| Innate immunity is an ancient host defense system that is widely distributed among organisms and serves as the first-line defenses against pathogen infections. Toll-like receptors (TLRs) are important pattern recognition receptors that play a crucial role in innate immunity in vertebrates. Toll-like receptors specifically recognize and bind conserved pathogen-associated molecular patterns (PAMPs) in pathogens to initiate an innate immune response and to prime the adaptive immune system. In contrast to the early concept of innate immunity, the discovery of toll-like receptors showed that pathogen recognition by the innate immune system was actually specific. Functional studies have found that vertebrate TLRs detect very diverse PAMP structures, including lipids, lipoproteins, glycan, proteins, and nucleic acids. However, it is now unclear that how do the vertebrate TLRs develop the diversity of ligand recognition.Toll-like receptors are type I transmembrane glycoproteins that are expressed on the cell surface or in intracellular compartments. They generally consist of an N-terminal extracellular ligand-binding domain, a single transmembrane helix and a C-terminal intracellular toll/interleukin-1 receptor (TIR) domain that mediates signaling. In this work, we investigated the mechanism of diversity ligand recognition of vertebrate Toll-like receptors by using minority known crystal structures and a large number of modeling structures obtained through threading method of TLR ectodomains. We proposed that the architectures of TLR ectodomains affect their ligand recognition specificities based on structural analysis, phylogenetic relationships, comparative genomics, evolutionary analysis, and so on. According to this theory, we further analyzed and predicted the potential ligands for those vertebrate TLRs with unclear functions. The primary contents and results are as follows:1. Ectodomain architecture affects the specificity of TLR pathogen recognitionThousands of members of the TLR multigene family have been identified in vertebrates. They specifically recognize diverse ligands that exhibit remarkably different structures. The TLR ectodomains (ECDs) that is responsible for ligand recognition possess characteristic horseshoe-shaped solenoid structures generated by a varying number of leucine-rich repeat (LRR) modules. The continuous hydrogen-bond network (asparagine ladder) formed among the asparagine residues on the concave surfaces of neighboring leucine-rich repeat modules assists in stabilizing the overall shape of TLR ectodomains. A structural analysis of the 28 types of major vertebrate TLRs showed that their ectodomains possessed three types of architectures: a single-domain architecture with an intact asparagine ladder, a three-domain architecture with the ladder interrupted in the middle, and a trans-three-domain architecture with the ladder broken in both termini. The 1428 known vertebrate TLRs were defined as the correct TLR types according to the phylogenetic analysis and they can be divided into eight families based on sequence and structural differences. TLR ectodomains in the same family possess the same architecture and recognize similar ligands. Based on a phylogenetic analysis, the three vertebrate TLR architectures arose during early evolution.The ligand specificities of TLRs are also affected by their ectodomain architectures. TLRs with three-domain architecture bind hydrophobic ligands, whereas TLRs with single-domain or trans-three-domain architectures mainly recognize hydrophilic ligands. With few exceptions, TLRs with three-domain architecture mainly recognize PAMPs from bacteria and fungi, whereas the majority of ligands for TLRs with single-domain architecture are derived from viruses. TLRs with trans-three-domain architecture chiefly respond to ligands from protozoan parasites. TLRs with three-domain architecture are mainly expressed on the cell surface, whereas TLRs with single-domain and trans-three-domain architectures are chiefly expressed in intracellular vesicles, such as endosomes. Furthermore, TLRs with three-domain and trans-three-domain architectures have been reported to possess some TLRs that work in the forms of functional heterodimers, whereas all known TLRs with single-domain architecture have been reported to form homodimers to bind ligands. Furthermore, the differences of TLR ectodomain architectures possibly deeply affect the evolution of vertebrates. The comparative analysis of 39 vertebrate genomes suggested that the number of single-domain TLR genes in terrestrial vertebrate genomes decreased by half compared to aquatic vertebrate genomes. The dN/dS values revealed that purifying selection was the dominant force driving vertebrate TLR evolution and that single-domain TLR genes had undergone stronger purifying selective pressures than three-domain TLR genes in mammals.2. Prediction of pathogens recognized by trans-three-domain TLR19Although more than ten years have passed since the first identification, the systematic knowledge about fish-specific TLR19 is still far