A YAP/TAZ-TEAD signalling module links endothelial nutrient acquisition to angiogenic growth

Angiogenesis, the process by which endothelial cells (ECs) form new blood vessels from existing ones, is intimately linked to the tissue’s metabolic milieu and often occurs at nutrient-deficient sites. However, ECs rely on sufficient metabolic resources to support growth and proliferation. How endothelial nutrient acquisition and usage are regulated is unknown. Here we show that these processes are instructed by Yes-associated protein 1 (YAP)/WW domain-containing transcription regulator 1 (WWTR1/TAZ)-transcriptional enhanced associate domain (TEAD): a transcriptional module whose function is highly responsive to changes in the tissue environment. ECs lacking YAP/TAZ or their transcriptional partners, TEAD1, 2 and 4 fail to divide, resulting in stunted vascular growth in mice. Conversely, activation of TAZ, the more abundant paralogue in ECs, boosts proliferation, leading to vascular hyperplasia. We find that YAP/TAZ promote angiogenesis by fuelling nutrient-dependent mTORC1 signalling. By orchestrating the transcription of a repertoire of cell-surface transporters, including the large neutral amino acid transporter SLC7A5, YAP/TAZ-TEAD stimulate the import of amino acids and other essential nutrients, thereby enabling mTORC1 activation. Dissociating mTORC1 from these nutrient inputs—elicited by the loss of Rag GTPases—inhibits mTORC1 activity and prevents YAP/TAZ-dependent vascular growth. Together, these findings define a pivotal role for YAP/TAZ-TEAD in controlling endothelial mTORC1 and illustrate the essentiality of coordinated nutrient fluxes in the vasculature.

Blood vessels form extensive tubular networks of arteries, capillaries and veins that nurture all body tissues. Endothelial cells (ECs) line the inner surface of these networks, where they are surrounded by diverse nutrients such as amino acids, glucose and lipids. In the resting state, ECs take up sufficient amounts of these nutrients to enable transport to perivascular tissues and to sustain their basal homeostatic needs 1,2 . However, when activated by growth factors to form new vessel branches, ECs must increase nutrient uptake and consumption to meet the metabolic demands of the angiogenic response 3,4 . In addition to adenosine triphosphate (ATP), growth factor-activated ECs need carbon, nitrogen and reducing agents to support the biosynthesis of macromolecules (for example, nucleic acids, proteins, lipids) necessary for vascular expansion 3,4 . Controlling nutrient acquisition and usage is, therefore, central to the function of the endothelium; yet, the mechanisms that regulate these processes are poorly understood.
We bred mice carrying floxed alleles of both cofactors (Yap fl/fl ;Taz fl/fl ) 20,21 and expressed the tamoxifen-inducible recombinase CreERT2 driven by the endothelial-restricted platelet derived growth factor subunit B (Pdgfb) promoter (Yap;Taz iEC-KO ). Analysis of Letters NATuRE METABOLISM the developing retinal vasculature in these mutants revealed severe angiogenic defects after 4-hydroxy-tamoxifen (4OHT) administration, as reported [5][6][7][8][9] (Extended Data Fig. 1a-c). Compared to controls, Yap;Taz iEC-KO mice had fewer and less proliferative ECs, giving rise to a sparse and mis-patterned vascular network with poor connectivity (Extended Data Fig. 1a-c). Deletion of Taz alone mimicked most of these vascular phenotypes, while deletion of Yap had little effect (Extended Data Fig. 1d-i). Compared to YAP, TAZ was also the more abundant transcript in various endothelial subtypes (Fig. 1a,b and Extended Data Fig. 2a,b), suggesting a critical role of TAZ for YAP/TAZ responses in the endothelium.
To study the role of TAZ in ECs further, we engineered a Taz reporter mouse, in which a green fluorescent protein (GFP) as well as FLAG and a biotin-labelling peptide were fused to the C terminus of the endogenous TAZ protein (Taz tag ) ( Fig. 1c and Extended Data Fig. 2c,d). Mice homozygous for this knock-in mutation (Taz tag/tag ) expressed the reporter-tagged fusion instead of the wild-type protein, had normal vascular patterning and transcript levels of prototypic YAP/TAZ target genes were unperturbed ( Fig.  1d and Extended Data Fig. 2e-g), indicating that the reporter tag does not alter TAZ function. Analysis of the GFP signal in platelet EC adhesion molecule (PECAM) (marking the surface of ECs) and ETS transcription factor ERG (marking endothelial nuclei) colabelled retinas confirmed high expression of TAZ in the endothelium and revealed differences in its subcellular localization (Fig.  1e,f). At the vascular front, where ECs actively divide, migrate and rearrange, TAZ showed a preferentially nuclear pattern, whereas in the central parts, where ECs are less active, TAZ was mostly cytoplasmic (Fig. 1e,f). Similar results were obtained in wild-type retinas labelled with a TAZ antibody (Extended Data Fig. 2h), suggesting dynamic regulation of TAZ subcellular localization during angiogenic growth.
To understand the functional consequences of this regulation, we generated Rosa26 knock-in mice expressing a nuclear-localized TAZ mutant on Cre-mediated recombination (Extended Data Fig. 3a,b). This mutant has the phospho-acceptor site serine 89 replaced by alanine (Rosa26-Taz S89A fl/fl ), rendering TAZ insensitive to cytoplasmic sequestration induced by the Hippo kinases LATS1/2 (refs. 14,18 ). Pdgfb-CreERT2-mediated expression of Taz S89A (Taz iEC-GOF ) in the retinal endothelium resulted in increased levels of nuclear TAZ as well as increased expression of YAP/TAZ target genes ( Fig. 1g and Extended Data Fig. 3c,d), which correlated with a dense and hyperplastic vascular network ( Fig. 1h-j). Endothelial 5-ethynyl-2′-deoxyuridine (EdU) incorporation was also increased in these mice (Fig. 1i,k and Extended Data Fig. 3e), suggesting that nuclear TAZ signalling is crucial for EC proliferation. Consistent with this finding, expression of Taz S89A in Yap;Taz iEC-KO mutants was sufficient to restore vascular density in these mice, while some of the patterning defects remained unchanged (Extended Data Fig. 3f,g).
