Genome-Wide Analysis of the Zebrafish ETS FamilyIdentifies Three Genes Required for Hemangioblast

Differentiation or AngiogenesisFeng Liu, Roger Patient

Abstract—ETS domain transcription factors have been linked to hematopoiesis, vasculogenesis, and angiogenesis.However, their biological functions and the mechanisms of action, remain incompletely understood. Here, we haveperformed a systematic analysis of zebrafish ETS domain genes and identified 31 in the genome. Detailed geneexpression profiling revealed that 12 of them are expressed in blood and endothelial precursors during embryonicdevelopment. Combined with a phylogenetic tree assay, this suggests that some of the coexpressed genes may haveredundant or additive functions in these cells. Loss-of-function analysis of 3 of them, erg, fli1, and etsrp, demonstratedthat erg and fli1 act cooperatively and are required for angiogenesis possibly via direct regulation of an endothelial celljunction molecule, VE-cadherin, whereas etsrp is essential for primitive myeloid/endothelial progenitors (hemangio-blasts) in zebrafish. Taken together, these results provide a global view of the ETS genes in the zebrafish genome duringembryogenesis and provide new insights on the functions and biology of erg, fli1, and etsrp, which could be applicableto higher vertebrates, including mice and humans. (Circ Res. 2008;103:1147-1154.)

Key Words: zebrafish � gene duplication � ETS transcription factors � hemangioblast � angiogenesis

Zebrafish has been recognized as an excellent genetic anddevelopmental biology model to study hematopoiesis

and vessel development. Large numbers of genetic mutantsand transgenic lines in both blood and endothelial lineageshave become available. More importantly, developmentalmechanisms, including the functions of many known impor-tant regulators involved in both blood and vessel develop-ment, are conserved between higher vertebrates and ze-brafish.1 It has been proposed that hematopoietic andendothelial cells share a common progenitor, termed thehemangioblast.2 This idea was initially conceived from theobservation that these 2 cell types develop in close proximitywithin the embryo.3 Coexpression of many important regula-tors of hematopoiesis and vasculogenesis in early embryosfurther supports the existence of a common progenitor forhematopoietic and endothelial cells. Furthermore, loss offunction of several of these regulators, including the zebrafishcloche (clo) mutant, affects both cell types.4–7 Recent workby us and others subsequently showed that several importanthematopoietic and endothelial regulators in mammals, suchas Scl/Tal1 and Lmo2, also play similar roles in zebrafish.8–10

ETS domain genes encode a super family of transcriptionfactors organized into 11 subfamilies based on domainstructures, which are conserved from worm and fly tomammals.11 These transcription factors are involved in cellfate specification, proliferation, migration, and differentiation

during embryonic development and adulthood and have beenlinked with diverse biological processes, from hematopoiesis,vasculogenesis, and angiogenesis to neurogenesis. Manyimportant blood and endothelial regulators have well-conserved and functional ETS binding sites in their promoterand/or enhancer regions.12–15 For instance, the stem cellleukemia gene, Scl, a key regulator during hemangioblastformation and hematopoiesis, is regulated by a complex thatconsists of ETS transcription factors (Fli1 and Elf1) and otherpartners.12 Moreover, many ETS genes are coexpressed inhematopoietic and endothelial progenitor cells, which sug-gests they may have redundant roles or are synergisticallyrequired during hematopoiesis and vasculogenesis.11 Defectsin human ETS counterpart genes, either by inappropriateexpression or by fusions with other proteins, cause carcino-genesis, such as leukemia.16

To comprehensively explore the functional roles of ETSfactors during hematopoiesis and vasculogenesis in zebrafish,we have used bioinformatics to systematically analyze theETS family from the zebrafish genome (Table I in the onlinedata supplement, available at http://circres.ahajournals.org).Interestingly, 12 of the 31 ETS genes identified showedexpression in blood and endothelial precursors during embry-onic development, which, together with phylogenetic analy-sis, suggested that some of them may have redundant oradditive functions. Functional analysis of 3 of them, etsrp,

Original received May 19, 2008; revision received September 17, 2008; accepted September 18, 2008.From the Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, University of Oxford, United Kingdom.Correspondence to Prof Roger Patient, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, University of Oxford, Oxford OX3 9DS,

United Kingdom. E-mail roger.patient@imm.ox.ac.uk© 2008 American Heart Association, Inc.

Circulation Research is available at http://circres.ahajournals.org DOI: 10.1161/CIRCRESAHA.108.179713

1147 by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from

erg and fli1, which are strongly expressed in blood andendothelial precursors/hemangioblasts, demonstrated thatthey are required for hemangioblast differentiation andangiogenesis.

Materials and MethodsFish Strains and EmbryosZebrafish embryos were obtained by natural spawning of adult ABstrain zebrafish. Embryos were raised and maintained at 28.5°C insystem water and staged as described.17 flk1:gfp/gata1:dsred trans-genic line and etsrpy11 mutant were kindly provided by StefanSchulte-Merker (Utrecht, The Netherlands) and Brant Weinstein(NIH), respectively.

BioinformaticsTo search for ETS genes, the sequences of the human and mouseETS domains were used to TBLASTN the zebrafish genome inNCBI (http://www.ncbi.nlm.nih.gov). The predicted zebrafish ETSgenes and proteins were named based on the alignment of theircoding regions to known entries in public databases. To find theconserved domain in the protein sequences, NCBI conserved domaindatabase was used (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi). To generate a phylogenetic tree, the full-length aminoacid sequences of ETS proteins were aligned using the CLUSTALWprogram. Chromosome locations of ETS genes and their surroundinggenes were obtained from the Ensembl zebrafish genome databaseas above.

MorpholinosAntisense morpholinos (MOs) were obtained from GeneTools(Philomath, Ore). The MOs used in this work were erg atgMO(5�-AGTGACACTCACTCTCTCTGAGGTA-3�), erg spliceMO(5�-AGACGCCGTCATCTGCACGCTCAGA-3�), and flil utrMO(5�-TTTCCGCAATTTTCAGTGGAGCCCG-3�). etsrp atgMO wasas reported previously.18 Stock solutions (1 mmol/L) in dH2O wereprepared, and 2 to 10 ng amounts were injected into 1- to 2-cell stageembryos at the yolk/blastomere boundary. Both erg atgMO andspliceMO gave similar phenotype and coinjection of both MOs didnot give a more severe phenotype than single injection.

