Homology Modeling of Toll-Like Receptor Ligand-Binding Domains

Toll-like receptors (TLRs) are in the front-line during the
initiation of an innate immune response against invading pathogens.
TLRs are type I transmembrane proteins that are expressed on the
surface of immune system cells. They are evolutionarily conserved
between insects and vertebrates. To date, 13 groups of mammalian
TLRs have been identified, ten in humans and 13 in mice. They share
a modular structure that consists of a leucine-rich repeat (LRR)
ectodomain, a single transmembrane helix and a cytoplasmic
Toll/interleukin-1 receptor (TIR) domain. Most TLRs have been shown
to recognize pathogen-associated molecular patterns (PAMPs) from a
wide range of invading agents and initiate intracellular signal
transduction pathways to trigger expression of genes, the products
of which can control innate immune responses. The TLR signaling
pathways, however, must be under tight negative regulation to
maintain immune balance because over-activation of immune responses
in the body can cause autoimmune diseases. The TLR ectodomains are
highly variable and are directly involved in ligand recognition. So
far, crystal structures are missing for most TLR ectodomains
because structure determination by X-ray diffraction or nuclear
magnetic resonance (NMR) spectroscopy experiments remains
time-consuming, and sometimes the crystallization of a protein can
be very difficult. Computational modeling enables initial
predictions of three-dimensional structures for the investigation
of receptor-ligand interaction mechanisms. Computational methods
are also helpful to develop new TLR agonists and antagonists that
have therapeutic significance for diseases. In this dissertation,
an LRR template assembly approach for homology modeling of TLR
ligand-binding domains is discussed. To facilitate the modeling
work, two databases, TollML and LRRML, have been established. With
this LRR template assembly approach, the ligand-binding domains of
human TLR5-10 and mouse TLR11-13 were modeled. Based on the models
of human TLR7, 8 and 9, we predicted potential ligand-binding
residues and possible configurations of the receptor-ligand complex
using a combined procedure. In addition, we modeled the cytoplasmic
TIR domains of TLR4 and 7, the TLR adaptor protein MyD88 (myeloid
differentiation primary response protein 88) and the TLR inhibitor
SIGIRR (Single immunoglobulin interleukin-1 receptor-related
molecule) to investigate the structural mechanism of TLR negative
regulation.