14
The convergence of Notch and MAPK signaling species the blood progenitor fate in the Drosophila mesoderm Melina Grigorian a, , Lolitika Mandal b , Manuel Hakimi a , Irma Ortiz a , Volker Hartenstein a a Department of Molecular Cell and Developmental Biology, University of California Los Angeles, Los Angeles, CA 90095, USA b Indian Institute of Science Education & Research, Mohali, India abstract article info Article history: Received for publication 9 December 2010 Revised 26 February 2011 Accepted 26 February 2011 Available online 5 March 2011 Keywords: Drosophila Blood Hemangioblast Notch MAPK EGF FGF Blood progenitors arise from a pool of pluripotential cells (hemangioblasts) within the Drosophila embryonic mesoderm. The fact that the cardiogenic mesoderm consists of only a small number of highly stereotypically patterned cells that can be queried individually regarding their gene expression in normal and mutant embryos is one of the signicant advantages that Drosophila offers to dissect the mechanism specifying the fate of these cells. We show in this paper that the expression of the Notch ligand Delta (Dl) reveals segmentally reiterated mesodermal clusters (cardiogenic clusters) that constitute the cardiogenic mesoderm. These clusters give rise to cardioblasts, blood progenitors and nephrocytes. Cardioblasts emerging from the cardiogenic clusters accumulate high levels of Dl, which is required to prevent more cells from adopting the cardioblast fate. In embryos lacking Dl function, all cells of the cardiogenic clusters become cardioblasts, and blood progenitors are lacking. Concomitant activation of the Mitogen Activated Protein Kinase (MAPK) pathway by Epidermal Growth Factor Receptor (EGFR) and Fibroblast Growth Factor Receptor (FGFR) is required for the specication and maintenance of the cardiogenic mesoderm; in addition, the spatially restricted localization of some of the FGFR ligands may be instrumental in controlling the spatial restriction of the Dl ligand to presumptive cardioblasts. © 2011 Elsevier Inc. All rights reserved. Introduction Cells of the blood, vascular and excretory system are develop- mentally related in all animals. In vertebrates, the lateral plate mesoderm gives rise to the progenitor cells of endothelia and blood cells (hemangioblasts) (Al-Adhami and Kunz, 1977; Cofn and Poole, 1988; Medvinsky et al., 1993; Cleaver et al., 1997; Eichmann et al., 1998; Gering et al., 1998; Crosier et al., 2002; Hartenstein, 2006; Xiong, 2008; Bertrand and Traver, 2009). Adjacent to the lateral plate is the intermediate mesoderm that produces the nephrocytes of the excretory system (pro-, meso-, metanephros). Hemangioblasts leave the lateral plate and migrate dorsally to differentiate into a system of endothelial tubes that pioneer the major blood vessels, including the aorta and cardinal veins. Progenitors of adult blood cells are specied within the walls of these primitive vessels; they are budded off into the vessel lumen and eventually settle in the blood forming organs (denitive hematopoiesis). In addition to these adult blood progenitors, blood cells populating the embryo are formed within the ventral blood islands and yolk sac (primitive hematopoiesis). In Drosophila, progenitors of the blood, vascular and excretory system originate from a lateral domain of the mesoderm, called the cardiogenic mesoderm (Hartenstein et al., 1992; Bate, 1993; Rugendorff et al., 1994; Zaffran et al., 2002; Mandal et al., 2004). The cardiogenic mesoderm is comprised of ten paired, segmentally repeated clusters of cells, three thoracic and seven abdominal clusters. Each cluster gives rise to a set of cardioblasts which assemble into a contractile, myo-endothelial tube, the dorsal vessel (or heart) of the y(Bodmer et al., 1997). In addition, the abdominal cardiogenic mesoderm produces groups of nephrocytes (pericardial nephro- cytes) which form part of the excretory system (Crossley, 1985; Rugendorff et al., 1994; Ward and Skeath, 2000). Progenitors of adult blood cells (prohemocytes) arise within the three thoracic cardio- genic mesodermal clusters. These prohemocytes aggregate together and form the so-called lymph gland that anks the anterior tip of the dorsal vessel in the late embryo and larva (Rugendorff et al., 1994; Mandal et al., 2004). Prohemocytes proliferate during larval stages and are released into the hemocoel (the body cavity of the insect) during metamorphosis (Lanot et al., 2001; unpublished data). Aside from this late embryonic/postembryonic denitivephase of blood development, there exists, just as in vertebrates, an early, primitivephase of blood formation that takes place in the mesoderm of the head of the early embryo (Tepass et al., 1994; Holz et al., 2003; Evans et al., 2003; de Velasco et al., 2006). Molecular factors specifying cells of the blood-vascular and nephrocyte lineages, as well as the signaling mechanisms controlling the expression of these factors, are highly conserved among all Developmental Biology 353 (2011) 105118 Corresponding author. E-mail address: [email protected] (M. Grigorian). 0012-1606/$ see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2011.02.024 Contents lists available at ScienceDirect Developmental Biology journal homepage: www.elsevier.com/developmentalbiology