insufficient, especially, the potential ligands recognized by TLR19 are still unclear. A phylogenetic analysis showed that TLR19 belonged to family 11, and clustered with TLR20 and TLR11/12 on the evolutionary tree. TLR20 is the closest paralogue of TLR19. The ectodomain of TLR19 contains 24 LRR modules. The structural analysis of TLR ectodomains showed that both TLR19 and the other members of family 11 possessed trans-three-domain architecture and the asparagine ladders in their ectodomains were interrupted in both termini. The electrostatic surface potential analysis indicated that the modeled structure of TLR19 ectodomain showed much stronger polarity on the ascending lateral surface than on the descending lateral surface. The ascending lateral surface with strong electrostatic surface potential possibly mainly participates in the ligand binding of TLR19 ectodomain. The trans-three-domain architecture and the electrostatic surface potential property of TLR19 ectodomain are similar to those of TLR11 and TLR12. Therefore, like TLR11 and TLR12, TLR19 also probably recognizes protein ligands from protozoon parasites. Furthermore, the quite small dN/dS value at the TLR19 locus showed that the evolution of TLR19 underwent strong functional constraints. Approximately one third codons in the coding sequence of TLR19 were subjected to significantly negative selection, whereas only 5 codons underwent significantly positive selection.3. Identification of three-domain TLR27 and the prediction for its ligand recognitionA novel Toll-like receptor was identified and named as TLR27 through a phylogenetic analysis using the sequences of all the known vertebrate TLRs in database. TL27 was only found in the three fish species, including coelacanth (Latimeria chalumnae), spotted gar (Lepisosteus oculatus), and elephant shark (Callorhinchus milii). TLR27 belonged to TLR family 1 and clustered with TLR14/18 and TLR25 on the evolutionary tree. The ectodomain of TLR27 is predicted to include 19 LRR modules. Structural modeling showed that the TLR27 ectodomain possessed three-domain architecture. The lack of conserved asparagines on the concave surface of the central subdomain causes a structural transition in the middle of TLR27 ectodomain, forming a distinct hydrophobic pocket at the border between the central subdomain and the C-terminal subdomain. This hydrophobic pocket is exactly located between LRR11 and LRR12. We infer that, like other functionally characterized TLRs in family 1, TLR27 is able to bind the hydrophobic ligands through the hydrophobic pocket in the ectodomain and further respond to pathogens. An evolutionary analysis showed that the dN/dS value at the TLR27 locus was very low. Approximately one quarter of the total number of TLR27 sites are under significant negatively selection pressure, whereas only two sites are under positive selection.4. Analysis of the unique activation mechanism for single-domain TLR15TLR15 is reported to have a unique activation mechanism. Unlike the other TLRs that require binding an extra ligand, the proteolytic cleavage of TLR15 ectodomain by the secreted virulence-associated fungal and bacterial proteases causes TLR15 activation without the requirement of an extra ligand. So far, the structural and evolution characterizations of the unusual TLR are still poorly understood. We identified 57 completed TLR 15 genes through a phylogenetic analysis based on massive whole-genomes recently sequenced. TLR 15 clustered into an individual clade and was closely related to TLRs in family 1 on the phylogenetic tree. Unlike three-domain TLRs in family 1 with the broken asparagine ladders in the middle, the ectodomains of all 57 TLR15s had an intact asparagine ladder, thus TLR15 was considered to be a signal-domain TLR. Therefore, we proposed that TLR15 should be designated to be an individual family, family 15. The amino acid evolutionary conservation analysis found that TLR 15 ectodomain had a highly evolutionarily conserved region on the convex surface of LRR11 module, which probably contains catalytic sites of the proteolytic cleavage of the ectodomain in TLR15 activation process. Furthermore, the protein-protein docking analysis indicated that TLR15 intracellular TIR domains were able to form homodimers, the predicted interaction interface of TIR dimer was formed mainly by residues from the BB-loops and aC-helixes. Although TLR15 mainly underwent purifying selection, we detected 27 sites under positive selection for TLR15,24 of which are located on its ectodomain.Overall, these findings mentioned above will contribute to understand that why the innate immunity system in vertebrates is able to recognize diverse pathogens through Toll-like receptors, and also have implications in seeking the potential ligands for those TLRs with unknown function. Meanwhile, these results can enhance the understanding for the interactions between the ligands from pathogens and Toll-like receptors, also help for the prophylaxis and treatment of the diseases caused by the variations of human TLR genes. |
PDF Full Text Request