Abstract
Animal Cells and Systems 13: 1-8, 2009 Mi Sun Jin and Jie-Oh Lee* Department of Chemistry , Department of Nanoscience and T echnoloogy , Korea Advanced Institute of Science and T echnology , Daejeon 305-701, Korea Abstract: Toll-like receptor (TLR) family proteins, type I reported since TLR4, the first TLR discovered, was transmembrane proteins, play a central role in human innate identified as a homolog of Dr osophila Toll. Toll has been immune response by recognizing common structural sh o wn to b e critical fo r embryo d evelopment and antifu ngal patterns in diverse molecules from bacteria, viruses and immune activity (Gay and Keith, 1991; Matsushima, 2007; fungi. Recently four structures of the TLR and ligand Medzhitov, 1997). TLRs are type I transmembrane complexes have been determined by high resolution x-ray glycoprotein composed of extracellular, transmembrane crystallographic technique. In this review we summarize reported structures of TLRs and their proposed activation and conserved intracellular domains (Gay and Gangloff, mechanisms. The structures demonstrate that binding of 2007). The ex tracellular domain s of TLRs characterized by agonistic ligands to the extracellular domains of TLRs repeated LRR (Leucine-R ich Repeat) modules are specific induces homo- or heterodimerization of the receptors. for pathogen-associated molecular patterns. Each TLR Dimerization of the TLR extracellular domains brings their binds to distinctive microbial components. For example, as two C-termini into close proximity . This suggests a plausible TLR2 complex with TLR1 or TLR6 recognizes lipoprotein mechanism of TLR activation: ligand induces dimerization of or lipopeptide, TLR3, TLR4, TLR5, TLR7 (or TLR8), and the extracellular domains, which enforces juxtaposition of i n trace ll u l a r si gn ali ng domai ns for r e crui tme n t of in tr acel lul a r TL R9 recog nize viral doub le s tranded RNA , lip opolys acc har id e adaptor proteins for signal initiation. (LPS), bacterial flagellin, single stranded RNA, and microbi al, res p ectively . Bindin g o f t hese lig ands to corr espo ndin g Key words: innate immune response, leucine rich repeat, pat- TLRs induces multimerization of the receptors and tern recognition receptor , toll-like receptor , hybrid LRR technique believed to trigger recruitment of intracellular adaptor proteins to the intracellular TIR (Toll/IL-1 Resistance) domains of TLR for signaling. TOLL-LIKE RECEPTOR F AMIL Y PROTEINS Although innate immune response against infectious microbes is an essential protective system, patients under Research of innate immunity is one of the most vigorously compromised immune condition may suffer from septic investigated field of immunology since Metchnikoff first syndrome due to unregulated and over-reactive immune introduced phagocytic theory of immunity (O’Neill, 2004; responses mediated by TLRs and other innate immune Silverstein, 2003). Innate immune system is characterized receptors. Moreover, genetic defects in some TLRs lead to by fast and immediate response against invading pathogens. severe defects toward infectious diseases such as meningitis, In contrast, highly specific adaptive imm une system operates tuberculosis and asthma (Echchannaoui, 2002; Ogus, 2004; later in infection. Innate immune response is initiated by the Texereau, 2005). Therefore, development of therapeutic limited number of receptors termed pattern recognition agents modulating activities of TLRs is urgently required. receptors (PRRs) present in several immune cell types To date, TLR agonists including imiquimod of TLR7, including dendritic cells, neutrophils and natural killer T monophosphorylated Lipid A (MPL) of TLR4 are cells. Among these receptors, the representative key player successfully used for clinical treatments of the diseases or is the TLR (Toll-Like Receptor) family. for enhancing vaccine responses (Beutner, 1998; Hemmi, Until now 10 human and 13 mouse TLRs have been 2002; Romagne, 2007). Furthermore, TLR antagonists that can bind to receptor but fail to initiate intracellular signal * T o whom correspondence should be addressed. are being developed to control microbial inflammations or T el: +82-42-350-2839; Fax: +82-42-350-2810 E-mail: jieoh.lee@kaist.ac.kr autoimmune diseases. For example, Eritoran, a potent ANIMAL CELLS AND SYSTEMS Vo l. 13 No. 1 1 Mi Sun Jin and Jie-O h Lee antagonist of TLR4 is currently under phase III clinical trialsubfamily proteins. against severe sepsis (Kanzler, 2007). Structure of TLR3, the first reported structure of TLR, Structural studies of TLR and ligand complexes haveshows the flat horseshoe-like shape with an uniform b sheet been attractive area of research because structural information conformation similar with those of other “typical” is critical for study of innate immune system as well as forsubfamily proteins (Bell, 2005; Choe, 2005). This structural development of novel drugs against several immune property is originated from well conserved sequence syndromes. In 2005, Choe et al. determined the first crystal pattern of TLR3 - conserved leucine residues point inward structure of human TLR in atomic detail (Choe, 2005). for hydrophobic