The convergence of Notch and MAPK signaling specifies … lateral plate and migrate dorsally to differentiate into a system of ... and are released into the hemocoel (the body cavity

Embed Size (px)

Citation preview

Developmental Biology 353 (2011) 105–118

Contents lists available at ScienceDirect

Developmental Biology

j ourna l homepage: www.e lsev ie r.com/deve lopmenta lb io logy

The convergence of Notch and MAPK signaling specifies the blood progenitor fate inthe Drosophila mesoderm

Melina Grigorian a,⁎, Lolitika Mandal b, Manuel Hakimi a, Irma Ortiz a, Volker Hartenstein a

a Department of Molecular Cell and Developmental Biology, University of California Los Angeles, Los Angeles, CA 90095, USAb Indian Institute of Science Education & Research, Mohali, India

⁎ Corresponding author.E-mail address: [email protected] (M. Gr

0012-1606/$ – see front matter © 2011 Elsevier Inc. Aldoi:10.1016/j.ydbio.2011.02.024

a b s t r a c t

a r t i c l e i n f o

Article history:Received for publication 9 December 2010Revised 26 February 2011Accepted 26 February 2011Available online 5 March 2011

Keywords:DrosophilaBloodHemangioblastNotchMAPKEGFFGF

Blood progenitors arise from a pool of pluripotential cells (“hemangioblasts”) within the Drosophilaembryonic mesoderm. The fact that the cardiogenic mesoderm consists of only a small number of highlystereotypically patterned cells that can be queried individually regarding their gene expression in normal andmutant embryos is one of the significant advantages that Drosophila offers to dissect the mechanismspecifying the fate of these cells. We show in this paper that the expression of the Notch ligand Delta (Dl)reveals segmentally reiterated mesodermal clusters (“cardiogenic clusters”) that constitute the cardiogenicmesoderm. These clusters give rise to cardioblasts, blood progenitors and nephrocytes. Cardioblasts emergingfrom the cardiogenic clusters accumulate high levels of Dl, which is required to prevent more cells fromadopting the cardioblast fate. In embryos lacking Dl function, all cells of the cardiogenic clusters becomecardioblasts, and blood progenitors are lacking. Concomitant activation of the Mitogen Activated ProteinKinase (MAPK) pathway by Epidermal Growth Factor Receptor (EGFR) and Fibroblast Growth Factor Receptor(FGFR) is required for the specification and maintenance of the cardiogenic mesoderm; in addition, thespatially restricted localization of some of the FGFR ligands may be instrumental in controlling the spatialrestriction of the Dl ligand to presumptive cardioblasts.

igorian).

l rights reserved.

© 2011 Elsevier Inc. All rights reserved.

Introduction

Cells of the blood, vascular and excretory system are develop-mentally related in all animals. In vertebrates, the lateral platemesoderm gives rise to the progenitor cells of endothelia and bloodcells (hemangioblasts) (Al-Adhami and Kunz, 1977; Coffin and Poole,1988; Medvinsky et al., 1993; Cleaver et al., 1997; Eichmann et al.,1998; Gering et al., 1998; Crosier et al., 2002; Hartenstein, 2006;Xiong, 2008; Bertrand and Traver, 2009). Adjacent to the lateral plateis the intermediate mesoderm that produces the nephrocytes of theexcretory system (pro-, meso-, metanephros). Hemangioblasts leavethe lateral plate and migrate dorsally to differentiate into a system ofendothelial tubes that pioneer the major blood vessels, including theaorta and cardinal veins. Progenitors of adult blood cells are specifiedwithin the walls of these primitive vessels; they are budded off intothe vessel lumen and eventually settle in the blood forming organs(“definitive hematopoiesis”). In addition to these adult bloodprogenitors, blood cells populating the embryo are formed withinthe ventral blood islands and yolk sac (“primitive hematopoiesis”).

In Drosophila, progenitors of the blood, vascular and excretorysystem originate from a lateral domain of the mesoderm, called

