Figures and data
Screening for small compound antagonists acting at the oligomerization site of CXCR4.
A) Cartoon and surface representation of CXCR4 and the cleft identified between TMV and TMVI. The protein structure is shown in gray, with TMV and TMVI colored in blue and pink, respectively. In green, residues involved in CXCL12 binding. The cavity identified by SurfNet is shown in orange. Volume for the cavity is measured and shown in Å3. T4 lysozyme inserted between TMV and TMVI in the crystallized version of CXCR4 (PDB: 3ODU) is also shown in yellow. B) Dose-response curve of the selected antagonists in Jurkat cell migration experiments in response to 12.5 nM CXCL12. Data are shown as percentage of migrating cells (mean ± SD; n = 5; * ≤ 0.05, *** p ≤ 0.001, **** p ≤ 0.0001). C) Chemical structure of the selected compounds (AGR1.131, AGR1.135 and AGR1.137). The differences between the lateral chains of the three compounds are shaded in green.
AGR1.135 and AGR1.137 alter CXCL12-mediated CXCR4 dynamics and nanoclustering.
Single-particle tracking analysis of JK−/− cells transiently transfected with CXCR4 (JK−/−-X4) treated with DMSO (control), AGR1.131, AGR1.135 or AGR1.137 on fibronectin (FN)- or FN+CXCL12-coated coverslips (DMSO: 581 particles in 59 cells on FN; 1365 in 63 cells on FN+CXCL12; AGR1.131: 1019 particles in 71 cells on FN; 1291 in 69 cells on FN+CXCL12; AGR1.135: 862 particles in 70 cells on FN; 1003 in 77 cells on FN+CXCL12; AGR1.137: 477 particles in 66 cells on FN; 566 in 64 cells on FN+CXCL12) n = 3. A) Frequency of CXCR4-AcGFP particles containing monomers plus dimers (≤2) or nanoclusters (≥3), ± SEM, calculated from mean spot intensity values of each particle as compared with the value of monomeric CD86-AcGFP (n = 3; n.s., not significant; ****p ≤ 0.0001). B) Intensity distribution (arbitrary units, a.u.) from individual CXCR4-AcGFP trajectories on unstimulated and CXCL12-stimulated JK−/−-X4 cells pretreated or not with the indicated compounds or vehicle (DMSO). Graph shows the distribution of all trajectories (gray outline), with the mean value of each experiment (black circles) and the mean of all experiments ± SD (red lines) (n = 3; n.s., not significant; ***p ≤ 0.001, ****p ≤ 0.0001). C) Percentage of mobile and immobile CXCR4-AcGFP particles at the membrane of cells treated as indicated. D) Diffusion coefficients (D1–4) of mobile single particle trajectories at the membrane of cells treated as indicated, represented by box-and-whiskers plots (Tukey method) with median (black line) indicated. (n.s., not significant, ****p ≤ 0.0001).
AGR1.135 and AGR1.137 treatments do not block CXCL12-mediated signaling pathways.
A) Cell surface expression of CXCR4 in Jurkat cells after stimulation with CXCL12 (12.5 nM) at different time points and analyzed by flow cytometry using an anti-CXCR4 antibody in nonpermeabilized cells. Results show mean ± SEM of the percentage of CXCR4 expression at the cell surface (n = 4; n.s. not significant; **p ≤ 0.01, ***p ≤ 0.001, **** p ≤ 0.0001). B) CXCL12-Atto-700 binding on untreated or Jurkat cells pre-treated with DMSO (vehicle), AGR1.131, AGR1.135, AGR1.137 or with AMD3100 as a control, followed by flow cytometry analysis. Results are expressed as mean fluorescence intensity (MFI) values (arbitrary units). Negative corresponds to basal cell fluorescence in the absence of CXCL12-Atto-700 (mean ± SD, n = 3 n.s. not significant; ****p ≤ 0.0001). C) Cells untreated or pretreated with the small compounds (50 μM, 30 min 37°C) or AMD3100 (1 μM, 30 min 37°C) were stimulated with CXCL12 (50 nM, 5 min, 37°C) followed by forskolin (10 μM, 10 min, 37°C). Cells were then lysed and cAMP levels were determined (mean ± SD; n = 3; n.s. not significant; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001). D) Western blot analysis of phospho (p)Akt and pERK in Jurkat cells pre-treated with DMSO, AGR1.131, AGR1.135 or AGR1.137, in response to CXCL12. As a loading control, membranes were re-blotted with an anti-Akt antibody. Representative experiments are shown (n = 4). E, F) Densitometry evaluation of the specific bands in D) (mean ± SD; n = 4).
AGR1.135 and AGR1.137 treatment alter CXCL12-mediated actin polymerization.