Reverse Transcription–Polymerase Chain ReactionRT-PCR analysis to determine efficacy of MO splicing inhibitionwas carried out with primers 5�-AGGCGCTGTCGGTGGTGA-GTGAG-3� (erg forward) and 5�-TCTGCTGGCGTCTGGAGG-GTAG-3� (erg reverse). cDNA template was produced from mRNAisolated from 24 hpf wild-type (wt) and erg spliceMO-injectedembryos. PCR products were analyzed on 1% agarose gel.

Whole-Mount In Situ HybridizationWhole mount in situ hybridization was performed as described.19

The following probes were used: fli1, scl, gata1, flk, pu.1,9,18 etsrp,18

pax2.1.20 erg, ets2, fev, elk4, gabpa, elf1, and elf2 were first reportedin this work and their EST clones were obtained from ImaGenes(Berlin, Germany).

PhotographyImages of live embryos were captured using a Nikon SMZ 1500dissecting microscope (�1 HR Plan Apo objective; numeric aper-ture, 0.13) with a Nikon DXM1200 digital camera driven by NikonAct-1 version 2.12 software (Nikon). Confocal images were acquiredwith a Zeiss LSM 510 META confocal laser microscope and 3Dprojections were generated using Zeiss LSM software (CarlZeiss Inc).

ResultsIdentification and Phylogenetic Analysis ofZebrafish ETS Domain GenesThe presence of ETS binding sites in the regulatory elementsof key hematopoietic and/or endothelial genes led us to

systematically search the whole zebrafish genome (NationalCenter for Biotechnology Information [NCBI]; Ensembl,zv7/Vega) for the potentially responsible ETS factors regu-lating hemangioblast development. Computational searchingof sequence databases revealed 31 ETS domain-containinggenes in the zebrafish genome sequence (supplemental TableI and data not shown), and their putative chromosomallocations were mapped according to the latest zebrafishgenome mapping information (supplemental Table I).

Phylogenetic analysis with the complete open readingframes of the ETS transcription factors, identified 11 subfam-ilies within the ETS gene family (supplemental Table I andFigure 1), which is consistent with the classification ofmammalian counterparts based on their structural composi-tion and their similarities in the DNA-binding ETS domains.Besides the conserved ETS domain, a subset of ETS familyproteins has another evolutionarily conserved domain calledthe POINTED (PNT) domain at their N-terminal regions,which is required for protein–protein interactions. Notably,because of genome duplication, there are several genes thatare closely related in this family (supplemental Table I). Forexample, there are 3 erf-like genes, erfl1, erfl2, and erfl3,

Figure 1. Phylogenetic analysis of zebrafish ETS genes. Thetree was generated from the alignment of the amino acidsequence of the whole protein including ETS DNA bindingdomain and Pointed domain (PNT) using the CLUSTAL Wmethod. Zebrafish genes with incomplete or highly divergentETS domains were not included in this analysis. Groups arenamed according to.16

1148 Circulation Research November 7, 2008

by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from

similar to erf itself, which are the orthologs of the mammalianerf gene. There are another 3 pairs of ETS genes, ie, elk4 andelk4l1, elf2 and elf2l1, spic and spicl1, which show highhomology to a single mammalian gene.

Developmental Expression Patterns of TwelveHematopoietic or Endothelial ETS GenesTo systemically characterize expression of these ETS genesin wt embryos, we performed a detailed expression patternanalysis at various developmental stages (Figure 2, supple-mental Figure I, and data not shown). We obtained cDNAclones for 27 of the 31 ETS genes (the other 4 clones areduplicate genes, ie, erfl2, elk4l1, elf2, and spicl1, whichshould give similar expression patterns because of highsimilarity of the sequences) (supplemental Table I) andsubjected zebrafish embryos to in situ hybridization at1-somite (1s), 6s, 10s, and 26-hpf stages when hemangio-blasts and erythroid/myeloid and endothelial lineages areclearly identifiable (Figure 2, supplemental Figure I, and datanot shown). Compared to the known hemangioblast marker,scl (supplemental Figure I, A through D), we found that 12ETS genes are expressed in the right place and early enoughto be involved in hemangioblast formation (Figure 2 andsupplemental Figure I). These are ets1, ets2, etsrp, fli1, fli1b,erg, fev, gabpa, elf1, elf2, pu.1, and elk4. Importantly, amongthem, ets2, erg, fev, gabpa, elf1, elf2, and elk4 are reported forthe first time in this study. Of note, etsrp is the earliest ETSgene strongly expressed in the anterior lateral plate mesoderm(ALM) (Figure 2A), which will give rise to myeloid andendothelial cells in zebrafish. In the posterior lateral platemesoderm (PLM), fli1 is the earliest and strongest expressedat 1s among these 12 genes (Figure 2E). erg is also expressedin the PLM at the 1s stage (Figure 2M). Later on, all of themexcept pu.1 are expressed in endothelial cells at 26 hpf(Figure 2 and supplemental Figure I), suggesting that theymight play roles in angiogenesis.

Expression Analysis in clo Mutants Confirms ThatTwelve ETS Genes Are Specifically Expressed inHematopoietic or Endothelial CellsTo confirm the expression of the 12 ETS genes in hemato-poietic and endothelial cells, we tested whether their expres-sion was affected in clo mutants, which lack both blood andendothelial cells. So far, clo mutants exhibit the earliesthematopoietic and endothelial lineage defects in zebrafish.7

As shown in Figure 3 and supplemental Figure II, endothelialexpression of 11 ETS genes, ie, ets1, ets2, etsrp, fli1, fli1b,erg, fev, gabpa, elf1, elf2, and elk4, and myeloid expressionof pu.1 were reduced or absent in clo embryos at 26 hpf,suggesting that they are indeed expressed in hematopoietic or

Figure 2. etsrp, fli1, fli1b, and erg are expressed in hematopoietic and/or endothelial cells during zebrafish embryogenesis. etsrp (Athrough D), fli1 (E through H), fli1b (I through L), and erg (M through P). Flat-mount view: A through C, E through G, I through K, and Mthrough O; lateral view, D, H, L, and P. Anterior to the left in all images.

Figure 3. Expression of etsrp, fli1, fli1b, and erg was reduced orlost in clo mutant embryos at 26 hpf. A, etsrp. B, fli1. C, erg. D,fli1b. Note that endothelial-specific expression was reduced orabsent in the clo embryos (arrows). Anterior to the left and dor-sal to the top.

