a–f Maximum intensity projection of confocal z-stacks of trunk vessels at different stages of zebrafish development. Cropped images are single-plane images of the z-stack. During lumen expansion of ISVs from 30 to 34 hpf embryos, higher levels of actin (a, Lifeact) and non-muscle myosin II (d, Myl9b) are assembled at the apical cortex of the posterior region of the lumen (iii in a, ii in d) compared to the expanding anterior region of the lumen (i and ii in a, i in d), which contains very little or no actomyosin. At 2 and 3 dpf, distinct actin (b, c) and non-muscle myosin II (e, f) are detected in the apical cortex of fully lumenised vessels. Images are representative of 6 (a, n = 2 independent experiments), 9 (b, n = 5 independent experiments), 7 (c, n = 4 independent experiments), 23 (d, n = 3 independent experiments), 8 (e, n = 3 independent experiments) and 4 (f, n = 2 independent experiments) embryos. g Quantification of apical Lifeact and Myl9b intensity levels in the anterior and posterior segments of the same lumen at 30–34 hpf. Each data pair represents one ISV (Lifeact: n = 4 ISVs from 4 embryos; Myl9b: n = 6 ISVs from 5 embryos). h, i Single-cell expression of Marcksl1a-EGFP in 34 hpf Tg(kdr-l:ras-mCherry)^s916 embryo (h, apical enrichment was observed in 5 out of 5 embryos from 3 independent experiments) and Marcksl1b-EGFP in 38 hpf Tg(fli1:myr-mCherry)^ncv1 embryo (i, apical enrichment was observed in 20 out of 20 embryos from 6 independent experiments). Arrows, apical cortex; arrowheads, basal cortex; dashed boxes, the magnified regions; DA dorsal aorta; DLAV dorsal longitudinal anastomotic vessel; ISV intersegmental vessel; L lumen; PCV posterior cardinal vein. Scale bars, 5 µm (a–f) and 10 µm (h, i). Source data are provided as a Source data file.

a, d Plasmid constructs encoding GAL4FF/UAS-driven expression of wildtype or mutated Marcksl1a (a) and Marcksl1b (d) proteins trailed by a self-cleaving T2A peptide fused to mKate2-CAAX. b, e Maximum intensity projection of confocal z-stacks of ISVs of 2 dpf Tg(fli1:GAL4FF)^ubs3;Tg(fli1ep:Lifeact-EGFP)^zf495 or Tg(fli1:GAL4FF)^ubs3;Tg(fli1:myr-EGFP)^ncv2 embryos expressing constructs encoding wildtype or mutated Marcksl1a (b) and Marcksl1b (e). Marcksl1 overexpressing cells are in magenta (cells expressing mKate2-CAAX). Arrows indicate dilated vessel. c Quantification of aISV, vISV and DLAV diameter in control and Marcksl1a-overexpressing blood vessels (control: n = 60 aISVs/51 vISVs/54 DLAVs from 85 embryos; wildtype Marcksl1a: n = 17 aISVs/16 vISVs/16 DLAVs from 25 embryos; Marcksl1aΔED: n = 16 aISVs/20 vISVs/16 DLAVs from 24 embryos; Marcksl1a-T124A: n = 43 aISVs/44 vISVs/35 DLAVs from 33 embryos; Marcksl1a-T124D: n = 43 aISVs/32 vISVs/24 DLAVs from 29 embryos). f Quantification of aISV, vISV and DLAV diameter in control and Marcksl1b-overexpressing vessels (control: n = 62 aISVs/35 vISVs/67 DLAVs from 47 embryos; wildtype Marcksl1b: n = 23 aISVs/14 vISVs/20 DLAVs from 24 embryos; Marcksl1bΔED: n = 21 aISVs/22 vISVs/25 DLAVs from 23 embryos; Marcksl1b-T162A: n = 31 aISVs/15 vISVs/22 DLAVs from 21 embryos; Marcksl1b-T162D: n = 54 aISVs/42 vISVs/34 DLAVs from 34 embryos). Violin plots represent the entire range of values, dotted lines indicate first and third quartiles, center lines are median. Mean values are indicated. Data are collected from 2 (Marcksl1bΔED), 3 (Marcksl1a-T124A, Marcksl1a-T124D, Marcksl1b-T162A and Marcksl1b-T162D) and 4 (Marcksl1aΔED, wildtype Marcksl1a and Marcksl1b) independent experiments and analyzed by ordinary one-way ANOVA with Sidak’s multiple comparisons test. ED effector domain; DLAV dorsal longitudinal anastomotic vessel; ISV intersegmental vessel; aISV arterial ISV; vISV venous ISV. Scale bars, 20 µm. Source data are provided as a Source data file.

