Zebrafish mutants and TEAD reporters reveal essential functions for Yap and Taz in posterior cardinal vein development

As effectors of the Hippo signaling cascade, YAP1 and TAZ are transcriptional regulators playing important roles in development, tissue homeostasis and cancer. A number of different cues, including mechanotransduction of extracellular stimuli, adhesion molecules, oncogenic signaling and metabolism modulate YAP1/TAZ nucleo-cytoplasmic shuttling. In the nucleus, YAP1/TAZ tether with the DNA binding proteins TEADs, to activate the expression of target genes that regulate proliferation, migration, cell plasticity, and cell fate. Based on responsive elements present in the human and zebrafish promoters of the YAP1/TAZ target gene CTGF, we established zebrafish fluorescent transgenic reporter lines of Yap1/Taz activity. These reporter lines provide an in vivo view of Yap1/Taz activity during development and adulthood at the whole organism level. Transgene expression was detected in many larval tissues including the otic vesicles, heart, pharyngeal arches, muscles and brain and is prominent in endothelial cells. Analysis of vascular development in yap1/taz zebrafish mutants revealed specific defects in posterior cardinal vein (PCV) formation, with altered expression of arterial/venous markers. The overactivation of Yap1/Taz in endothelial cells was sufficient to promote an aberrant vessel sprouting phenotype. Our findings confirm and extend the emerging role of Yap1/Taz in vascular development including angiogenesis.

The reporter construct was used to develop a Hsa.CTGF-based YAP1/TAZ zebrafish reporter. Hsa.CTGF promoter fragment was cloned into a Gateway 5′ entry clone (p5E-MCS) and used to create the pDest(Hsa.CTG-F:nlsmCherry) and pDest(Hsa.CTGF:eGFP)) (Fig. 1B). One-cell stage zebrafish embryos were injected with each destination vector together with Tol2 transposase mRNA. Mosaic transgenic fish displaying a strong fluorescence at 24 hpf were selected and raised to adulthood. An average of 76% (13/17) of the injected fish prescreened for mosaic fluorescence were found transmitting the transgene to their offspring. All the offspring from different founder fish exhibited a similar reporter protein expression pattern, displaying strong fluorescence in identical anatomical districts, such as the lens and otic vesicles, the heart, the pharyngeal arches and the vasculature (Figs 1C and S2). Two founders were selected to establish stable reporter lines containing a single allele in their germline: Tg(Hsa.CTGF:eGFP) ia48 and Tg(Hsa.CTGF:nlsmCherry) ia49 .
Hsa.CTGF-based zebrafish transgenic lines are bona fide Yap1/Taz reporter. To validate whether the Hsa.CTGF-based zebrafish reporter lines can be used to visualize the endogenous activity of YAP1/TAZ in vivo, we knocked down zygotic expression of both yap1 and taz by co-injecting one-cell stage embryos with two splice morpholinos targeting respectively yap1 and taz pre-mRNAs. mCherry expression of the Tg(Hsa.CTGF:nlsmCherry) ia49 reporter line was significantly reduced throughout the entire embryo compared to control morpholino-injected embryos ( Fig. 2A). The knockdown of endogenous Yap1 and Taz proteins was confirmed by Western blotting using whole embryo extracts (Fig. S1B).
To test whether the reporter line was able to visualize increased YAP1/TAZ activities, we injected in one-cell stage embryos of the Tg(Hsa.CTGF:nlsmCherry) ia49 line the mRNAs coding for a constitutively active form of YAP1, TAZ or TEAD (YAP-5SA, TAZ-4SA, and TEAD-VP16 respectively). As expected, the injection of any of these mRNAs significantly increased Hsa.CTGF:nlsmCherry expression ( Fig. 2B-D). A similar responsiveness of the reporter signal upon Yap1/Taz activity modulation was found for Tg(Hsa.CTGF:eGFP) ia48 (data not shown). We also validated more specifically that the reporter signal is dependent on Yap/Taz in a number of different tissues, such as the eye, the heart, the floorplate, and the major axial vessels (Fig. S3).