interaction and asparagine ladders interact However many important questions about TLRs and their via hydrogen bonds with the amide and carbonyl groups in roles in innate immune response had remained elusive the neighboring LRR module. In contrast, TLR1, TLR2 because the structure did not contain bound ligands and and TLR4, although they clearly are members of typical therefore could not provide direct insights into ligand subfamily, show unusual structural features different from specificity or receptor activation. Recently four structures that of TLR3 (Jin, 2007; Kim, 2007b; Park, 2009). Due to of the TLR-ligand complexes have been determined bysharp structural transitions in the β sheet, structures of these high resolution x-ray crystallography; TLR1-2-lipopeptide, TLR proteins can be divided into N-terminal, central, and TLR3-dsRNA, TLR4-MD-2-Eritoran, and TLR4-MD-2- C-terminal subdomains. Both N-terminal and C-terminal LPS. In this review we summarize (1) common structural domains have conserved structure pattern common to other characteristics of the TLR family , (2) structural patterns intypical family proteins. However, the central domains adopt ligands that are responsible for TLR binding and activation, the unusual β sheet conformation with smaller inner radius and (3) structures of the TLR dimers induced by agonistic and larger twist angles. These unusual structural alterations ligands. appear to b e caused b y irregular LRR motiv es concentrated on the central domains: central domains of TLR1, TLR2 and TLR4 have no conserved asparagine ladders that STRUCTURE OF THE TLR F AMIL Y maintain overall curvature of the horseshoe-like shape by LRR is a relatively large protein family that includes more fo rming con ti nu ou s h yd ro ge n-b on din g in te ra cti on s. M oreo ve r , than 6000 proteins in Pfam data base. LRR domains exist in the LRR modules of the central domains have considerable about 300 human proteins (Kobe and Deisenhofer, 1994; variations of amino acid lengths, ranging from 20 to 33 Kobe and Deisenhofer, 1995; Kobe and Kajava, 2001; residues. Generally LRR modules containing fewer amino Matsushima, 2007). T o date, approximately seventy structures acid residues have loops in their variable convex regions, of three-dimensional structures of LRR superfamily have but those containing more residues tend to form helices in been reported. All known structures of LRR superfamily the variable regions. Central domains of TLR1, TLR2 and proteins demonstrate an arc or horseshoe-like shape. This TLR4 have mixture of a helices containing regions and protein architecture is derived from unique sequence loop containing regions. This appears to distort the pattern in the LRR proteins. All LRR modules in the otherwise regular structures of the convex region and proteins are composed of the highly conserved and variablemakes inner radii smaller and substantial deviation from the parts (Kajava, 1998; Kobe and Kajava, 2001). The conserved principal plane of the horse-shoe like structure. Protein “LxxLxLxxN” motif part forms the inner concaves regions with unusual structural alterations often play containing the continuous parallel β-strands. On the otherimportant role in function because they can provide unusual hand, the variable sequences form the outer convexes shape or flexibility necessary for activity . Consistently , the surfaces fo rmed by α-h elices and/or loops. Almost all LRR central domains of TLR1, TLR2 and TLR4 have essential proteins have the capping modules termed LRRNT and roles in ligand recognition or binding of accessory protein LRRCT at their both termini (Kajava, 1998). These(discussed later) (Jin, 2007; Kim, 2007b; Park, 2009). capping modules containing cysteine clusters shield theOther TLRs can be divided into these two classes although hydrophobic core composed of conserved leucines fromtheir structures are not known yet. TLR5, TLR7, TLR8 and exposure to outside. The reported structures of TLRs TLR9, which have strictly conserved asparagine ladders, demonstrate that TLR family proteins belong to the probably have the flat single domain-like architectures “typical” subfamily of the LRR superfamily (Bell, 2005; similar with those of the traditional “typical” subfamily Choe, 2005; Jin, 2007; Kim, 2007b; Liu, 2008; proteins (Matsushima, 2007). Conversely, TLR10 that has Matsushima, 2007). They all contain the horseshoe-like the asparagine ladders broken and the irregular LRR shapes similar with the structures of other LRR containing modules with longer lengths in the central part of the proteins. However, some TLRs demonstrate unusualprotein is likely to have distorted multi-domain architecture structural features distinctive from those of other “typical”present in structures of TLR1, TLR2 and TLR4. 