a Maximum intensity projection of confocal z-stacks of zebrafish trunk vessels of 2 dpf wildtype, marcksl1a^rk23, marcksl1b^rk24 and marcksl1a^rk23;marcksl1b^rk24 embryos in Tg(fli1ep:Lifeact-EGFP)^zf495;Tg(kdr-l:ras-mCherry)^s916 background. b, c Quantification of aISV, vISV and DLAV (b) and DA and PCV (c) in wildtype and marcksl1 mutant embryos (wildtype: n=26 aISVs/27 vISVs from 15 embryos; marcksl1a^rk23: n = 43 aISVs/63 vISVs from 29 embryos; marcksl1b^rk24: n = 79 aISVs/57 vISVs from 39 embryos; marcksl1a^rk23;marcksl1b^rk24: n = 74 aISVs/46 vISVs from 28 embryos). Data are collected from 2 (marcksl1a^rk23;marcksl1b^rk24 embryos), 3 (marcksl1a^rk23 and marcksl1b^rk24 embryos) and 4 (wildtype embryos) independent experiments. Violin plots represent the entire range of values, dotted lines indicate first and third quartiles, center lines are median. Mean values are indicated. Statistical significance was determined by ordinary one-way ANOVA with Tukey’s multiple comparisons test. d–f Microangiography in wild type and marcksl1 mutant embryos. Images are representative of wildtype, marcksl1a^rk23, marcksl1b^rk24 and marcksl1a^rk23;marcksl1b^rk24 embryos at 2 dpf. Quantification of lumenisation of ISVs in wildtype and marcksl1 mutant embryos at 2 dpf (e, n = 1472 ISVs from 32 wildtype embryos; n=1058 ISVs from 23 marcksl1a^rk23 embryos; n = 1150 ISVs from 25 marcksl1b^rk24 embryos; n=1564 ISVs from 34 marcksl1a^rk23;marcksl1b^rk24 embryos) and at 3 dpf (f, n = 1518 ISVs from 33 wildtype embryos; n = 1104 ISVs from 24 marcksl1a^rk23 embryos; n = 782 ISVs from 17 marcksl1b^rk24 embryos; n = 1748 ISVs from 38 marcksl1a^rk23;marcksl1b^rk24 embryos). ISVs no. 4-26 were analyzed on both sides of the embryo. Data are collected from three independent experiments and analyzed by ordinary one-way ANOVA with Dunnett’s multiple comparisons test. DA dorsal aorta; DLAV, dorsal longitudinal anastomotic vessel; ISV intersegmental vessel; aISV arterial ISV; vISV venous ISV; PCV posterior cardinal vein. Scale bars, 50 µm (a) and 100 µm (d). Source data are provided as a Source data file.

a Live imaging of an ISV composed of EC with ectopic expression of Marcksl1b from 48 hpf reveals changes in cell shape and fluctuation in vessel diameter. (n = 5 embryos, three independent experiments). Dashed line, region of vessel measured. b–d Overexpression of Marcksl1 increases EC size while depletion of marcksl1b or marcksl1a and marcksl1b decreases EC size in vivo. Methodology to perform in vivo single-cell shape analysis (b). Quantification of cell area (c) and cell aspect ratio (d) of lynEGFP- or Marcksl1b-EGFP-expressing ECs surrounded by wildtype (WT) cells, marcksl1^rk23;marcksl1b^rk24 ECs transplanted into WT embryo and ECs of marcksl1b^rk24 or marcksl1^rk23;marcksl1b^rk24 embryos at 2 dpf. e Diameter of ISVs composed of single marcksl1^rk23;marcksl1b^rk24 EC transplanted into WT embryo at 2 dpf (n = 10 vessels, n = 5 independent transplantations). f–m Overexpression and knockdown of Marcksl1 increases and decreases, respectively, HUVEC size (n = 3 independent transfections). Maximum intensity projection of confocal z-stacks of control (EGFP) and Marcksl1-overexpressing (Marcksl1-EGFP) HUVECs 1 and 2 days post transfection (dpt, f). Quantification of cell area (g), cell spikiness index (h) and cell aspect ratio (i, EGFP: n = 423/363 cells at 1/2 dpt; Marcksl1-EGFP: n = 358/298 cells at 1/2 dpt). Maximum intensity projection of confocal z-stacks of control and MARCKSL1 knockdown HUVECs stained for DAPI and phalloidin at 1 and 2 dpt (j). Plasmid expressing U6 promoter-driven MARCKSL1 shRNA was used with CMV promoter-driven EGFP-CAAX expression as an internal marker. Plasmid expressing scrambled shRNA was used as a control. Quantification of cell area (k), cell spikiness index (l) and cell aspect ratio (m) after MARCKLS1 knockdown (shControl, n = 223/278 at 1/2 dpt; shMARCKSL1, n = 474/342 at 1/2 dpt). Statistical significance was assessed by ordinary one-way ANOVA with Sidak’s multiple comparisons test (c, d) and two-tailed unpaired t-test (e, g–i, k–m). Violin plots in (c–e, g–i and k–m) represent the entire range of values, dotted lines indicate first and third quartiles, center lines are median. Mean values are indicated. Scale bars, 20 µm (a), 50 µm (f, j) and 25 µm (cropped images in f, j). Source data are provided as a Source data file.

