All animals critically rely on their immune system for protection against infectious microorganisms and disruption of homeostasis. An animal’s immune response typically starts with receptors that detect a danger signal and subsequently initiate the expression and/or release of effectors. The effectors then clear the source of the danger signal to restore homeostasis. One family of receptors that play a key role in the immune system are the Toll-like receptors (TLRs). TLRs are expressed in endolysosomal compartments or on the surface of cells where they scan the extracellular environment for the presence of molecules from microorganisms or damaged tissues. Upon detection of such molecules, TLRs activate transcription factors that drive the expression of immune system effectors. As TLRs can detect both microbial and endogenous molecules, these receptors are strongly involved in many infectious and non-infectious diseases including auto-immune diseases and cancer. This has marked TLRs as attractive targets for pharmacological modulation. Yet despite the interest in TLRs, multiple aspects of these receptors, for example their evolutionary history, mode of activation and interaction with ligand, are not yet fully understood. The incomplete knowledge about multiple aspects of TLR biology impairs comprehending the precise role that TLRs play in diverse diseases and may be one of the reasons why therapeutics aimed at altering the function of TLRs are not yet effective in mitigating diseases. To improve this situation it is important to gain a deeper understanding of the fundamental principles that underlie the biology of these receptors.
TLRs are present in almost all animals and during animal evolution the TLR gene repertoire has expanded and diversified. This has resulted in a large family of receptors of which the size and composition differ between animal species. New species are formed by an evolutionary force that is driven by the necessity to adapt functions to changing environmental challenges. In the case of TLRs this means that receptors of different species may have undergone species-specific adaptations that alter the receptor’s function to meet a specific challenge. Comparing TLRs of different species therefore holds discriminatory power to discover features of the receptor that are important for its function. In this thesis we have implemented this concept by studying TLR biology using a species-comparative strategy. With this strategy our work has resulted in novel insights in multiple aspects of fundamental TLR biology.
Understanding the evolutionary history of a protein can greatly aid in interpreting its role in the biology of an organism. In Chapter 2 we reviewed the evolution of TLRs. These receptors, and their orthologs called Toll receptors, are ancient as the ancestral receptor originated in the eumetazoan ancestor roughly 600 million years ago. TLRs and Toll receptors have been highly conserved ever since in all deuterostomian (e.g. vertebrates) and protostomian (e.g. insects) animals. In almost all animals, TLRs as well as some of the Toll receptors are involved in the sensing of molecules derived from microorganisms. The ancestral form of these receptors therefore likely evolved to play a role in the immune system. While the general structure of the different TLRs and Toll receptors has remained highly similar, species-specific evolutionary requirements have resulted in large differences in the number of receptor genes present in animals. For example, the nematode C. elegans only has one Toll receptor while the purple sea urchin has more than 250 TLRs encoded in its genome. The expansion and diversification of TLRs throughout animal evolution is partly driven by the microbes that co-evolve with the animal host. Some bacteria have evolved strategies to evade detection by TLRs and thus avoid alerting the host immune system. Such microbial strategies, as well as animal-specific co-evolution with distinct microbes, forces a selective pressure onto TLRs that drives either adaptive evolution, seen in TLRs that detect structures specific to bacteria, or purifying evolution seen in TLRs that detect nucleic acids which are not specific to microorganisms. Finally, in protostomian animals some Toll receptors are involved in for instance embryogenesis. Evidence is accumulating that TLRs in deutrostomian animals, by detecting endogenous ligands, are also involved in physiological processes other than immune responses.