A) Actin polymerization in response to CXCL12 as determined by F-actin (phalloidin-TRITC) staining in Jurkat cells treated with DMSO (vehicle) or the indicated modulators. Statistical significance of the different time points in comparison with the control (DMSO) is shown in the table (mean ± SD; n = 3; n.s. not significant; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001). B) Percentage of polarized T cell blasts adhered to fibronectin and treated or not with CXCL12 in the presence of the indicated antagonists, as analyzed by immunostaining with anti-ICAM3-Alexa fluor 488 and phalloidin-TRITC. More than 500 cells were analyzed in each condition. Data are presented as percentage of polarized cells (mean ± SD; n = 3; n.s. not significant; ****p < 0.0001). C) Representative images of T cell blasts adhered to fibronectin and treated or not with CXCL12 in the presence of the indicated antagonists, as analyzed by immunostaining with anti-ICAM3-Alexa fluor 488 and phalloidin-TRITC as in B). D) CD4+ T cells pretreated with AGR1.131, AGR1.135 or AGR1.137 were perfused in flow chambers coated with ICAM-1-containing lipid bilayers, alone or CXCL12-coated, and analyzed for cell contacts with the substrate. Data are presented as percentage of adhered cells (mean ± SD; n = 3; n.s. not significant; ****p ≤ 0.0001). E) Cells in D) were analyzed for cell migration. Data are presented as percentage of migrating cells (mean ± SD; n = 3; n.s. not significant; ****p ≤ 0.0001).
Ribbon and sticks representation of the CXCR4 modulators bound to the receptor.
In the ribbon and sticks representations, the TMV helix is colored in blue and TMVI in pink. Figures show the binding of ligands AGR1.135 (A) and AGR1.137 (B) Zoom images showing the residues involved in the interaction of both compounds are also shown (right panels). C) Binding of AGR1.131 represented as sticks with carbon atoms in yellow, oxygen in red, nitrogen in blue and fluorine in green. CXCR4 residues involved in CXCL12 engagement and initial signal transmission are represented as green spheres.
The antagonistic behavior of AGR1.135 and AGR1.137 depends on specific residues of CXCR4.
A) CXCL12-induced migration of AGR1.131-AGR1.135- or AGR1.137-pretreated JK−/− cells transiently transfected with CXCR4wt or the different mutants described. Data are shown as the mean percentage ± SD of input cells that migrate (n = 4; n.s. not significant, **p ≤ 0.01, ****p ≤ 0.0001). B) Cartoon representation of Y256 and its intramolecular interactions in the CXCR4 X-ray solved structure 3ODU. C, D) Cartoon representation of the interaction of CXCR4 F249L mutant with AGR1.135 and AGR1.137, respectively. The two most probable conformations of leucine rotamers are represented in cyan. Van der Waals interactions are depicted in cyan dashed lines and hydrogen bonds in black dashed lines.
AGR1.137 reduces tumorigenesis and metastasis in a zebrafish model.
(A–D) Migration of HeLa cells treated with vehicle (DMSO) or with the selected small compounds or inhibitor as indicated, on μ-chambers in response to a CXCL12 (A, B) or CXCL2 (C, D) gradient (n = 2, in duplicate, with at least 50 cells tracked in each condition). Figures A) and C) show representative spider plots with the trajectories of tracked cells migrating along the gradient (black) or moving in the opposite direction (red). Black and red dots in the plots represent the final position of each single tracked cell. (B, D) Quantification of the Forward Migration Index of experiments performed in A) and C) (mean ± SD; n = 3; n.s. not significant, ****p ≤ 0.0001). E) Representative fluorescent images of DiI-labeled HeLa cells in zebrafish larvae treated with vehicle (DMSO), AGR1.131 50 μM, AGR1.137 50 μM or AMD3100 10 μM at 0 or 3 days post-implantation and treatment. Quantitation of the relative tumor size at day 3 compared with that of day 0 normalized to the relative tumor size in the DMSO control group, is shown for each experimental group (mean ± SD; n = 20; n.s. not significant, **p ≤ 0.01). F) Representative fluorescent images of the caudal hematopoietic plexus of larvae from the same groups as shown in E) at 3 days postimplantation. Quantitation of the relative amount of metastasized cells in each group relative to DMSO-controls is shown (mean ± SD; n = 20, n.s. not significant, * p ≤ 0.05, *** p ≤ 0.001).
Small compound-mediated inhibition of CXCL-12-induced cell migration.
A, B). Migration of Jurkat cells, untreated or treated with the selected small compounds (50 μM) or DMSO (vehicle) in the previous in silico screening, in response to 12.5 nM CXCL12. Data shown as percentage of migrating cells (mean± SD; n = 5; *p ≤0.05, **p ≤0.01, ***p ≤0.001, ****p ≤0.0001, ns: not significant).
Effect of small compounds on Jurkat cell cycle.