the cardiogenic mesoderm (Hartenstein et al., 1992; Bate, 1993;Rugendorff et al., 1994; Zaffran et al., 2002; Mandal et al., 2004). Thecardiogenic mesoderm is comprised of ten paired, segmentallyrepeated clusters of cells, three thoracic and seven abdominal clusters.Each cluster gives rise to a set of cardioblasts which assemble into acontractile, myo-endothelial tube, the dorsal vessel (or “heart”) of thefly (Bodmer et al., 1997). In addition, the abdominal cardiogenicmesoderm produces groups of nephrocytes (“pericardial nephro-cytes”) which form part of the excretory system (Crossley, 1985;Rugendorff et al., 1994; Ward and Skeath, 2000). Progenitors of adultblood cells (prohemocytes) arise within the three thoracic cardio-genic mesodermal clusters. These prohemocytes aggregate togetherand form the so-called lymph gland that flanks the anterior tip of thedorsal vessel in the late embryo and larva (Rugendorff et al., 1994;Mandal et al., 2004). Prohemocytes proliferate during larval stagesand are released into the hemocoel (the body cavity of the insect)during metamorphosis (Lanot et al., 2001; unpublished data). Asidefrom this late embryonic/postembryonic “definitive” phase of blooddevelopment, there exists, just as in vertebrates, an early, “primitive”phase of blood formation that takes place in themesoderm of the headof the early embryo (Tepass et al., 1994; Holz et al., 2003; Evans et al.,2003; de Velasco et al., 2006).

Molecular factors specifying cells of the blood-vascular andnephrocyte lineages, as well as the signaling mechanisms controllingthe expression of these factors, are highly conserved among all

Fig. 1. Partitioning of the mesoderm by the Wingless (Wg) and Decapentaplegic (Dpp)signals. cm, cardiogenic mesoderm; fb, fat body sm, somatic mesoderm; vm, visceralmesoderm. Combined Notch and MAPK activity defines smaller clusters (“myogenicand cardiogenic clusters”) in the somatic and cardiogenic mesoderm (light coloredovals).

106 M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

animals (reviewed in Zaffran et al., 2002; Evans and Banerjee, 2003;Hartenstein and Mandal, 2006; Hartenstein, 2006; Crozatier andMeister, 2007; Martinez-Agosto et al., 2007). Hemangioblasts of thevertebrate lateral plate and Drosophila cardiogenic mesoderm areinduced from naive mesoderm by several signaling pathways thatinclude Bone morphogenetic protein (BMP)/Decapentaplegic (Dpp)and fibroblast growth factor (FGF)/Heartless (Htl). The activity ofthese signals is required for upregulating blood and vascular/heartdeterminants such as the GATA factors/Serpent (Srp) or Nkx2.5/Tinman (Tin), and thereby specifies and maintains the fate ofhemangioblasts (Baron, 2003; 2005; Crosier et al., 2002; Maeno,2003). Notch signaling plays an important role in specifying bloodprogenitors from the bi-potential (vascular/blood) hemangioblasts.Thus, Notch activation in the endothelial hemangioblasts of vertebrateembryos results in the expression of blood cell determinants such asGATA2, stem cell leukemia (SCL) and acutemyeloid leukemia 1 (AML1),placing the Notch signaling pathway high up in the molecular networkinitiating hematopoiesis (Kumano et al., 2003; Hadland et al., 2004;Robert-Moreno et al., 2005; Burns et al., 2005). Again, the same switchbetween vascular cells and blood progenitors is under the control ofNotch signaling in Drosophila: loss of Notch activity during the phasewhen the fate of cardiogenic mesodermal cells is specified results inthe absence of blood progenitors and an excess of cardioblasts (Mandalet al., 2004). Whereas the important role of Notch signaling for bloodprogenitor fate in Drosophila, as in vertebrates, is well established, thesignal that activates the Notch receptor, and the mechanism by whichthe signal itself becomes active in the restricted spatial pattern thatguaranteeing the proper number and distribution of blood progenitors,has not so far been elucidated. In this paper we address these questionsin Drosophila, taking advantage of the fact that the Drosophilacardiogenic mesoderm is comprised of only a few cells, many ofwhich can be individually marked by molecular reagents.

The Drosophila mesoderm arises as a single cell layer duringgastrulation. Subsequently, mesodermal cells undergo two parasyn-chronous mitotic divisions (Campos-Ortega and Hartenstein, 1985;Bate, 1993). During this early phase, genes which later demarcatespecific mesodermal lineages, such as tin orMyocyte Enhancer Factor 2(Mef2), are expressed in the entire mesoderm (Azpiazu and Frasch,1993; Bodmer, 1993; Lilly et al., 1994; Taylor et al., 1995), suggestingthat mesodermal cells are equipotent and have not yet split up intodiscrete lineages. The specification of different lineages is initiated by aseries of signaling steps,which include theWingless (Wg), Notch, Dpp,Epidermal Growth Factor (EGF), and FGF pathways (Bate, 1993;Zaffran et al., 2002; Hartenstein, 2006). As a result, the mesoderm ofeach segment becomes subdivided into four quadrants (Borkowski etal., 1995; Riechmann et al., 1997;Mandal et al., 2004;Martinez-Agostoet al., 2007; Fig. 1): high levels of Tin expression, dependent on bothfibroblast growth factor receptor (FGFR) activity and an ectodermallyderived Dpp signal, define the dorsal mesoderm. The anterior, Wg-positive dorsal mesoderm forms the cardiogenic mesoderm while theWg-negative posterior dorsal mesoderm becomes the visceral meso-derm. The two ventral quadrants give rise to the somatic musculature(Wg-positive, anteriorly) and fat body (Wg-negative, posteriorly). Theexpression of cell type specific genes at this stage indicates that fatbody, somatic mesoderm and visceral mesoderm have becomemolecularly distinct entities. By contrast, the cardiogenic mesodermat this stage still represents a mixed population of cells with differentfates. These fates include cardioblasts, pericardial nephrocytes and atleast two other, different types of “pericardial” cells (Even-skipped(Eve) and Tin-positive pericardial cells), blood progenitors, and dorsalmuscle cells (Ward and Skeath, 2000; Knirr and Frasch, 2001).