Liu and Patient ETS Genes in Zebrafish 1149

by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from

endothelial cells. Of note, neural crest expression of fli1, ets1,gabpa, elf1, and elf2 were relatively normal in clo embryos(Figure 3B and supplemental Figure II, A, D, E, and F),compared to wt siblings.

etsrp Is Required for Anterior Hemangioblast(Endothelial/Myeloid Progenitor) DifferentiationGene expression profiling showed that fli1, erg, and etsrp arestrongly expressed in the putative hemangioblast populationin 10s stage embryos and later on are restricted to endothelialcells (Figure 2). We reasoned that they might be involved inblood/endothelial progenitor specification and differentiation,as well as in angiogenesis. erg has previously been implicatedin angiogenesis and in some leukemia cases in mammaliancell lines.21 Because no knockout mice for erg had beenreported (but see Discussion), we decided to use the zebrafishas a model to study the function of erg in vivo. To knock-down erg expression in zebrafish, we designed 2 antisenseMOs (Figure 4A). The erg atgMO was designed to target theerg ATG and inhibit translation. The erg spliceMO wasdirected against the boundary of exon 4 and intron 4,spanning part of the PNT domain, to skip exon 4 and generate

a truncated protein containing only a small portion of thePNT domain and, because of a frameshift, no downstreamETS DNA binding domain (Figure 4A). To test MO effi-ciency, RT-PCR analysis of erg spliceMO-injected embryoswas carried out, revealing the formation of aberrantly splicedRNAs (Figure 4B, lower arrow in spliceMO-injected em-bryos), instead of the wt product, which was virtuallyabolished (Figure 4B, upper arrow in control embryos). Thus,endogenous erg activity was substantially knocked downwith the splice MO. spliceMO-injected embryos developednormally without any obvious morphological defects, apartfrom a hemorrhage in the head area between 2.5 and 3 dayspostfertilization (see below). The atgMO gave a similarhemorrhage phenotype (data not shown), indicating that thehemorrhage is a specific phenotype caused by ergknockdown.

To determine the consequences of loss-of-function of fli1,we designed an antisense MO against the 5� untranslatedregion immediately upstream of the fli1 translation initiationcodon. This fli1 utrMO completely abolished the greenfluorescent protein (GFP) expression in fli1:gfp embryos,suggesting that this antisense MO knockdown strategy wasefficient (Figure 4D). The fli1 morphants also developed ahemorrhage in the head around 2.5 days postfertilization(supplemental Figure IV).

etsrp has been demonstrated to be involved in endothelialdifferentiation but is dispensable for erythroid developmentin zebrafish embryos.18 However, its role in the earliestprecursor, the hemangioblast, has not yet been tested. Knock-down experiments showed that whereas fli1MO and ergMOhad no effects on the early hemangioblast program in bothALM and PLM, etsrpMO caused significant reduction of scland an absence of pu.1 and flk1 expression in the ALM,which gives rise to myeloid and endothelial lineages in wtembryos (supplemental Figure III). In contrast, the erythroidmarker gata1 was not affected in the PLM of all 3 morphants(supplemental Figure III, U through X).

To confirm the etsrpMO phenotype, we used the availableetsrp mutant, y11, in zebrafish.22 The predicted Etsrppolypeptides synthesized in the homozygous mutant containonly a quarter of the wt Etsrp protein and do not have thewell-conserved ETS DNA binding domain essential for thefunction of ETS genes. Therefore, it is very likely to be a nullallele, which should phenocopy the loss-of-function pheno-type induced by etsrp atgMO injection. Whole-mount in situhybridization shows that the hemangioblast marker, scl, wassignificantly reduced or absent in the ALM of y11 embryos(Figure 5A). The myeloid marker pu.1 was completelyabsent, and the endothelial marker flk1 was also absent inboth the ALM and the PLM (Figure 5B, 5C, and 5E). In thePLM, scl expression was downregulated in y11 embryos(Figure 5D), suggesting that etsrp is also involved in posteriorhemangioblast differentiation into the endothelial lineage.The erythroid marker gatal was not affected at either the 10sstage in the PLM (Figure 5F) or at 24 hpf in the intermediatecell mass (data not shown), confirming that etsrp is notrequired for erythroid differentiation.18

Figure 4. MO efficiency test. A, Partial erg gene structure show-ing the targets of erg atgMO and spliceMO and the positions offorward (F) and reverse (R) primers for RT-PCR to verify thesplicing defects. The ETS DNA binding domain is located inexon 8/9. B, RT-PCR analysis revealed formation of an alterna-tive splice product following spliceMO injection (446 bp, lowerblack arrow), indicating knockdown of wt transcrpit (656 bp,upper black arrow) caused by exon 4 skipping. erg plasmid andefla primers were used as controls to verify primers andRT-PCR, respectively. C and D, 3 ng of fli1 utrMO can effi-ciently knockdown GFP expression in fli1:gfp transgenicembryos (D).

1150 Circulation Research November 7, 2008

by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from

Knockdown of erg and fli1 Causes AngiogenesisDefects and Downregulation of VE-CadherinTo further characterize the erg morphant phenotype, weinjected ergMOs into flk1:gfp-gata1:dsRed transgenic em-bryos to visualize the whole vasculature and blood circula-tion. The hemorrhage phenotype was reproduced in �40% ofinjected embryos. Confocal microscopy clearly showed thatthe hemorrhage occurred in the head vasculature from 60 hpfonward in erg spliceMO-injected embryos (Figure 6Athrough 6F, arrows). In addition, we also found that inter-somitic vessel (ISV) patterning was slightly affected (Figure

6H). In the controls, ISVs were arranged in very regularchevron shapes (Figure 6G). However, in the erg morphants,ISVs often did not follow this chevron shape and grew eitherstraight or reversely (Figure 6H).

