Behavior of Chemokine Receptor 6 (CXCR6) in Complex with CXCL16 Soluble form Chemokine by Molecular Dynamic Simulations: General Protein‒Ligand Interaction Model and 3D-QSAR Studies of Synthetic Antagonists

Abstract

The CXCR6‒CXCL16 axis is involved in several pathological processes, and its overexpression has been detected in different types of cancer, such as prostate, breast, ovary, and lung cancer, along with schwannomas, in which it promotes invasion and metastasis. Moreover, this axis is involved in atherosclerosis, type 1 diabetes, primary immune thrombocytopenia, vitiligo, and other autoimmune diseases, in which it is responsible for the infiltration of different immune system cells. The 3D structure of CXCR6 and CXCL16 has not been experimentally resolved; therefore, homology modeling and molecular dynamics simulations could be useful for the study of this signaling axis. In this work, a homology model of CXCR6 and a soluble form of CXCL16 (CXCR6‒CXCL16s) are reported to study the interactions between CXCR6 and CXCL16s through coarse-grained molecular dynamics (CG-MD) simulations. CG-MD simulations showed the two activation steps of CXCR6 through a decrease in the distance between the chemokine and the transmembrane region (TM) of CXCR6 and transmembrane rotational changes and polar interactions between transmembrane segments. The polar interactions between TM3, TM5, and TM6 are fundamental to functional conformation and the meta-active state of CXCR6. The interactions between D77-R280 and T243-TM7 could be related to the functional conformation of CXCR6; alternatively, the interaction between Q195-Q244 and N248 could be related to an inactive state due to the loss of this interaction, and an arginine cage broken in the presence of CXCL16s allows the meta-active state of CXCR6. A general protein‒ligand interaction supports the relevance of TM3‒TM5‒TM6 interactions, presenting three relevant pharmacophoric features: HAc (H-bond acceptor), HDn (H-bond donator), and Hph (hydrophobic), distributed around the space between extracellular loops (ECLs) and TMs. The HDn feature is close to TM3 and TM6; likewise, the HAc and Hph features are close to ECL1 and ECL2 and could block the rotation and interactions between TM3‒TM6 and the interactions of CXCL16s with the ECLs. Tridimensional quantitative structure-activity relationships (3D-QSAR) models show that the positive steric (VdW) and electrostatic fields coincide with the steric and positive electrostatic region of the exo-azabicyclo[3.3.1]nonane scaffold in the best pIC50 ligands. This substructure is close to the E274 residue and therefore relevant to the activity of CXCR6. These data could help with the design of new molecules that inhibit chemokine binding or antagonize the receptor based on the activation mechanism of CXCR6 and provoke a decrease in chemotaxis caused by the CXCR6‒CXCL16 axis.

Keywords: 3D-QSAR; CG-MD simulations; CXCL16s; CXCR6; docking.

Figure 1

Activation mechanism of chemokine receptors. (A) Initial interaction of a chemokine with NH₂-T and the movement into the transmembrane region, promoting a conformational change in the receptor; (B) interaction of the chemokine with NH₂-T and movement into the transmembrane region, promoting a conformational change in the receptor and triggering the separation of the βγ complex from the α_i/0 subunit in the G-protein.

Figure 2

Structures of CXCR6 and CXCL16s: (A) CXCR6-M1. NH₂-T is depicted in turquoise, TM1 in red, TM2 in blue, TM3 in green, TM4 in purple, TM5 in orange, TM6 in yellow, TM7 in brown, and COOH-T in pale yellow; (B) CXCR6 reorientated; (C) CXCL16s; and (D) CXCL16s cluster1.

Figure 3

CXCR6‒CXCL16 molecular docking. (A) CXCR6‒CXCL16s complex; (B) interaction diagram, where R is CXCR6 and Q is CXCL16s; the hydrophobic environment is represented by residues in red and pink, the ionic interactions are represented by red dotted lines, and the H-bonds are represented by green dotted lines.

Figure 4

RMSD of CXCR6 CG-MD. (A) RMSD CXCR6, CXCR6 in complex with CXCL16s, and CXCL16s; (B) RMSD TMs.

Figure 5

RMSD of CXCR6 CG-MD. (A) RMSD loops and (B) rotation of TMs.

Figure 6

Distance between CXCR6‒CXCL16s CG-MD. (A) Structure of complex at 500 ns, with CXCL16s depicted in turquoise and CXCR6 in blue; (B) distance between center of mass of CXCL16s and CXCR6; we can observe a decrease in distance between CXCL16s and CXCR6, with an equilibrium around 500 ns.

Figure 7

RMSD and rotation angles of CXCR6‒CXCL16s CG-MD simulation. (A) RMSD of system, CXCR6, and CXCL16; (B) RMSD of loops; (C) RMSD of TMs.

Figure 8

General protein–ligand interaction model based on pIC50. (A) General model: hydrogen bond acceptors (HAc) depicted in magenta, hydrogen bond donors (HDn) in gray, and hydrophobic groups (Hph) in yellow. (B) Residues around 5 Å; (C) 3D conformation of ligand 82; (D) ligand interaction diagram of ligand 82.

Figure 9

3D-QSAR model. (A) Pharmacophoric features in the QSAR model (in yellow: negative VdW field; in green: positive VdW field); (B) pharmacophoric features in the QSAR model (in red: negative electrostatic field; in light blue: positive electrostatic field); (C) superposition of general protein‒ligand interaction model and VdW fields; (D) superposition of general protein‒ligand interaction model and electrostatic fields; (E) the three ligands with the best pIC50 (82, 81, and 78) in VdW fields; (F) the three ligands with the best pIC50 in electrostatic fields; (G) the three ligands with the worst pIC50 (51, 56, and 55) in VdW fields; (H) the three ligands with the worst pIC50 in electrostatic fields; (I) the three-dimensional structure of the QSAR model.