In the nucleus, YAP and TAZ can interact with various transcription factors to control the expression of their target genes 14,15,18 . To identify such transcriptional regulators in ECs, we first examined proteins coimmunoprecipitating with TAZ. To this end, we transduced human umbilical vein ECs (HUVECs) with FLAG-tagged TAZ S89A (AdTAZ S89A ) or GFP as a control (AdCtrl) and performed immunoprecipitations with a FLAG antibody. Mass spectrometry analysis revealed that the TEAD proteins were the most significantly enriched transcription factors in the TAZ S89A interactome under these conditions (log 2 FC > 1, FDR < 0.05) ( Fig. 2a and Supplementary Table 1). Similar results were obtained when we analysed proteins that interact with a nuclear form of YAP (YAP S127A ) ( Fig. 2b and Supplementary Table 1) or when the interactions were assessed between endogenous proteins (Fig. 2c,d), suggesting that TEADs are a central transcriptional platform through which endothelial YAP/TAZ signal.
To confirm this hypothesis, we conditionally eliminated Tead1, Tead2 and Tead4 in ECs of mice: the three TEAD family members that interact with YAP/TAZ (Fig. 2a,b and Supplementary Table 1) and are expressed by the endothelium (Fig. 2e,f) 22 . In accordance with a functional overlap 23  To explain how endothelial YAP/TAZ-TEAD promote endothelial proliferation and angiogenic growth, we performed RNA-sequencing (RNA-seq) in HUVECs transduced with doxycycline-inducible lentiviruses encoding for YAP S127A , TAZ S89A or GFP as a control (Extended Data Fig. 5a,b). Comparative gene expression analyses at 48 hours of doxycycline treatment revealed that both transcriptional regulators induced profound gene expression changes, with 1,395 genes being up-or downregulated in YAP S127A -expressing ECs and 1,410 genes being altered in TAZ S89A -transduced cells (log 2 fold change ≥ 1 and FDR ≤ 0.05) (Fig. 3a,b and Supplementary Table 2). Roughly 78% of the regulated genes overlapped in both transcriptomes (Fig. 3a,b and Supplementary Table 2), indicating that YAP and TAZ can signal redundantly when overexpressed and localized in the nucleus. Among the upregulated genes were prototypical YAP/TAZ targets (for example, ANKRD1, CTGF, AXL, CYR61) as well as numerous genes linked to mechanistic target of rapamycin complex 1 (mTORC1) signalling, including genes involved in nutrient uptake (for example, SLC7A5, SLC1A4, SLC7A11, SLC1A5, SLC2A1), Fig. 1 | Nuclear TAZ signalling drives vascular growth. a, Yap/Taz transcript levels in ECs sorted from postnatal day (P) 6 mouse retinas as determined by RNA-seq (n = 4 independent samples). b, YAP/TAZ immunoblot analysis of ECs isolated from murine brains and lungs. c, Schematic of the Taz tag reporter containing GFP, FLAG and a biotin-labelling peptide. d, Expression of the reporter-tagged TAZ protein in whole lung lysates of wild-type, heterozygous and homozygous Taz tag mice. VEGFR2, endothelial marker. e, Quantification of TAZ subcellular localization in ECs of Taz tag/tag P6 retinas. N, preferentially nuclear; NC, nucleo-cytoplasmic; C, preferentially cytoplasmic (n = 10 independent samples). f, Images of GFP, ERG and PECAM-labelled P6 retinas derived from Taz tag/tag mice. The grey images (lower panels) show the isolated GFP signal. The small (white) boxed area is shown at higher magnification in the upper right corner. Scale bars, 50 μm. g, Immunolabelling for TAZ, GFP and PECAM in P6 retinas of Taz iEC-GOF (Pdgfb-CreERT2;Rosa26 -Taz S89A fl/fl ) and control (Ctrl; Rosa26-Taz S89A fl/fl ) mice. The grey images (right panels) show the isolated TAZ signal. Scale bar, 100 μm. h, Confocal images of PECAM-labelled P6 mouse retinas of Ctrl and Taz iEC-GOF mice. A, artery; V, vein. Scale bar, 200 μm. i, Quantification of vascular parameters in Ctrl and Taz iEC-GOF mutants as indicated (EC area, n = 16 (Ctrl) and 14 (Taz iEC-GOF ) independent samples; EC number/field, n = 12 (Ctrl) and 10 (Taz iEC-GOF ) independent samples; EC proliferation, n = 13 (Ctrl) and 6 (Taz iEC-GOF ) independent samples). j, ERG and PECAM-labelled retinas at P6 showing a hyperplastic vasculature in Taz iEC-GOF mice. k, Immunofluorescence images of the angiogenic front in P6 retinas of Ctrl and Taz iEC-GOF mice labelled for EdU, ERG and PECAM-. Scale bar, 200 μm. Western blot data in b and d are from the respective experiment, processed in parallel and are representative of three independent experiments. For a,d,e and i, data represent mean ± s.e.m.; two-tailed unpaired t-test was used. **P < 0.01; ***P < 0.001; ****P < 0.0001; NS, not significant. The numerical data, unprocessed western blots and P values are provided as source data.
To confirm these observations in vivo, we labelled P6 mouse retinas with two different S6 antibodies recognizing distinct mTORC1-dependent phosphorylation sites (serine 235/236 → p-S6 Ser235/236 and serine 240/244 → p-S6 Ser240/244 ). The retinas were also labelled with PECAM and vascular endothelial cadherin (VECAD) antibodies to identify ECs. Both p-S6 antibodies gave similar results; in controls, the endothelial p-S6 signal was strongly enriched at sites of active angiogenic growth, particularly at the vascular front and the peri-venous region (Fig. 3i,j and Extended Data Figs. 7a,b and 8a). This labelling pattern was diminished in both Yap/Taz iEC-KO and Tead1/2/4 iEC-KO mice, yet intensified in Taz iEC-GOF mutants (Fig. 3i,j and Extended Data Fig. 7a-h). Crucially, the mTOR inhibitor rapamycin extinguished the retinal p-S6 signal (Extended Data Fig. 8a,b), demonstrating the specificity of the antibody labelling. Moreover, rapamycin treatment prevented the overgrowth of ECs in the Taz S89A mutant (Extended Data Fig.  8c). Together, these results indicate not only a critical role of YAP/ TAZ-TEAD in the regulation of endothelial mTORC1 but also of mTORC1 as an effector of the YAP/TAZ-TEAD signalling network.
