zCxcl8 Is Necessary and Sufficient to Instruct Neutrophil Migration(A) Diagram of zebrafish larva at 2 dpf. Arrow indicates site of E. coli injection. Square denotes area viewed in (B).(B) DsRed-expressing E. coli was detected by immunofluorescence at 2 hr postinfection (hpi). Fluorescence is shown alone or overlaid on bright-field image (BF). Scale bars represent 25 μm.(C) In situ hybridization of zcxcl8 mRNA at 2 hpi. Scale bars represent 100 μm (left) and 20 μm (right). mb, midbrain; hb, hindbrain.(D) In vitro chemotactic response of adult zebrafish neutrophils to conditioned media from zCxcl8 or mock-transfected HEK293T cells, in a transwell assay. Number of transmigrated neutrophils was measured by flow cytometry and normalized to the same number of calibration beads. Data are means of triplicate samples, representative of four experiments. Error bar represents SEM.(E) SDCM images of neutrophils within live larvae treated or not treated with splice-blocking zcxcl8 morpholino (sMO) at 1 hpi. Y, yolk sac. Scale bars represent 100 μm.(F) RT-PCR of zcxcl8 mRNA, from larvae treated or not treated with splice-blocking zcxcl8 morpholino (sMO), performed at 2 hpi. ef1a is a loading control.(G) Neutrophil recruitment to E. coli in larvae treated or not treated with splice (sMO) or combined splice and translation-blocking (sMO+tMO) zcxcl8 morpholinos, quantified at 2 hpi (control: n = 29 larvae; sMO: n = 19; sMO+tMO: n = 20 larvae, three independent experiments). Error bars indicate SEM. p < 0.01; p < 0.0001 (Mann-Whitney test).(H) zCxcl8-V5 and mCherry expression constructs.(I) Transplantation site of transfected HEK293T cells. The square indicates the area imaged.(J) SDCM images of neutrophils within live larvae 1 day after transplantation with mCherry-transfected HEK293T cells, cotransfected (zCxcl8) or not cotransfected (mock) with zCxcl8. Y, yolk sac. Scale bars represent 100 μm.(K) Percentage of neutrophil tracks reaching the transplant out of all tracks in the imaged area. Mock: n = 568; zCxcl8: n = 1,378; mock+SB225002: n = 156; zCxcl8+SB225002: n = 182. p < 0.0001 (Fisher′s exact test for cells reaching or not transplant). Data are pooled from three to four larvae per group originating from two or three independent experiments.See also Figure S1 and Movie S1.

zCxcl8 Imposes an Orthotactic Bias on Neutrophil Random Walk(A) Evolution of the propagator, i.e. the probability density function (Pdf) of neutrophil displacement r(t) at different times t, in larvae with mock or zCxcl8 transplants. Points are the results of the experiments and lines are Gaussian fits. Colors are associated with times: black for t = 1 min, red for t = 2 min, green for t = 4 min, blue for t = 8 min, and cyan for t = 16 min. Plots are in log-log scale. Insets are the collapsed curves, P(ρ), into a unique Gaussian. Black line is the Gaussian fit. Data are from four larvae in four experiments.(B and E) Left panels: representative SDCM images of neutrophils (green) within the area of transplanted cells (red). Representative 50 × 50 × 50 μm cubes indicate example areas covering <100–200 μm (B) or <100 μm (E) distance from transplant, such as were used for motility analysis shown in (C) and (D) and in (F) and (G), respectively. E, eye; Y, yolk sac. Scale bars represent 100 μm. Right panels: tracks of neutrophils within the indicated representative cube. Blue arrows indicate the direction toward the transplant.(C and F) Proportion of tracks within <100–200 μm or <100 μm from the transplant, respectively, that moved toward the transplant. For (C), mock: n = 141, zCxcl8: n = 109; for (F), mock: n = 155; zCxcl8: n = 154 (Fisher′s exact test for tracks moving toward or away from transplant).(D and G) Straightness (displacement/path length) of tracks within ~100–200 μm or <100 μm from transplant, respectively, that migrated toward the transplant. For (D), mock n = 68; zCxcl8: n = 60; for (G), mock: n = 71; zCxcl8: n = 95 (Mann-Whitney test). Error bars indicate SEM. ns, not significant; p < 0.05; p < 0.01.(H) Approach angle θ toward transplant for neutrophil steps made in time interval, δt = 30 s.(I) Average approach angles for individual neutrophil steps. Mock: n = 2,585; zCxcl8: n = 2,138 (Mann-Whitney test). Error bars indicate SEM. ns, not significant.(J) Average time during which a cell maintained orientation toward (θ < 90°) the transplant. zCxcl8: n = 380; mock: n = 500. (Mann-Whitney test). ns, not significant.(K) Average speed of neutrophil steps with approach angles to the transplant <45°. Mock: n = 702; zCxcl8: n = 564 (Mann-Whitney test). Error bars indicate SEM. p < 0.0001.Pooled data from several cubes in three larvae per case, from three independent experiments, are shown in (C), (D), (F), (G), and (I)–(K). See also Figure S2.