Cell–cell interactions depending on Notch, EGF and FGF signalingsingle out from within the Wg-positive somatic and cardiogenicmesoderm 19 small clusters of cells, called C1-C19 (Carmena et al.,1995; 1998; 2002; Buff et al., 1998; Fig. 5A). These are revealed by theexpression of the bHLH transcription factor Lethal of scute (L'sc), as

well as the expression of phosphorylated Mitogen Activated ProteinKinase (pMAPK), a reporter for EGF and FGF signaling activity.Clusters of the somatic mesoderm (C1, C3–C13, C17) give rise to theso-called muscle founder cells, which act as pioneers to specify thepattern of somatic muscles (Bate, 1993; Carmena et al., 1995). WhileC18 and C19 have not been as thoroughly characterized as the rest, itis known that C18 expresses higher levels of L'sc in thoracic than inthe abdominal segments, while C19 is found only in thoracic segments(Carmena et al., 1995). We show, in this study, that the dorsal clusters(C2; C14–C16) constitute the definitive cardiogenic mesoderm thatgives rise to cardioblasts, blood progenitors and nephrocytes. As thecardioblasts emerge they accumulate high levels of the Notch ligandDl, which is required to restrict the number of cells adopting thecardioblast fate. Loss of Dl converts all cells of the cardiogenicmesoderm into cardioblasts. Activation of theMAPK pathway by EGFRand FGFR is required for the specification and maintenance of thecardiogenic mesoderm, and is also likely to be involved in the spatialrestriction of the Dl ligand to cardioblasts.

Materials and methods

Fly stocks

Delta9p, Df(1)scB57, dppH46, dpp-lacZ, EGFRf2, htlAB42, spitz-lacZ,UAS-EGFR DN, UAS-htl Constitutively activated, UAS-RasV12, UAS-dpp,UAS-wg, wgCX4 (Bloomington Stock Center). 12X Su(H)-lacZ (con-struct with 12 Su(H) binding sites upstream of a lacZ; Go et al.,1998),Mef2-gal4 (Ranganayakulu et al. 1996), UAS-EGFR activated #3(UAS-λ top 4.4; Queenan et al. 1997, kindly gifted by T. Schupbach),UAS-l'sc (Carmena et al. 1995), UAS-l'sc Rnai (NIG: Japan). htl-lacZ(Stathopoulos et al. 2004), thisbe-lacZ #2 (Stathopoulos et al. 2004),Df(2R)ths238 and Df(2R)pyr36 (Kadam et al. 2009), wg-lacZ (NIG:Japan N755).

Antibodies

Antibodies used for immunohistochemistry include: α-mouse-β-galactosidase (1:50) (Promega), α-rabbit-β-galactosidase (1:2500)(Cappel), α-guinea pig-delta (1:1000)(Huppert et al., 1997), α-rat-EGFR (1:500) (kindly gifted by B. Shilo), α-rat-L'sc (1:800)(Martin-Bermudo et al., 1991, kindly gifted by A. Carmena), α-Rabbit-Mef2(1:1500) (Nguyen et al., 1994), α-rabbit-Odd (1:500) (Ward andSkeath, 2000), α-Mouse-Pericardin (1:3) (Developmental HybridomaBank), α-Mouse-pERK (pMAPK) (1:50) (Sigma), α-Rabbit-Tinman(1:800) (Venkatesh et al., 2000).

107M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

Embryo fixation

RegularFlies were placed on fruit juice plates for overnight egg laying

collection. Embryos/eggs were treated with 100% bleach for 6 minutesand then washed with water. Embryos were then fixed in a solution of3.5 ml PEMS, 3.5 ml Heptane and 1.5 mL 37% Formaldehyde for35 minutes. PEMS (pH 7) is a solution consisting of 100 mL PIPES(0.2 M solution, pH 7), 90 mL of Deionized water, 8 mL EGTA (0.05 Msolution) and 2 mL MgSo4 (0.1 M solution). The bottom phase of thesolution (PEMS)was then removed andMeOHwas added in an amountcomparable to the PEMS removed. Vial was shaken vigorously andsettled embryos were collected and stored in EtOH at −20 °C till use.