To decipher how YAP/TAZ-TEAD regulate mTORC1, we searched for YAP/TAZ-induced expression changes in the mTORC1 pathway. YAP S127A and TAZ S89A did not affect the transcripts of the mTORC1 complex MTOR, MLST8, AKT1S1 and RPTOR nor of its downstream effectors (Supplementary Table 2). Instead, both cofactors mice (EC area, n = 8 (Ctrl) and 6 (Tead1/2/4 iEC-KO ) independent samples; EC number/field, n = 6 (Ctrl) and 5 (Tead1/2/4 iEC-KO ) independent samples; EC proliferation, n = 7 (Ctrl) and 5 (Tead1/2/4 iEC-KO ) independent samples). Western blot data in c and d are from the respective experiment, processed in parallel and are representative of at least three independent experiments. For e, f and i, data represent mean ± s.e.m.; two-tailed unpaired t-test. ***P < 0.001. The numerical data, unprocessed western blots and P values are provided as source data.

NATuRE METABOLISM
induced the transcription of several cell-surface transporters involved in the regulation of mTORC1 by nutrients (Fig. 3k). In particular, a cluster of amino acid transporters was upregulated, including solute carrier family member (SLC) 7A5 (SLC7A5), SLC38A5, SLC7A1, SLC1A5 and SLC3A2, which have been shown to promote mTORC1 signalling 29,30 (Fig. 3k). Notably, SLC7A5 and SLC3A2 form a heterodimeric amino acid transporter for large neutral amino acids 31 that cooperates with SLC1A5 to activate mTORC1 (refs. 29,30 ), suggesting a highly coordinated transcriptional response.
To further assess whether the regulated transporters are direct YAP/TAZ-TEAD target genes, we performed chromatin immunoprecipitation-sequencing (ChIP-seq) studies. We found that endogenous TAZ bound to the genomic regions of canonical YAP/TAZ targets (for example, ANKRD1, AXL, CTGF, CYR61) as well as to several of the regulated transporters, including SLC7A5, SLC3A2, SLC1A5, SLC7A1 and SLC7A11 (Fig. 3q,r, Extended Data Fig. 9a,b and Supplementary Table 3). The TEAD consensus sequence was highly enriched in the TAZ ChIP-seq peaks (Fig. 3p), suggesting a critical role for TEADs in recruiting TAZ to chromatin 35 . To confirm this, we performed ChIP-seq for TEAD1, the most abundant TEAD family member in HUVECs (Fig. 2f), and found that TAZ peaks overlapped with TEAD1 peaks. Genes occupied by TAZ and TEAD1 comprised not only canonical targets (for example, ANKRD1, AXL, CTGF and CYR61) but also nutrient transporters such as SLC7A5, SLC3A2, SLC7A1, SLC7A11 (  Table 3). Similar results were also obtained for endogenous YAP, although fewer candidate genes were    Table 3). Together, these data indicate a model whereby YAP/ TAZ-TEAD orchestrate a transcriptional program that facilitates the transport of amino acids and other essential nutrients, thereby enabling mTORC1 pathway activation.
To study the functional relevance of the proposed mechanism, we first tested whether altered transporter expression is sufficient to regulate endothelial mTORC1. To this end, we inactivated SLC7A5, one of the most abundant endothelial transporters (Fig. 4a) highly sensitive to YAP/TAZ signalling (Fig. 3k-r and Extended Data Fig.  9c,d). Depletion of SLC7A5 by clustered regulatory interspaced short palindromic repeats (CRISPR)-Cas9 in vitro (gSLC7A5) or Pdgfb-CreERT2-induced recombination of a floxed allele in vivo (Slc7a5 iEC-KO ) lowered endothelial mTORC1 signalling and reduced retinal angiogenic growth (Fig. 4b-g and Extended Data Fig. 9e-h). Moreover, inhibiting SLC7A5 function-either by siRNA or the pharmacological inhibitor JPH203-disabled YAP S127A or TAZ S89A to activate endothelial mTORC1 (Fig. 4h,i), underscoring the importance of SLC7A5 for YAP/TAZ signalling responses. Of note, overexpression of SLC7A5 alone was not sufficient to restore mTORC1 activity in YAP/TAZ-depleted HUVECs (Extended Data Fig. 9i), presumably because the levels of SLC7A5's complex partner SLC3A2 were not restored. To further corroborate the hypothesis that altered nutrient/ amino acid acquisition links YAP/TAZ-TEAD to mTORC1, we next inactivated Rag GTPase signalling in ECs. Rags (RagA-D) recruit mTORC1 to the surface of lysosomes when sufficient amino acids are available, thereby allowing full pathway activation [36][37][38] . To disrupt their signalling and bypass compensatory effects, we inactivated two essential family members, RagA and RagB, simultaneously (  Fig. 10f,g).
Finally, to directly test the requirement of Rag-mediated mTORC1 activity for the anabolic and proliferative functions of endothelial YAP/TAZ, we inactivated Rag signalling in ECs with activated YAP/TAZ. Depletion of RagA/B in HUVECs stalled YAP S127A -or TAZ S89A -induced anabolism and proliferation ( Fig. 4t and Extended Data Fig. 10h,i). Moreover, depletion of RagA/B in mice overexpressing Taz S89A in the endothelium (Taz iEC-GOF ;RagA/B iEC-KO ) was sufficient to prevent vascular overgrowth in these mutants (Fig.  4u). Collectively, these results establish nutrient-driven mTORC1 signalling as a crucial regulatory pathway in ECs that determines YAP/TAZ-induced vascular expansion, although it seems likely that other RagA/B-independent mechanisms contribute as well.
