JBL generate luminal connectivity.

a: High resolution time lapse of EGFP-ZO1 and mRuby2-UCHD during junctional rearrangements. Yellow asterisks point JBL. yellow double-headed arrows indicate the distance between the junction and a reference point within the cell cytoplasm (Supplementary movie 1). b: Time-lapse stills of EGFP-Podxl2 and mRuby2-UCHD during luminal fusion, starting at approximately 32 minutes. The leading edges of the converging junctions are indicated by white arrowhead. Yellow arrowheads point to JBL. Scale bar 10 µm (Supplementary movie 2) c: Graphic depiction of blood vessel lumenization by cell convergence. Top: At first, cells 1 and 3 and cell 2 and 3, respectively, are forming cell-cell interfaces (cell junctions: green), which enclose local lumens (apical membrane: purple). Middle: JBL drive convergent cell movements, which lead to the formation of a novel cell contact (cells 1 and 2) and the merging of the two lumens into one (bottom). Graphic depiction of a lumenized DLAV, displaying the multicellular architecture. d: timelapse movie (Supplementary movie 3) showing the oscillatory behavior of JBL (yellow arrowhead) during DLAV formation (30hpf). The white dashed line surrounds the UCHD-labeled JBL domain. Scale bar 5 µm. Schematic reproduction of the time lapse in the bottom panel.

JBL form new junctions at the distal e nd of the actin protrusion.

a: Time-lapse of VE-cad-Venus imaged at rate of 1stack/12s during distal junction formation. (supplementary movie 4) New distal junctions emerge in clusters at the distal end of the JBL. Ve-cad is diffusely localized during early step of JBL, while it accumulates in big foci at later time points. Similar observations were made in 9 separate movies. Scale bar 2 µm. On the right panels we present a simplified reproduction of labeled VE-cad spatiotemporal distribution during distal junction formation. b: Time-lapse of mRuby2-UCHD and VE-cad-Venus during JBL (0s) and distal junction formation (30s). (supplementary movie 5). The dashed line encircles the protrusion. White arrow heads point distal junction foci at the distal tip of the protrusion. Scale bar 2 µm. Similar observations were made in 15 movies. c: Time-lapse of EGFP-ZO1 and mRuby2-UCHD during the presence of proximodistal junction. After distal junction formation, f-actin gradually diminishes from proximal junction and interjunctional space, while maintaining strong localization at the distal junction. Similar observations were made in 5 movies. White arrowheads confine UCHD expression domain. Scale bar 2 µM. d,e Still image of EGFP-ZO1 and mRuby2-UCHD (D), and mRuby2-UCHD and VE-cad-Venus (E) respectively, in proximodistal junction. Supplementary video 6) While junctional proteins localize more strongly at the proximal junction, respect to the new, immature distal junction, f-actin faintly localize at the proximal junction and strongly at the distal. Similar observations were made in 10+11 videos. Scale bar 2 µm. f Schematic representation the spatiotemporal distribution of Cdh5, F-actin and ZO-1 during formation of the distal junction in top and side view: Initially Cdh5 appears diffusely dispersed throughout the membrane protrusion. Distal junction foci form at the distal side of the lamellipodia. Actin network along the distal junction. Finally, junctional actin network along the proximal junction dissolves, as pointed by orange arrow heads.

Arp2/3 localization oscillates at the distal end of JBL.

a Time-lapse (Supplementary movie 8) of Arpclb-Venus and ZO1-tdTomato at about 30 hpf, showing deposition of Arp2/3 at the distal side of the junctional ring during 2 JBL cycles. Scale bars 2 µm. b Time-lapse (Supplementary movie 9) Of Arpc1b-Venus and mRuby2-UCHD during a JBL event. Scale bar 2 µm. c and c’ Plots showing the average intensity of Arpc1b-Venus and mRuby2-UCHD along rectangular ROIs drawn along the JBL at 30 and 60 seconds, respectively. d Schematic model of top and side view of spatiotemporal localization of ARP2/3, f-actin and ZO1 during JBL formation and extension for 1.5 JBL cycle.

Arp2/3 is required for junctional elongation and JBL formation.

a Still-images from a movie (Supplementary movies 10, 11) of JBL labelled with EGFP-UCHD, at around 32 hpf, in the presence of DMSO (1%) or 1h incubation with CK666 (200 µM). Similar observations were made in 10 movies. Scale bar 5 µm. Similar observations were made in more than 12 movies. Scale bar 5 µm. b Quantification of number of filopodia/JBL in DMSO (1%), n=10 JBL and CK666 (200 µM), n = 10 JBL events. Unpaired t-test was used for statistical analysis. (P value < 0.0001). c Quantification JBL filopodia length in DMSO (1%), n=15 JBL and CK666 (200 µM), n = 34 JBL events. d Still images from time-lapse movies (Supplementary movies 12, 13) of an Zo1-tdTomato labeled junctional rings in the presence of DMSO (1%), CK666 (200 µM). Top panels t =0 and bottom panels after 1 h incubation. Scalebar 5 Scale bar 10 µm. e Quantification of the junctional elongation velocity in DMSO (1%), n =21 junctions (10) embryos) and CK666 (200 µM), n = 24 (10 embryos). Dotted line indicated no movement observed, black lines are medians. Unpaired t-test was used for statistical analysis. (P value = 0.0047).