zCxcl8 Locally Restricts Neutrophil Motility(A and E) SDCM example image of neutrophils (green) within the area of transplanted mCherry/zCxcl8 cotransfected cells (red). Areas between circles indicate zones at different distances from transplant center (indicated with a white x) that were used for the motility analysis shown in (B) and (F).(B and F) Distribution of instantaneous speed of individual neutrophil steps in the presence (black bars; n = 16,138 for B; n = 13,646 for F) or absence (white bars; n = 9,486 for B; n = 3,234 for F) of zCxcl8 (two-tailed chi-square test). Data were binned every 2 μm/min.(C and G) SDCM example images of neutrophils (green) within the area of transplanted mCherry/zCxcl8 cotransfected cells (red). Representative cubes indicate example areas at different distances from transplant, such that were used for motility analysis shown in (D) and (H), respectively.(D and H) Average track speed in the presence or absence (mock) of zCxcl8. For (D), mock: n = 92, zCxcl8: n = 97; for (H), mock: n = 66, zCxcl8: n = 290 tracks (Mann-Whitney test). Y, yolk sac. ns, not significant; p < 0.0001.(I) Representative images of larvae infected with E. coli and treated or not with zCxcl8 sMO. x, center of infection.(J) Evolution of neutrophil speed with the distance from the center of the infection site.For (D), (H), and (J) error bars indicate SEM. For (B), (F), (D), (H), and (J), data are pooled from three to four larvae per case in three or four experiments. For (A), (C), (E), (G), and (I), scale bars represent 100 μm. See also Figure S3 and Movies S2 and S3.

zCxcl8 Directs and Restricts Neutrophil Motility as HSPG-Bound Gradients(A–I) Extracellular distribution of sec-mCherry (A–C), zCxcl8-mCherry (D–F), and zCxcl8-DM-mCherry (G–I) secreted from transplanted cells. Top: 3 dpf larvae. Squares indicate imaged regions.(A, D, and G) Projections of confocal image stacks showing immunodetection of sec-mCherry, zCxcl8-mCherry, and zCxcl8-DM-mCherry at the site of secretion. Left: grayscale images with green dotted lines outlining the secreting cells. Right: same images color coded for fluorescence intensity.(B, E, and H) Quantification of mCherry staining as a function of the distance from the nearest producing cell. Fluorescence intensity signals along various lines starting from the implants is shown. Colors are associated to different larvae, thin curves are the intensity evolutions along individual lines, and thick curves are the average of the thin curves. Data are from six or seven larvae in two or three independent experiments.(C, F, and I) Projections of confocal image stacks showing immunodetection of sec-mCherry, zCxcl8-mCherry, and zCxcl8-DM-mCherry in the vasculature. CA, caudal artery; CVP, caudal vein plexus; dCV, definitive caudal vein. Data are representative of six or seven larvae in two or three independent experiments. For (A), (C), (D), (F), (G), and (I), scale bars represent 25 μm.(J) Percentage of neutrophil tracks reaching transplanted cells expressing or not expressing zCxcl8 or zCxcl8-DM. Mock: n = 568; zCxcl8: n = 1,378; zCxcl8-DM: n = 772 (Fisher′s exact test for tracks reaching or not transplant). p < 0.0001.(K) Evolution of the average normalized speed with the orientation of motion (cosine of the angle θ, with θ = 0 the direction of the transplant center) within a radius of 150 μm from the transplant center. Speed values are normalized by the average value of the speed, to allow comparison across different experiments. Data are from four larvae in two to four independent experiments. Error bars indicate SEM.(L) Average speed of neutrophil steps inside a radius of 100 μm (left) and between radii of 100 μm and 200 μm (right). Mock: n = 2,293 (left) and n = 6,924 (right); zCxcl8: n = 9,208 (left) and n = 14,375 (right); zCxcl8-DM: n = 3,269 (left) and n = 6,589 (right). Data are from four larvae in two to four independent experiments (Mann-Whitney test). Error bars indicate SEM. p < 0.0001.See also Figure S4 and Movie S4.