Our study identifies an essential metabolic link between YAP/ TAZ and angiogenic growth, which involves the master regulator of cellular anabolism, mTORC1. We demonstrate that YAP/TAZ form a transcriptional module with TEADs to orchestrate the expression of a cluster of cell-surface transporters, importing amino acids and other metabolic fuels. Signalling by this module supplies ECs with resources for growth and proliferation and enables mTORC1 pathway activation, thereby promoting anabolic processes such as protein and DNA synthesis. By placing nutrient acquisition under the control of YAP/TAZ-TEAD-which integrates mechanical, metabolic and growth factor signals 14,15,18,19 -ECs ensure a coherent angiogenic response that is coupled to the tissue environment. Such coupling is crucial for angiogenic growth as ECs are confronted with changing tissue milieus when forming new vessel branches. The metabolic challenges that arise in these environments might also explain the susceptibility of ECs to perturbed nutrient-regulated mTORC1 signalling, whose disruption causes angiogenic arrest.
Our data indicate that among the regulated cell-surface transporters, SLC7A5 plays a central role. The requirement of this large neutral amino acid carrier for YAP/TAZ-induced mTORC1 activation aligns with previous findings in cancerous cells 32,33 and highlights the importance of intracellular amino acids for YAP/ TAZ-mediated growth responses. Other modes of nutrient acquisition might contribute to the maintenance of sufficient intracellular amino acids levels during angiogenic growth. Macropinocytosis, for instance, is a non-selective endocytic mechanism for bulk ingestion of extracellular macromolecules (proteins) 3 , which has recently been involved in YAP/TAZ signalling 40 and endothelial biology 41 . These considerations raise the intriguing possibility that YAP/TAZ activity may play a broader role in determining cellular nutrient acquisition strategies and underscore the need to understand these processes in more detail. Delineating the mechanisms that determine how ECs take up, transport and use nutrients will provide new insights not only into vascular (patho-)physiology but also into the role of specific nutrients in instructing vascular growth and function.

RNA interference.
HUVECs were transfected with 50 nM of ON-TARGETplus SMARTpool siRNAs (Dharmacon) listed in Supplementary Table 4 using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's recommendations.
CRISPR-Cas9 genome editing of HUVECs. For each target gene, three independent guide RNAs (Supplementary Table 5) were cloned into the plentiCRISPRv2 plasmid (Addgene, no. 52961) and cotransfected with the packaging vectors for lentivirus production as described above. Scramble guide RNA sequences were used as a control (gCtrl).

NATuRE METABOLISM
Coimmunoprecipitations. Cells were lysed in IPLS buffer (50 mM Tris-HCL pH7.5, 120 mM NaCl, 0.5 mM EDTA and 0.5% Nonidet P-40) freshly supplemented with protease inhibitor mix (Roche) and 1 mM phenylmethylsulfonyl fluoride 45 . Samples were cleared by centrifugation and protein concentrations determined by the Bradford method. Equal amounts of total lysates were precleared using A/G agarose beads (Santa Cruz Biotechnology, SC-2003) under gentle rotation at 4 °C for 45 min. Epitope tag immunoprecipitations were performed using the anti-FLAG M2 affinity gel beads (Sigma, no. F2426) at 4 °C with gentle rotation for 2 h. Collected beads were washed five times in IPLS buffer and bound proteins eluted in Laemmli sample buffer for subsequent immunoblot analysis. Immunoprecipitations of endogenous YAP and TAZ proteins were performed similarly using antibodies recognizing YAP (Cell Signaling, no. 4912) or TAZ (Santa Cruz Biotechnology, no. SC-48805). Rabbit immunoglobulin G (IgG) (Diagenode, no. C15410206) was used as a control. Precleared lysates were incubated with the antibodies at 4 °C overnight followed by incubation with protein G Sepharose 4 Fast Flow beads (GE Healthcare, no. 17-0618-05) for 2 h. Beads were washed with IPLS buffer five times and eluted proteins subjected to subsequent immunoblotting.
Mass spectrometry. HUVECs were infected with AdCtrl, AdYAP S127A and AdTAZ S89A and immunoprecipitations performed with FLAG M2 beads 16 h after transduction. Eluates were separated by SDS-PAGE (NuPAGE 4-12% BisTris gel, Invitrogen) and stained with colloidal Protein Staining Solution (Invitrogen). Gel pieces were excised for in-gel digestion using trypsin after reduction and alkylation. After washes with 50% 50 mM NH 4 HCO 3 /50% ethanol for 20 min, gel pieces were dehydrated with 100% ethanol for 10 min and vacuum dried. Samples were reduced with 10 mM dithiothreitol for 45 min at 56 °C and alkylated with 55 mM iodocetamide (BioUltra, Sigma-Aldrich) for 30 min at room temperature in the dark. After washing/dehydration, gel pieces were dehydrated twice with 100% ethanol for 15 min, vacuum dried and digested overnight at 37 °C in 50 μl of digestion buffer containing 12.5 ng μl −1 of Sequencing Grade Modified Trypsin (Promega Corp.). Released peptides were extracted once by adding 100 μl of 30% acetonitrile liquid chromatography-mass spectrometry (LC-MS) grade (Thermo Scientific)/3% trifluoroacetic acid (Sigma-Aldrich) in water, and twice by adding 70% acetonitrile, followed by two final extractions with 100% acetonitrile. Extracts were vacuum dried to remove acetonitrile and subsequently acidified with 0.5% trifluoroacetic acid 46 . Peptides were purified by stop and go extraction tips 47 and analysed by LC-MS using an EASY-nLC chromatograph and a QExactive mass spectrometer (Thermo Fisher Scientific). Peptide/spectrum matching and label-free quantification were performed by the MaxQuant suite of algorithms [48][49][50] and data were postprocessed using Perseus 51 .