We further tested the Hsa.CTGF:nlsmCherry transgene by treating the reporter embryos with Dexamethasone (DEX), a synthetic glucocorticoid recently shown to promote the nuclear activity of YAP1 in vitro 39 . 24 hpf Tg(Hsa.CTGF:nlsmCherry) ia49 embryos were exposed to a solution containing 25 μM of DEX and analyzed after 24 h. DEX treated embryos displayed a significant increase of fluorescence when compared with controls, confirming in-vivo that glucocorticoids activate Yap1/Taz/Tead mediated transcription. Interestingly, we also observed that DEX treatment does not increase the reporter signal in Yap1/Taz knockdown embryos, thus confirming that the response induced by glucocorticoids is mediated by Yap/Taz activation. (Fig. S4A-A").
Recently, it has been shown that blood flow can induce nuclear import of YAP1 and its transcriptional activity 18 . We thus tested Hsa.CTGF zebrafish reporter with silent heart (sih/tnnt2a) morpholino to block cardiac contractility and blood flow; sih morphants showed a significant reduction of mCherry transgene expression in endothelial cells when compared to control morpholino-injected embryos (Fig. S4B).
Taken together, these experiments demonstrate that the Tg(Hsa.CTGF:nlsmCherry) ia49 and Tg(Hsa.CTGF:eGFP) ia48 lines can be used to analyze Yap1/Taz activity in vivo.

Tg(Hsa.CTGF:nlsmCherry) ia49 zebrafish reveals in vivo the spatio-temporal activation of Yap1/Taz.
We next sought to identify the embryonic stages and tissues in which the reporter is active. Due to maternal expression, fluorescence is ubiquitously detectable at the dome stage and during epiboly in embryos derived from Tg(Hsa.CTGF:nlsmCherry) ia49 females crossed to wild-type males (Fig. S5A). Consistently, mCherry expression is prominent in the adult ovary (the organ with the strongest reporter signal in adults), where the mCherry protein is clearly visible in the nuclei of the oocytes and the accompanying follicle cells (Fig. S5B,C).
The first zygotic expression of the reporter protein in transgenic embryos (derived from Tg(Hsa.CTGF:nlsmCherry) ia49 males crossed to wild-type females) is detectable at late somitogenesis. At 20 hpf fluorescence is found widely across the developing embryo, with the strongest signal localized in the mesenchyme of the tail bud (Fig. 3A). By 24 hpf, fluorescence is observed in many tissues and organs, such as eyes, heart, midbrain-hindbrain boundary (MHB), rhombencephalon, epidermis, muscles, neural tube, notochord, floorplate and vasculature. Reporter expression is persistent in those districts even during later developmental stages ( Fig. 3B-H). In the eye the signal is strong in the lens and is also present in the neural retina (Fig. 3B). The lens remains strongly fluorescent until adulthood (Fig. 3C). In the dorsal portion of the head, two regions display transgene activation by 24 hpf: the MHB and the rhombencephalon. In the rhombencephalic region Hsa.CTGF:nlsmCherry fluorescence is present in six stripes of cells that follow the metameric organization of rhombomeres ( Fig. 3D). At 72 hpf, besides the lens, the reporter signal is particularly strong in the pharyngeal arches (mainly in the mandibular one), otic vesicles, pectoral fins and heart (Fig. 3I,J). A time series of the reporter expression during embryonic and early larval stage is shown in Fig. S6.
