Pathogen Recognition: TLRs Throw Us a Curve
2005; Cell Press; Volume: 23; Issue: 4 Linguagem: Inglês
10.1016/j.immuni.2005.09.008
ISSN1097-4180
AutoresPeter Kirk, J. Fernando Bazán,
Tópico(s)Respiratory viral infections research
ResumoToll-like receptors (TLRs) are the archetypal pattern recognition receptors (PRRs) envisioned by Janeway, 1989Janeway C.A. Cold Spring Harb. Symp. Quant. Biol. 1989; 54: 1-13Crossref PubMed Google Scholar as innate sensors of pathogen attack and host triggers of an adaptive immune response. Two recent papers (Choe et al., 2005Choe J. Kelker M.S. Wilson I.A. Science. 2005; 309: 581-585Crossref PubMed Scopus (470) Google Scholar, Bell et al., 2005Bell J.K. Botos I. Hall P.R. Askins J. Shiloach J. Segal D.M. Davies D.R. Proc. Natl. Acad. Sci. USA. 2005; 102: 10976-10980Crossref PubMed Scopus (302) Google Scholar) reveal the distinctive architecture of a TLR sensor domain and hint at how this structural design facilitates the recognition of a wide array of pathogen molecules. Toll-like receptors (TLRs) are the archetypal pattern recognition receptors (PRRs) envisioned by Janeway, 1989Janeway C.A. Cold Spring Harb. Symp. Quant. Biol. 1989; 54: 1-13Crossref PubMed Google Scholar as innate sensors of pathogen attack and host triggers of an adaptive immune response. Two recent papers (Choe et al., 2005Choe J. Kelker M.S. Wilson I.A. Science. 2005; 309: 581-585Crossref PubMed Scopus (470) Google Scholar, Bell et al., 2005Bell J.K. Botos I. Hall P.R. Askins J. Shiloach J. Segal D.M. Davies D.R. Proc. Natl. Acad. Sci. USA. 2005; 102: 10976-10980Crossref PubMed Scopus (302) Google Scholar) reveal the distinctive architecture of a TLR sensor domain and hint at how this structural design facilitates the recognition of a wide array of pathogen molecules. Molecules and mechanisms that perfectly fit the PRR concept have emerged with great force in the past few years, reshaping our view of vertebrate immunology and showing deep evolutionary links with the defensive measures of more primitive immune systems (Flajnik and Du Pasquier, 2004Flajnik M.F. Du Pasquier L. Trends Immunol. 2004; 25: 640-644Abstract Full Text Full Text PDF PubMed Scopus (194) Google Scholar, O'Neill, 2004O'Neill L.A. Trends Immunol. 2004; 25: 687-693Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). Chief among these elucidated systems is the Toll pathway of Drosophila, first studied for its role in embryonic patterning, then unmasked in the adult fly as a critical component of host defense against infection by fungi and gram-positive bacteria (Roach et al., 2005Roach J.C. Glusman G. Rowen L. Kaur A. Purcell M.K. Smith K.D. Hood L.E. Aderem A. Proc. Natl. Acad. Sci. USA. 2005; 102: 9577-9582Crossref PubMed Scopus (874) Google Scholar). And yet, the fly Toll receptor is defiantly not a PRR. Instead, upstream molecules perform the recognition task and thereupon activate proteolytic cascades in the hemolymph that cleave the secreted cytokine Spaetzle to a form that binds Toll and triggers signaling (Weber et al., 2005Weber A.N.R. Moncrieffe M.C. Gangloff M. Imler J.-L. Gay N.J. J. Biol. Chem. 2005; 280: 22793-22799Crossref PubMed Scopus (55) Google Scholar). The discovery of genes encoding eight additional Toll-like receptors in the fly genome—along with seven extra Spaetzle-like ligands—fueled the idea of a broader defensive shield triggered likewise, at a distance, by other fly PRRs. To date, however, the majority of fly Toll paralogs appear to work as developmental regulators, and only Toll9 retains an immune signaling role (Weber et al., 2005Weber A.N.R. Moncrieffe M.C. Gangloff M. Imler J.