Fli1 and Erg belong to the same subfamily in the ETSfamily (supplemental Table I and Figure 1), and interestinglyfli1 and erg MOs gave similar hemorrhage phenotypes in thehead at around 60 to 72 hpf (supplemental Figure IV, Athrough C). Moreover, fli1 knockout mice also showed ahemorrhage phenotype in the head.23,24 These data suggestthat erg and fli1 are involved in establishing and/or maintain-

Figure 5. etsrp is essential foranterior hemangioblast forma-tion in zebrafish embryos. ALMexpression of scl was nearlyabsent in the etsrp mutant (A,6/6). pu.1 expression in ALMwas completely lost in etsrpmutants (B, 6/6). flk1 expres-sion in ALM was absent inetsrp mutants (C, 5/5). PLMexpression of scl was reducedin etsrp homozygous mutants(D, 5/5), whereas flk1 expres-sion in PLM was completelylost in etsrp mutants (E, 6/6).gata1 expression in PLM wasnormal in etsrp mutants (F,5/5). Dorsal view and anterior tothe top; wt and y11 embryoswere at 6s (A through F).

Figure 6. erg and fli1 are addi-tively required for angiogene-sis. A through I, Confocalmicroscopy at 72 hpf showedthat hemorrhage occurred spe-cifically in the head vascula-ture by using flk1:GFP/gata1:dsRed (A through F,arrows) and slightly disorga-nized ISVs (H, red arrows),compared to control embryos(A through C and G). Doubleknockdown by ergMO andfli1MO gave more severelydisorganized ISV patterning(K, 22/32, arrows). The dorsallongitudinal anastomotic vesselis indicated by dotted boxes.Lateral views with anteriortoward left (A through I).

Liu and Patient ETS Genes in Zebrafish 1151

by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from

ing vessel integrity. Because the 2 genes have similar expres-sion patterns in mice and also in zebrafish, these 2 ETStranscription factors could either be equally or redundantlyrequired for angiogenesis and vessel remodeling after theendothelial cells have been specified (vasculogenesis). How-ever, double knockdown of both fli1 and erg gave a moresevere phenotype than the individual knockdowns, namely amore severe ISV patterning defect and higher penetrance ofthe hemorrhage (71% versus 41%, Figure 6I versus 6H anddata not shown). In the double knockdown, ISVs were moredisorganized and they did not grow properly toward thedorsal side of the embryo (Figure 6I, arrows). In addition, thedorsal longitudinal anastomotic vessel did not develop prop-erly in the double knockdown morphants (Figure 6I, dottedbox), which contrasts with the individual knockdowns (Fig-ure 6H, dotted box). Thus, these data demonstrate that ergand fli1 are additively required for vessel integrity andangiogenic remodeling.

VE-cadherin, a key regulator of endothelial intercellularjunctions, has recently been shown, by chromatin immuno-precipitation and functional assays, to be a direct target of ergin human umbilical vein endothelial cells.25 We, therefore,looked at VE-cadherin mRNA expression in the erg, fli1, anddouble morphants. At 24 hpf, VE-cadherin was equallyexpressed in both controls and the morphants (data notshown). However, there was a reduction of VE-cadherinexpression in both the trunk and the head at 50 hpf in themorphants (supplemental Figure IV, D through G, arrows),just before the hemorrhage occurred, suggesting that down-regulation of VE-cadherin by erg and fli1 knockdown mightbe responsible for the hemorrhage and ISV patterning defectsobserved in the morphants.

DiscussionTo initiate a comprehensive evaluation of ETS gene functionin hematopoiesis and endothelial cells, we have systemati-cally characterized the ETS family in an organism ideallysuited to functional testing, the zebrafish. Bioinformaticanalysis identified a total of 31 complete ETS genes in thezebrafish genome. Gene expression profiling showed that 12of them are expressed in blood and/or endothelial lineagesduring embryogenesis. Functional analysis of erg and fli1demonstrated that they are required for angiogenesis, possiblyvia direct regulation of VE-cadherin, a key regulator ofendothelial intercellular junctions. Combined use of antisenseknockdown and a genetic mutant clearly demonstrated thatetsrp plays an early role during hemangioblast developmentto drive both myeloid and endothelial differentiation.

The ETS Family and Genome DuplicationOur phylogenetic tree analysis indicates that the zebrafishgenome contains 3 pairs of duplicated ETS genes, ie, elk4 andelk4l1, elf2 and elf2l1, spic and spicl1, and 1, erf, that has 3diverged copies (supplemental Table I and Figure 1). Thezebrafish-specific ETS gene duplicates may result fromzebrafish-specific gene duplication or mammalian-specificgene loss, or both. Several lines of evidence suggest thatwhole-genome duplication and subsequent massive loss ofduplicated genes occurred in the teleost.26–29 Syntenic anal-

ysis further confirmed the likelihood of a genome-wideduplication. For example, the ets1-fli1 pair on chromosome18 has been duplicated and their duplicates, etsrp and fli1bare located on chromosome 16. In the most extreme case, erfhas 3 additional duplicates. These duplicated ETS genes,together with their surrounding genes, often form duplicatedchromosome segments located on either the same chromo-some or a different chromosome, further suggesting genomeduplication occurred. Of note, their close neighboring genesalso have putative mammalian orthologs located near thehuman counterpart genes, highlighting highly conserved syn-tenies between zebrafish and mammals such as human andmouse. Taken together, our data and published data stronglysupport the view that the zebrafish-specific ETS genes arosefrom genome-wide duplication.26–29

A Role for Etsrp in Hemangioblast FormationThe list of genes involved in both blood and endotheliallineages is expanding, although the existence of a bipotentialprogenitor, the hemangioblast, is still contentious. Scl, Lmo2and Gata2 have all been demonstrated to be critical for thehemangioblast program.8,9,20,30–32 Using binding and transac-tivation assays, Pimanda et al have implicated Fli1 in thecontrol of scl and gata2 expression; however, no perturba-tions were performed to support such a role in hemangio-blasts.32 We show here that, although erg and fli1 areexpressed in the hemangioblast population, loss of theirfunctions in zebrafish has no effect on hemangioblast devel-opment, suggesting that they are either not required orpossibly redundantly required. We have recently found thatfli1 is indeed essential for blood and endothelial developmentin Xenopus embryos and that gain-of-function of fli1 caninduce many hemangioblast markers in zebrafish, consistentwith a role for this family at the top of gene regulatoryhierarchy.33 However, even a quadruple knockdown of erg,fli1, fli1b, and fev (belonging to the same subfamily) did notgive a hemangioblast phenotype (supplemental Figure V),indicating that if redundancy is the explanation, it involvesETS factors outside this subfamily in zebrafish.