2 ANIMAL CELLS AND SYSTEMS Vol. 13 No. 1 T LR structures ST UCTURES O F THE TLR-L IGAND C O MPLEXES only a catalyst for lipopeptide binding by TLR2 (Nakata, 2006). The crystal structure shows that the lipopeptide by To date, four structures of TLR complexes with their itself is sufficient to form the TLR1-TLR2 heterodimer ligands are reported by different laboratories (Jin, 2007; without CD14 (Jin, 2007). As CD14 can induce transfer of Kim, 2007b; Liu, 2008; Park, 2009). This allows nice the monomerized LPS to the TLR4-MD-2 complex for answers to controversial questions about TLRs. For proinflammatory responses (Fujihara, 2003; Miyake, instance, it is clear that most TLRs except TLR4 directly 2003), it is conceivable that CD14 facilitates lipopeptide bind to their ligands, that MD-2 is absolutely necessary for transfer as monomeric form to TLR1-TLR2 complex. ligand binding by TLR4, and that dimerization of TLRs is Moreover, we found that CD14 interacts only weakly with induced by the ligand binding. Moreover, it offers clues to TLR2 (unpublished data). These results suggest that CD14 how other TLRs interact with their ligands to be activated is a catalyst, not a co-receptor, for lipopeptide recognition although the related structural information remains to bein the TLR1 and TLR2 complex. further in vestigated. Structure of TLR1-TLR2-lipopeptide provides us with information to predict binding mode of LTA(lipoteichoic Structure of the TLR1-TLR2-lipopeptide complex acid) to TLR1-TLR2 and di-acylated lipopeptide to TLR2- TLR2 in association with TLR1 is essential for recognizing TLR6. LTAs have been reported as potent inducers of bacterial lipoproteins, especially tri-acylated lipoproteins innate immune response (Deininger, 2003; Han, 2003; (Takeuchi, 2001; Takeuchi, 2002). Crystal structure ofSchroder, 2003). LTAs have two lipid chains attached to the human TLR1-TLR2-PamCSK shows that overall shape glycerol backbone and phosphate containing units repeated 3 4 of the complex is similar with alphabet “m” where two N- ranging from 4 to 25 times (Morath, 2001). S tructure of the terminal domains outstretch toward opposite ends and theTLR2-PamCSKcomplex shows that the PamCSK binds 24 2 4 C-terminal domains converge in the middle (Jin, 2007). in similar fashion with that of PamCSK : two lipid chains 3 4 PamCSK is a synthetic peptide containing the conserved of the PamCSK are inserted into the TLR2 pocket and 3 4 2 4 N-terminal cysteine acylated by three palmitate groups glycerol and peptide backbo ne is held by conserved protein (Hioe, 1996). It can bind and activate TLR1-TLR2 complex residues (Jin, 2007). Because of structural homology with like natural lipoproteins. Surprisingly, the lipopeptide- di-acylated PamCSK lipopeptides, two lipid chains of 2 4 binding site is found in a highly unexpected region (Jin,LTAs are likely to be inserted into the TLR2 binding pocket 2007). It is located at the boundary of the central and C- like PamCSK . Di-acylated lipopeptide is recognized by 2 4 terminal domain in th e conv ex regions o f TLR1 and TLR2 .the TLR2-TLR6 complex (Buwitt-Beckmann, 2005; Takeuchi, The two hydrophobic pockets form a continuous ligand 2001). The structure of TLR6 can be predicted by homology binding channel by forming a ligand induced heterodimer. modeling (Jin, 2007). The structural model suggests a clue T wo of the three lipid chains of the PamCSK are inserted to why TLR2-TLR6 complex is unable to interact with the into the TLR2 pocket, and the remaining amide-bound lipid tri-acylated lipoproteins. It is the two bulky phenylalanine chain is inserted into the narrower TLR1 channel. Moreover , residues blocking the potential lipid binding channel in the complex of TLR1 and TLR2 is further stabilized TLR6, and thus the amide-bound lipid chain of the tri- through protein-protein interactions in the ligand bindingacylated lipopeptide may not be inserted into TLR6. pocket. Structural study of TLR1-TLR2-lipopeptide offers an Structure of the TLR3-dsRNA complex op portunity to re so lve the question s whether CD14 as a co- TLR3 has been shown to recognize double stranded RNA receptor of TLR2. CD14, which is one of the accessory produced by viral replication (Alexopoulou, 2001). In a proteins of TLR4, is also a LRR family protein (Fujihara, recent issue, Liu et al. reported structure of mouse TLR3 2003; Miyake, 2006). CD14 forms a homodimer through bound to dsRNA (Liu, 2008). Overall structure of the interaction between the two C-terminal regions (Kim, TLR3 homodimer induced by dsRNA resembles “m” 2005). The CD14 dimer represents the horseshoe-likeshaped complex of the TLR1-TLR2-lipopeptide. The dsRNA structure like TLRs. Surprising ly CD14 has the hydrophobic interacts with both N-terminal (LRRNT and LRR1-3) and pocket in the convex region of the LRRNT and LRR C-terminal (LRR19-21) site on the glycan-free convex modules similar with the ligand binding pocket of TLR1- region of each TLR3. The positively charged residues of TLR2 complex. This amino terminal pocket of CD14 is both termini of mTLR3 are the major determinants for predicted to be the LPS binding pocket of CD14. It has interaction with the sugar-phosphate backbones of dsRNA. been reported that CD14 enhances the binding efficiency of Interestingly, while the glycan-free surface is essential for lipoprotein to the TLR1-TLR2 complex (Iwaki, 2005; recognition of dsRNA, the N-glycosyl moiety of Asn413 Manukyan, 2005). However , there are controversies whether located on the concave surface of mTLR3 extends toward CD14 is necessary for ligand binding as a co-receptor or the dsRNA and