-L. Gay N.J. J. Biol. Chem. 2005; 280: 22793-22799Crossref PubMed Scopus (55) Google Scholar, Roach et al., 2005Roach J.C. Glusman G. Rowen L. Kaur A. Purcell M.K. Smith K.D. Hood L.E. Aderem A. Proc. Natl. Acad. Sci. USA. 2005; 102: 9577-9582Crossref PubMed Scopus (874) Google Scholar). But the story in humans is very different. Thirteen TLRs have been unearthed from mammalian genomes (ten of which persist in humans; Roach et al., 2005Roach J.C. Glusman G. Rowen L. Kaur A. Purcell M.K. Smith K.D. Hood L.E. Aderem A. Proc. Natl. Acad. Sci. USA. 2005; 102: 9577-9582Crossref PubMed Scopus (874) Google Scholar), and they are all genuine PRRs in the Janeway mold, capable of discriminating among a harrowing array of protozoan, fungal, bacterial, and viral invaders and of signaling alarm into the cell via a core NF-κB pathway (O'Neill, 2004O'Neill L.A. Trends Immunol. 2004; 25: 687-693Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar, Kawai and Akira, 2005Kawai T. Akira S. Curr. Opin. Immunol. 2005; 17: 338-344Crossref PubMed Scopus (448) Google Scholar). How are these immunological feats accomplished? Great strides have been made in elucidating the signaling aspects, and we now understand that there is a dedicated contingent of adaptor molecules (headlined by the peripatetic MyD88) that combinatorially engage TLR cytoplasmic domains and link them to special IRAK family kinases, which in turn drive the NF-κB activation cascade (Kawai and Akira, 2005Kawai T. Akira S. Curr. Opin. Immunol. 2005; 17: 338-344Crossref PubMed Scopus (448) Google Scholar). Although we do know the crystal structure of the characteristic TIR domain (named to mark this region of homology between Toll and Interleukin-1 Receptors; Xu et al., 2000Xu Y. Tao X. Shen B. Horng T. Medzhitov R. Manley J.L. Tong L. Nature. 2000; 408: 111-115Crossref PubMed Scopus (217) Google Scholar) that adorns both TLR intracellular segments and adaptor molecules and dominates their binding, we do not have in hand—yet!—a heteromeric TIR-TIR complex that reveals the structural basis of specificity for the initiating event of the signaling cascade. Still, experimental mapping of TLRs to private or shared adaptors (TRIF, TRAM, TIRAP, and others) is proceeding apace and has importantly led to the discovery of secondary signaling pathways that spring from receptor complexes and mobilize Interferon-regulatory factor (IRF)-class transcription factors alongside (or sometimes, instead of) NF-κB. The convergence of these TLR signals induces a profusion of genes encoding proinflammatory and immune defense cytokines, completing the task of alerting and shaping the adaptive immune response to follow (Kawai and Akira, 2005Kawai T. Akira S. Curr. Opin. Immunol. 2005; 17: 338-344Crossref PubMed Scopus (448) Google Scholar). Now for the recognition part, whose cunning we posit is more biochemical than immunological in nature. From the first announcement that TLR4 was the likely receptor for gram-negative bacterial lipopolysaccharide (LPS) in mice and humans, there has been a growing sense of amazement at how this family of germline-encoded, nonclonal receptors could encompass or evolve the sufficient structural diversity to specifically recognize a broad swath of chemically dissimilar ligands (O'Neill, 2004O'Neill L.A. Trends Immunol. 2004; 25: 687-693Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). The current (and incomplete; see Kawai and Akira, 2005Kawai T. Akira S. Curr. Opin. Immunol. 2005; 17: 338-344Crossref PubMed Scopus (448) Google Scholar) tally includes molecules like bacterial lipopeptides, peptidoglycans, glycolipids, lipoteichoic acid and flagellin, fungal zymosan, viral CpG DNA, single- and double-stranded (ds)RNA, and synthetic or drug-like compounds that mimic natural ligands, like the antiviral imiquimod or the plant-derived taxol. Gene-deletion experiments for 10 of the 12 mouse TLRs have speeded the precise mapping of these pathogen-derived molecular inducers to their respective binding receptors (Kawai and Akira, 2005Kawai T. Akira S. Curr. Opin. Immunol. 