CCL5 phase separates with heparin in solution.

(A) Confocal images of assembling status of 20 μM CCL5-Cy3 mixed with heparin at different ratios and 5% PEG. Scale bar=10 μm. (B) Turbidity changes with increase of heparin concentration were measured at 620 nm. (C) Fusion of phase-separated droplets formed at the ratio of CCL5-Cy3: heparin=1:2. The white arrows indicate the dynamic fusion of two adjacent droplets. (D) Representative FRAP results of droplets formed by CCL5-Cy3: heparin=1:2 depicted in Figure 1A, showing the intensity of fluorescence pre- and after photobleaching. The images of representative droplets in different recovery stages are shown. (E) Confocal images of assembling status of 44AANA47-CCL5 or CCL5 in the presence of heparin. Scale bar=10 μm. (F) Comparison of CCL5 and 44AANA47-CCL5 turbidity in the presence of heparin. Data are mean ± s.d. n=3 (for B, F). P values were determined by unpaired two-tailed t-tests.

Co-phase separation of CCL5 and heparin establishes chemokine gradient

(A) CCL5-EGFP or 44AANA47-CCL5-EGFP were bound to heparin beads or Ni-NTA beads, respectively, and were placed in Matrigel in 96-well plate. After incubation for 12 hr images were taken to quantify the fluorescence intensity. (B) Quantification of the fluorescence signals along the lines with arrows indicated in Figure 2A. (C) Illustration of in vitro chemotaxis assay.

(D) Heparin beads or Ni-NTA beads bound with CCL5 or 44AANA47-CCL5 were placed in the lower chamber. THP-1 cells (3×105 cells) were added to upper chamber. After 5 hr, THP-1 in the lower chamber was collected and counted. Data are mean ± s.d. n=3. P values were determined by unpaired two-tailed t-tests.

CCL5 phase separates with heparan sulfate on the cell surface

(A) Microscopy images of CCL5-Cy3 on the surface CHO-K1 and 677 as well as CHO-K1 treated with 1 mg/mL heparin. From left to right, the images are fluorescent-field, bright-field and overlay of two illuminations. Scale bar=10 μm. (B) Z-stack scanning of CCL5-Cy3 phase separation on CHO-K1 cell surface. The cell was imaged by confocal microscope with the Z-stack method. Scale bar=10 μm. (C) FRAP of the condensates formed by CCL5-Cy3, showing the intensity of fluorescence pre- and after photobleaching. The size of the representative droplets in different recovery stages are shown above of the graph. (D) Graphical illustration of the cell-based chemotaxis assay. CHO cells (1×105 cells/well) were plated onto the lower chamber for 24 h (fully attached) and CCL5 or 44AANA47-CCL5 or heparin were added as indicated. THP-1 cells (3×105) were placed on upper chambers. After 5 h, a small volume of medium in the lower chamber was aspirated to count THP-1 transmigrated through the membrane. (E) Quantification of THP-1collected from the lower chamber. Data are mean ± s.d. n=3. P values were determined by unpaired two-tailed t-tests.

Formation of chemokine gradient on the cell surface by phase separation

(A) HUVEC cells were seeded on the plates for 24 h and stained with Dil (red). After washing with PBS, CCL5-EGFP transfected CHO-K1 were added and co-cultured for 24 hr before taking the images (Scale bar=10 μm). (B) Fusion events of droplets formed by CCL5-EGFP on HUVEC cell surface. The white arrows in cropped images indicate the fusion of two droplets with time. (C) FRAP of the condensates formed by CCL5-EGFP in HUVEC cell surface, showing the intensity of fluorescence pre and after photobleaching. The images of representative droplets in different recovery stages are shown. (D) HUVEC cells were seeded on the plate and firmly adhered before adding the CCL5-EGFP transfected CHO-K1placed in 50% of Matrigel. After 1 hr co-culture, confocal images were taking showing CCL5-EGFP diffusion on the surface of HUVEC. Dotted while circle indicates the source cell of CCL5-EGFP transfected CHO-K1. Scale bar=100 μm. (E) Quantification of the fluorescence intensity shows decreased signals of CCL5-EGFP as the distance from the source cells increased. (F) The chemotaxis assay (same experimental conditions as described in Figure 1) shows higher activity of CCL5-EGFP transfected CHO-K1 cells than wild type CHO-K1 cells. Data are mean ± s.d. n=3. P values were determined by unpaired two-tailed t-tests.

Transmigration of human peripheral blood cells

The cells (CHO-K1, CHO-677, or HUVEC) (1×105 cells/well) as indicated were plated onto the lower chamber for 24 h (fully attached) and CCL5 was added. PBMC isolated from three healthy volunteer donors were placed in the upper chamber (total number of cells: Sample 1: 8.65^105; Sample 2: 2.31^106; Sample 3: 2.23^106). After culturing for 5 hr, the total transmigrated cells in the lower chamber were counted.

Chemokines phase separation promotes cell recruitment in vivo.

Mice were treated by intraperitoneal injection of CCL5 or the reagents as indicated. After 18h, the animals were sacrificed and peritoneal lavage was collected. Total cell number was counted by automated cell counter. Data are mean ± s.d. n=4. P values were determined by unpaired two-tailed t-tests.

Schematic model of CCL5 phases separation along with heparan sulfate on the cell surface.

The inflammatory cells secrete chemokines that are immobilized by heparan sulfate in the form of Liquid-Liquid phase separation and subsequently diffused to form a concentration gradient.