Embryo fixation for pMAPK stainingAbove fixation technique was carried out with the following

modification: 10 mM of Sodium Orthovanadate and 1:500 of 30%hydrogen peroxide (H2O2) were added to the PEMS solution. Also, forpMAPK staining, fixed embryos were stained immediately withpMAPK antibody.

Staging of embryos

We followed the system of stages introduced by Campos-Ortegaand Hartenstein (1985). To recognize smaller steps within stages 11(late extended germband) and 12 (germband retraction), each ofwhich lasts for about 2 h, we defined morphological criteria that areeasily recognizable without specific markers. Stages 11 and 12 weresubdivided into three roughly equal intervals called 11A (early 11),11B (mid 11), 11C (late 11), 12A (early 12), 12B (mid 12) and 12C(late 12). Defining criteria for these stages:

– 11A: Tracheal pits start to form and progress to hemisphericalinvaginations.

– 11B: Tracheal pits form vertical slits; salivary placode can bedistinguished as a thickened part of the ventral ectoderm of thelabial segment.

– 11C: Salivary placode starts to invaginate.– 12A: Salivary placode invaginates to form a straight, vertical, tube-

shaped primordium. Germband starts to retract (83% to 75%).– 12B: Salivary primordium elongates and is bent posteriorly at

posterior tip. Germband retracts to 50%.– 12C: Stage of rapid germband retraction (50%–0%).

Each of these stages can be further subdivided into intervals of 10–15 minutes length, each defined by clear (if subtle) morphologicalchanges (V. Hartenstein, unpublished data).

Immunohistochemistry

RegularSamples were washed with 0.1% PBT three times for 15 minutes

each at room temperature while rotating. Samples were then blockedfor 30 minutes in a solution of 10% NGS made in 0.1% PBT. Samplesthen placed in 1° antibody solutionmade in 10% NGS and left to rotateovernight at 4 °C. Samples were then washed and blocked as des-cribed above and placed in a 2° antibody conjugated to a fluorescentmarker. Samples were left in 2° antibody to rotate overnight at 4 °C.Samples were then washed as described above and mounted inVectashield (Vector Laboratories) containing a 1:500 concentration ofToto-3-iodide (Invitrogen).

L'sc antibody stainingUsed Perkin-Elmer Life Sciences Renaissance Tyramide Signal

Amplification Biotin System (Catalog #NEL700A).

In situ hybridization

L'sc RNA probe synthesisL'sc clone (RE59335) was received from Berkeley Drosophila

Genome Project (BDGP). Plasmid was digested with appropriaterestriction enzyme. Then treated with 300 μL of TE or water. Phenol/chloroform extraction was carried out by adding 150 μL of each agentand then centrifuging DNA twice at 12,000 rpm for 5 minutes each.The third time 300 μL of chloroformwas added and centrifugationwascarried out as described above. The upper phase was saved each timeand moved to a new centrifuge tube. After final spin, 270 μL of theupper phase was transferred to a new tube and 27 μL of 3 M sodiumacetate and 210 μL of isopropanol was added. Sample was mixed wellbetween addition of each agent. Sample was then incubated for20 minutes at room temperature. The sample was then spun at15,000 rpm for 15 minutes. The supernatant was discarded and thepellet was washed with 500 μL of 70% EtOH. Sample was spun againfor 5 minutes at 15,000 rpm. The supernatantwas again discarded andthe pellet was allowed to air dry for 10 minutes. Pellet was thendissolved in 16 μL TE. Probe was then synthesized by placing 1 μg ofprepared template DNA with 4 μL of 5X transcription buffer, 2 μL of100 mM DTT, 2 μL Dig-NTP mix, 2 μL RNA Polymerase, 1 μL RNaseinhibitor, 1 μL50 mM MgCl2 and filling up to 20 μL with ddH2O. Themixture was incubated for 2 h at 37 °C. 2 μL of DNaseI (RNase free)was then added and sample was incubated for an additional25 minutes at 37 °C. 80 μL of RNase free water was then added. TheRNA probe was then purified using a Qiagen RNeasy Kit (Cat #74106).Protocol from kit was used to purify RNA. Ethanol precipitation wasthen carried out by adding 6 μL of 3 M sodium acetate and 15 μL of100% EtOH to the sample and leaving it at either −20 °C overnight orat −80 °C for 30 minutes. Sample was then spun at 15,000 rpm for15 minutes at 4 °C. Supernatant was discarded and pellet was rinsedwith 300 μL of 70% EtOH. Sample was spun again at 15,000 rpm for5 minutes at 4 °C. EtOH was then discarded and RNA was dissolved in40 μL of RNAse free water. The quality of the probe was checked byelectrophoresis and 120 μL of hybridization buffer was added to itbefore it was stored at −20 °C.