An ancient TLR family member that has remained highly conserved throughout animal evolution is TLR5. In mammals and birds, TLR5 detects the bacterial protein flagellin and in mammals this receptor is important for detecting pathogenic bacteria and shaping the microbial community in the intestine. Prior to this thesis it was not known whether the direct detection of flagellin by TLR5 was conserved among species other than mammals and birds. The common ancestors of mammals and birds were early reptiles and reptiles thus take a central position in vertebrate evolution. Studying reptiles may hence aid in understanding host-microbe co-evolution but the immune system of reptiles is strongly understudied compared to that of other vertebrates. In Chapter 3 we provided the first functional characterization of TLR5, and with that of any TLR, from a reptile. We observed that cells of an iguana lizard activated the NF-κB transcription factor in response to bacterial flagellin. In another reptilian species, the green anole lizard, gene expression of the flagellin receptor TLR5 was demonstrated throughout the body. Functional studies with recombinant anole TLR5 showed that this receptor detects bacterial flagellin and induced activation of NF-κB in reptile and human cells, indicating strong evolutionary conservation of both TLR5 ligand binding and signaling capacity. The anole TLR5 was found to recognize the D1 domain in flagellin, just as human TLR5. Yet, given their long independent evolution we questioned whether anole and human TLR5 had developed a differential sensitivity to bacterial flagellins. While anole TLR5 was similarly sensitive as human TLR5 to flagellin of Salmonella, the reptile receptor proved more sensitive to flagellin of Pseudomonas, which is an opportunistic pathogen to reptiles and humans. These findings indicate that the recognition of a conserved domain in flagellin by TLR5 has remained important for more than 320 million years of vertebrate evolution. In addition, our findings illustrate that a selective pressure exerted by distinct microbes drove different adaptations among reptiles and mammals thereby leading to species-specific recognition of the danger signal flagellin.
The species-specific recognition of flagellin indicates that there are functionally important differences between animals in the flagellin-binding site of their TLR5. Flagellin directly binds to a site in the extracellular domain of TLR5 and for the rational design of strategies to therapeutically target TLR5 in the future it is imperative to understand the exact molecular interactions of flagellin binding to TLR5. A previously published crystal structure of flagellin in complex with TLR5b of the zebrafish is used as a model to understand the interaction of flagellin with TLR5 from other species, including human TLR5. However, activation of zebrafish TLR5b by flagellin had never been reported. In addition, while the zebrafish is a widely used model animal in biomedical science, it is an exception in terms of TLRs as the zebrafish carries several additional copies of TLR genes including two tlr5 genes. The role of these two zebrafish TLR5 paralogs in flagellin detection was unknown. In Chapter 4 we discovered that the zebrafish TLR5 paralogs TLR5a and TLR5b, unlike TLR5 of any other species, do not detect flagellin as conventional homodimers but instead evolved to physically cooperate and form a heterodimeric flagellin receptor. We also identified that the TLR chaperone UNC93B1 of the zebrafish strongly contributed to TLR5 heterodimerization by facilitating transport of both receptor paralogs to intracellular vesicles. To better understand what heterodimerization of zebrafish TLR5 meant for the receptor interaction with flagellin, we performed a detailed functional analysis using chimeric receptors based on the TLR5b crystal structure. This analysis showed that there are subtle but functionally important differences in the structure of TLR5a, TLR5b and human TLR5 which cannot be explained using the TLR5b crystal structure as a model. Further analysis with the use of constructed chimeric receptors revealed that to be functional, zebrafish TLR5a and TLR5b must complement each other across multiple regions of the dimeric complex suggesting that a transregional conformational change, possibly via rotation, underlies the TLR5 activational mechanism.
Within the structure of each TLR multiple distinct regions can be identified and for most regions a function has been assigned. The TLR extracellular domain binds ligands, the transmembrane region is necessary for embedment in the membrane and involved in receptor dimerization, and the TIR domain facilitates signal induction. In Chapter 5 we discovered that the C-terminal tail region of human TLR5, a region without prior known function, is necessary for receptor localization and ligand-induced signaling. TLRs are localized to different cellular compartments but the receptor features that direct transport towards these compartments are poorly defined. When we removed the C-terminal tail of human TLR5 the receptor no longer localized at the plasma membrane and did not respond to flagellin stimulation. Surprisingly, the somewhat evolutionarily conserved charged amino acids in the TLR5 tail were not involved in receptor localization and function. To determine whether potential localization motifs were hidden in the tail, we randomly scrambled its sequence which resulted in the blocking of receptor trafficking and function, indicating that the role of the tail was sequence dependent. However, replacement of the human TLR5 tail with the tail from zebrafish TLR5b, which is also a considerably different sequence, still enabled the chimeric receptor to reach the plasma membrane and respond to flagellin. Interestingly, both the human and zebrafish TLR5 tail sequence were predicted to be phosphorylated at threonine residues and scrambling of the human TLR5 tail sequence weakened this prediction. These novel findings reveal a critical contribution of the TLR5 tail region to receptor localization and function and point towards evolutionarily conserved threonine phosphorylation as a potential mechanism.
