Structural basis for chemokine recognition and receptor activation of chemokine receptor CCR5

Abstract

The chemokine receptor CCR5 plays a vital role in immune surveillance and inflammation. However, molecular details that govern its endogenous chemokine recognition and receptor activation remain elusive. Here we report three cryo-electron microscopy structures of G_i1 protein-coupled CCR5 in a ligand-free state and in complex with the chemokine MIP-1α or RANTES, as well as the crystal structure of MIP-1α-bound CCR5. These structures reveal distinct binding modes of the two chemokines and a specific accommodate pattern of the chemokine for the distal N terminus of CCR5. Together with functional data, the structures demonstrate that chemokine-induced rearrangement of toggle switch and plasticity of the receptor extracellular region are critical for receptor activation, while a conserved tryptophan residue in helix II acts as a trigger of receptor constitutive activation.

Fig. 1. Overall structures of the CCR5–chemokine and CCR5–G_i1 complexes.

a Cryo-EM structures of CCR5–G_i1, MIP-1α–CCR5–G_i1, and RANTES–CCR5–G_i1 and crystal structure of CCR5–MIP-1α. The structures are shown in cartoon representation. The receptor CCR5 in the four structures is colored green, blue, cyan, and gold, respectively. The chemokines MIP-1α and RANTES are colored magenta and orange, respectively. The three subunits in G_i1 are colored light pink, light gray, and light cyan, respectively. Disulfide bonds are shown as yellow sticks. b Comparison of the transmembrane helical bundle conformation in the CCR5 structures. The helical bundles in the structures of CCR5–G_i1, MIP-1α–CCR5–G_i1, RANTES–CCR5–G_i1, and CCR5–MIP-1α and the previously determined structure of CCR5–maraviroc (PDB ID: 4MBS) are colored green, blue, cyan, gold, and gray, respectively. The red arrows indicate the movements of the intracellular tips of helices V, VI, and VII in the G_i1-bound CCR5 structures compared to those in the structures of CCR5–MIP-1α and CCR5–maraviroc. c Comparison of the G protein-binding pocket in the structures of MIP-1α–CCR5–G_i1 and other class A GPCR–G_i/o complexes. The helical bundles and the C termini of Gα α5-helix in the MIP-1α–CCR5–G_i1 structure and the structures of CCR6–G_i, CXCR2–G_i, μOR–G_i, A₁R–G_i, CB1–G_i, and M2R–G_o (PDB IDs: 6WWZ, 6LFO, 6DDE, 6D9H, 6N4B, and 6OIK) are colored blue, dark red, pink, purple, light green, brown, and yellow, respectively. d Comparison of the extracellular regions in the structures of CCR5–G_i1, MIP-1α–CCR5–G_i1, RANTES–CCR5–G_i1, and CCR5–MIP-1α. The N-terminal regions in MIP-1α and RANTES (residues 1–9) are shown in cartoon representation and colored magenta and orange, respectively. The red arrows indicate movements of the extracellular tips of helices I and II in the RANTES–CCR5–G_i1 structure compared to those in the MIP-1α-bound structures.

Fig. 2. Overall chemokine binding mode in CCR5 and receptor-chemokine recognition in CRS1.

a Overall chemokine binding mode in CCR5. The structures of CCR5–MIP-1α, MIP-1α–CCR5–G_i1, and RANTES–CCR5–G_i1 are shown in cartoon representation. The receptors in the three structures are colored gold, blue, and cyan, respectively. The chemokines MIP-1α and RANTES are colored magenta and orange, respectively. The regions of CRS1, CRS1.5, and CRS2 are indicated by gray dashed lines. b The receptor-chemokine interactions in the CRS1 of the CCR5–MIP-1α structure. The key residues of CCR5 and MIP-1α that form contacts are shown as sticks with gold and magenta carbons, respectively. The two interaction cores other than the interactions between the tyrosines in CCR5 and the basic residues in MIP-1α are indicated by green dashed circles. c The free energy surface (FES) of the GaMD simulations of the CCR5–MIP-1α and MIP-1α–CCR5–G_i1 complexes. The FES profile was calculated along with the collective variables of the RMSDs of the CCR5 helices and N terminus. The experimentally identified binding states between MIP-1α and the CCR5 N terminus are highlighted with white-colored stars. Source data are provided as a Source Data file.

