endogenous ligand, Spatzle, that is generated as a result of a series of proteolytic processing steps by extracellular serine proteases. Recently, a certain protein family has been demonstrated to participate in the direct recognition of pathogens [62], [63], [64],

References (85)

  • B. Lemaitre et al.

    Cell

    (1996)
  • O. Takeuchi et al.

    Immunity

    (1999)
  • I. Chambaud et al.

    Trends Microbiol.

    (1999)
  • H. Wagner

    Curr. Opin. Microbiol.

    (2002)
  • W. Chu et al.

    Cell

    (2000)
  • L. Gomez-Gomez et al.

    Mol. Cell

    (2000)
  • Z. Bauer et al.

    J. Biol. Chem.

    (2001)
  • R.L. Miller et al.

    Int. J. Immunopharmacol.

    (1999)
  • T. Kawai et al.

    Immunity

    (1999)
  • T. Kaisho et al.

    Trends Immunol.

    (2001)
  • B. Opitz et al.

    J. Biol. Chem.

    (2001)
  • R.M. Vabulas et al.

    J. Biol. Chem.

    (2001)
  • R.M. Vabulas et al.

    J. Biol. Chem.

    (2002)
  • K. Kawasaki et al.

    J. Biol. Chem.

    (2000)
  • Y. Okamura et al.

    J. Biol. Chem.

    (2001)
  • S. Akira et al.

    Nat. Immunol.

    (2001)
  • A. Aderem et al.

    Nature

    (2000)
  • C.A. Janeway et al.

    Annu. Rev. Immunol.

    (2002)
  • J.A. Hoffmann et al.

    Nat. Immunol.

    (2002)
  • B. Lemaitre et al.

    Proc. Natl. Acad. Sci. USA

    (1995)
  • A. Poltorak et al.

    Science

    (1998)
  • S.T. Qureshi et al.

    J. Exp. Med.

    (1999)
  • K. Hoshino et al.

    J. Immunol.

    (1999)
  • R.B. Yang et al.

    Nature

    (1998)
  • C.J. Kirschning et al.

    J. Exp. Med.

    (1998)
  • R.B. Yang et al.

    J. Immunol.

    (1999)
  • R. Shimazu et al.

    J. Exp. Med.

    (1999)
  • H. Heine et al.

    J. Immunol.

    (1999)
  • M. Hirschfeld et al.

    J. Immunol.

    (2000)
  • C. Werts et al.

    Nat. Immunol.

    (2001)
  • M. Hirschfeld et al.

    Infect. Immun.

    (2001)
  • A. Yoshimura et al.

    J. Immunol.

    (1999)
  • O. Takeuchi et al.

    Int. Immunol.

    (2001)
  • O. Takeuchi et al.

    J. Immunol.

    (2000)
  • H.D. Brightbill et al.

    Science

    (1999)
  • A.O. Aliprantis et al.

    Science

    (1999)
  • M. Hirschfeld et al.

    J. Immunol.

    (1999)
  • O. Takeuchi et al.

    J. Immunol.

    (2000)
  • M.A. Campos et al.

    J. Immunol.

    (2001)
  • D.M. Underhill et al.

    Proc. Natl. Acad. Sci. USA

    (1999)
  • T.K. Means et al.

    J. Immunol.

    (1999)
  • D.M. Underhill et al.

    Nature

    (1999)
  • Cited by (0)

    View full text