In the heart, fluorescence is already visible at early stages and persists throughout adulthood ( Fig. 3K-M). To confirm the expression in the cardiac progenitors and specifically in the cardiomyocytes, we outcrossed the Tg(Hsa.CTGF:nlsmCherry) ia49 line to the Tg(myl7:GFP) line 40 . As indicated by the co-localization between mCherry and eGFP from 22 hpf, the Yap1/Taz reporter is active in cardiac progenitors as well as in differentiated cardiomyocytes (Fig. 3K,L), in agreement with previously described zebrafish reporter lines 41,42 and with genetic data in mammals, where YAP1 is active together with TEADs to promote the growth of embryonic and fetal cardiomyocytes 11,[43][44][45][46] . Hsa.CTGF:nlsmCherry fluorescence is visible also in the intestinal epithelium along the whole intestine ( Fig. 3N-P). Specific expression in the epithelium of the intestine is demonstrated by fluorescence co-localization at cellular resolution in Tg(Hsa.CTGF:nlsmCherry) ia49 /Tg(gut:GFP) s854 47 double transgenic embryos (Fig. 3P).   To study the role played by Yap1/Taz in zebrafish development, and more specifically in the vasculature, yap1 and taz mutants (yap1 bns19 and taz bns35 , respectively) generated by CRISPR/CAS9 were used 48,49 . The yap1 bns19 allele contains a 41 bp deletion in exon 1 of yap1, while the taz bns35 allele contains a 29 bp insertion in exon 2 of taz. Both frame-shift mutations are predicted to encode a truncated protein (Fig. S7A,B). The reduction of Yap1/Taz activity in the mutants was confirmed by the analysis of the Tg(Hsa.CTGF:nlsmCherry) ia49 reporter line. As expected, we observed a decrease of Hsa.CTGF:nlsmCherry expression in the endothelium of 48 hpf yap1 bns19 and taz bns35 mutant embryos (Fig. S7C,C').
While we did not observe striking vascular phenotypes in taz mutants, the primary vascular phenotypes of yap1 mutants have been described recently 18 . Here, we report additional alterations in the cranial and ocular vasculature: the mesencephalic vein (MsV) and the dorsal longitudinal vein (DLV) were truncated in more than 50% (5/9) of yap1 −/− larvae analyzed at 72 hpf; at 5 dpf, fewer hyaloid vessels could be observed in the eyes (Fig. S8), a number further reduced in colobomatous eyes.
We aimed to investigate whether these PCV alterations might be associated with defects of arterial/venous specification during axial vessel development. In situ hybridization for expression of the arterial marker efnb2a showed that it was not affected in yap1 −/− ;taz +/− embryos at 24 and 32 hpf. On the contrary, expression of mrc1a (a marker for veins and lymphatics 51 ) was altered in yap1 −/− ;taz +/− embryos when compared to WT siblings (Fig. 5E). During development, mrc1a is initially expressed in the presumptive venous progenitors localized in the axial vessel, becoming restricted to the PCV and other veins as well as in lymphatic vessels 51 . In 70% (7/10) of 32 hpf yap1 −/− ;taz +/− embryos, some mrc1a expression was observed in the DA, unlike in WT animals ( Fig. 5E), suggesting a defect in DA specification.
Taken together, these findings confirm the role of Yap1/Taz in vascular development and reveal its specific requirement for the cranial and ocular vascular networks, as well as for the specification and organization of the axial vessels.

The upregulation of Yap1/Taz/Tead-mediated transcription causes aberrant vessel sprouting.
The vascular developmental defects observed in yap1;taz mutants and the emerging role of Yap1/Taz in developmental angiogenesis [16][17][18] , prompted us to explore the effect of the upregulation of Yap1/Taz activity on embryonic angiogenesis in zebrafish. Therefore, we injected a constitutively active form of Taz (TAZ-4SA mRNA) in one-cell stage Tg(kdrl:GFP) s843 embryos. While no alteration was observed in PCV development, at 32 hpf we observed the appearance of abnormal sprouts emerging horizontally from some ISVs in TAZ-4SA injected embryos (Fig. 6A,B-D). These abnormal sprouts pointed mostly toward the adjacent ISVs, and often gave rise to complete anastomosis between two adjacent ISVs (Fig. 6B'). Despite the low frequency of aberrant vessel sprouting in TAZ-4SA injected embryos, this phenomenon was never observed in control fish. In particular, complete anastomosis between adjacent ISVs was never detected in more than 50 injected control embryos analysed (Fig. 6A,B' ,E). To further confirm the aberrant sprouting phenotype, the experiment was repeated using YAP-5SA mRNA, obtaining similar results (data not shown).