Etsrp, an ETS domain protein discovered in zebrafish, hasbeen shown to be necessary and sufficient for the initiation ofvasculogenesis.18 Here, we have clearly demonstrated, byusing antisense MO knockdown and a genetic homozygousmutant, that Etsrp is absolutely required for anterior heman-gioblast formation in zebrafish. This requirement could becell-autonomous, because etsrp is coexpressed with scl andflk1 in the same cell population, the anterior lateral platemesoderm, during embryogenesis. As etsrp is dispensable forerythroid differentiation, it suggests that the anterior heman-gioblast, only giving rise to myeloid and endothelial lineages,and the posterior hemangioblast population, giving rise toerythroid and endothelial lineages, are differentially pro-grammed. It is highly possible that both cell-intrinsic differ-ences and local environments account for the differentialregulation of these 2 hemangioblast populations. Our studieshave identified etsrp as a critical regulator required forgeneration of the hemangioblasts that differentiate into my-eloid and endothelial lineages. Consistent with our finding,mice carrying a mutation in Er71, the homolog of etsrp in

1152 Circulation Research November 7, 2008

by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from

mammals, have very recently been shown to lack bothhematopoietic and endothelial lineages.34

Sumanas et al recently reported the dependence of myeloidgene expression on etsrp in the anterior hemangioblastpopulation in zebrafish embryos.35 They showed, by etsrpMOknockdown that myeloid markers, such as lcp1, mpx, andpu.1, were reduced and that overexpression of etsrp couldupregulate lcp1 and pu.1 expression. However, anteriorexpression of endothelial markers, such as flk1, was notanalyzed. Together with our data, we can conclude that etsrplies at the top of the genetic hierarchy programming bothblood and endothelium in the anterior hemangioblast popu-lation. A recent study by Xiong et al36 showing that etsrp canpartially rescue both hematopoietic and endothelial defects incloche embryos is consistent with our finding here.

Erg, Fli1, and AngiogenesisAlthough erg is dispensable for the hemangioblast program,its specific expression in endothelial cells at later stages ofdevelopment suggests that it might be involved in angiogen-esis. In addition, fli1 knockout mice showed a central nervoussystem hemorrhage phenotype, suggesting that it is notredundant in angiogenesis.23,24 In contrast, the very recentlyreported ErgMld mutation did not cause a vasculature pheno-type in mice.37 A possible explanation is that this Mld2mutation, which caused a substitution in the ETS DNA-binding domain but did not affect DNA binding affinity,might not therefore be a null allele. Many ETS genes havebeen implicated in angiogenesis.11 Pham et al have demon-strated a combinatorial function for ETS transcription factors,fli1, fli1b, ets1, and etsrp, in the developing vasculature in thezebrafish.22 The additive interaction of erg and fli1 discov-ered in this study supports such combinatorial functionsamong ETS factors in vessel development.

Downregulation of VE-cadherin, a key molecule involvedin cell–cell junctions and cell survival, by erg and fli1knockdown could explain why a hemorrhage occurred in bothmorphants. Targeted null mutation of VE-cadherin impairsthe organization of vascular-like structures in embryoidbodies.39 More importantly, electrophoretic mobility shiftassays and chromatin immunoprecipitation analyses showedthat Erg binds to the VE-cadherin promoter.25,40 The strongestexpression of VE-cadherin and erg/fli1 in the head vascula-ture indicates that they might be particularly essential for thevessel integrity and stabilities in that region. However, wehave not been able to detect any discernible defects in thehead vasculature, even using confocal microscopy imaging ofthe flk1:GFP/gata1:dsRed double transgenic line, althoughthe hemorrhage always occurs in the head. The hemorrhagephenotype in the erg morphants is very similar to thehemorrhaged mutants recently identified in zebrafish, such asbubblehead/�pix and redhead/pak2a.41,42 It was suggestedthat �Pix-Pak2a signaling pathways regulate cerebral vascu-lar stability, possibly via the p38 mitogen-activated proteinkinase pathway.41 Nevertheless, it would be intriguing toknow whether expression of VE-cadherin was affected inthese mutants, as in the erg and fli1 morphants describedhere. As a first test for effects on FGF or VEGF signaling, 2pathways known to be important in angiogenesis, we asked

whether activated forms of Erg or Fli1 could induce expres-sion of the ligands (supplemental Figure VI, D through I).However, we found that vegf expression was unaffected and,even though fgf8 expression was upregulated, it was in the tailbud and not the vasculature. In contrast, both activated ETSfactors, upregulated flk1 expression (supplemental Figure VI,A through C), raising the possibility that they might sensitizeendothelial cells to VEGF signaling.

In conclusion, by using systematic bioinformatic analysisand gene expression profiling, we have comprehensivelycharacterized the whole ETS gene family in zebrafish andidentified 12 of the 31 ETS genes as expressing in hemato-poietic or endothelial tissues. Functional assays have identi-fied 3 genes, etsrp, erg, and fli1, that are specifically requiredfor hemangioblast development and angiogenesis. Our stud-ies provide new insights into how tissue-specific differentia-tion and maintenance is achieved by the ETS domain tran-scription factors during hematopoiesis, vasculogenesis, andangiogenesis.

AcknowledgmentsWe thank Aldo Ciau-Uitz for helpful comments.

Sources of FundingThis work was funded by the Medical Research Council (UnitedKingdom).

DisclosuresNone.

References1. Davidson AJ, Zon LI. The ‘definitive’ (and ‘primitive’) guide to zebrafish

hematopoiesis. Oncogene. 2004;23:7233–7246.2. Keller G. The Hemangioblast. In: Marshak DR, Gardner RL, Gottlieb D,

eds. Stem Cell Biology. New York: Cold Spring Harbor Laboratory Press;2001:329–348.

3. Sabin FR. Studies on the origins of blood-vessels and of red blood-corpuscles as seen in the living blastoderm of chicks during the secondday of development. Contrib Embryol. 1920;36:213–262.

4. Shalaby F, Rossant J, Yamaguchi TP, Gertenstein M, Wu XF, BreitmanML, Schuh AC. Failure of blood-island formation and vasculogenesis inFlk-1 deficient mice. Nature. 1995;376:62–66.

5. Ferrara N, Carver-Moore K, Chen H, Dowd M, Lu L, O’Shea KS,Powell-Braxton L, Hillan KJ, Moore MW. Heterozygous embryoniclethality induced by targeted inactivation of the VEGF gene. Nature.1996;380:439–442.