The conservation of TLR5 in almost all vertebrate animals indicates that TLR5 evolved very stably and that the detection of flagellin by TLR5 has likely remained an important feature throughout vertebrate evolution. Not all TLRs evolved as stably as TLR5. Members of the TLR1 subfamily evolved much more dynamically across vertebrates and show losses or duplications of TLR genes in diverse clades of animals. TLR15 is a member of the TLR1 subfamily that was previously found to be exclusively present in birds and reptiles. Yet, the evolutionary history of this receptor and whether its function as a receptor for microbial proteases remained conserved between birds and reptiles was still unclear. Chapter 6 describes our investigation of the functional evolution of avian and reptilian TLR15. After analyzing a large collection of TLR sequences from diverse vertebrate species, we were surprised to find a TLR15 ortholog to be present in a shark species. This finding reversed the understanding of TLR15 evolution as it indicated that TLR15 is actually an ancient TLR and not an invention in the bird and reptile lineage. Although ancient, the function of TLR15 was likely redundant multiple times throughout evolution as we found that the tlr15 gene had been lost from the genomes of many vertebrates, including turtles. Still, functional analysis with the TLR15 of a lizard and two crocodile species showed that these receptors detected fungal proteolytic activity, just like the chicken TLR15 ortholog, indicating conservation of TLR15 function through more than 280 million years of evolution. We observed a peculiar difference in protein expression efficiency among the functional reptile and chicken TLR15s which revealed another interesting feature about TLR15 evolution; the codon usage among the unstably evolving tlr15 genes was highly variable, much more than among tlr genes that show a more stable evolution such as TLR5. Combined, these findings indicate that TLR15 evolved far more dynamically than most TLRs and that species-specific codon bias is an important determinant in TLR expression and potentially useful as a prediction parameter for the evolutionary fate of TLRs.
All TLRs recognize conserved microbial ligands that are important for microbe viability. The identification of bird and reptilian TLR15 as a receptor for secreted microbial proteases is therefore unusual and raised questions about the range of micro-organisms that can be detected and whether the detection of microbial proteolytic activity by TLR15 is species-specific. In Chapter 7 we observed that TLR15 of an alligator could respond to secreted fractions of fungal, yeast as well as bacterial pathogens while TLR15 of the anole lizard only responded to the fungal pathogen. Chicken and crocodile TLR15 also showed variable responses to these pathogens suggesting that unlike most other TLRs, ligand detection by TLR15 developed highly variable throughout host-microbe co-evolution. Testing of the secreted fraction of the pathogenic bacterium Pseudomonas aeruginosa revealed an unexpected difference in sensitivity between the highly similar alligator and crocodile TLR15, suggesting that only very few, possibly even one, amino acid conveys species-specificity in TLR15 function. Finally, in search for a specific agonist that activated TLR15 we tested several P. aeruginosa protease mutants. This identified the multifunctional virulence factor LasB as a likely activator of TLR15. As a receptor to microbial proteases, TLR15 is unique among the TLRs which generally detect highly conserved microbial structures. The unstably evolving nature of its ligand, which include pathogen-specific proteases, may partly explain the observed high level of species-specific functionality and evolutionary regression of TLR15.
In Chapter 8 our novel findings are integrated in a general discussion about different aspects of TLR biology. One major topic includes the factors that shape receptor-ligand evolution, with TLR5 and TLR15 as attractive examples due to their opposite evolutionary histories (stable vs. dynamic). Other topics are discussed from a cell biological perspective following the cellular life cycle of a TLR, i.e. receptor expression, maturation, transport and ligand interaction. Lastly, the evolution driven species-specificity of ligand recognition and its diverse implications are discussed and future perspectives on TLR research are presented.
Collectively, the work described throughout the chapters of this thesis demonstrate the strength of evolution based, species-comparative research for better understanding the principles behind the function of a protein. Hopefully, the novel insights in TLR biology gained through this approach and described herein will serve future research into TLR biology.