directly contacts it. The protein-protein ANIMAL CELLS AND SYSTEMS Vo l. 13 No. 1 3 Mi Sun Jin and Jie-O h Lee interactions occur only at the LRRCT modules of TLR3 in (Gibbard, 2006). This result agrees well with insertion the TLR3-dsRNA complex, bringing the C-termini of specificity model proposed by Bell et al (Bell, 2003). TLR3 homodimer into contact. Collectively, like TLR3, LRR motives with inserted Regardless of apparent similarity of the overall shape, sequence located at the glycan-free surface of TLR7-9 the TLR3 homodimer has the completely different ligandappears to be the sites for ligand binding. binding mode compared to that of the TLR1-TLR2 heterodimer (Jin, 2007; Liu, 2008). Although ligand Structure of the TLR4-MD-2-ligand complex binding pocket in TLR1-TLR2 complex is located at theTLR4 in company with its co-receptor MD-2 is responsible boundary between the central and the C-terminal domains, for LPS recognition (Shimazu, 1999; Viriyakosol, 2001). dsRNA interaction sites in TLR3 are placed in evolutionally LPS is a major component of the outermembrane of Gram- well-conserved N- and C-termini. Furthermore, TLR3 negative bacteria (Erridge, 2002). They are composed of interacts with the sugar-phosphate backbones by mainly lipid A, core and O-antigen sugars. Previous research charge interactions, not by hydrophobic interactions as demonstrated that the lipid A region with 5~7 acyl chains shown in TLR1-TLR2 heterodimer. The protein-protein and phosphorylated di-glucosamine is sufficient for interaction of the TLR1-TLR2 complex is localized near immunological toxicity of LPS (Tanamoto and Azumi, the ligand binding pocket, but that of TLR3 homodimer is 2000). In an article by Kim et al., it has been reported that separated far from the main ligand binding site of the one molecule of mouse MD-2 interacts with mouse TLR4 TLR3-dsRNA complex. at the concave surface provided by the N-terminal and the Other intracellularly localized TLRs, TLR7, TLR8 and central domains (Kim, 2007b). MD-2 adopts a β cup fold TLR9 interact with nucleic acid-derived ligands as wellwith two antiparallel β-sheets. The LPS binding pockets are (Akira and Takeda, 2004; Gay and Gangloff, 2007; West, generated by separating one side of the β-sandwich and 2006). Their ligand bound structures can be predicted based exposing the hydrophobic core residues. The LPS binding on the structure of the TLR3-ligand complex because they pocket of the MD-2 is lined with hydrophobic residues, share substantial sequential and functional similarities except the opening region that h as many positiv ely char ged (Matsushima, 2007). As previously proposed, the glycan- residues. Therefore, the overall shape and electrostatic free surface of TLR3 is a critical requisite for dsRNA behavior of MD-2 seem to be suitable for binding of the binding (Liu, 2008). TLR3 has the fifteen N-linked amphipathic and negatively charged ligand such as LPS. glycosylation sites. Among them, four glycosylations of theThe interaction between TLR4 and MD-2 is formed mainly inner concave surface seem to inhibit binding of dsRNA. by hydrogen bonds in two opposite char ged patches named All TLR7, TLR8 and TLR9 also have large number of A and B patches. The negatively char ged and evolutionarily potential N-linked glycosylation sites, ranging from 13 to well conserved A patch of TLR4 interacts with the basic 21. There are four or five glycosylation sites in the innerresidues in MD-2. On the other hand, the positively charged concave of TLR7-9 that can interfere with ligand binding. B patch is located in less conserved area of TLR4 and To predict existence of glycan-free surface in these TLRs,interacts with acidic residues in MD-2. structure of human TLR7 was homology-modeled usingKim et al. successfully crystallized human TLR4-VLR the human TLR3 structure as a template. The model shows hybrid complexed with MD-2 bound to its antagonistic that TLR7 also contains the glycan-free surface like TLR3. ligand Eritoran using a novel technique termed the “Hybrid Potential N-glycosylation sites of TLR8 and TLR9 are well LRR technique” (Jin and Lee, 2008; Kim, 2007b). Eritoran con serv ed with th ose of TLR3 . The plausib le glyco sylation containing four acyl chains is a lipid A analog and a potent analysis strongly suggests that ligand binding site in TLR7- antagonist of TLR4. Crystallographic analysis of TLR4- 9 would be the face without any glycosylations in the ligand complexes has been hampered by several technical con vex r e gio n. T h e L R R20 mo d ule w i t h sh or t i n s e r t s e quen ce difficulties. Simple truncation of non essential parts of of TLR3 is one of the sites essential for interaction with protein often improve crystallization behavior of many dsRNA (Liu, 2008). Basic residues in LRR20 protruded proteins but this method is not inappropriate for TLR from the convex region directly bind sugar-phosphate family because they are composed of continuous LRR backbone of dsRNA. Surprisingly, all eighth LRRs of modules. Simple removal of the LRRNT or LRRCT TLR7, TLR8 and TLR9 contain nonconsensus LRR modules protecting the hydrophobic cores of the TLR modules with additional 16 amino acid residues, which