RNA-seq analysis and gene set enrichment analysis. RNA was isolated from
HUVECs using the miRNeasy Micro Kit (Qiagen) combined with on-column DNase digestion (DNase-Free Set, Qiagen). RNA integrity was verified using LabChip Gx Touch 24 (Perkin Elmer). Then 4 μg of total RNA input was used for Truseq Stranded messenger RNA Library preparation following the low sample protocol (Illumina) and subjected to 1 × 75 bp single end setup sequencing (Illumina NextSeq500) using v.2 chemistry, resulting in minimum of 32 million reads per library. Data quality was assessed using the FastQC v.0.10.1 quality-control tool 52 for high throughput sequence data. RNA-seq reads were mapped to the human reference genome version hg19 (GRCh37) with STAR software 53 . For gene set enrichment analysis 54,55 , gene set collections from the Molecular Signatures Database (MSigDB) v.4.0 (http://www.broadinstitute.org/gsea/msigdb/) were used. Heat maps were generated using Morpheus, a publicly available program from the Broad Institute (https://software.broadinstitute.org/morpheus/). Quantitative PCR with reverse transcription. RNA was isolated using RNeasy Kit (Qiagen) and cDNA synthesized using M-MLV reverse transcriptase (Thermo Fisher). Quantitative PCR with reverse transcription analysis was carried out using the StepOnePlus system (Applied Biosystems). Relative gene expression was calculated with the comparative Ct method and normalized to ACTB or Actb expression. TaqMan probes used are listed in Supplementary Table 6.
ChIP. HUVECs were fixed with 1% formaldehyde for 15 min and quenched with 0.125 M glycine. Chromatin was isolated by the addition of lysis buffer, followed by disruption with a dounce homogenizer. Lysates were sonicated and the DNA sheared to an average length of 300-500 bp. Genomic DNA (input) was prepared by treating aliquots of chromatin with RNase, proteinase K and heat for reverse-crosslinking, followed by ethanol precipitation. Pellets were resuspended and the resulting DNA was quantified on a NanoDrop spectrophotometer. Extrapolation to the original chromatin volume allowed quantitation of the total chromatin yield. Sheared chromatin (30 μg) was precleared with protein A agarose beads (Invitrogen). Genomic DNA regions of interest were isolated using ChIP-grade antibodies against YAP1 (Abcam, no. ab52771), TAZ (Sigma, no. HPA007415) or TEAD1 (CST, no. 12292BF). Complexes were washed, eluted from the beads with SDS buffer and subjected to RNase and proteinase K treatment. Crosslinks were reversed by incubation overnight at 65 °C and ChIP DNAs were purified by phenol-chloroform extraction and ethanol precipitation.
ChIP-seq and analysis. Illumina sequencing libraries were prepared from ChIP and input DNAs by the standard consecutive enzymatic steps of end-polishing, dA-addition, and adaptor ligation. After a final PCR amplification step, the resulting DNA libraries were quantified and sequenced on Illumina's NextSeq500 (75-nt reads, single end). Reads were aligned to the human genome (hg38) using the Burrows-Wheeler alignment algorithm (default settings). Duplicate reads were removed and only uniquely mapped reads (mapping quality ≥25) were used for further analysis. Alignments were extended in silico at their 3′-ends to a length of 200 bp, which is the average genomic fragment length in the size-selected library, and assigned to 32-nt bins along the genome. The resulting histograms (genomic 'signal maps') were stored in bigWig files. Peak locations were determined using the model-based analysis of ChIP-seq (MACS) algorithm (v.2.1.0) 56 with a cut-off P = 1 × 10 −7 . MACS2 peak regions that overlapped with any of the ENCODE blacklist of known false ChIP-seq regions (by a minimum of 1-bp) were removed using a custom-made Perl script (Active Motif). Signal maps and peak locations were used as input data. Binding Motifs were identified with the findMotifsGenome program of the HOMER package 57 using default parameters and input sequences comprising ±100-bp from the centre of the top 1,000 peaks. All profiles were plotted on a normalized reads-per-million basis. The processed data were plotted and visualized using software of the R project for statistical computing.
Western blot analysis and antibodies. HUVECs were lysed in RIPA buffer (Sigma, no. R0278) supplemented with 1× EDTA-Free Complete Protease Inhibitor Cocktail (Roche) and 1 mM phenylmethylsulfonyl fluoride 58 . Proteins were resolved by SDS-PAGE using Criterion TGX Precast gels (Bio-Rad) and transferred onto nitrocellulose membranes using the Trans Turbo Blot system (Bio-Rad). Membranes were blocked in 5% BSA or 5% milk + 0.01% Tween-20 in TBS 1× for 1 h at room temperature. Primary antibodies in blocking buffer were incubated overnight at 4 °C. Peroxidase-conjugated secondary antibodies were incubated for 1 h at room temperature. Immunoblots were visualized using Clarity Western ECL kit (Bio-Rad) and the ChemiDoc MP Imaging System (Bio-Rad). Band intensities were quantified using the Image Lab software (Bio-Rad). Antibodies used are listed in Supplementary Table 7.
Proliferation assays. HUVECs were seeded on six-well plates at 2.5 × 10 4 cells per well and allowed to attach overnight. Next day (0 h), the total cell number was counted with a hemocytometer. Cell counts were repeated every 24 h and culture medium replaced every 48 h. For Metabolic flux assays. To determine protein synthesis, HUVECs were incubated with medium containing 1 μCi per ml [ 3 H]-tyrosine (Perkin Elmer) for 6 h. Cells were washed with ice-cold PBS, proteins precipitated with 10% trichloroacetic acid overnight and collected by centrifugation at 21.000g for 5 min. The protein pellet was resuspended in 0.5 M NaOH with 0.1% (v/v) Triton X-100 and the amount of [ 3 H]-tyrosine incorporated into protein was measured by scintillation counting and subsequently normalized to protein content 60 . Glucose-dependent DNA synthesis was measured by assessing the incorporation of 14 C into DNA using 2.9 mCi per mmol [U 14 C]-glucose (Perkin Elmer). Incorporation was analysed at 48 h in triplicate and measured by scintillation counting. Counts were normalized to the total amount of DNA per sample. Total DNA was isolated using Trizol.