A). Cell cycle analysis using propidium iodide incorporation and flow cytometry of Jurkat cells treated with AGR1.131, AGR1.135 and AGR1.137 (2 hours, 37°C) or with DMSO (vehicle), H2O2 (10%) and staurosporine (10 μM) as controls. The percentage of cells in each cycle stage is shown. B). Summary of the percentage of treated cells in each cell cycle phase.
Flow cytometry analysis of the binding of CXCL12-ATTO700 to JK cells.
Cells were pre-incubated for 30 min at 37°C with the indicated antagonists, followed by incubation with CXCL12-ATTO700 (30 min 37°C). CXCL12 binding was determined in flow cytometry. A representative experiment of 4 performed is shown.
Western blot analysis of AMD3100-treated JK cells.
A) Cells were pre-incubated for 30 min at 37°C with AMD3100 or DMSO (vehicle) as control, stimulated with CXCL12 and analyzed by western blot using specific antibodies. B) Densitometry analysis of the western blot images in A) using Image J software.
AGR1.135 and AGR1.137 do not alter anti-CD3-induced actin polimerization.
F-actin (Phalloidin-TRITC) staining of Jurkat cells treated with the indicated compounds or DMSO (vehicle), adhered to fibronectin and stimulated with anti-CD3 antibody, as indicated (mean ± SD; n = 3).
AGR1.131, AGR1.135 and AGR1.137 are not CXCR4 agonists.
Migration of Jurkat cells in response to 12.5 nM CXCL12, 50 μM AGR1. 131, 50 μM AGR1. 135 or 50 μM AGR1. 137. Data shown as percentage of migrating cells (mean ± SD; n = 5; n.s. not significant, **** p ≤0.0001).
FRET analysis of CXCR4 in the presence of AGR1.131, AGR1.135 and AGR1.137.
FRET efficiency in HEK-293 cells transiently transfected with CXCR4-YFP/ CXCR4-CFP (ratio 15:9), in the presence of 50 μM AGR1.131, 50 μM AGR1.135, 50 μM AGR1. 137 or vehicle (DMSO). Data shows FRET efficiency (a.u.) (mean± SD; n = 6; n.s. not significant, **** p ≤0.0001).
Cartoon and surface representation of CXCR4 and the clefts identified between TMV and TMVI.
The protein structure is shown in gray, with TMV and TMVI colored in blue and pink, respectively. Cavities associated to AGR1. 135 (A) and AGR1. 137 (B) binding were identified by SurfNet software are shown in orange and green. Volumes for each cavity are measured (Å3).
Expression and function of CXCR4 and CXCR4 mutants coupled to AcGFP in transfected JK-/- cells.
A) Flow cytometry analysis of cells were transiently transfected with the indicated CXCR4 mutants or wt CXCR4, coupled to AcGFP and their expression was determined 24 hours after transfection by GFP detection in flow cytometry. A representative experiment of 4 performed is shown. B) CXCL12-induced migration of CXCR4-defficient Jurkat cells (JK-/-) transiently transfected with CXCR4wt or its mutants, CXCR4I204K, CXCR4G207I, CXCR4L208K, CXCR4R235L, CXCR4F249L, CXCR4Y256F and CXCR4S260A. Data are shown as the mean ± SD percentage of input cells that migrate (n = 3).
Characterization of HeLa cells.
A) CXCR4 expression on the surface of HeLa cells, analyzed by flow cytometry using a specific anti-CXCR4 antibody. Data show a representative experiment of 6 performed. B) Effect of the indicated compounds on HeLa cells growth. Cells (0.5×106) were plated on 48 well plates in the presence of 100 mM of the compounds, which were added freshly every 24 hours. Cells were counted at every 24 hours and maintained for 72 h. The mean ± SD (n=3) is shown.
Residue conservation among all GPCR class A chemokine receptors for TMV and TMVI.
Graphs were produced in the WebLogo server after aligning all receptors sequence in the sequence alignment tool included in GPCRdb. Font size for each residue represented in one letter code stands for the degree of conservation among all chemokine receptors. Residues are colored by their chemistry being green for polar, orange for hydrophobic, red for acidic, blue for basic and purple for neutral.
Ligplot representation of antagonists interactions with CXCR4.
Ligplot representations of AGR1. 135 (A) and AGR1. 137 (B), with hydrogen bonds represented in green dotted lines and non-bonded contacts in red spoke arcs. Residues labeled in blue font are present in the interface of both AGR1. 135 and AGR1. 137. CXCR4 residues forming a hydrogen bond with the ligands are shown in gray.
TMVI transition from inactive to active position.
Ribbon representation of TMVI of CXCR4 structures superimposed for inactive (PDB code 3ODU), intermediate active (AGR1. 135, AGR1. 137) and active (PDB 8U4N) upon interaction with no, negative allosteric modulators and CXCL12-induced Gi-protein interaction, respectively. Arginine 235 is depicted in stick representation as a reference for the movement of the protein.
Compounds with minimal interaction energy in the area of interest