family could expose the hydrophobic core to make the include two cysteine-rich “CXXC” box (Matsushima, structure of TLR very unstable. Hybrid LRR technique 2007). Site-directed mutagenesis conducted by Gibbard et allows fusion of truncated fragments of TLRs with the al. demonstrated that modification of any of these four LRRNT or LRRCT domains from other LRR-containing cysteine residues to alanine completely abolished signaling proteins at the highly conserved “LxxLxLxxN” site (Kim, by TLR8 when activated by both resiquimod and ssRNA 2007b). Variable Lymphocyte Receptors (VLRs) were 4 ANIMAL CELLS AND SYSTEMS Vol. 13 No. 1 T LR structures chosen as fusion partners because all VLR proteins have patch in TLR4. Hydrophilic interaction between MD-2 and canonical LRR modules with almost infinite sequence TLR4 surrounds this core hydrophobic interaction, and diversity (Kim, 2007a). VLRs mediate adaptive immune ionic interaction between the phosphate group of LPS and responses in jawless fish and have almost unlimitedTLR4-MD-2 further stabilizes the TLR4 dimer. The TLR4- sequence diversity (Pancer and Cooper, 2006). Among MD-2-LPS dimer shows the “m” shaped architecture, fifteen constructs of TLR4-VLR hybrids generated, seven which further supports the hypothesis that all TLR family of them were successfully expressed as soluble proteins. In receptors have similar “m” shaped dimeric structure when our laboratory TLR1, TLR2, TLR5, TLR7, TLR8 hybridsbound to agonistic ligands. with VLR fragments were produced with the similar or better rate of success (unpublished data). VLR fusion did LIG A N D -IN D U C ED D I ME RIZA TIO N AN D ACTI V A T IO N not change native structure of TLR and the TLR-VLR OF TLR hybrids retain the function. It is likely that hybrid proteins failed producing soluble and folded proteins probably have P rev io us c r ystal lo grap hi c stu di es de mo n s t rate d th at ho mo d i me r serious collision of side chains at the boundary of fusion or heterodimer of TLRs can be induced by binding of site. Although VLRs are shown to be a successful fusion agonistic ligands (Jin, 2007; Kim, 2007b; Liu, 2008; Park, partner for TLRs, it is yet not clear if VLR is the best fusion 2009). A model of the TLR activation mechanism is partner for other LRR family proteins. More research is proposed based on the following structural and biological required to find optimal fusion partner for LRR hybridization. observations (Fig. 1). (1) It has been known that several Eritoran binds to the hydrophobic pocket in MD-2 in a TLRs form homo- or hetero-multimer before binding of fashion that acyl chains occupy most of the MD-2 pockettheir ligands (Akira and Takeda, 2004; Latz, 2007; Ozinsky, ( K i m, 2 007b ) . T he ma i n hy d r opho bic int er ac ti on is r e inf or c e d 2000; Triantafilou, 2006). These pre-ligand complexes by charge interaction between the ligand and positively cannot induce the signaling cascade for immune responses charged residues in MD-2. There is no direct interaction presumably because the intracellular TIR domains have between Eritoran and TLR4. Ohto et al reported crystal unproductive orientation for signaling. (2) Bindings of structure of human MD-2 with lipid IV a (Ohto, 2007). The ligands to TLRs do not result in major conformational tetraacylated lipid IV a is a derivative of Lipid A and acts as changes (Jin, 2007; Liu, 2008; Park, 2009). Overall an antagonist in human cells like Eritoran (Kusumoto, structures of TLRs in the TLR1-TLR2 heterodimer, TLR3 2003; Mullarkey, 2003). Although their chemical structuresand TLR4-MD-2 homodimers are nearly identical with that have large difference, Eritoran and Lipid IV a show highly of the monomers. (3) The two TIR domains should be homologous structure when bound to MD-2 (Kim, 2007b;brought into close proximity in the ligand induced homo- or Ohto, 2007). As noted, both antagonistic Eritoran and lipid heterodimer of TLRs. In TLR1-TLR2 complex, the C- IV a have four lipid chains although agonistic LPS contains termini of the extracellular domains of TLR1 and TLR2 more or longer lipid chains. Since lipid chains of Eritoran conver g e in the middle (Jin, 2007). Similarly , the C-termini or lipid IV occupy almost all the available volume in MD-2 of the extracellular do mains o f TLR3 are only ~7 Å ap art in pocket, LPS is proposed to bind TLR4-MD-2 complex with the ligand induced homodimer (Liu, 2008). (4) Sequence different structure. It is thus conceivable that LPS may be alignment shows that there are only a few residues able to able to make the MD-2 pocket larger because one side of act as a long and flexible linker between the extracellular MD-2 pocket is opened with no any disulfide bonds. and transmembrane domain or between the transmembrane Alternatively, four chains of LPS are inserted into MD-2and intracellular domain (Matsushima, 2007). Hence the C- pocket and the remaining two chains protrude from MD-2 termini of TLRs into close apposition appear to induce pocket for interaction with another TLR4. Both proposed juxtaposed intracellular TIR domains for signaling. Above binding modes of LPS to MD-2 pocket need to provoke the results provide clues to signaling