Embryo preparation and in situ hybridizationEmbryos fixed as stated above, except in 4% Formaldehyde (EM

grade)/PBS with same amount of heptane for 30–45 minutes.Embryos are stored in MeOH. To rehydrate, embryos were treatedwith 75%, 50%, and 25% MeOH made in PBS for 5 minutes each.Embryos were then washed three times for 5 minutes each in 0.1%PBST. Embryos were then treated with 300 μL of 10 μg/mL proteinaseK for 4 minutes (did not rotate). Proteinase K was then discarded andglycine (2 mg/ml) made in PBST added twice for 1 minute each.Sample was then fixed again with 4% EM Grade Formaldehyde for5 minutes. Sample then washed with PBST 5 times for 5 minutes each.Embryos were then pre-hybridized by adding a 1:1 solution of pre-hybridization buffer and PBST to them for 5 minutes. Embryos werethen treated with Pre-hybridization buffer alone for 5 minutes.Embryos were then rinsed with 300 μL of hybridization buffer before300 μL of hybridization buffer was added and embryos wereincubated for more than 1 h at 60 °C. To carry out hybridization,diluted 16 μL of labeled probe with 100 μL hybridization buffer. Thediluted probe was heated at 80 °C for 10 minutes and then chilled onice. Extra solution was removed from tube containing embryos,leaving only 100 μL of hybridization buffer. Denatured probe was thenadded to the tube and hybridization was allowed to ensue at 60 °Covernight. The embryos are then washed with preheated (60°)washing reagents. Washed initially with 50% formamide and 2× SSCTtwice for 20 minutes each at 60 °C. Followedbywashing twicewith 2×SSCT for 10 minutes each. Finally, washed twice with 0.2× SSCT for30 minutes each. Immunoreaction was then carried out. Embryoswere rinsedwith PBST. Embryoswere then blockedwith 0.2% blocking

Fig. 2. Delta (Dl) activity is localized and required in the cardiogenic mesoderm for thespecification of blood progenitors and pericardial nephrocytes. A–F: lateral views ofstage 13 (A–B, E–F) and 14 (C–D) embryos. Left column (A, C, E) are wild type, rightcolumn (B, D, F) Dl mutants. Upper panels (A–B) are labeled with anti-Dl (green), anti-Mef2 (red) and Sytox (blue). Mef2 labels both cardioblasts (cb) and somatic musclecells (sm). Middle panels (C–D) are labeled with anti-Tinman (Tin; red) and anti-Delta (green). Tin labels cardioblasts and Tin positive pericardial cells (tpc). Lowerpanels (E, F) show labeling with Sytox (blue) and anti-Odd skipped (red), which labelsthe lymph gland progenitors (lg) and the Odd positive pericardial nephrocytes (pcn). Dlis expressed in cardioblasts in wild type embryos (A). In the absence of Dl (B, D),cardioblasts are strongly increased in number, while Odd-positive pericardialnephrocytes and lymph gland blood progenitors are missing (F). G–I: Highmagnification of the cardiogenic mesoderm of a stage 13 wild-type embryo, showingDl expression in the cardioblasts (G), and expression of a lacZ gene found downstreamof 12 Suppressor of hairless [Su(H)] binding sites [reporter for Notch activity (N. rep.)]in the adjacent pericardial nephrocytes and lymph gland (H, I). Other abbreviations: deOdd-positive dorsal ectoderm cells Bars: 25 μm (A–D); 10 μm (E–G).

108 M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

reagent (Roche) made in PBST for 2 h at room temperature. Embryoswere then placed in anti-DIG antibody made in 0.2% blocking reagentovernight at 4 °C or for 2 h at room temperature. Antibody was thenwashed 6 times with PBST for 10 minutes each. Embryos then washedtwice with AP buffer (with 0.01% Tween20) for 5 minutes each.Finally, to stain NBT/BCIP tablets were dissolved in water and 500 μLof it was added to the embryos. Signal was then developed in the dark.To stop the reaction, embryos were washed 5 times with PBST (shortwashes initially, followed by a final 5 minute wash). To removebackground, PBST containing 1% tween20 was added to embryosovernight at room temperature. Embryos were then rinsed andmounted in 50% glycerol/PBS initially and finally in 100% glycerol.

Solutions included hybridization buffer (20 mL Formamide, 10 mLof 20× SSC, 40 mg RNA (ribosomal), 200 μL 10 mg/mL Heparin, 200 μL20%Tween20 andddH20 (autoclaved) up to 40 mL), Pre-Hybridizationbuffer (20 mL Formamide, 10 mL of 20× SSC, 200 μL 10 mg/mLHeparin, 200 μL 20% Tween20 and ddH2O (autoclaved) up to 40 mL),20× SSC (175.3 g NaCl and 88.2 g SodiumCitrate in 1 L of ddH2Owith apH of 7.0), AP buffer (5 mL Tris–HCl (pH 9.5), 1 mL of 5 N NaCl, 2.5 mLof 1 M MgCl2 and up to 50 mL of ddH2O), AP buffer with Tween20(added 0.5 μL of 20% Tween20 to each 1 mL of AP buffer), PBST (PBScontaining 0.1% Tween20).