zcxcl8 Is Upregulated upon Infection, and zcxcl8 Knockdown Does Not Affect Larval Development or Neutrophil Distribution (Related to Figure 1)
(A and B) qPCR analysis of zcxcl8-chr 1 (A and B), zcxcl8-chr 7a (A), zcxcl8-chr 7b (A) mRNA expression 0.5, 2 and 4 hours post injection (hpi) of E.coli. NI, non injected. Data are normalized to ef1a expression. n=3 qPCR samples. Error bars: SEM. Data in A and B are from two independent experiments, each with pools of 15-20 larvae per condition.
(C) Bright field and wide-field fluorescence images of whole Tg(mpx:gfp) larvae at 2.5 dpf treated or not (control) with splice-blocking (sMO) or combined splice and translation-blocking (sMO+tMO) zCxcl8 morpholinos. E, eye. Y, yolk sac.
(D) Flow cytometry of whole Tg(mpx:gfp) larvae, treated or not with splice-blocking (sMO) or combined splice and translation-blocking (sMO+tMO) zCxcl8 morpholinos. GFP⁺ neutrophils are boxed (green) and their percentage over the total cell number is shown. Pools of 15 larvae. Representative of four independent experiments.

zCxcl8 Imposes an Orthotactic Bias on Neutrophil Random Walk (Related to Figure 2)
(A) Evolution of the Mean Square Displacement (MSD) with time in log-log scale. Blue color is associated to cells inside a radius of 100μm and green color is associated to the complete set of cells in the image field. Red line slope is 1.2 and black line slope is 1. Note that the MSD starts super diffusive with a slope (log-log scale) close to 1.2 but eventually returns to a slope close to 1 in both cases.
(B) Evolution of the normalized displacement correlation K(τ,t) with time τ. Color code is identical to A. Solid lines are exponential fits.
(C) Evolution of the nematic order parameter S with the distance from the center of the transplant.
(D) (Left) Representative reference image showing the mCherry/zCxcl8-expressing transplanted cells in red, GFP+ neutrophils in green and a 150μm radius in white. (Right) Evolution of the average normalized speed with the orientation of the motion. Color code is identical to A. Note the strong discrepancy between the green and blue curve in the zCxcl8 case, which is not observed in the control case.
(E) Evolution of the propagator, i.e. the probability density function (Pdf) of neutrophil displacement r(t) at different times t, in larvae treated or not with zCxcl8 sMO. Colors are associated with times: black for t=1 min, red for t=2 min, green for t=4 min, blue for t=8 min and cyan for t=16 min. Plots are in log-log scale. Inserts are the collapsed curves, P(ρ), into a unique Gaussian. Black line is the Gaussian fit.
(F) Evolution of the Mean Square Displacement (MSD) with time in log-log scale. Blue color is associated to cells inside a radius of 100μm and green color is associated to the complete set of cells. Red line slope is 1.2 and black line slope is 1. Note that the MSD starts super diffusive with a slope (log-log scale) close to 1.2 but returns to a slope close to 1 in both cases.
(G) Evolution of the normalized displacement correlation K(τ,t) with time τ.. Color code is identical to A. Solid lines are exponential fits.
(H) Evolution of the nematic order parameter S with the distance from the center of the infection site.
(I) (Left) Representative image showing DsRed+ E. coli in red, GFP+ neutrophils in green and a 150μm radius in white. (Right) Evolution of the average normalized speed with the orientation of the motion. Color code is identical to A. Note the strong discrepancy between the green and blue curve in the control case, which is less marked in the case of the zCxcl8 knockdown.
(A-I) Data are from 4 larvae in 2-4 experiments. (A, B, C, D, F, G, H, I) Error bars, SEM