mechanism via TLRs. structural changes of MD-2 that should play important roles TLRs without their ligands already exist as the preformed for dimerization of the TLR4-MD-2. and weakly bound homo- or heterodimers in the cellular Recently reported structure of the TLR4-MD-2-LPS membrane (Fig. 1). The cytoplasmic TIR domains of these complex supports the second hypothesis that the LPS complexes are inactive because the distance or orientation binding does not change overall size of the MD-2 pocket between the TIR domains is not suitable for signal (Park, 2009). Instead, LPS moves up to the solvent area by initiation. Binding of agonistic ligands to TLRs induce the ~5Å, which generates room enoug h fo r two add itional lipidrearrangement of the extracellular domains and bring the C- chains in the MD-2 pocket. Dimerization of the TLR4-MD- termini of extracellular domains into close distance, which 2 complex is mediated by several interactions. The major leads to juxtaposition of the TIR domains for signal contribution is derived from hydrophobic interactioninitiation. between the sixth lipid chain of LPS and a hydrophobic In so m e T LRs, it h as b e e n pro po s e d th at th e co nform at io nal ANIMAL CELLS AND SYSTEMS Vo l. 13 No. 1 5 Mi Sun Jin and Jie-O h Lee Fig. 1. Model of TLR activation by agonistic ligands . The "m"-shaped TLR dimers are induced by binding of their agonistic ligands. S tru cture o f t h e pre- ligand c o mplex is unk n o wn . T h e c r ys t a l s t r u ct ures of T L R 1 -T LR2- PamCSK (lef t), T L R 3 -RN A (middle) and T L R 4 -MD- 2-LPS (r ight) ar e 3 4 drawn as proposed by Jin, Liu and Prak et al. (Jin, 2007; Liu, 2008; Park, 2009). TLR dimers are colored blue and orange, and l igands are colored black, respectively . Asterisks are used to mark the second TLRs or TIRs in the receptor complex. chang es indu ced by ligand s bind ing are critical for receptor FUTURE PERSPECTIVES activation. Latz et al. provided evidences that only stimulatory DNAs lead to dramatic conformational changesIn this review, we summarized current structural understanding in the TLR9 extracellular domain and these alterationsof TLR-ligand interaction and models of receptor activation bring the TIR domains of TLR9 closer for activation (Latz, by the homo- or heterodimeriztion. Future structural 2007). TLR9 has the unusual LRR15 motif with 58 amino research of other TLRs will confirm this hypothesis. The acid residues located in the central part of the extracellular TLR family and its adaptor proteins, MyD88, MAL, TRIF domain (Matsushima, 2007). Long LRR motif appears to and TRAM, interact each other through evolutionary confer flexibility and allows conformational change in the conserved TIR domains (O'Neill L and Bowie, 2007). extracellular domain of TLR9. Like TLR9, both TLR7 and Several structure models of TIR complexes are contradictory TLR8 have the long LRR motifs, containing 73 residues ineach other probably due to lack of high resolution structural TLR7 and 64 in TLR8, in the central parts of the extracellular information (Dunne, 2003; Gautam, 2006; Núñez Miguel, do mains. Therefo re, it appears that co nformational changes 2007). Therefore determining experimental structures of induced by ligand binding controls signaling by TLR7 and the TIR multimers will be crucial for clearer understanding TLR8. Similarly, the conformational change of accessory of signaling mechanism. The Hybrid LRR technique protein, not TLRs, seems to be important for signal proved its usefulness in crystallographic research of the transduction in the TLR4 system.TLR family. This technique can have broader application in 6 ANIMAL CELLS AND SYSTEMS Vol. 13 No. 1 T LR structures si gnal i ng . J Bi ol C h e m 2 81: 3 0132 -3 014 2. structural analysis of other LRR proteins as well as generation of artificial proteins with beneficial therapeutic Gay NJ and Gangloff M (2007) Structure and function of Toll receptors and their ligands. Annu Rev Biochem 76: 1 41- 165 . activities. Gay NJ and Keith FJ (1991) Drosophila Toll and IL-1 receptor. Na tur e 351: 355-356. ACKNOWLEDGMENTS Gibbard RJ, Morley PJ, and Gay NJ (2006) Conserved features in the extracellular domain of human toll-like receptor 8 are This work was supported by the Creative Research Initiatives essential for pH-dependent signaling. J Biol Chem 281: (Center for Membrane Receptor Research) of MEST and KOSEF. 2 750 3- 2751 1. Han SH, Kim JH, Martin M, Michalek SM, and Nahm MH (2003) Pneumococcal lipoteichoic acid (L T A) is not as potent REFERENCES as staphylococcal LTA in stimulating Toll-like receptor 2. I n f ect I mmu n 71: 5541-5548. Akira S and Takeda K (2004) Toll-like receptor signalling. Nat Hemmi H, Kaisho T, Takeuchi O, Sato S, Sanjo H, Hoshino K, Rev Immunol 4: 499-511. Horiuchi T, Tomizawa H, Takeda K, and Akira S (2002) Alexopoulou L, Holt AC, Medzhitov R, and Flavell RA (2001) Small anti-viral compounds activate immune cells via the Recognition of double-stranded RNA and activation of NF- TLR7 MyD88-dependent signaling pathway . Nat Immunol 3: kappaB by Toll-like receptor 3. Na tur e 413: 732-738. 