2005; 17: 338-344Crossref PubMed Scopus (448) Google Scholar, Roach et al., 2005Roach J.C. Glusman G. Rowen L. Kaur A. Purcell M.K. Smith K.D. Hood L.E. Aderem A. Proc. Natl. Acad. Sci. USA. 2005; 102: 9577-9582Crossref PubMed Scopus (874) Google Scholar). What has not emerged from this arduous process is the clarifying intuition that could reason or even predict a given TLR's ligand binding proclivities. So it is from this perspective that we can appreciate the impact of two papers, one from the Scripps Research Institute (Choe et al., 2005Choe J. Kelker M.S. Wilson I.A. Science. 2005; 309: 581-585Crossref PubMed Scopus (470) Google Scholar) and the other from the National Institutes of Health (NIH) (Bell et al., 2005Bell J.K. Botos I. Hall P.R. Askins J. Shiloach J. Segal D.M. Davies D.R. Proc. Natl. Acad. Sci. USA. 2005; 102: 10976-10980Crossref PubMed Scopus (302) Google Scholar), that independently reveal the crystal structure of the human TLR3 ectodomain with sufficiently rich detail (at 2.1 and 2.4 Å resolution, respectively) that we can perhaps start to make sense of the TLR recognition puzzle. Initial analyses of human TLR sequences suggested the presence of degenerate leucine-rich repeats (LRRs) across the 500–800 amino acid length of their extracellular segments, a defining feature of the fly Toll receptors (Roach et al., 2005Roach J.C. Glusman G. Rowen L. Kaur A. Purcell M.K. Smith K.D. Hood L.E. Aderem A. Proc. Natl. Acad. Sci. USA. 2005; 102: 9577-9582Crossref PubMed Scopus (874) Google Scholar). Molecular models refined the idea that the TLR ectodomains would fold as an uninterrupted solenoidal array of 19–26 LRRs with a pronounced curvature (Bell et al., 2003Bell J.K. Mullen G.E.D. Leifer C.A. Mazzoni A. Davies D.R. Segal D.M. Trends Immunol. 2003; 24: 528-533Abstract Full Text Full Text PDF PubMed Scopus (560) Google Scholar, Weber et al., 2004Weber A.N.R. Morse M.A. Gay N.J. J. Biol. Chem. 2004; 279: 34589-34594Crossref PubMed Scopus (96) Google Scholar). And indeed, that is what the TLR3 crystal structures show: the 670 amino acid ectodomain flaunts a smooth arrangement of 23 LRRs flanked by characteristic cysteine-rich N- and C-terminal capping modules (Choe et al., 2005Choe J. Kelker M.S. Wilson I.A. Science. 2005; 309: 581-585Crossref PubMed Scopus (470) Google Scholar, Bell et al., 2005Bell J.K. Botos I. Hall P.R. Askins J. Shiloach J. Segal D.M. Davies D.R. Proc. Natl. Acad. Sci. USA. 2005; 102: 10976-10980Crossref PubMed Scopus (302) Google Scholar). The curvature of the horseshoe-shaped LRR fold is most reminiscent of the ectodomain structures of the Nogo receptor (He et al., 2003He X.L. Bazan J.F. McDermott G. Park J.B. Wang K. Tessier-Lavigne M. He Z. Garcia K.C. Neuron. 2003; 38: 177-185Abstract Full Text Full Text PDF PubMed Scopus (167) Google Scholar; superposition by Choe et al., 2005Choe J. Kelker M.S. Wilson I.A. Science. 2005; 309: 581-585Crossref PubMed Scopus (470) Google Scholar) and follicle-stimulating hormone receptor (FSHR; Fan and Hendrickson, 2005Fan Q.R. Hendrickson W.A. Nature. 2005; 433: 269-277Crossref PubMed Scopus (442) Google Scholar; our comparison), smaller 10 LRR arrays that share a particular 24 residue repeat motif with TLR3. This LRR type sports economical loops on the outer (convex) face that lace together the inner, parallel β strands to form a concave surface for the array. There are two main protrusions from the regular TLR3 fold, one a radially extended loop from LRR20, and the other a sideways elbow jutting out from LRR12; other TLRs exhibit similar insertions in other repeats (Bell et al., 2003Bell J.K. Mullen G.E.D. Leifer C.A. Mazzoni A. Davies D.R. Segal D.M. Trends Immunol. 