Metabolomics and CoRe analysis. HUVECs (2.5 × 10 5 ) were plated onto six-well plates and transfected with the indicated siRNAs. The culture medium was replaced 4 h after transfection (t = 0) and cells were incubated for additional 36 h. A control plate with culture medium (no cells) was prepared at t = 0 and processed in parallel at the experimental end-point. The collected media were centrifuged at 4 °C for 10 min at 21,000g and 50 µl of the supernatant was extracted in 750 µl of cold metabolite extraction solution (50% methanol, 30% acetonitrile, 20% water). Extracts were placed for 15 min over dry ice, vortexed and incubated in a Thermomixer (1.400 r.p.m.) at 4 °C for 15 min. Samples were incubated for 1 h at −20 °C and metabolite extracts were cleared by centrifugation at 4 °C for 10 min at 21,000g, transferred into autosampler vials and stored at −80 °C until further analysis. Cell culture media extracts from cell cultures were analysed for each condition. Samples were randomized to avoid bias due to machine drift and processed blindly. LC-MS analysis was carried out using a Vanquish Horizon UHPLC system coupled to a QExactive HF mass spectrometer (both Thermo Fisher Scientific). Sample extracts (5 µl) were injected onto a Sequant ZIC-pHILC column (150 × 2.1 mm, 5 µm) and guard column (20 × 2.1 mm, 5 µm, Merck Millipore) kept at 45 °C. The mobile phase was composed of 20 mM ammonium carbonate and 0.1% ammonium hydroxide in water (solvent A) and acetonitrile (solvent B). The mobile phase was composed of Letters NATuRE METABOLISM 20 mM ammonium carbonate with 0.1% ammonium hydroxide in water (solvent A) and acetonitrile (solvent B). The flow rate was set at 200 μl min −1 with the following gradient: 0 min 80% B, 2 min 80% B, 17 min 20% B, 17.1 min 80% B and a hold at 80% B for 5 min (ref. 61 ). The mass spectrometer was operated in full MS and polarity switching mode. The acquired spectra were analysed using XCalibur Qual Browser and XCalibur Quan Browser software (Thermo Fisher Scientific) by referencing to an internal library of compounds. To obtain a relative measure of the metabolite consumption/release (CoRe), the background levels of each metabolite in the media controls were subtracted to the levels measured in the cell-conditioned media samples and adjusted to the average cell number.
Genetic mouse models and pharmacological treatments. The following published mouse lines were used in this study: Yap fl/fl (ref. 20 )., Taz fl/fl (ref. 21 ), Tead4 fl/fl (ref. 62 ), Rraga fl/fl (ref. 39 ), Rragb fl/fl (ref. 39 ), Slc7a5 fl/fl (refs. 63,64 ) and Pdgfb-iCreERT2-IRES-EGFP (Pdgfb-creERT2) 65 . For the construction of the Cre-activated Taz gain-of-function allele (Taz GOF ), a 3xFLAG-TAZ S89A -IRES-nEGFP sequence preceded by a floxed Neomycin-STOP cassette was knocked into the Rosa26 locus. Cre-mediated removal of the STOP sequence results in CAG promoter-driven expression of 3xFLAG-TAZ S89A and nuclear-localized enhanced GFP (nEGFP). To generate a Taz knock-in reporter mouse (Taz tag ), a fusion tag consisting of GFP, FLAG and a biotin-labelling peptide was inserted in-frame upstream of the stop codon of the endogenous Wwtr1 (Taz) locus. The conditional Tead1 knockout allele was generated by flanking exons 3 to 5 with loxP sites, while the straight knockout allele of Tead2 (Tead2 ko ) was generated by deleting exons 1 to 4. The conditional Taz GOF and Taz tag knock-in alleles were developed together with genOway. All mice were bred on a C57BL/6J genetic background. Floxed mice were crossed to mice expressing the tamoxifen-inducible Pdgfb promoter-driven CreERT2 recombinase. Littermates that were negative for CreERT2 were used as controls. For the combined inactivation of Tead1, Tead2 and Tead4, Tead1 fl/fl ;Tead2 wt/ ko mice (Tead1 and Tead2 are both located on chromosome 7) were interbred with Tead4 fl/fl and Pdgfb-CreERT2 mice. To activate CreERT2, pups were administered 25 μl of 4-hydroxytamoxifen (4OHT; 2 mg ml −1 ) intraperitoneally from P1 to P4. Animals were euthanized and retinas harvested at P6 (Supplementary Table 8).
To inhibit mTOR signalling, animals were randomly divided into two groups and injected with vehicle or 2 µg g −1 rapamycin (LC Laboratories) from P1 to P5. To detect proliferating cells, pups were administered 25 µl of 5-ethynyl-2′-deoxyuridine (EdU) (6 mg ml −1 ; Invitrogen, A10044) intraperitoneally 3 h before euthanasia. Both male and female animals were used. Animal experiments were performed in accordance with institutional guidelines and protocols approved by the Committee for Animal Rights Protection of the State of Hessen (Regierungspraesidium Darmstadt) with the project numbers B2/1061 and B2/1230.
Immunofluorescence. HUVECs were plated on glass bottom dishes (Mattek) and fixed with 4% paraformaldehyde for 30 min at room temperature. Permeabilization and blocking was performed in 1% BSA, 10% FBS and 0.5% Tween-20 in PBS. Primary antibodies were incubated in blocking buffer at 4 °C overnight. Following washes with PBST (0.1% Tween-20 in PBS), cells were incubated with AlexaFluor-conjugated secondary antibodies (Invitrogen) for 2 h at room temperature and mounted using VectaShield (Vector Laboratories, no. H-100). For mouse retina immunostaining, eyes were fixed in 4% paraformaldehyde on ice for 2.5 h (ref. 66 ). After dissection, retinas were incubated in blocking buffer (3% FBS, 1% BSA, 0.25% Tween-20 and 0.25% Triton X-100 in PBS) for 1 h at room temperature. Primary antibodies were incubated in blocking buffer diluted in PBS (1:1) overnight at 4 °C. After washing in PBST, retinas were incubated with Alexa-conjugated secondary antibodies for 2 h at room temperature, washed and flat-mounted with ProLong Gold (Life Technologies). To detect EdU-labelled DNA, an additional step was performed before mounting using the Click-It EdU kit (Invitrogen, no. C10338). Immunostainings were performed in tissues from littermates and processed under the same conditions except for studies with the Tazi EC-GOF ;RagA/B iEC-KO mice given the very low allele frequency. Images were acquired with a Leica confocal microscope SP8. For the comparisons of phenotypes or signal intensities, setting for laser excitation and detector were kept constant between groups. Volocity (Perkin Elmer), Fiji/ImageJ, Photoshop (Adobe) and Illustrator (Adobe) software were used for image acquisition and processing.