We next asked whether the effect on ISVs angiogenesis was due to enhanced Yap1/Taz/Tead-mediated transcription, or to other non-transcriptional functions of Yap1 and Taz proteins. To answer this question, Tg(kdrl:GFP) s843 embryos were injected with TEAD-VP16 mRNA, which constitutively activates Tead target gene transcription independently of Yap1/Taz. Aberrant vessel sprouting and anastomosis between ISVs were found also after the injection of TEAD-VP16 mRNA, thus confirming the vessel sprouting-promoting ability of Yap1/ Taz/Tead-mediated transcription (Fig. 6E).
To address whether the effects of Yap1 and Taz on ISVs was specifically due to their activity in endothelial cells, we designed a new construct, placing TAZ-4SA expression under the control of the fli1a endothelial-specific promoter. To perform this experiment, we used the pDestTol2CG2 transposon backbone, containing the cardiac is depicted. (K) Confocal sagittal section of the heart region of a Tg(Hsa.CTGF:nlsmCherry) ia49 /Tg(myl7:GFP) double transgenic embryo at 22 hpf. (K'-K"') Inset of K: mCherry channel (K'), GFP channel (K"), merge (K"').  52 . We speculated that with these experimental conditions we would have been able to detect the presence of aberrant sprouting phenotype only if the vessel sprouting-promoting ability of Yap1/Taz had been due to its transcriptional activity in endothelial cells. After the injection of pDestTol2CG2-fli1a-TAZ-4SA plasmid in one-cell stage Tg(kdrl:GFP) embryos, we observed that about 44% (7/16) of the mosaic embryos bearing endothelial-specific TAZ overexpression exhibited the aberrant vessel sprouting effect induced by YAP-5SA or TAZ-4SA mRNAs injection (Fig. 6C,E). Moreover, by injecting pDestTol2CG2-fli1a-TAZ-4SA plasmid in Tg(Hsa.CTGF:nlsmCherry) ia49 ; Tg(kdrl:GFP) s843 embryos, we detected a strong induction of Hsa.CTGF:nlsmCherry reporter expression (i.e. Yap1/Taz nuclear activity) specifically in endothelial cells undergoing sprouting (Fig. 6F). These data indicate that the upregulation of Yap1/ Taz activity in single endothelial cells is sufficient to promote aberrant ISVs sprouting.
We further confirmed the autonomous nature of the phenotype and ruled out the possibility that mosaic expression outside of the promoter-specific domains upon DNA injection might be partially responsible for the observed sprouting. We carried out an endothelium-specific Yap1/Taz upregulation experiment by transiently expressing an activated form of Yap1 (yap1-1β S87A or CAYAP1) 53 fused with the fluorescent protein mKate. The injection of the PCS2-fli1a-mKate-CAYAP1 plasmid in one-cell stage Tg(kdrl:GFP) embryos allowed us to detect, with one-cell resolution, the vessel domains in which we were actually upregulating the pathway and to demonstrate the correlation between vascular abnormalities and vascular-specific CAYAP1 expression. Similar to the pDestTol2CG2-fli1a:TAZ-4SA system, in the mosaic PCS2-fli1a-mKate-CAYAP1 injected embryos we observed a specific expression of the mKate-CAYAP1 in conjunction with the ISVs undergoing anomalous endothelial sprouting (Fig. 6G).

Discussion
YAP1/TAZ signaling has recently gained much attention in developmental, cancer and regeneration biology. However, most data in adult mammalian tissues indicate that YAP1/TAZ are dispensable for homeostatic self-renewal and become required only upon genetic deletion of inhibitory cues such as Hippo, or upon inflammatory and tumorigenic stimuli; this led to the still unresolved issue of whether YAP1/TAZ are active or not in most adult tissues. Here, we described the generation, validation and characterization of a Yap1/Taz zebrafish reporter, which represents a powerful tool to follow this signaling activity in a living vertebrate, offering also interesting applications in drug screening, cancer and regenerative biology.