a, b Marcksl1 regulates filopodia formation. Increased expression of Marcksl1a in ECs (magenta) increases filopodia formation when compared to wildtype EC (green) at 31 hpf (a). Dashed box, region of vessel measured. Quantification of wildtype EC or ECs overexpressing Marcksl1b (Marcksl1b OE) or depleted of marcksl1a and marcksl1b at 1 and 2 dpf (b, 1 dpf: wildtype, n = 87 ISVs/ 29 embryos; Marcksl1b OE, n = 32 ISVs/19 embryos; marcksl1a^rk23;marcksl1b^rk24, n = 67 ISVs/16 embryos. 2 dpf: wildtype, n = 29 ISVs/10 embryos; Marcksl1b OE, n = 18 ISVs/15 embryos; marcksl1a^rk23;marcksl1b^rk24, n = 126 ISVs/28 embryos). Statistical significance was assessed by ordinary one-way ANOVA with Sidak’s multiple comparisons test. Mean values are indicated. c Overexpression of Marcksl1a in ECs causes ectopic membrane blebbing during lumen expansion. Apical membrane (arrow) of an endothelial tip cell expands until the leading edge of the tip cell, at which point ectopic basal blebs (arrowhead) protrude into the surrounding tissue. Movies were taken from 31 hpf. d Overexpressing of Marcksl1b in EC (magenta) induces ectopic basal blebs (arrowhead) in the ISV of 54 hpf embryo. e Quantification of membrane blebbing shows a significant increase in blebbing in ECs with increased expression of Marcksl1b when compared to wildtype EC of ISVs in 2 dpf Tg(kdr-l:ras-mCherry)^s916 embryos (wildtype, n = 21 cells; Marcksl1b OE, n = 20 cells). Statistical significance was assessed by two-tailed unpaired t test with Welch’s correction. f, g Marcksl1-induced blebs are comprised of apical and basal membranes (green). Newly formed basal blebs (arrowhead) are filled with Dextran (magenta). 00:00; minutes:seconds. Dashed box in f, a magnified region. Violin plots in b, e represent the entire range of values, dotted lines indicate first and third quartiles, center lines are median. Scale bars, 5 µm (f) and 10 µm (a, c, d). Source data are provided as a Source data file.

The spatiotemporal dynamics Lifeact (actin, a, b; representative images of 17 embryos from 4 independent experiments; secondary bleb formation was observed in 27/100 blebs, retraction in 52/100 blebs and persistent blebs in 27/100 blebs) and Myl9b (non-muscle myosin II, c representative image of 21 retracting blebs from 5 embryos from 2 independent experiments) at the cortex of Marcksl1b-induced bleb are depicted in time-lapse images (i and ii in a, i in b, c), kymographs (iii in a, ii in b, c) and as an average intensity over time (iv in a, iii in b, c). Negative correlation between actin (a) or non-muscle myosin II (c) intensity and bleb area is observed in retracting blebs while little change in actin intensity is observed in blebs that do not retract (b). *, secondary bleb. Arrow, increase in actin or non-muscle myosin II assembly. Scale bar, 2 µm.