Quantitative analysis of retinal vasculature. Endothelial coverage was quantified from confocal fields behind the angiogenic front in between arteries and veins using Volocity (Perkin Elmer). Endothelial coverage (EC area) was measured as the ratio of PECAM-positive area to total area of vascularized field (sized 200 × 200 μm 2 ). EC proliferation was scored as the ratio of EdU and ERG double-positive cells to the total number of ERG-positive cells per field. All parameters were quantified from at least three vascularized fields per sample. To quantify p-S6 levels in retinal ECs, the absolute intensity of the p-S6 and PECAM double-positive area was quantified using Imaris (Bitplane) and expressed as fold change relative to controls.

Statistics and reproducibility.
For quantitative analyses, a minimum of three biological replicates were analysed. Western blot data are from the respective experiment, processed in parallel and are representative of at least three independent experiments. Images from immunofluorescence studies are representative of the respective phenotype observed in samples from at least three independent experiments/litters. Statistical analyses were performed by unpaired, two-tailed Student's t-test unless indicated otherwise. For all bar graphs, data are represented as mean ± s.e.m. A value of P < 0.05 was considered significant. No statistical method was used to predetermine sample size. Calculations were performed using the Prism v.9.0 software (GraphPad Software Inc.). Numerical data and exact P values are provided as source data.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
RNA-and ChIP-seq datasets have been deposited in National Center for Biotechnology Information Gene Expression Omnibus with the accession number GSE163459. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD026872. Source data are provided with this paper. All other data supporting the findings of this study are available from the corresponding author upon reasonable request.

NATuRE METABOLISM
as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/. © The Author(s) 2022 this work. Primary handling editor: Christoph Schmitt, in collaboration with the Nature Metabolism team.
Reprints and permissions information is available at www.nature.com/reprints.  Fig. 2 | Expression and subcellular localization of TAZ in angiogenic ECs. a, RNA-seq analysis of YAP and TAZ transcript levels in different human EC cultures (n = 3 independent samples). HAECs, human aortic ECs; HMVECs, human microvascular ECs; HUVECs, human umbilical vein ECs; HDLEC, human dermal lymphatic ECs. b, Immunoblot analysis of YAP and TAZ protein expression in different endothelial subtypes. c, Targeting strategy to generate a Taz (Wwtr1) knock-in allele encoding for a GFP, FLAG, and biotin-labelling peptide tagged TAZ protein. A schematic representation of the wild-type Taz (Wwtr1) locus, the targeting vector, the recombined as well as the excised locus is shown. E5 − 7, exons 5 to 7. Triangles denote loxP sites. Neo, neomycin positive selection cassette. GFB, GFP-FLAG-biotin labelling peptide fusion tag. DTA, diphteria toxin negative selection marker. d, PCR analysis of genomic DNA obtained from wild-type, heterozygote and homozygote Taz tag mice. Lane 1, DNA marker. e, RT-qPCR analysis of the canonical YAP/TAZ target gene Ctgf in retinal ECs derived from wild-type (Ctrl) and homozygous Taz tag/tag mice, showing no expression difference between the two genotypes (n = 2 (Ctrl) and 3 (Taz tag/tag ) independent samples). f, g, ERG-and PECAM-immunofluorescence staining (f) and quantification of vascular parameters (g) of P6 Ctrl and Taz tag/tag retinas, revealing no gross difference in vascular morphogenesis between the two genotypes (EC area: n = 4 (Ctrl) and 8 (Taz tag/tag ) independent samples; Number of ECs: n = 3 (Ctrl) and 9 (Taz tag/tag ) independent samples). h, Images of the vascular front and plexus in TAZ, ERG and PECAM labelled P6 mouse retinas derived from C57BL/6 wild-type mice. The images in grey (right panels) show the isolated TAZ signal. Western blot data in b are from the respective experiment, processed in parallel, and are representative of at least three independent experiments. For a, e and g, data represent mean ± s.e.m.; two-tailed unpaired t-test. **P < 0.01; ****P < 0.0001; NS, not significant. The numerical data, unprocessed western blots and P values are provided as source data.  Fig. 5 | Forced nuclear expression of YAP or TAZ induces canonical YAP/TAZ signaling in ECs. a, Immunoblots of HUVECs transduced with doxycycline (Dox)-inducible control (Ctrl), YAP S127A (iYAP S127A ) and TAZ S89A (iTAZ S89A ) encoding lentiviruses, showing expression of FLAG-tagged YAP S127A and TAZ S89A upon Dox treatment. Samples were analyzed 48 h after treatment with Dox or vehicle. b, RT-qPCR analysis of the canonical YAP/TAZ target genes ANKRD1, CTGF, and CYR61 in iYAP S127A and iTAZ S89A expressing HUVECs. Expression changes are shown relative to Ctrl (n = 8 independent samples). c, d, Gene set enrichment analysis (GSEA) showing an enrichment of the YAP-conserved target gene expression signature in the transcriptomes of iYAP S127A (c) and iTAZ S89A (d) expressing HUVECs. ES, enrichment score; NES, normalized enrichment score. e, f, FLAG immunoprecipitation studies in HUVECs overexpressing FLAG-tagged iYAP S127A and iYAP S94A/S127A (e) or iTAZ S89A and iTAZ S51A/S89A (f). The mutation of serine 94 to alanine in YAP and serine 51 to alanine in TAZ disrupts the interaction of (nuclear) YAP and TAZ with TEADs. g, Heatmap of mRNA expression changes in HUVECs expressing Ctrl, iYAP S127A , iYAP S94A/S127A , iTAZ S89A or iTAZ S51A/S89A as determined by RNA-seq. Canonical YAP/TAZ target genes are shown (n = 3 independent samples). Western blot data in a, e and f are from the respective experiment, processed in parallel, and are representative of at least three independent experiments. For b, data represent mean ± s.e.m.; two-tailed unpaired t-test. For c, d, Kolmogorov-Smirnov test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. The numerical data, unprocessed western blots and P values are provided as source data.  