6. Carmeliet P, Ferreira V, Breier G, Pollefeyt S, Kieckens L, GertsensteinM, Fahrig M, Vandenhoeck A, Harpal K, Eberhardt C, Declercq C,Pawling J, Moons L, Collen D, Risau W, Nagy A. Abnormal blood vesseldevelopment and lethality in embryos lacking a single VEGF allele.Nature. 1996;380:435–439.

7. Stainier DY, Weinstein BM, Detrich HW III, Zon LI, Fishman MC.Cloche, an early acting zebrafish gene, is required by both the endothelialand hematopoietic lineages. Development. 1995;121:3141–3150.

8. Patterson LJ, Gering M, Eckfeldt CE, Green AR, Verfaillie CM, EkkerSC, Patient R. The transcription factors Scl and Lmo2 act together duringdevelopment of the hemangioblast in zebrafish. Blood. 2007;109:2389–2398.

9. Patterson LJ, Gering M, Patient R. Scl is required for dorsal aorta as wellas blood formation in zebrafish embryos. Blood. 2005;105:3502–3511.

10. Dooley KA, Davidson AJ, Zon LI. Zebrafish scl functions independentlyin hematopoietic and endothelial development. Dev Biol. 2005;277:522–536.

11. Oikawa T, Yamada T. Molecular biology of the Ets family of tran-scription factors. Gene. 2003;303:11–34.

12. Gottgens B, Nastos A, Kinston S, Piltz S, Delabesse EC, Stanley M,Sanchez MJ, Ciau-Uitz A, Patient R, Green AR. Establishing the tran-scriptional programme for blood: the SCL stem cell enhancer is regulated

Liu and Patient ETS Genes in Zebrafish 1153

by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from

by a multiprotein complex containing Ets and GATA factors. Embo J.2002;21:3039–3050.

13. Chan WY, Follows GA, Lacaud G, Pimanda JE, Landry JR, Kinston S,Knezevic K, Piltz S, Donaldson IJ, Gambardella L, Sablitzky F, GreenAR, Kouskoff V, Gottgens B. The paralogous hematopoietic regulatorsLyl1 and Scl are coregulated by Ets and GATA factors, but Lyl1 cannotrescue the early Scl-/- phenotype. Blood. 2007;109:1908–1916.

14. Pimanda JE, Silberstein L, Dominici M, Dekel B, Bowen M, Oldham S,Kallianpur A, Brandt SJ, Tannahill D, Gottgens B, Green AR. Transcrip-tional link between blood and bone: the stem cell leukemia gene and its�19 stem cell enhancer are active in bone cells. Mol Cell Biol. 2006;26:2615–2625.

15. Landry JR, Kinston S, Knezevic K, Donaldson IJ, Green AR, Gottgens B.Fli1, Elf1, and Ets1 regulate the proximal promoter of the LMO2 gene inendothelial cells. Blood. 2005;106:2680–2687.

16. Sharrocks AD. The ETS-domain transcription factor family. Nat Rev MolCell Biol. 2001;2:827–837.

17. Westerfield M. The Zebrafish Book: A Guide for the Laboratory Use ofZebrafish (Brachydanio rerio). Eugene, Ore: University of Oregon Press;1993.

18. Sumanas S, Lin S. Ets1-related protein is a key regulator of vasculo-genesis in zebrafish. PLoS Biol. 2006;4:e10.

19. Jowett T, Yan YL. Double fluorescent in situ hybridization to zebrafishembryos. Trends Genet. 1996;12:387–389.

20. Gering M, Yamada Y, Rabbitts TH, Patient RK. Lmo2 and Scl/Tal1convert non-axial mesoderm into haemangioblasts which differentiateinto endothelial cells in the absence of Gata1. Development. 2003;130:6187–6199.

21. Sato Y. Role of ETS family transcription factors in vascular developmentand angiogenesis. Cell Struct Funct. 2001;26:19–24.

22. Pham VN, Lawson ND, Mugford JW, Dye L, Castranova D, Lo B,Weinstein BM. Combinatorial function of ETS transcription factors in thedeveloping vasculature. Dev Biol. 2007;303:772–783.

23. Hart A, Melet F, Grossfeld P, Chien K, Jones C, Tunnacliffe A, Favier R,Bernstein A. Fli-1 is required for murine vascular and megakaryocyticdevelopment and is hemizygously deleted in patients with thrombocy-topenia. Immunity. 2000;13:167–177.

24. Spyropoulos DD, Pharr PN, Lavenburg KR, Jackers P, Papas TS, OgawaM, Watson DK. Hemorrhage, impaired hematopoiesis, and lethality inmouse embryos carrying a targeted disruption of the Fli1 transcriptionfactor. Mol Cell Biol. 2000;20:5643–5652.

25. Birdsey GM, Dryden NH, Amsellem V, Gebhardt F, Sahnan K, HaskardDO, Dejana E, Mason JC, Randi AM. Transcription factor Erg regulatesangiogenesis and endothelial apoptosis through VE-cadherin. Blood.2008;111:3498–3506.

26. Woods IG, Wilson C, Friedlander B, Chang P, Reyes DK, Nix R, KellyPD, Chu F, Postlethwait JH, Talbot WS. The zebrafish gene map definesancestral vertebrate chromosomes. Genome Res. 2005;15:1307–1314.

27. Postlethwait JH, Yan YL, Gates MA, Horne S, Amores A, Brownlie A,Donovan A, Egan ES, Force A, Gong Z, Goutel C, Fritz A, Kelsh R,Knapik E, Liao E, Paw B, Ransom D, Singer A, Thomson M, Abdul-jabbar TS, Yelick P, Beier D, Joly JS, Larhammar D, Rosa F, WesterfieldM, Zon LI, Johnson SL, Talbot WS. Vertebrate genome evolution and thezebrafish gene map. Nature Genet. 1998;18:345–349.

28. Amores A, Force A, Yan YL, Joly L, Amemiya C, Fritz A, Ho RK,Langeland J, Prince V, Wang YL, Westerfield M, Ekker M, Postlethwait

JH. Zebrafish hox clusters and vertebrate genome evolution. Science.1998;282:1711–1714.

29. Sun XJ, Xu PF, Zhou T, Hu M, Fu CT, Zhang Y, Jin Y, Chen Y, ChenSJ, Huang QH, Liu TX, Chen Z. Genome-wide survey and developmentalexpression mapping of zebrafish SET domain-containing genes. PLoSONE. 2008;3:e1499.