Generation of 3D digital models

Staged Drosophila embryos labeled with anti-Delta and othermarkerswere viewed aswholemounts by confocalmicroscopy (BioradMRC 1024ES microscope using Biorad Lasersharp version 3.2software; lenses: 40× oil; 60× oil). Complete series of optical sectionswere taken at 1 mm intervals for at least five individuals per stage.Tif files of the confocal stacks were generated using Image J (Abramoffet al., 2004), and imported into the software package TrakEM2(Cardona et al., 2010), which allows for efficiently segmentingstructures by outlining them or, in case of simple shapes likenuclei, simply clicking them. We used the built in 3D viewer ofTrakEM2 to generate three dimensional models of the segmentedstructures.

Results

Delta signaling within the cardiogenic mesoderm specifiesblood precursor fate

Previous studies had shown that Notch activity is required withinthe cardiogenic mesoderm during 8–10 h after egg laying (AEL) forthe correct specification of the different cardiogenic lineages,including cardioblasts (vascular cells), blood progenitors, and peri-cardial nephrocytes. Specifically, a high level of Notch activity isnecessary for the specification of Odd-skipped (Odd)-positivepericardial nephrocytes and blood progenitors (Mandal et al., 2004).Drosophila Notch has two ligands, Dl and Serrate (Ser). To testwhich one of these acts in the cardiogenic mesoderm, we analyzedtheir loss of function phenotype. Embryos mutant for the Ser genedid not exhibit any detectable abnormalities among the cell typesformed from the cardiogenic mesoderm (data not shown). On theother hand, null mutations in Dl caused a complete loss of pericardialcells and hematopoietic precursors (Fig. 2E–F). In contrast, cardioblastnumbers were dramatically increased, phenocopying the Notchmutant phenotype (Mandal et al., 2004; Fig. 2A–D). Two markers,Mef2 and Tin were used to label cardioblasts. Mef2 in addition tobeing found in cardioblasts is also found in somatic muscle cells(Fig. 2A). Meanwhile Tin is expressed in cardioblasts as well as in theTin positive pericardial cells (Fig. 2C). Dl mutants labeled with eithermarker showed an increase in the number of cells being labeled incomparison to their WT counterparts (Fig. 2B, D).

A construct with 12× Su(H) binding sites upstream of a lacZ wasused to report activated Notch signaling (Go et al., 1998). This inconjunction with Dl expression studies provided further insight intothe Notch/Dl-dependent signaling step acting in the cardiogenicmesoderm. As shown in the following section, in more detail,expression of Dl is widespread and dynamic in the mesoderm aroundthe time when the fate of cardiogenic cells is specified (stage 12; 7–9 h AEL). Toward the end of this period, expression of Dl becomesrestricted to the cardioblasts, where it stays on until stage 14(approximately 11 h AEL; Fig. 2G). 12× Su(H)-lacZ is also expressedin the cardiogenic mesoderm of stage 12–14 embryos, but becomesrestricted to the pericardial cells and blood progenitors flankingthe Dl expressing cardioblasts (Fig. 2H–I). These findings indicatethat the expression of Dl in the cardiogenic mesoderm provides thesignal controlling the proper balance between the different cardio-genic lineages.

Origin and morphogenesis of the cardiogenic mesoderm

To learn more about the signaling mechanism that distinguishesbetween the different cardiogenic fates we next carried out a detailedanalysis of the dynamic changes that occur in cell position and

109M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

proliferation within the mesoderm.We then correlated these changesto the expression pattern of Dl, as well as other factors known to beinvolved in the control of Dl, notably the proneural gene l'sc, andmembers of the EGF and FGF signaling pathways.

The different cardiogenic lineages split from each other duringembryonic stage 12 (7–8.5 h AEL). Prior to the beginning of this phase,the cardiogenic mesoderm, defined by the expression of Tin, formssegmentally reiterated, elongated clusters of approximately 30–35cells each (“early cardiogenic mesoderm”; Fig. 3A, Q). This numberincreases to about 60 cells at the end of stage 12 (Fig. 3N, T), followingthe fourth round of mitosis (Bate, 1993) that takes place during thistime period (Fig. 3F, L, O). Based on cell size/shape and the expressionlevels of Dl, one can distinguish a dorsal and ventral subdomainwithin the cardiogenicmesoderm of the stage 12 embryos. Cells of theventral domain are small and express moderate levels of Dl and Tin;dorsally, we see larger cells, expressing higher levels of Tin (Fig. 3E, R).Dl expression is highly dynamic in the dorsal domain; at most timepoints during stage 12, we see an anterior and posterior cluster ofstrongly Dl-positive cells, flanking a central cluster with lower Dllevels (Fig. 3E', R; schematically shown in Fig. 4). This central clusterexpresses Even-skipped (Eve; Fig. 3F'; Fig. 4). As a matter of fact, Eve-expression defines a dorso-central cluster already in the earlycardiogenic mesoderm (Fig. 3B, C and see below); from within thiscluster, the pair of Eve-positive pericardial cells is specified (Carmenaet al., 1998). The later Eve-positive/low-Dl dorso-central cluster givesrise to a dorsal muscle. The anterior and posterior, high-Dl clustersform the “definitive” cardiogenic mesoderm (cmd in Fig. 3); asdetailed below, they will give rise to the cardioblasts of the dorsalvessel and the Odd-positive blood progenitors and pericardialnephrocytes. The ventral domain showing low levels of Tin and Dl(colored green in Fig. 3Q–T) contribute to the dorsal musculature; anumber of these cells seem to undergo apoptosis at a later stage (datanot shown).