Validation through knockdown and overexpression approaches demonstrated that the Hsa.CTGF-based zebrafish transgenic lines faithfully report in vivo Yap1/Taz activity, being able to reveal significant corresponding variations. A first Yap1/Taz zebrafish reporter, named 4xGTIIC, has been previously published by Miesfeld and Link 41 . During embryogenesis, a strong Yap1/Taz activity was observed in the same regions of the Hsa.CTGF-based reporter presented here, such as the epidermis, heart, otic and lens vesicles, midbrain-hindbrain boundary (MHB) region and striated muscle of the trunk. However, we described a much wider expression pattern for the Hsa.CTGF reporter with respect to that described for the 4xGTIIC, although the latter was mainly described with a destabilized GFP-expressing line, which makes hard to do comparisons. The Hsa.CTGF transgenic line allowed us to point out regions of Yap1/Taz nuclear activity previously not described in the other reporters for the pathway (rhombencephalon, neural tube, notochord, floorplate, pharyngeal arches, and pectoral fin). The Hsa.CTGF reporter activation in the myl7 positive cardiac precursor cells and in the heart is consistent with that of the Tg(eef1a1l1:Gal4db-TEAD2ΔN-2A-mCherry);(UAS:GFP) reporter developed by Fukui and colleagues 42 . The activity in the endothelium, not shown in the 4xGTIIC reporter, is also consistent with the Tg(eef1a1l1:Gal4db-TEAD2ΔN-2A-mCherry);(UAS:GFP) reporter 18 , as well as with the endothelium-specific reporter described by Nagasawa-Masuda and Terai 20 . Yap1 and TEAD2 transcriptional activity has been shown to be modulated in vivo by blood flow 18 and, similarly, we confirmed the positive regulation exerted by the circulation on Yap1/Taz activity by showing that the also the expression of Hsa.CTGF:nlsmCherry in endothelial cells is modulated by blood flow. Notably, the Hsa.CTGF reporter is responsive to the synthetic glucocorticoid Dexamethasone, confirming in vivo the recent findings of Sorrentino and colleagues 39 , and highlighting the potential application of Hsa.CTGF reporter lines in drugs screenings.
A potential criticism might be that the Hsa.CTGF Yap1/Taz transgenic lines are reporting the expression pattern of the zebrafish Yap1/Taz target genes ctgfa rather than the global TEAD-dependent Yap1/Taz signaling activity. There are several reasons ruling out this hypothesis: (i) the reporter expression is driven by a 200 bp fragment of the human CTGF promoter, that represents only a minimal part with respect to the whole promoter regulating the expression of the CTGF gene. Pfefferli and Jazwinska analyzed all the transcription factor binding sites within the 3.18 kb upstream regulatory sequence of ctgfa, showing that the vast majority of binding sites are outside the 200 bp fragment that was used to drive the Hsa.CTGF transgene expression 54 . (ii) The zebrafish Tg(ctgfa:EGFP) reporter based on the 3.18 kb ctgfa promoter described by Pfefferli and Jazwinska showed a different expression pattern from the Hsa.CTGF reporter, being limited to the notochord, the heart and the connective tissue of regenerating fins 54,55 . (iii) In spite of a partial overlap (lens, otic vesicles, heart, pharyngeal arches, pectoral fin, and floorplate), the expression of the ctgfa/b gene and the Hsa.CTGF:nlsmCherry transgene is different. For instance, in situ hybridization performed to detect the mCherry expression on 48 and 72 hpf Tg(Hsa.CTGF:nlsmCherry) reporter embryos (data not shown) clearly labels the MHB and the rhombencephalic regions, which are not positive for ctgfa/b expression 56 . On the contrary, ctgfa is expressed in the pancreatic bud (https://zfin.org/ZDB -FIG-060130-1737) while the Hsa.CTGF reporter doesn't exhibit transgene activity in this region.