30. Gering M, Rodaway AR, Gottgens B, Patient RK, Green AR. The SCLgene specifies haemangioblast development from early mesoderm.EMBO J. 1998;17:4029–4045.

31. Lugus JJ, Chung YS, Mills JC, Kim SI, Grass J, Kyba M, Doherty JM,Bresnick EH, Choi K. GATA2 functions at multiple steps in heman-gioblast development and differentiation. Development. 2007;134:393–405.

32. Pimanda JE, Ottersbach K, Knezevic K, Kinston S, Chan WY, WilsonNK, Landry JR, Wood AD, Kolb-Kokocinski A, Green AR, Tannahill D,Lacaud G, Kouskoff V, Gottgens B. Gata2, Fli1, and Scl form a recur-sively wired gene-regulatory circuit during early hematopoietic devel-opment. Proc Natl Acad Sci U S A. 2007;104:17692–17697.

33. Liu F, Walmsley M, Rodaway A, Patient R. Fli1 acts at the top of thetranscriptional network driving blood and endothelial development. CurrBiol. 2008;18:1234–1240.

34. Lee D, Park C, Lee H, Lugus JJ, Kim SH, Arentson E, Chung YS, GomezG, Kyba M, Lin S, Janknecht R, Lim DS, Choi K. ER71 acts downstreamof BMP, Notch, and Wnt signaling in blood and vessel progenitor spec-ification. Cell Stem Cell. 2008;2:497–507.

35. Sumanas S, Gomez G, Zhao Y, Park C, Choi K, Lin S. Interplay amongEtsrp/ER71, Scl, and Alk8 signaling controls endothelial and myeloid cellformation. Blood. 2008;111:4500–4510.

36. Xiong JW, Yu Q, Zhang J, Mably JD. An acyltransferase controls thegeneration of hematopoietic and endothelial lineages in zebrafish. CircRes. 2008;102:1057–1064.

37. Loughran SJ, Kruse EA, Hacking DF, de Graaf CA, Hyland CD, WillsonTA, Henley KJ, Ellis S, Voss AK, Metcalf D, Hilton DJ, Alexander WS,Kile BT. The transcription factor Erg is essential for definitive hemato-poiesis and the function of adult hematopoietic stem cells. Nat Immunol.2008;9:810–819.

38. Carrere S, Verger A, Flourens A, Stehelin D, Duterque-Coquillaud M.Erg proteins, transcription factors of the Ets family, form homo, het-erodimers and ternary complexes via two distinct domains. Oncogene.1998;16:3261–3268.

39. Vittet D, Buchou T, Schweitzer A, Dejana E, Huber P. Targeted null-mutation in the vascular endothelial-cadherin gene impairs the organi-zation of vascular-like structures in embryoid bodies. Proc Natl Acad SciU S A. 1997;94:6273–6278.

40. Gory S, Dalmon J, Prandini MH, Kortulewski T, de Launoit Y, Huber P.Requirement of a GT box (Sp1 site) and two Ets binding sites for vascularendothelial cadherin gene transcription. J Biol Chem. 1998;273:6750–6755.

41. Buchner DA, Su F, Yamaoka JS, Kamei M, Shavit JA, Barthel LK,McGee B, Amigo JD, Kim S, Hanosh AW, Jagadeeswaran P, GoldmanD, Lawson ND, Raymond PA, Weinstein BM, Ginsburg D, Lyons SE.pak2a mutations cause cerebral hemorrhage in redhead zebrafish. ProcNatl Acad Sci U S A. 2007;104:13996–14001.

42. Liu J, Fraser SD, Faloon PW, Rollins EL, Vom Berg J, Starovic-SubotaO, Laliberte AL, Chen JN, Serluca FC, Childs SJ. A betaPix Pak2asignaling pathway regulates cerebral vascular stability in zebrafish. ProcNatl Acad Sci U S A. 2007;104:13990–13995.

1154 Circulation Research November 7, 2008

by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from

Supplemental Table 1. Zebrafish ETS domain genes

Gene ID Gene Name Chromosome Description Subfamily Human Homologue

BC092935 ets1 18 ETS1 ETS ETS1 NM_001023580 ets2 10 ETS2 ETS ETS2 NM_001037375 etsrp 16 ETS1 related protein (Etsrp) ETS ER71 NM_001008616 erg 10 Transcripional regulator ERG ERG ERG NM_131348 fli1 18 Friend leukemia integration 1 transcription factor ERG FLI1 NM_001008780 fli1b 16 Friend leukemia integration 1b transcription factor ERG FLI1 XM_001341479 fev 9 ETS oncogene family ERG FEV NM_131587 gabpa 9 GA-binding protein alpha chain (GABP-alpha subunit) ELG GABPa NM_131832 tel1/etv6 6 translocation ets leukemia/ets variant gene 6 TEL TEL1/ETV6 XM_001340202 tel2 n.a. translocation ets leukemia 2 TEL TEL2 NM_131425 pea3 12 polyomavirus enhancer-binding activator 3 PEA PEA3 NM_131205 erm 6 ets related protein PEA ERM NM_001030182 er81/etv1 15 ets variant gene 1 PEA ER81/ETV1 NM_214720 etv5 9 ets variant gene 5 PEA ETV5 NM_131159 elf1 14 E74-like factor 1 ELF ELF1 NM_199219 elf2 1 E74-like factor 2 ELF ELF2 NM_001004116 elf2l1 14 E74-like factor 2-like 1 ELF ELF2 XM_684268 elf3 22 E74-like factor 3 ESE ELF3 NM_001039822 ehf 25 ets homologous factor ESE EHF XM_001343041 pdef 6 Prostate Epithelium-specific Ets Transcription Factor PDEF PDEF

NM_198062 spi1/pu.1 7 spleen focus forming virus (SFFV) proviral integration oncogene spi1 SPI SPI-1/PU.1

NM_001004621 spic 4 spleen focus forming virus (SFFV) proviral integration oncogene spi-c SPI SPI-C

XM_001341111 spicl 1 n.a. spleen focus forming virus (SFFV) proviral integration oncogene spi-c-like SPI SPI-C