2003; 24: 528-533Abstract Full Text Full Text PDF PubMed Scopus (560) Google Scholar, Bell et al., 2005Bell J.K. Botos I. Hall P.R. Askins J. Shiloach J. Segal D.M. Davies D.R. Proc. Natl. Acad. Sci. USA. 2005; 102: 10976-10980Crossref PubMed Scopus (302) Google Scholar). Another layer of complexity is the visible N-glycosylation: both the Scripps and NIH teams solved their TLR3 structures from glycoproteins produced in insect cells, showing 8 and 11 glycan chains, respectively, of 15 possible (Choe et al., 2005Choe J. Kelker M.S. Wilson I.A. Science. 2005; 309: 581-585Crossref PubMed Scopus (470) Google Scholar, Bell et al., 2005Bell J.K. Botos I. Hall P.R. Askins J. Shiloach J. Segal D.M. Davies D.R. Proc. Natl. Acad. Sci. USA. 2005; 102: 10976-10980Crossref PubMed Scopus (302) Google Scholar). An interesting finding is that the inner cavity of the TLR3 horseshoe has quadrants occluded by the glycosylation, and parts of the convex surface are covered as well. Sequence analysis of the glycosylation patterns of other TLRs suggests that this is a general phenomenon with the family (Weber et al., 2004Weber A.N.R. Morse M.A. Gay N.J. J. Biol. Chem. 2004; 279: 34589-34594Crossref PubMed Scopus (96) Google Scholar). So where do these TLR3 structures place the likely binding site for the natural viral ligand, dsRNA, or its mimic, poly(I:C)? The dogma of structures past would immediately flag the cavity created by the concave β sheet: molecular images of ribonuclease inhibitor curled around angiogenin, internalin around an E-cadherin domain, and platelet glycoprotein-IBα grasping a von Willebrand factor A1 module convincingly intimate that LRR arrays preferentially use their inner surfaces to bind target proteins (Bell et al., 2003Bell J.K. Mullen G.E.D. Leifer C.A. Mazzoni A. Davies D.R. Segal D.M. Trends Immunol. 2003; 24: 528-533Abstract Full Text Full Text PDF PubMed Scopus (560) Google Scholar). Even Nogo receptor and FSHR utilize their concave LRR faces to bind their respective ligands (Schimmele and Pluckthün, 2005Schimmele B. Pluckthün A. J. Mol. Biol. 2005; 352: 229-241Crossref PubMed Scopus (31) Google Scholar, Fan and Hendrickson, 2005Fan Q.R. Hendrickson W.A. Nature. 2005; 433: 269-277Crossref PubMed Scopus (442) Google Scholar). But in the present case, have the PRR demands forced a novel solution onto the TLR family? Sequence conservation, hydrophobic analysis, and electrostatic potential mapped to the TLR3 surface paint a flat, glycan-free side face on the C-terminal half of the LRR array (following the LRR12 elbow) as a likely interaction epitope; by contrast, the concave β sheet appears markedly acidic and hence repulsive to nucleic acids. Here is where things get interesting, as both the Scripps and NIH structures capture a symmetric TLR3 dimer in the crystal with a modest buried surface area of ∼1150 Å2 partly encompassing the above interaction site, with locked LRR12 elbows. This forges a novel arrangement of reverse-facing horseshoes that are dimerized at their lower (C-terminal) ends (see Figure 1), parsimoniously recreating a membrane bound tableau that could be physiologically relevant to signaling. This last point is important, because experiments with chimeric receptors that take advantage of heterologous ectodomains (CD4 in the case of TLR4; Medzhitov et al., 1997Medzhitov R. Preston-Hurlburt P. Janeway C.A. Nature. 1997; 388: 394-397Crossref PubMed Scopus (4208) Google Scholar, Ozinsky et al., 2000Ozinsky A. Underhill D.M. Fontenot J.D. Hajjar A.M. Smith K.D. Wilson C.B. Schroeder L. Aderem A. Proc. Natl. Acad. Sci. USA. 2000; 97: 13766-13771Crossref PubMed Scopus (1616) Google Scholar) or intracellular segments (FcγRIIa for TLR3; de Bouteiller et al., 2005de Bouteiller, O., Merck, E., Hasan, U.A., Hubac, S., Benguigui, B., Trinchieri, G., Bates, E.E.M., and Caux, C. (2005). J. Biol. Chem., in press. Published online September 6, 2005. 