The analysis of the spatio-temporal activation of Hsa.CTGF reporter revealed a wide activation of Yap1/Taz signaling during early embryonic development, with a stronger signal in the proliferating and undifferentiated tail bud mesenchyme. This is likely reflecting the main function for YAP1/TAZ as transducers of the Hippo pathway: promotion of cell proliferation and organ growth during development 44,57,58 . In adulthood, YAP1/TAZ expression is strongly limited, being enriched in the stem/progenitor cells niches 59 . Consistently, the Hsa.CTGF reporter is largely silenced in the fully-grown fish with respect to the embryonic and larval development. Highly positive organs or tissues in the adult include the ovary, the lens, the heart and the endothelium. While in the lens the presence of the reporter protein could be simply due to almost absent protein turnover of these cells 60 , the persistence of Yap1/Taz activity in the cardiovascular system suggests its important role in the maintenance of cardiac and vascular functions. This is in agreement with the requirement of Yap1 for the maintenance of blood vessels during zebrafish development 18 .
yap1 and taz single knockouts, unlike the double mutants, do not exhibit significant vascular phenotypes, implying a functional redundancy of the two genes during vascular development. Nevertheless, we described slight defects in the cranial and ocular vasculatures of the yap1 −/− embryos, phenotypes not reported in previous works. While the specific reduction of the number of hyaloid vessels in yap1 mutants might be a consequence of coloboma, yap1 −/− ;taz +/− animals exhibited an unusual phenotype during the formation of the axial vessels: the PCV deviates intermittently from the midline, and occasionally exhibits two distinct lumens. Analysis of arterial/ sprouting (arrowheads in F) with respect to the other normal ISVs. (G) In mosaic embryos injected with the PCS2-fli1a:CAYAP-mKate plasmid a specific expression of the mKate was reported in conjunction with the anomalous endothelial sprouts (arrowhead). The plasmid is endothelium-specific, as highlighted by the colocalization (arrows) between the mosaic mKate and the GFP of the stable transgenic line Tg(kdrl:EGFP). Lateral view, anterior to the left, dorsal to the top. ***p < 0.001. ISV: intersegmental vessel; DLAV: dorsal longitudinal anastomotic vessel. Scale bar: 50 µm.
SCIeNtIFIC RepoRts | (2018) 8:10189 | DOI:10.1038/s41598-018-27657-x venous-specific markers revealed that defective specification of the DA is also observed. Specifically, the expression of mrc1a in 32 hpf yap1 −/− ;taz +/− embryos is not restricted to the PCV and venous/lymphatic vessels as in WT 51 , but is still observed in the DA.
Notably, the overactivation of Yap1/Taz in the endothelium was sufficient to cause an abnormal vessel sprouting phenotype with formation of atypical anastomosis between adjacent ISVs, a result consistent with recent evidences in vitro and in mouse models on the role of YAP1/TAZ in angiogenesis 16,17,[61][62][63] .
Altogether, our results present a novel comprehensive in vivo view of Yap1/Taz activity during development and adulthood at the whole organism level. Together with the vascular phenotypes displayed by yap1/taz mutants and upon endothelium-specific upregulation of Yap1/Taz/Tead-mediated transcription, it confirms and further extends the emerging role of Yap1/Taz signaling in vessel maintenance and developmental angiogenesis.

Material and Methods
Animals. All live animal procedures were approved by the institutional ethics committee for animal testing of the University of Padua and the Max Planck Society as well as in accordance with the relevant guidelines and regulations of Italy, Germany and European Union.