Fig. 3. Binding modes of MIP-1α and RANTES in CRS2.

a Comparison of the binding poses of MIP-1α and RANTES in CCR5. The structures of MIP-1α–CCR5–G_i1 and RANTES–CCR5–G_i1 are shown in cartoon representation and colored blue (CCR5)/magenta (MIP-1α) and cyan (CCR5)/orange (RANTES). The N-terminal regions of the chemokines are highlighted by a green dashed box. b, d Binding mode of the N-terminal residues of MIP-1α. b Binding mode of the MIP-1α residues S1–A3; d binding mode of the MIP-1α residues A4-T6. The MIP-1α–CCR5–G_i1 structure is shown in cartoon representation. The MIP-1α and CCR5 residues that are involved in interactions are shown as sticks with magenta and blue carbons, respectively. The polar interactions are displayed as green dashed lines. c, e Binding mode of the N-terminal residues of RANTES. c Binding mode of the RANTES residues S1–Y3; e binding mode of the RANTES residues S4-D6. The RANTES–CCR5–G_i1 structure is shown in cartoon representation. The RANTES and CCR5 residues that are involved in interactions are shown as sticks with orange and cyan carbons, respectively. The polar interactions are displayed as green dashed lines. f Comparison of interactions formed by the chemokine residues 29 and 33. The structures of MIP-1α–CCR5–G_i1 and RANTES–CCR5–G_i1 are shown in cartoon representation. The MIP-1α residues E29 and Q33 and the CCR5 residues R168 and K191^5.35 form polar interactions in the MIP-1α–CCR5–G_i1 structure and the counterparts in the RANTES–CCR5–G_i1 structure are shown as sticks.

Fig. 4. Chemokine-induced and constitutive activation of CCR5.

a Chemokine-induced conformational change of W248^6.48 and Y251^6.51. The structure of MIP-1α–CCR5–G_i1 is shown in cartoon representation and colored blue (CCR5) and magenta (MIP-1α). The CCR5–maraviroc structure is shown in cartoon representation and colored gray. The ligand maraviroc is shown as yellow sticks. The receptor residues W248^6.48 and Y251^6.51 in the two structures and the MIP-1α residue S1 are shown as sticks. The hydrogen bond between Y251^6.51 and S1 in the MIP-1α–CCR5–G_i1 structure is displayed as a green dashed line. The red arrows indicate the conformational changes of W248^6.48 and Y251^6.51 and the outward movement of helix VI in the MIP-1α–CCR5–G_i1 structure relative to those in the CCR5–maraviroc structure. b Structural comparison of the RANTES–CCR5–G_i1 and CCR5–[5P7]RANTES (PDB ID: 5UIW) complexes. The receptors in the two structures are colored cyan and red, respectively. The chemokines RANTES and [5P7]RANTES are colored orange and purple, respectively. The RANTES residues S1 and S4–T7 and the [5P7]RANTES residues Q0 and L4–L7 are shown as sticks. The hydrogen bond between Y251^6.51 and S1 in the RANTES–CCR5–G_i1 structure is displayed as a green dashed line. The red arrow indicates the movement of the extracellular end of helix VII in the CCR5–[5P7]RANTES structure compared to that in the RANTES–CCR5–G_i1 structure. c Structural comparison of the CCR5–G_i1 (CCR5-apo), MIP-1α–CCR5–G_i1, and CCR5–maraviroc complexes. The receptors in the three structures are colored green, blue, and gray, respectively. MIP-1α is colored magenta. The receptor residues W86^2.60, Y108^3.32, W248^6.48, and Y251^6.51 as well as the MIP-1α residue A3 are shown as sticks. The red arrows indicate the conformational changes of the four receptor residues in the CCR5–G_i1 structure relative to those in the inactive CCR5–maraviroc structure. The residue T82^2.56 in the CCR5–G_i1 structure is also displayed as sticks, showing close proximity to W86^2.60 and Y108^3.32. d Basal activity of the wild-type CCR5 (WT) and mutants of residues involved in constitutive activation. The IP accumulation assay was performed in parallel with the measurement of the IP production using the cells only transfected with the chimeric Gα protein Gα_Δ6qi4myr as a control. The basal activity was calculated by subtracting the IP production measured in the control for the WT receptor and all the mutants and is shown as per cent of the WT activity. The numbers of independent experiments (n) performed in triplicate for the WT and mutants are shown in the parentheses. ***P < 0.0001 by one-way ANOVA followed by Dunnett’s post-test, compared with the basal activity of the WT (WT-maraviroc, W86F, W86A, Y108F, and Y251F: P < 0.0001). See Supplementary Table 1 for detailed statistical evaluation and expression level. Source data are provided as a Source Data file.