Computational Approaches to Toll-Like Receptor 4 Modulation

Abstract

Toll-like receptor 4 (TLR4), along with its accessory protein myeloid differentiation factor 2 (MD-2), builds a heterodimeric complex that specifically recognizes lipopolysaccharides (LPS), which are present on the cell wall of Gram-negative bacteria, activating the innate immune response. Some TLR4 modulators are undergoing preclinical and clinical evaluation for the treatment of sepsis, inflammatory diseases, cancer and rheumatoid arthritis. Since the relatively recent elucidation of the X-ray crystallographic structure of the extracellular domain of TLR4, research around this fascinating receptor has risen to a new level, and thus, new perspectives have been opened. In particular, diverse computational techniques have been applied to decipher some of the basis at the atomic level regarding the mechanism of functioning and the ligand recognition processes involving the TLR4/MD-2 system at the atomic level. This review summarizes the reported molecular modeling and computational studies that have recently provided insights into the mechanism regulating the activation/inactivation of the TLR4/MD-2 system receptor and the key interactions modulating the molecular recognition process by agonist and antagonist ligands. These studies have contributed to the design and the discovery of novel small molecules with promising activity as TLR4 modulators.

Keywords: MD simulations; TLR4/MD-2 modulators; Toll-like receptor 4; computational chemistry; docking; drug design; homology modeling; molecular recognition; virtual screening.

Figure 1

Schematic representation of the LPS-induced dimerization of the TLR4/MD-2 complex leading to immune system activation. Red arrows indicate motion and mutual recognition. (a) Two LPS (yellow) are engaged by two distinct TLR4/MD-2 systems (blue and red, MD-2 in pale blue and pale red); (b) two TLR4/MD-2/LPS complexes dimerize by protein-protein interactions at the dimerization interface; (c) dimerization brings together the two intracellular TIR-containing domains providing a suitable molecular surface for recruiting downstream adaptors.

Figure 2

Representation of the 3D structure of TLR4/MD-2/LPS. (a) Large-scale representation showing the intracellular, transmembrane and extracellular domains of TLR4/MD-2 in complex with E. coli LPS. 3D Structures correspond to the X-ray crystallographic structure for the extracellular domain (PDB ID 3FXI) and homology modeling for the transmembrane and intracellular domains. (b) Close-up look at the TLR4 extracellular domain (purple) along with MD-2 (yellow) and LPS (CPK colors with C atoms in green) from PDB ID 3FXI.

Figure 3

Representation of the LPS in complex with TLR4/MD-2. (a) Detail of the 3D structure of the complex between TLR4/MD-2 and E. coli LPS (CPK colors with C atoms in green and R2 C atoms in magenta) from the X-ray crystallographic structure (PDB ID 3FXI); (b) chemical structure of E. coli lipid A. The R2 FA chain (magenta) placed at the channel of MD-2 completes the dimerization interface.

Figure 4

Superimposition of the X-ray crystallographic structures of the agonist (magenta) and the antagonist (green) conformations of MD-2 from, respectively, PDB ID 3FXI and 2E56. Bound ligands have been hidden for the sake of clarity (E. coli LPS in 3FXI; three myristic acids in 2E56). Conformational change of the molecular switch Phe126 is marked.

Figure 5

Lipid A and synthetic lipid A analogues with activity as TLR4 modulators. Activity is referred to hTMR4/MD-2.

Figure 6

Synthetic LPS mimetics studied by computational approaches.

Figure 7

Non-LPS-like TLR4/MD-2 modulators studied by computational approaches.

Figure 8

Novel TLR4/MD-2 modulators found by VS approaches.

Figure 9

Intracellular TIR domain of TLR4. (a) 3D Representation of the homology model; (b) FASTA sequence.

Figure 10

Representation of the different ways the dimer is proposed by published computational strategies to be assembled in the literature by computational strategies. (a) First reported by Miguel et al. [83]; (b) reported by Gong et al. [82]; (c–e) reported by Guven-Maiorov et al. [85]. The monomer has been built by homology modeling, and the secondary structure representation has been altered to resemble the other models. The dimers have been assembled manually, fitting as best as possible the schemes present in each paper, to provide an overview of the variety of binding poses reported so far. The dimmers shown do not have the pretention of being as precise as those shown in the original papers and should be considered schematic.