To inhibit pigment formation, embryos and larvae were incubated in 0.003% 1-phenyl-2-thiourea (PTU). The following fish lines were used and outcrossed either to wild-type fish or to the Tg(Hsa.CTGF:nlsmCherry) ia49 Yap1/Taz reporter line: Tg(myl7:GFP) 64  Hsa.CTGF-for (5′-TCTAGAAGATCTTCTGTGAGCTGGAGTGTGC-3′) and Hsa-CTGF-rev (5′-AAGCTTCCATGGAGCGGGGAAGAGTTGTTGT-3′). The CTGF promoter fragment was subcloned in the Gateway 5′ entry vector pME-MCS (Invitrogen) using the BglII and HindIII restriction enzymes. The resulting p5E-Hsa.CTGF entry vector was recombined with the nlsmCherry, and eGFP-carrying middle entry vectors and the p3E-polyA entry clone containing the SV40 late polyA signal (Invitrogen). Entry plasmids were recombined into the Tol2 destination vector pDestTol2pA2 (Invitrogen) through a MultiSite Gateway LR recombination reaction as previously described 52  YAP-5SA is a constitutively active version of the human YAP1 protein, which has been mutated in its five key serine residues, resulting insensitive to LATS1/2-dependent phosphorylation and sequestration in the cytoplasm. Analogously, TAZ-4SA is the constitutively active version of the murine TAZ protein, mutated in its four serine residues recognized by LATS1/2. TEAD-VP16 is a fusion protein of the N-terminal region of TEAD transcription factor and the activation domain of herpes simplex virus VP16. TEAD-VP16 does not need any transcriptional co-activator to work, leading again to a constitutive transcription of its target genes 74 . pCS2-Flag-mTAZ-4SA, pCS2-TEAD-VP16 and pCSP1-Flag-YAP-5SA were digested with a specific restriction enzyme (NotI for Flag-TAZ-4SA and TEAD-VP16, AscI for Flag-YAP-5SA) and transcribed using the SP6 polymerase (AM1340, Lifetechnology). In the overexpression experiments, one cell-stage Tg(Hsa.CTGF:nlsmCherry) ia49 embryos, Tg(kdrl:GFP) embryos or Tg(Hsa.CTGF:nlsmCherry) ia49 /Tg(kdrl:GFP) double transgenic embryos were injected with 0,2/0,4 pg of TAZ-4SA and TEAD-VP16 mRNAs and 5/10 pg of YAP-5SA mRNA per embryo. To avoid phenotypic alterations associated with RNA toxicity we injected the lowest concentration of constitutively active forms of YAP and TAZ showing biological activity. The injections of the same amount of wild type YAP/TAZ RNA (used as injection control) did not induce induced overexpression effects in zebrafish embryos. Flag-mTAZ-4SA 75 was subcloned in pME-MCS (Invitrogen) from a pCS2 vector.MultiSite Gateway LR recombination reaction was performed to recombine the obtained pME-TAZ-4SA, the p5E-fli1a and the p3E-polyA (Invitrogen) into the Tol2 destination vector pDestTol2CG2 (Invitrogen). 20-40 pg of the recombined Tol2 destination vectors were co-injected into one cell-stage Tg(kdrl:GFP) zebrafish embryos. The effect of Yap1/Taz transient overactivation in the endothelium by TAZ-4SA expression was analysed at 32 hpf by confocal microscopy. Mkate-CAYAP1 plasmid 53 was injected into one cell-stage Tg(kdrl:GFP) embryos that were analysed by confocal microscopy at 32 hpf.
Image acquisition and analysis. The fluorescence was visualized using 488 nm (for GFP) and 561 nm (for mCherry) lasers and 20x or 40x immersion objectives (Nikon). Fluorescence quantification of the images acquired either with the conventional fluorescence or the confocal microscope was carried out using Fiji software, by quantifying the fluorescent signal as integrated density as described elsewhere 76 . For the quantification of the reporter signal specific of the endothelium, Hsa.CTGF:nlsmCherry reporter fluorescence was acquired with the confocal microscope together with the kdrl:GFP fluorescence in 32 hpf double transgenic embryos. In order to isolate the reporter expression in the endothelium, Fiji software was used to filter the mCherry signal using the kdrl:GFP as a mask. For each embryo, a confocal Z-stack projection was realized with the obtained filtered images, and then the fluorescent signal was quantified.