10.1074/jbc.M507163200.Google Scholar) to respectively drive or measure association argue that TLRs form signaling dimers. This is consistent with our notion that TLRs operate like other cell-surface receptor complexes, with their bases in close proximity on the membrane (Stroud and Wells, 2004Stroud R.M. Wells J.A. Sci. STKE. 2004; 2004: re7PubMed Google Scholar). As promising as the TLR3 dimer appears in first light, is it potentially an artifact of the common crystal form used by both the Scripps and NIH groups to solve their X-ray structures? Alternative packing schemes are embedded in the crystal lattice: for example, both data sets disclose another symmetric TLR3 dimer that centers on a convex epitope around LRR8 and buries an equivalent amount of surface area (∼1170 Å2), but this scheme seems biologically implausible, as the horseshoe C termini are splayed far apart. The present dimer interface shows poor surface complementarity, intrusion of water molecules, and juxtaposed charged residues—arguably features of a transient or forced encounter. Still, we should not necessarily expect an unbound receptor structure to fully recapitulate the ligand-assembled version (Stroud and Wells, 2004Stroud R.M. Wells J.A. Sci. STKE. 2004; 2004: re7PubMed Google Scholar); in the case of TLRs, the critical test at this moment should be how well this unorthodox architectural motif fits the emerging body of data on TLR association (Ozinsky et al., 2000Ozinsky A. Underhill D.M. Fontenot J.D. Hajjar A.M. Smith K.D. Wilson C.B. Schroeder L. Aderem A. Proc. Natl. Acad. Sci. USA. 2000; 97: 13766-13771Crossref PubMed Scopus (1616) Google Scholar, de Bouteiller et al., 2005de Bouteiller, O., Merck, E., Hasan, U.A., Hubac, S., Benguigui, B., Trinchieri, G., Bates, E.E.M., and Caux, C. (2005). J. Biol. Chem., in press. Published online September 6, 2005. 10.1074/jbc.M507163200.Google Scholar), ligand binding epitopes (Mizel et al., 2003Mizel S.B. West A.P. Hantgan R.R. J. Biol. Chem. 2003; 278: 23624-23629Crossref PubMed Scopus (113) Google Scholar, Grabiec et al., 2004Grabiec A. Meng G. Fichte S. Bessler W. Wagner H. Kirschning C.J. J. Biol. Chem. 2004; 279: 48004-48012Crossref PubMed Scopus (59) Google Scholar, Omueti et al., 2005Omueti, K.O., Beyer, J.M., Johnson, C.M., Lyle, E.A., and Tapping, R.I. (2005). J. Biol. Chem., in press. Published online August 29, 2005. 10.1074/jbc.M504320200.Google Scholar), and coreceptor usage (Nishitani et al., 2005Nishitani C. Mitsuzawa H. Hyakushima N. Sano H. Matsushima N. Kuroki Y. Biochem. Biophys. Res. Commun. 2005; 328: 586-590Crossref PubMed Scopus (29) Google Scholar, Jiang et al., 2005Jiang Z. Georgel P. Du X. Shamel L. Sovath S. Mudd S. Huber M. Kalis C. Keck S. Galanos C. et al.Nat. Immunol. 2005; 6: 565-570Crossref PubMed Scopus (500) Google Scholar). This implies a broader utility of the TLR3 monomer (and proposed dimer) fold as a template for the remainder of the receptor family, with the conviction that closely related molecules in sequence and structure tend to conserve their interaction modalities (Aloy et al., 2003Aloy P. Ceulemans H. Stark A. Russell R.B. J. Mol. Biol. 2003; 332: 989-998Crossref PubMed Scopus (251) Google Scholar). While the NIH group contemplates a binding groove for dsRNA within the inner cavity of the TLR3 monomer (threading the nucleotide backbone through the positions of two bound sulfate ions; Bell et al., 2005Bell J.K. Botos I. Hall P.R. Askins J. Shiloach J. Segal D.M. Davies D.R. Proc. Natl. Acad. Sci. USA. 2005; 102: 10976-10980Crossref PubMed Scopus (302) Google Scholar), the most intriguing proposal—voiced more strongly by the Scripps contingent (Choe et al., 2005Choe J. Kelker M.S. Wilson I.A. Science. 