CAK04878 erf 19 ETS2 repressor factor ERF ERF XM_681312 erfl1 16 ETS2 repressor factor-like 1 ERF ERF NM_001044927 erfl2 19 ETS2 repressor factor-like 2 ERF ERF XM_689171 erfl3 16 ETS2 repressor factor-like 3 ERF ERF XM_691202 elk1 8 Ets-like protein 1-like 1 TCF ELK1 NM_001030108 elk3 4 Ets-like protein 3 TCF ELK3 NM_130930 elk4 11 Ets-like protein 4 TCF ELK4 XM_001332642 elk4l1 11 Ets-like protein 4-like 1 TCF ELK4

1

Supplemental Figure Legends

Figure S1. ets1, ets2, fev, gabpα, elf1, elf2, pu.1, elk4 are expressed in

haematopoietic and/or endothelial cells during zebrafish embryogenesis. Scl, the

known hemangioblast marker in zebrafish was used for comparison (A-D). ets1 (E-

H), ets2 (I-L), fev (M,N), gabpa (O,P), elf1 (Q,R), elf2 (S,T), pu.1 (U,V) and elk4

(W). Of note, elk4 is not expressed at 6s stage. Anterior to the left in all panels.

Figure S2, Expression of ets1, ets2, fev, gabpα, elf1, elf2, pu.1, elk4 was reduced or

lost in cloche mutant embryos at 26 hpf. (A) ets1, (B) ets2, (C) fev, (D) gabpa, (E)

elf1, (F) elf2, (G) pu.1 and (H) elk4. Note that endothelial or blood specific expression

was reduced or absent in the cloche embryos (arrows). Anterior to the left and dorsal

to the top.

Figure S3. etsrp but not fli1 and erg is essential for anterior hemangioblast

formation in zebrafish embryos. ALM expression of scl was not affected in fli1- (B,

10/10) and erg- morphants (C, 10/10), but was significantly reduced in etsrp-

morphants (D, 12/12). pu.1 expression in ALM was not affected in fli1- (F, 10/10) and

erg- morphants (G, 10/10), but was completely lost in etsrp morphants (H, 10/10).

flk1 expression in ALM was only slightly affected in fli1- (J, 6/10) and erg-

morphants (K, 4/9), but was absent in etsrp morphants (L, 10/10). PLM expression of

scl was not affected in erg- (N, 11/12) and fli1- (O, 10/10) morphants, but was

reduced in etsrp- morphants (P, 10/10), whereas flk1 expression in PLM was slight

affected in erg- (R, 4/10) and fli1- morphants (S, 6/11), but completely lost in etsrp

morphants (T, 10/10). gata1 expression in PLM was normal in erg- (V, 9/9), fli1- (W,

2

10/11) and etsrp- (X, 10/10) morphants. Dorsal view and anterior to the top; controls

and MO-injected embryos were at 10s (A-X).

Figure S4. Knockdown of erg and fli1 causes downregulation of ve-cadherin.

(A,B, C) Live embryo at 3 dpf showed a hemorrhage in the head of erg morphants (B,

arrows, 24/58) and of the fli1 morphants (C, arrow, 8/24), with otherwise perfectly

normal overall morphology. Downregulation of ve-cadherin was seen in erg- (E,

arrows, 10/13), fli1- (F, arrows, 14/14) and the double morphants (G, arrows, 23/24)

at 50 hpf. Lateral views with anterior towards left (A-G).

Figure S5. Hemangioblast gene expression is unaffected in fev or quadruple (erg,

fli1, fli1b and fev) knockdowns. scl expression was normal in the fev morphant (B,

10/10) and the quadruple knockdown (C, 9/10) at 10s. Dorsal views, anterior to the

top.

Figure S6. (A-I) Activated erg (B, 10/10) and fli1 (C, 10/12) upregulate flk1

expression, and also fgf8 (E, 10/12; F, 13/14) but not vegf (H, 10/10; I, 9/10).

Embryos are at 24 hours postfertilization (hpf), anterior to the left. CA-Erg and CA-

Fli1 were constructed by subcloning the full length inserts into pBUT2VP16 vector

made in the lab 1. Fli1 and Erg coding sequences were fused at their C-terminus with

the viral constitutively active transactivation domain, VP16. The constructs were

verified by sequencing. Capped mRNAs (50-100 ng/embryo) for injection were

synthesized using the mMessage mMachine kit according to the manufacturer’s

protocol (Ambion, Austin, USA).

3

4

Supplemental References 1. Gering M, Rodaway AR, Gottgens B, Patient RK, Green AR. The SCL gene

specifies haemangioblast development from early mesoderm. Embo J. 1998;17:4029-4045.

fev

elf1

pu.1

VU

NM

gabpa

PO

Q

Liu_FigS1

6s 26h6s 26h

elf2

S TR

elk4

W

1s 6s 10s 26hCB D

scl

A

ets1

ets2

E GF

I KJ

H

L

Liu_FigS2

ets1

ets2

A

B

elk4

wt

clo

H

elf2

wt

clo

F

gabpa

wt

clo

D

elf1

wt

clo

E

fev

wt

clo

C

pu.1

G

wt

clo

wt

clo clo

wt

Liu_FigS3

Control fli1MO ergMO etsrpMO

scl

flk1

E F G H

I J K L

M N O P

Q R S T

U V W X

scl

pu.1

flk1

gata1

ALM

PLM

A B C D

A BWT ergMO fli1MOC

fli1MO ergMO+fli1MO

E

G

ergMO

F

D WT

ve-cadherin

Liu_FigS4

A B C

scl

WT fevMO 4MOs

Liu_FigS5

control CA-Erg CA-Fli1

flk1

fgf8

vegf

Liu_Fig

S6

A B

D E F

G H I

C

Feng Liu and Roger PatientHemangioblast Differentiation or Angiogenesis

Genome-Wide Analysis of the Zebrafish ETS Family Identifies Three Genes Required for

Print ISSN: 0009-7330. Online ISSN: 1524-4571 Copyright © 2008 American Heart Association, Inc. All rights reserved.is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation Research

doi: 10.1161/CIRCRESAHA.108.1797132008;103:1147-1154; originally published online October 2, 2008;Circ Res. 

http://circres.ahajournals.org/content/103/10/1147World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://circres.ahajournals.org/content/suppl/2008/10/03/CIRCRESAHA.108.179713.DC1.htmlData Supplement (unedited) at:

  http://circres.ahajournals.org//subscriptions/

is online at: Circulation Research Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. Further informationEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theCirculation Researchin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

by guest on November 25, 2015http://circres.ahajournals.org/Downloaded from