Quantifying Hsa.CTGF reporter signal in the endothelium of yap1/taz mutants. An incross of yap1 +/− ;taz +/− fish were performed to obtain embryos for this experiment. Double homozygous mutants die by 30 hpf and are excluded from analysis. Remaining siblings that are TgBAC(etv2:EGFP) and Tg(Hsa.CTGF:nlsmCherry) positive were embedded in 1% low melting agarose and images were acquired with spinning disk confocal microscope (25x objective) at 48 hpf. Image analyses were done with Imaris software. Firstly, EGFP positive cells were selected to delineate endothelial cell nuclei expressing mCherry. Only endothelium on the side of the embryo closest to the objective lens was analyzed. For each nucleus, average signal intensity of mCherry channel was normalized to average signal intensity of EGFP. The normalized values for all selected nuclei were averaged per animal. The calculation can be summarized by the following equation: where X j is the readout for fish j = 1…N; mCherry and EGFP are average intensities of respective channels for nuclei i = 1…n.
Characterization of the PCV phenotype. yap1 −/− ;taz +/− embryos and randomly sampled siblings were embedded in 1% low melting agarose for image acquisition with an LSM800 confocal microscope (25x objective) at 30 and 48 hpf. The orthogonal projection function on the Zen software was utilized to obtain transverse sections of the trunk. For cryosection of 72 hpf, yap1 −/− ;taz +/− larvae and randomly sampled siblings animals, were fixed with 4% PFA and genotyped using a small piece of the tail. Larvae positive for the TgBAC(etv2:EGFP) transgene were selected. Only yap1 −/− ;taz +/− and WT sibling larvae were embedded in OCT using standard procedures. Each section is 12 μm thick. Sections were permeabilized with 0.1% Triton X-100 followed by blocking with 5% sheep serum. Standard immunohistochemistry was performed with anti-EGFP antibody (GFP-1020, Aves Labs, Portland, OR) and Alexa-568 conjugated phalloidin (A12380, Thermo Fisher). Sections were counterstained with DAPI and imaged with LSM800.
Characterizing cranial and hyaloid vasculature phenotype in yap1 mutants. Scoring of the cranial vasculature phenotypes in yap1 mutants (obtained from an incross of yap1 heterozygous adults) was performed blind under a Nikon SMZ25 stereomicroscope. Images of the cranial vasculature in 72 hpf yap1 mutants ( Figure S8A) were obtained with an LSM700 confocal microscope (20x objective). Images of hyaloid vessels of both eyes from 5 dpf animals were taken with spinning disk confocal microscope (40x objective) from the dorsal side of the eye (Fig. S8B).
Whole-mount in situ hybridization (WISH). Standard WISH procedure was performed as described previously 78 . yap1 −/− ;taz +/− embryos from a cross between yap1 +/− ;taz +/− and yap1 +/− fish were identified by their notochord or tail phenotype (see above) and fixed with 4% PFA at 24 and 32 hpf. In parallel, WT embryos from WT fish crosses were fixed with 4% PFA at 24 and 32 hpf. Both yap1 −/− ;taz +/− and WT embryos at each developmental stage were mixed into the same reaction tube after Proteinase K permeabilization. Images were acquired with Nikon SMZ25 stereomicroscope followed by genotyping. The primers used to synthesize probes for efnb2a and mrc1a are found in Table S1.
Statistical analyses. In Figs 2, 6, S4 and S7 data are presented as mean ± SEM and statistical comparison between groups were performed using a two-tailed Student's t-test. Statistical analyses were carried out with Prism software (GraphPad). Statistical tests for Fig. 5C,D were performed using Poisson regression with glm function in R. The cranial vasculature phenotype (Fig. S8A') has been evaluated using a chi-square test. The number of hyaloid vessel (Fig. S8B') was tested using standard Student T Test. Boxplots were generated with ggplot2 package in R.