zCXCL8 Overexpression Inhibits Neutrophil Motility (Related to Figure 3)
(A) Representative SDCM images of larvae injected with farnesylated (cell membrane-associated) mCherry (fmCherry) mRNA, or co-injected with fmCherry and zCxcl8 mRNA, or injected with zCxcl8-mCherry fusion mRNA.
(B) Average neutrophil step speed in mRNA-injected larvae. fmCherry: n=25258; zCxcl8+fmCherry: n=29128, zCxcl8-mCherry: n=15183. Mann Whitney test. Data are from 5-7 larvae in 2 experiments. Error bars, SEM.

zCxcl8 Guides Neutrophils as HSPG-Bound Gradients (Related to Figure 4)
(A) zCxcl8-mCherry and sec-mCherry expression constructs. SS, zCxcl8 signal sequence (amino acids 1-22).
(B) Western blot detection of mCherry in culture supernatants of HEK293T cells, transfected with mCherry, sec-mCherry and zCxcl8-mCherry.
(C) In vitro chemotactic response of adult zebrafish neutrophils to concentrated conditioned media from sec-mCherry or zCxcl8-mCherry-transfected HEK293Ts.
(D) Western blot detection of mCherry in extracts of 2 dpf larvae ubiquitously expressing injected zCxcl8-mCherry or sec-mCherry mRNA. Supernatants of dissociated larvae represent „soluble fractions whilst residual larval extracts represent „tissue bound fractions. Pools of 15 larvae. *, mCherry degradation product.
(E) Alignment of amino acid sequences of human CXCL8, zebrafish zCxcl8, mouse CXCL1 and CXCL2, indicating conserved HSPG-binding residues (blue asterisks) and the residues that were mutated to alanines for construction of double mutant zCxcl8 (in red boxes).
(F) In vitro chemotactic response of adult zebrafish neutrophils to wild-type zCxcl8 (zCxcl8-WT) and double mutant zCxcl8 (zCxcl8-DM). Data representative of three experiments.
(G and H) Sensograms showing the binding of increasing concentrations (brown, 1 μM; violet, 0.75 μM; turquoise, 0.5 μM; yellow, 0.37 μM; pink, 0.25 μM; blue, 0.18 μM) of zCxcl8-DM (G) or zCxcl8 (H) to immobilized heparan sulfate. Data representative of three experiments.
(I) (Left) Representative SDCM reference image of neutrophils (green) within the area of transplanted cells (red). Cube indicates example area, such that was used for motility analysis shown right. (Right) Average track speed in the presence or absence of zCxcl8 or zCxcl8-DM. mock: n=66; zCxcl8: n=290; zCxcl8-DM=110 (Mann Whitney test). Y, yolk sac.
(J) (Left) Representative SDCM image of neutrophils (green) within the area of transplanted cells (red). Cubes indicate example areas within < 100 μm from the transplant, such that were used for the analysis shown right. (Right) Average track speed in the presence of absence of zCxcl8 or zCxcl8-DM. mock: n=92; zCxcl8: n=97; zCxcl8-DM=95 (Mann Whitney test). Y, yolk sac.
(K and L) Neutrophil recruitment to E. coli within larvae treated with Heparinase III (Hep III) or buffer (control). (K) SDCM images at 75 min after infection. (L) Neutrophil recruitment to E. coli quantified at 2 hpi. Control: n= 19; Hep III: n= 21 larvae (Student′s t test). Data representative of four independent experiments.
M) Evolution of the average normalized speed with the orientation of motion. Blue color is associated to cells inside a radius of 100 μm and green color is associated to the complete set of cells. Note the strong differences between the green and blue curve for the control (left) and the almost identical evolution for heparinase-treated larvae (Hep III) (right).
(N) Average speed values of the cells inside a radius of 100μm (left) and between radius of 100 μm and 200 μm (right).
Error bars, SEM. ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.0001. Data in I, J, M, N are pooled from 3-4 larvae per group in 2-4 experiments. Scale bars, 100 μm (I and J) and 50 μm (K).