Fig. 9 | Characterization of direct YAP/TAZ-TEAD target genes in ECs. a,b, ChIP-seq signals of TAZ, YAP and TEAD1 at the genomic loci of the canonical target genes ANKRD1 (a) and AXL (b). ChIP-seq signals are represented as reads per kilobase per million mapped reads (RPKMs). c,d, Slc7a5 expression in ECs isolated from the lungs of Rosa26-Taz S89A fl/fl (c) or Yap fl/fl ;Taz fl/fl (d) mice followed by transduction with control (AdCtrl) or Cre-encoding (AdCre) adenoviruses. Slc7a5 transcript levels were determined by RT-qPCR (c, n = 6 independent samples; d, n = 6 independent samples). e, Reduced cell proliferation in SLC7A5-deficient HUVECs (gSLC7A5) as measured by 3 H-thymidine DNA incorporation (n = 8 independent samples). Values are represented as fold change relative to control (gCtrl). f, Quantification of vascular parameters in Ctrl and Slc7a5 iEC-KO mice as indicated (EC area: n = 8 (Ctrl) and 10 (Slc7a5 iEC-KO ) independent samples; EC number / field: n = 9 (Ctrl) and 11 (Slc7a5 iEC-KO ) independent samples; EC proliferation: n = 9 (Ctrl) and 11 (Slc7a5 iEC-KO ) independent samples). g, Immunofluorescence staining for EdU, ERG and PECAM in P6 Ctrl and Slc7a5 iEC-KO retinas. h, Confocal images of VECAD, p-S6 Ser235/236 and PECAM stained P6 retinas of Ctrl and Slc7a5 iEC-KO mice, suggesting reduced mTORC1 signaling in Slc7a5 mutants. The isolated p-S6 Ser235/236 signal is shown in grey at the bottom of the panel. i, Immunoblots of HUVECs transduced with Ctrl or SLC7A5 encoding lentivirus, showing that overexpression of SLC7A5 is insufficient to restore mTORC1 activity in YAP/TAZ-depleted HUVECs (siYAP/TAZ). Western blot data in i are from the respective experiment, processed in parallel, and are representative of at least three independent experiments. For c, d, e-f, data represent mean ± s.e.m.; two-tailed unpaired t-test. **P < 0.01; ***P < 0.01; ****P < 0.0001. The numerical data, unprocessed western blots and P values are provided as source data.  Fig. 10 | Suppression of endothelial RAG GTPase signaling inhibits mTORC1 and proliferation in ECs. a, RT-qPCR analysis in gCtrl and gRagA/B HUVECs, showing that RagA/B deficiency does not alter MTOR transcript levels (n = 3 independent samples). b, c, Immunoblot analysis (b) and quantification (c) of mTOR protein levels in control (gCtrl) or in RagA/B-depleted (gRagA/B) HUVECs, confirming the RT-qPCR analysis (n = 4 independent samples). d, Analysis of mTOR/LAMP2 co-localization in RagA/B-deficient ECs as assessed by immunofluorescence imaging (n = 5 independent samples). e, Immunofluorescence images of gCtrl and gRagA/B HUVECs stained for the lysosomal protein LAMP2, mTOR, phalloidin (PHAL) and DAPI, showing reduced mTOR/LAMP2 co-localization in gRagA/B ECs when compared to gCtrl. f, Higher magnification images of P6 Ctrl and RagA/B iEC-KO mutant mice labelled for p-S6 Ser235/236 , VECAD and PECAM. The images in the lower panel show the isolated p-S6 Ser235/236 signal in grey. g, Higher magnification confocal images of P6 Ctrl and RagA/B iEC-KO retinas labelled for EdU, ERG and PECAM demonstrating a reduced number of proliferating ECs in the mutants. h, i, Suppression of DNA (h) and protein (i) synthesis in Ctrl, iYAP S127A and iTAZ S89A HUVECs subjected to simultaneous depletion of RagA/B. DNA synthesis was measured by assessing the incorporation of 14 C-glucose into DNA (n = 5 independent samples). Protein synthesis was measured by assessing the incorporation of 3 H-tyrosine into protein (n = 9 independent samples). Western blot data in b are from the respective experiment, processed in parallel, and are representative of three independent experiments. For a, c, d, h and i, data represent mean ± s.e.m.; two-tailed unpaired t-test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; NS, not significant. The numerical data, unprocessed western blots and P values are provided as source data.

nature research | reporting summary
April 2020 Corresponding author(s): Michael Potente Last updated by author(s): May 2, 2022 Reporting Summary Nature Research wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Research policies, see our Editorial Policies and the Editorial Policy Checklist.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one-or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information.

Data
Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A list of figures that have associated raw data -A description of any restrictions on data availability The data supporting the findings of this study are available within the paper. Genome wide transcriptome data have been deposited in NCBI Gene Expression

April 2020
Omnibus under the accession numbers GSE163456 and GSE199858. ChIP-seq data have been deposited under the accession number GSE163458. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD026872.
The following publicly available databases were used: GSEA pathway analysis: MsigDB (https://www.gsea-msigdb.org/gsea/msigdb); gRNA design: Genetic Perturbation Platform (https://portals.broadinstitute.org/gpp/public/); Any additional information required to interpret, replicate or build upon the findings of this study are available from the corresponding author upon reasonable request.
Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences
Behavioural & social sciences Ecological, evolutionary & environmental sciences For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design
All studies must disclose on these points even when the disclosure is negative.