2005; 309: 581-585Crossref PubMed Scopus (470) Google Scholar)—is that the TLR3 binding site for dsRNA is decidedly not on the concave LRR surface. Instead, the ligand interaction site is suggested to lie in the V-shaped valley between dimerized structures, bridging basic patches of amino acids partly occluded by the present dimer contact (Choe et al., 2005Choe J. Kelker M.S. Wilson I.A. Science. 2005; 309: 581-585Crossref PubMed Scopus (470) Google Scholar, Bell et al., 2005Bell J.K. Botos I. Hall P.R. Askins J. Shiloach J. Segal D.M. Davies D.R. Proc. Natl. Acad. Sci. USA. 2005; 102: 10976-10980Crossref PubMed Scopus (302) Google Scholar). This simple solution may offer a molecular explanation for the varied PRR activities of TLRs, as these receptors could be utilizing the more divergent loops, variable in sequence and length, on the convex top and flat sides of the iconic horseshoe-shaped fold to create diversity—and allowing different receptor (homo- or hetero-) dimers to bind completely different pathogen ligands. A corollary of this proposal is that the sequence determinants responsible for ligand recognition and receptor oligomerization likely reside in the C-terminal half of the LRR arrays of TLRs, leaving the N-terminal hook of the LRR fold free to potentially interact with accessory molecules or coreceptors. Famously, for the best-studied TLR4-LPS signaling complex, this offers sites for interaction with another LRR array coreceptor, CD14, and MD2, an obligate accessory molecule that solubilizes LPS (Nishitani et al., 2005Nishitani C. Mitsuzawa H. Hyakushima N. Sano H. Matsushima N. Kuroki Y. Biochem. Biophys. Res. Commun. 2005; 328: 586-590Crossref PubMed Scopus (29) Google Scholar, Jiang et al., 2005Jiang Z. Georgel P. Du X. Shamel L. Sovath S. Mudd S. Huber M. Kalis C. Keck S. Galanos C. et al.Nat. Immunol. 2005; 6: 565-570Crossref PubMed Scopus (500) Google Scholar). This structural scheme is perhaps evolutionarily diverged from the recognition of Spaetzle dimers by Drosophila Toll, whereby the receptors form an inactive pair in the absence of ligand (utilizing their C-terminal LRR halves) but rearrange to sequentially bind the cytokine dimer (with their N-terminal LRR arrays) with a final stoichiometry of 2:2 receptor:ligand subunits (Weber et al., 2005Weber A.N.R. Moncrieffe M.C. Gangloff M. Imler J.-L. Gay N.J. J. Biol. Chem. 2005; 280: 22793-22799Crossref PubMed Scopus (55) Google Scholar). In the absence of Spaetzle equivalents in mammals, TLRs may have coopted the inactive state of the Toll dimer as a starting pose for pathogen recognition with their convex surfaces (the arrangement captured by the TLR3 structures); subsequent rearrangements that pivot open the current dimer (for example, allowing dsRNA access to the basic patch above the contact point) could likewise favor N-terminal LRR interactions with host accessory proteins. TLR3 is a member of a specialized subset of TLRs that recognize unmethylated viral nucleic acids in acidic endosomal compartments (Kawai and Akira, 2005Kawai T. Akira S. Curr. Opin. Immunol. 2005; 17: 338-344Crossref PubMed Scopus (448) Google Scholar, Kariko et al., 2005Kariko K. Buckstein M. Ni H. Weissman D. Immunity. 2005; 23: 165-175Abstract Full Text Full Text PDF PubMed Scopus (831) Google Scholar). Of these receptors, persuasive measurements have been made of the direct binding of poly(I:C) to TLR3 (de Bouteiller et al., 2005de Bouteiller, O., Merck, E., Hasan, U.A., Hubac, S., Benguigui, B., Trinchieri, G., Bates, E.E.M., and Caux, C. (2005). J. Biol. Chem., in press. Published online September 6, 2005. 