1 96- 20 0. Bell JK, Botos I, Hall PR, Askins J, Shiloach J, Segal DM, and Hioe CE, Qiu H, Chend PD, Bian Z, Li ML, Li J, Singh M, Davies DR (2005) The molecular structure of the Toll-like K u ebl e r P , McGee P , O' Ha gan D, Z a mb T , Kof f W , Allso pp C, receptor 3 ligand-binding domain. Proc Natl Acad Sci USA Wang CY, and Nixon DF (1996) Comparison of adjuvant 102: 10976-10980. formulations for cytotoxic T cell induction using synthetic Bell JK, Mullen GE, Leifer CA, Mazzoni A, Davies DR, and peptides. Vaccine 14: 412-418. Segal DM (2003) Leucine-rich repeats and pathogen Iwaki D, Nishitani C, Mitsuzawa H, Hy akushima N, Sano H, and recognition in Toll-like receptors. Trends Immunol 24: 528- Kuroki Y (2005) The CD14 region spanning amino acids 57- 53 3. 64 is critical for interaction with the extracellular Toll-like Beutner KR, Tyring SK, Trofatter KF, Jr., Douglas JM, Jr., receptor 2 domain. Biochem Biophys Res Commun 328: 173- Spruance S, Owens ML, Fox TL, Hougham AJ, and Schmitt 1 76. KA (1998) Imiquimod, a patient-applied immune-response Jin MS, Kim SE, Heo JY, Lee ME, Kim HM, Paik SG, Lee H, and modifier for treatment of external genital warts. An tim icr o b Lee JO (2007) Crystal structure of the TLR1-TLR2 Agents Chemother 42: 789-794. heterodimer induced by binding of a tri-acylated lipopeptide. Buwitt-Beckmann U, Heine H, Wiesmuller KH, Jung G, Brock R, Cel l 130: 1071-1082. Akira S, and Ulmer AJ (2005) Toll-like receptor 6- Jin MS and Lee JO (2008) Application of hybrid LRR technique independent signaling by diacylated lipopeptides. Eur J to protein crystallization. BMB Rep 41: 353-357. I mmun o l 35: 282-289. Kajava AV (1998) Structural diversity of leucine-rich repeat Choe J, Kelker MS, and Wilson IA (2005) Crystal structure of proteins. J Mo l Bio l 277: 519-527. human toll-like receptor 3 (TLR3) ectodomain. Sc ienc e 30 9: Kanzler H, Barrat FJ, Hessel EM, and Coffman RL (2007) 58 1- 585. Therapeutic targeting of innate immunity with Toll-like Deininger S, Stadelmaier A, von Aulock S, Morath S, Schmidt receptor agonists and antagonists. Nat Med 13: 552-559. RR, and Hartung T (2003) Definition of structural prerequisites Kim HM, Oh SC, Lim KJ, Kasamatsu J, Heo JY, Park BS, Lee H, for lipoteichoic acid-inducible cytokine induction by synthetic Y o o OJ, Kasah ara M, a nd L ee JO ( 20 07a) S tr uctu ra l di ver sity de ri vati ves. J Immunol 170: 4134-4138. of the hagfish variable lymphocyte receptors. J Biol Chem Dunne A, Ejdeback M, Ludidi PL, O'Neill LA, and Gay NJ 282: 6726-6732. (2003) Structural complementarity of Toll/interleukin-1 receptor Kim HM, Park BS, Kim JI, Kim SE, Lee J, Oh SC, Enkhbayar P , domains in Toll-like receptors and the adaptors Mal and Matsushima N, Lee H, Y oo OJ, and Lee JO (2007b) Crystal MyD88. J Bio l C h e m 2 78: 4 144 3- 4145 1. structure of the TLR4-MD-2 complex with bound endotoxin Echchannaoui H, Frei K, Schnell C, Leib SL, Zimmerli W , and antagonist Eritoran. Cell 130: 906-917. Landmann R (2002) Toll-like receptor 2-deficient mice are Kim JI, Lee CJ, Jin MS, Lee CH, Paik SG , Lee H, and Lee JO highly susceptible to Streptococcus pneumoniae meningitis (2005) Crystal structure of CD14 and its implications for because of reduced bacterial clearing and enhanced lipopolysaccharide signaling. J Biol Chem 280: 11347-11351. in fl amma ti on. J Infect Dis 186: 798-806. Kobe B and Deisenhofer J (1994) The leucine-rich repeat: a Erridge C, Bennett-Guerrero E, and Poxton IR (2002) Structure versatile binding motif. Trends Biochem Sci 19: 415-421. and function of lipopolysaccharides. Microbes Infect 4: 837- 85 1. Kobe B and Deisenhofer J (1995) Proteins with leucine-rich r e p eats. Curr O p in S t ru ct B i ol 5: 409-416. Fujihara M, Muroi M, Tanamoto K, Suzuki T, Azuma H, and Ikeda H (2003) Molecular mechanisms of macrophage Kobe B and Kajava AV (2001) The leucine-rich repeat as a activation and deactivation by lipopolysaccharide: roles of theprotein recognition motif. Curr Opin Struct Biol 11: 725-732. receptor complex. P harma col T h er 100: 171-194. Kusumoto S, Fukase K, Fukase Y , Kataoka M, Y oshizaki H, Sato Gautam JK, Ashish, Comeau LD, Krueger JK, and Smith MF , Jr . K, Oikawa M, and Suda Y (2003) Structural basis for (2006) Structural and functional evidence for the role of the e ndot oxi c an d ant ago nisti c act iv iti es: i nvest igat io n wi th n ovel TLR2 DD loop in TLR1/TLR2 heterodimerization andsynthetic lipid A analogs. J Endotoxin Res 9: 361-366. ANIMAL CELLS AND SYSTEMS Vo l. 13 No. 1 7 Mi Sun Jin and Jie-O h Lee Latz E, Verma A, Visintin A, Gong M, Sirois CM, Klein DC, Ozinsky A, Underhill DM, Fontenot JD, Hajjar AM, Smith KD, Monks BG, McKnight CJ, Lamphier MS, Duprex WP, Wilson CB, Schroeder L, and Aderem A (2000) The Espevik T, and Golenbock DT (2007) Ligand-induced repertoire for pattern recognition of pathogens by the innate conformational changes allosterically activate Toll-like immune system is defined by cooperation between toll-like receptor 9. Nat Immunol 8: 772-779. r ece pto r s. Proc Natl Acad Sci U S A 97: 1 376 6- 1377 1. Liu L, Botos I, W ang Y , Leonard JN, Shiloach J, Segal DM, and Pancer Z and Cooper MD (2006) The evolution of adaptive Davies DR (2008) Structural basis of toll-like receptor 3immunity. A nnu R ev I mmu nol 24: 