a Local weakening of EC cortex induces basal bleb. Laser ablation (magenta line) of an arterial ISV was performed in a 3 dpf Tg(kdr-l:ras-mCherry)^s916 embryo. Bleb formation was observed in 18 out of 35 ISVs (n = 15 embryos, 2 independent experiments). Arrow, basal bleb. b, c Short-term inhibition of actin polymerisation in 2 dpf Tg(kdr-l:ras-mCherry)^s916 embryo leads to increased membrane bleb generation at both apical (arrow) and basal (arrowheads) membranes (b). Quantification of basal blebs along the length of ISVs in 49–52 hpf Tg(kdr-l:ras-mCherry)^s916 embryos after treatment with 0.3 µg/ml Latrunculin B for 2.5–3 h (c, 0.03% DMSO treatment: aISVs, n = 37 ISVs/25 embryos; vISVs, n = 39 ISVs/25 embryos; Latrunculin B treatment: aISVs, n = 36 ISVs/22 embryos; vISVs, n = 39 ISVs/24 embryos). Data were collected from three independent experiments. Violin plots represent the entire range of values, dotted lines indicate first and third quartiles, center lines are median. Statistical significance was determined by two-tailed unpaired t-test. d Basal blebs (arrowhead) induced by increased expression of Marcksl1b are suppressed upon decreasing blood flow using 4X tricaine (i and ii, reduced blebbing was observed in 7 out of 8 embryos from 6 independent experiments). Reestablishment of normal blood flow (1X tricaine) increases apical membrane blebbing (arrow in i) and vessel diameter (iii). Triangle, filopodia. 00:00, hh:mm after start of treatment. Dashed boxes, magnified region in b, d. ISV intersegmental vessel; aISV arterial ISV; vISV venous ISV. Scale bars, 5 µm (a, b and iii in d) and 10 µm (i in d). Source data are provided as a Source data file.

a–c EC cortex contains a dynamic meshwork of actomyosin. Maximum intensity projection of Airyscan images of actin in EC from an ISV of a 2 dpf Tg(fli1ep:GAL4FF)^ubs3; Tg(UAS:EGFP-UCHD)^ubs18 embryo (a, actin meshwork is observed in 48 out of 63 images from 7 embryos from 5 independent experiments) and the apical cortex of HUVEC stained with Phalloidin (b, 21 cells from 4 independent experiments). Time-lapse imaging of actin and non-muscle myosin II in EC of 2 dpf Tg(fli1:Lifeact-mCherry)^ncv7;Tg(fli1ep:EGFP-myl9b)^rk25 embryo (c). Lifeact and Myl9b intensity in boxed region was quantified. Dynamic F-actin intensity was observed in 23 ISVs/12 embryos from 5 independent experiments. Dynamic Myl9b intensity was observed in 14 ISVs/6 embryos from 2 independent experiments. d Maximum intensity projection of 2 z-slices from the apical cortex of HUVEC expressing Marcksl1-EGFP and stained with Phalloidin reveals co-localisation of Marcksl1 and actin bundles (arrowheads, 23 cells from 5 independent experiments). e, i Analysis of actin density and bundle width after overexpression of Marcksl1, Marcksl1-AAA and Marcksl1-DDD (e–g) and knockdown of MARCKSL1 (h, i) in HUVECs. Single slice Fast Airyscan images of the apical cortex of HUVEC stained with Phalloidin reveal decreased actin density surrounding actin bundles (arrow) in Marcksl1-EGFP-transfected cells 1 day post transfection (e). f–i Mean values are indicated. Three independent experiments were performed (EGFP, 51 ROIs from 28 cells; Marcksl1-EGFP, 86 ROIs from 29 cells; Marcksl1-AAA-EGFP, 76 ROIs from 28 cells; Marcksl1-DDD-EGFP, 77 ROIs from 32 cells; shControl, 66 ROIs from 24 cells; shMARCKSL1, 74 ROIs from 29 cells). Violin plots represent the entire range of values, dotted lines indicate first and third quartiles, center lines are median. Data was analysed by ordinary one-way ANOVA with Sidak’s multiple comparisons test (f, g) and two-tailed Mann–Whitney test (h, i). j Model illustrating Marcksl1 favours the formation of linear actin bundles at EC cortex. Scale bars, 2 µm (a–e). Source data are provided as a Source data file.

a Endothelial-specific ectopic expression of Fascin1a (Fscn1a)-T2A-mKate2CAAX (magenta) at 30 hpf. Box, region of DLAV undergoing lumenisation. Arrow, expanding apical membrane (00:08). Arrowheads, basal blebs (from 02:16). Blebbing was observed in 7 out of 8 ECs with overexpression of Fscn1a (n = 6 embryos from two independent experiments). b Treatment of a 2 dpf Tg(kdr-l:ras-mCherry)^s916 embryo with 400 µM CK666 for 1 hour induces basal blebs (i and ii, arrowheads) and apical blebs (iii, arrow; 4 out of 4 embryos from one experiment). Boxes are observed regions. c Increase in linear actin bundle formation modifies EC membrane behavior and response to haemodynamic forces. 00:00, hours:minutes. Scale bars, 10 µm (a) and 5 µm (b).