10.1074/jbc.M507163200.Google Scholar) and CpG oligodeoxyribonucleotides to TLR9 (Cornelie et al., 2004Cornelie S. Hoebeke J. Schacht A.-M. Bertin B. Vicogne J. Capron M. Riveau G. J. Biol. Chem. 2004; 279: 15124-15129Crossref PubMed Scopus (92) Google Scholar, Rutz et al., 2004Rutz M. Metzger J. Gellert T. Luppa P. Lipford G.B. Wagner H. Bauer S. Eur. J. Immunol. 2004; 34: 2541-2550Crossref PubMed Scopus (434) Google Scholar). In the present case, the Scripps and NIH groups, respectively, show intriguing glimpses at data of poly(I:C) and dsRNA binding to their purified TLR3 ectodomains (Choe et al., 2005Choe J. Kelker M.S. Wilson I.A. Science. 2005; 309: 581-585Crossref PubMed Scopus (470) Google Scholar, Bell et al., 2005Bell J.K. Botos I. Hall P.R. Askins J. Shiloach J. Segal D.M. Davies D.R. Proc. Natl. Acad. Sci. USA. 2005; 102: 10976-10980Crossref PubMed Scopus (302) Google Scholar), so in coming months we may hope to view an exact solution to this TLR recognition puzzle. Completing the circle, ongoing biochemical and crystallographic studies of the Toll-Spaetzle complex by a group of Cambridge investigators (Weber et al., 2005Weber A.N.R. Moncrieffe M.C. Gangloff M. Imler J.-L. Gay N.J. J. Biol. Chem. 2005; 280: 22793-22799Crossref PubMed Scopus (55) Google Scholar) may also offer a glimpse into the wholesale evolutionary reformation of TLRs from cytokine receptors to true PRRs. There are other LRR sensor proteins in our innate immune arsenal (Flajnik and Du Pasquier, 2004Flajnik M.F. Du Pasquier L. Trends Immunol. 2004; 25: 640-644Abstract Full Text Full Text PDF PubMed Scopus (194) Google Scholar), notably the intracellular NLRs (for NACHT-LRRs, which collectively group the NOD and NALP families, vertebrate counterparts to plant Resistance proteins) that may represent new PRR clans (Martinon and Tschopp, 2005Martinon F. Tschopp J. Trends Immunol. 2005; 26: 447-454Abstract Full Text Full Text PDF PubMed Scopus (493) Google Scholar). Taking cues from the TLR3 complex structure, it may be useful to consider the unorthodox TLR dimer architecture as a possible solution for other vexing LRR recognition problems in nature; in addition, protein design efforts that are aimed at remodeling the inner, concave surface of LRR arrays as all-purpose binding platforms (Schimmele and Pluckthün, 2005Schimmele B. Pluckthün A. J. Mol. Biol. 2005; 352: 229-241Crossref PubMed Scopus (31) Google Scholar) may broaden their guiding principle of a single, evolutionarily malleable face to LRRs. A parallel, oft-neglected family of innate immune PRRs are the peptidoglycan recognition proteins (PGRPs), that, among other things, sense infection by gram-negative bacteria and trigger the Immune Deficiency (Imd) defense pathway in Drosophila (Flajnik and Du Pasquier, 2004Flajnik M.F. Du Pasquier L. Trends Immunol. 2004; 25: 640-644Abstract Full Text Full Text PDF PubMed Scopus (194) Google Scholar). Crystallographic studies of human and fly PGRP sensor domains directly in complex with peptidoglycan analogs are illuminating another primordial mechanism of pathogen recognition and ligand-driven PRR signaling (Chang et al., 2005Chang C.-I. Ihara K. Chelliah Y. Mengin-Lecreulx D. Wakatsuki S. Deisenhofer J. Proc. Natl. Acad. Sci. USA. 2005; 102: 10279-10284Crossref PubMed Scopus (70) Google Scholar). Together, the developing stories of TLRs and PGRPs attest to the magnificent insight and immunological impact of Charles Janeway's PRRs (Janeway, 1989Janeway C.A. Cold Spring Harb. Symp. Quant. Biol. 1989; 54: 1-13Crossref PubMed Google Scholar). We thank S.-A. Kellermann, C. Kirk, numerous colleagues, and an anonymous reviewer for critical comments and suggestions. Our apologies in advance to the authors of papers we have failed to mention for the sake of brevity.
Referência(s)