497-518. signaling with double-stranded RNA. Science 320: 379-381. Par k BS , So ng D H , Kim HM, Choi BS, L ee H, and Le e JO ( 2 0 09) Manukyan M, Triantafilou K, Triantafilou M, Mackie A, Nilsen The structural basis of lipopolysaccharide recognition by the N, E sp e vi k T , W ie smu ll er K H , U lme r AJ, and Hei n e H ( 2005 ) T L R 4- MD -2 c o mp l e x. Natu r e. Binding of lipopeptide to CD14 induces physical proximity of Romagne F (2007) Current and future drugs targeting one class of CD14, TLR2 and TLR1. Eur J Immunol 35: 911-921. innate immunity receptors: the Toll-like receptors. Drug Matsushima N, Tanaka T, Enkhbayar P, Mikami T, Taga M,Discov Today 12: 80-87. Yamada K, and Kuroki Y (2007) Comparative sequence Schroder NW, Morath S, Alexander C, Hamann L, Hartung T, analysis of leucine-rich repeats (LRRs) within vertebrate toll- Zahringer U, Gobel UB, Weber JR, and Schumann RR (2003) like receptors. B M C G e nomi c s 8: 124. Lipoteichoic acid (LTA) of Streptococcus pneumoniae and Medzhitov R, Preston-Hurlburt P , and Janeway CA, Jr . (1997) A Staphylococcus aureus activates immune cells via Toll-like human homologue of the Drosophila Toll protein signals receptor (TLR)-2, lipopolysaccharide-binding protein (LBP), activation of adaptive immunity. Nat u r e 388: 394-397.and CD14, whereas TLR-4 and MD-2 are not involved. J Biol Ch em 278: 15587-15594. Miyake K (2003) Innate recognition of lipopolysaccharide by CD14 and toll-like receptor 4-MD-2: unique roles for MD-2. Shimazu R, Akashi S, Ogata H, Nagai Y , Fukudome K, Miyake Int Immunopharmacol 3: 1 19- 128 . K, and Kimoto M (1999) MD-2, a molecule that confers lipopolysaccharide responsiveness on Toll-like receptor 4. J Miyake K (2006) Roles for accessory molecules in microbial Ex p Me d 1 89: 1 777- 17 82. recognition by Toll-like receptors. J Endotoxin Res 12: 195- 20 4. Silverstein AM (2003) Darwinism and immunology: from Met c hni kof f t o Bur n et . Nat Immunol 4: 3-6. Morath S, Geyer A, and Hartung T (2001) Structure-function relationship of cytokine induction by lipoteichoic acid from Takeuchi O, Kawai T, Muhlradt PF, Morr M, Radolf JD, Staphylococcus aureus. J Exp Med 193: 393-397. Zychlinsky A, T akeda K, and Akira S (2001) Discrimination of bacterial lipoproteins by T oll-like receptor 6. Int Immunol Mullarkey M, Rose JR, Bristol J, Kawata T, Kimura A, 13: 933-940. Ko bay a shi S, Pr zeta k M, Chow J, G u sovsky F , Chr ist W J, and Rossignol DP (2003) Inhibition of endotoxin response by T akeuchi O, Sato S, Horiuchi T , Hoshino K, T akeda K, Dong Z, e5564, a novel Toll-like receptor 4-directed endotoxinModlin RL, and Akira S (2002) Cutting edge: role of Toll-like an tago nist . J Pharmacol Exp Ther 304: 1093-1102. receptor 1 in mediating immune response to microbial l i p opr ote i ns. J Immunol 169: 10-14. Nakata T , Y asuda M, Fuj ita M, Kataoka H, Kiura K, Sano H, and Shibata K (2006) CD14 directly binds to triacylated T anamoto K and Azumi S (2000) Salmonella-type heptaacylated lipopeptides and facilitates recognition of the lipopeptides by lipid A is inactive and acts as an antagonist of the receptor complex of Toll-like receptors 2 and 1 without lipopolysaccharide action on human line cells. J Immunol binding to the complex. Cell M i c r ob iol 8: 1899-1909.164: 3149-3156. Núñez Miguel R, Wong J, Westoll JF, Brooks HJ, O'Neill LA, Texereau J, Chiche JD, Taylor W , Choukroun G , Comba B, and Gay NJ, Bryant CE, and Monie TP (2007) A dimer of the Mira JP (2005) The importance of Toll-like receptor 2 Toll-like receptor 4 cytoplasmic domain provides a specific polymorphisms in severe infections. Clin Infect Dis 41 Suppl scaffold for the recruitment of signalling adaptor proteins.7: S408-415. PLoS ONE 2: e788. Triantafilou M, Gamper FG , Haston RM, Mouratis MA, Morath Ogus AC, Yoldas B, Ozdemir T, Uguz A, Olcen S, Keser I, S, Hartung T , and T riantafilou K (2006) Membrane sorting of Coskun M, Cilli A, and Yegin O (2004) The Arg753GLn toll-like receptor (TLR)-2/6 and TLR2/1 heterodimers at the polymorphism of the human toll-like receptor 2 gene in cell surface determines heterotypic associations with CD36 tuberculosis disease. Eur Respir J 23: 219-223.and intracellular targeting. J Biol Chem 281: 31002-31011. Ohto U, Fukase K, Miyake K, and Satow Y (2007) Crystal V iriyakosol S, T obias PS, Kitchens RL, and Kirkland TN (2001) str u ct ur es of human MD- 2 an d it s comp lex w ith ant iend o to xicMD-2 binds to bacterial lipopolysaccharide. J Biol Chem 276: lipid IVa. Sci e nce 316: 1632-1634. 3 804 4- 3805 1. O’Neill LA and Bowie AG (2007) The family of five: TIR- West AP , Koblansky AA, and Ghosh S (2006) Recognition and domain-containing adaptors in Toll-like receptor signalling. signaling by toll-like receptors. Annu Rev Cell Dev Biol 22: Nat R ev I mmu nol 7: 353-364. 4 09- 43 7. O’Neill LA (2004) TLRs: Professor Mechnikov, sit on your hat. [Received February 16, 2009; accepted March 20, 2009] Trends Immunol 25: 687-693. 8 ANIMAL CELLS AND SYSTEMS Vol. 13 No. 1
Journal
Animal Cells and Systems
– Taylor & Francis
Published: Jan 1, 2009
Keywords: innate immune response; leucine rich repeat; pattern recognition receptor; toll‐like receptor; hybrid LRR technique