MLC is enriched at the junction pole.

a: Time lapse (Supplementary movie 14) of VE-cad-Venus during junctional merging. Blue and red arrowheads point proximal and distal junction respectively. The 2 junctions are gradually moving closer until they merge. Similar observations were made in 13 movies. Scale bars 2. b: Confocal stack of ISV and DLAV of Myl9b-mCherry and ZO1-EGFP expressing embryo at around 32 hpf. Yellow arrow heads point JBLs. Dashed lines underline Myl9b-mCherry distal expression domain at the junction pole. Scale bars 2 µm. c: Confocal stack of DLAV and ISV immunostaining against GFP (green), VE-cad (magenta) and ZO1 (cyan) of an embryo expressing myl9a-GFP at around 32hpf in a case of mosaic expression: the left cell has a strong Myl9-GFP expression (as in the schematic). c’ and c” dashed squares delimitate JBL regions of the 2 cells. e quantification of Myl9-EGFP average intensity the yellow dashed squares ROIs d’and d” drawn on the JBL domain. The average intensity inside the ROIs was corrected subtracting the cytoplasmic background. The cytoplasmic average background was calculated drawing 5 squares in non-junctional cytoplasmic region of the highly Myl9-GFP expressing cell. f. Still images of a Myl9a-GFP and mRuby2-UCHD during JB event anastomosis in PCeV at around 60 hpf Imaged with spinning disc (Supplementary movie 15). White dashed lines delimitate JBL region. Yellow arrowheads point distal junction. Myl9-GFP is enriched inside the lamellipodia and localize at the distal junction in later time points. Scale bars 2 µm.

Myosin light-chain dynamics correlates with junctional merging.

a Time-lapse (Supplementary movie 16) of Cdh5ΔC-EGFP and Myl9b-mCherry during junctional merging. Blue dash lines confine the applied mask. Blue and black arrowheads point to MLC accumulation on proximal and distal junction. Similar observations were done in 6 videos. Grays arrowheads point interjunctional newly recruited Myl9-mCherry. Similar observations were made in 9 videos. Scale bars 2 µm. b Dashed blue lines delimitate 3 regions: proximal junction region (P), distal junction region (D) and interjunctional space (I). c Plot of average intensity of Myl9b-mCherry in the 3 regions, over time. The intensities have been corrected for background which has been evaluated as average of average intensities in 5 rectangles in the cell cytoplasm.

Actomyosin contractility drives junctional conversion.

a S till images from movies (Supplementary movies 21 and 22) EGFP-UCHD labelled junctional rings around 32 hpf, in the presence of DMSO (1%) or Y27632 (45 µM). Top panels t = 0 and bottom panels 60 min incubation. b Quantification of the junctional elongation velocity in DMSO (1%), n =11 junctions and CK666 (200 µM), n = 8. Dotted line indicated no movement observed, black lines are medians. Unpaired t-test was used for statistical analysis. P value = 0.0081. c Time-lapses (supplementary movies 23 and 24) VE-cad-Venus labelled proximodistal junction, during junctional merging in the presence of DMSO (1%) or Y27632 (75 µM). White and yellow arrowheads are pointing distal and proximal junctions respectively. d Tracking of proximal-distal junction distance over time of single junctional merging events in DMSO (1%) (green lines), and Y27632 (75 µM) treated embryos (magenta lines). e Quantification of the persistence of proximal and distal junction in DMSO (1%), n =13 junctional merging events and Y27632 (75 µM) n = 20. Unpaired t-test was used for statistical analysis. P value = 0.0027.

Schematic representation of different steps during JBL oscillation.

(see supplementary Movie 25) JBL formation is initiated by Arp2/3 activation. The JBL is pushed forward by F-actin polymerization. At the distal end a new cell-cell junction is formed. MyosinII is recruited to the interjunctional space. Actomyosin contraction merges the proximal and distal junctions resulting in an overall elongation of the junctional ring.

The distal junctions form de novo.

a: Schematic representation of the experimental design. At first mClav2 labeled VE-cad is photoconverted in one pole of the junctional ring. Then time lapse imaging is performed on the half-converted ring. b: Time-lapse (Supplementary movie 7) of a DLAV junctional ring of an embryo expressing Cdh5-mClav2 before and right-after photoconversion. The red dashed square delimitates the photoconverted area. White dashed square demarcates the zoomed-in area of c. c: Time-lapse of the zoomed in junctional ring pole. right-after photoconversion and 4 min later. White and yellow arrowheads point distal and proximal junction respectively. Distal junction is labeled by green, non-photoconverted VE-cad; but not by red photoconverted VE-cad.

CK666 disrupts Arp2/3 localization in DLAV anastomotic junctions.

Stills of Arpclb-Venus and ZO1-td tomato before (a, c) and after 30 min treatment with DMSO 1% c (n=17) and CK666 200 µM d (n=25). T0 around 30 hpf n>20. White arrow heads are pointing at junctional poles. a’, b’, c’, d’ are magnifications of the white dashed square delimited areas.

Myosin light-chain recruitment at the interjunctional space during merging.

a, a’: Time-lapse of Cdh5ΔC-EGFP and Myl9b-mCherry during doble junction state (Supplementary movie 17). a: Still-image showing polarized accumulation of MLC during “double junction” state. a’: Still-images showing recruitment of MLC between proximal and distal junction. b: Still-images from a time-lapse (Supplementary movie 18) of VE-cad-Venus and Myl9b-mCherry showing increase of MLC in the junctional space prior to proximodistal junction mergence. c, d: Time-lapses of Cdh5ΔC-EGFP and Myl9b-mCherry (Supplementary movies 19, 20) showing recruitment of MLC between proximal and distal junction during “double junction” state.