Individual lineages derived from the cardiogenic mesodermsegregate from each other and begin to differentiate during thesecond half of stage 12 (12B and 12C; 8–9 h AEL). Morphogeneticevents that unfold during this time interval are complex; they includecell division (the fourth mesodermal mitosis), cell movements withineach segment, as well as cell movements occurring as part of germband retraction. The anterior and posterior high-Dl cluster of eachsegment move closer together (compare Fig. 3E/E' with 3K/K'), whilethe central, low-Dl cluster moves “out of the way,”migrating laterallyand ventrally (compare Fig. 3F/F' with Fig. 3L/L'). During the phasewhen this cell rearrangement takes place andmost cells undergo theirfinal (4th) mitosis, Dl expression becomes restricted to the cardio-blasts. In the thoracic segments, each of the high-Dl clusters forms twocardioblasts (Fig. 3K/K', Fig. 3N/N'; Fig. 4). In abdominal segments,three cardioblasts arise per cluster. However, only two cells of eachabdominal cluster maintain a high level of Dl and Tin. The third celldown-regulates these genes and expresses Seven-up (Svp; Gajewskiet al., 2000; Ward and Skeath, 2000). Expression of Svp is alreadyapparent during the beginning of stage 12when the high-Dl cells forma separate anterior and posterior cluster in each segment (Fig. 3G).Like most other cells within the high-Dl clusters, the Svp-positive cellsalso divide once, forming an anterior and posterior pair in eachsegment (Fig. 3G, H/H'). When high-Dl clusters are pushed togetherduring germ band retraction, the posterior Svp-positive pair of a givensegment meets the anterior Svp-positive pair of the posteriorlyadjacent segment (Fig. 3M/M'; 3P/P'; schematically shown in Fig. 4).The dorsal sibling of each Svp-positive pair will become a cardioblast;the ventral sibling a pericardial nephrocyte (Ward and Skeath, 2000;see below). In this manner, the pattern of alternating sets of four Tin/Dl-positive and two Svp-positive cardioblasts characteristic of theabdominal dorsal vessel is generated (Fig. 3P/P').

Pericardial nephrocytes and blood progenitors also becomediscernable by size, position and expression of the gene odd by late

stage 12. Based on our detailed reconstructions of embryos expressingodd in combination with tin or Dl we propose that nephrocytes andblood progenitors derive from the high-Dl clusters (Figs. 3, 4).Abdominal segments produce four pairs of nephrocytes; two pairs aresvp-positive (siblings of the Svp-positive cardioblasts; see above), theother two are Svp-negative. Meanwhile, the three thoracic segmentseach give rise to approximately eight to ten pairs of blood progenitors,all of which are Svp-negative. Similar to the abdominal nephrocytes, aconsiderable fraction of the blood progenitors are also siblings of thecardioblasts (Mandal et al., 2004). What accounts for the highernumber per segment of blood progenitors vs. nephrocytes? The high-Dl clusters of the thoracic segments of stage 12B embryos (prior toseparation of cardioblasts and blood progenitors) are not significantlylarger than their abdominal counterparts (Fig. 3F/F'; Fig. 4). However,the fraction of cells undergoing a fourth round ofmitosis appears to behigher in the thoracic segments than the abdominal segments(Fig. 3F/F'), which may be in part responsible for the higher numberof blood progenitors. A second contributing factor may be the fact thatthe two pairs of cells that express Svp and produce cardioblasts andnephrocytes in the abdominal segments generate only bloodprogenitors in the thoracic segments.

To sum up, we show that two high-Dl clusters located within thelateral mesoderm of the stage 11/early 12 embryo, called “cardiogenicclusters” from hereon onward, represent the definitive cardiogenicmesoderm fromwhich cardioblasts, blood progenitors and pericardialnephrocytes are derived. Notch/Dl dependent interactions within thecardiogenic clusters separate these cell types, whereby the signal Dlbecomes restricted to cardioblasts, and high levels of Notch activityappear in the adjacent blood progenitors/nephrocytes.

The cardiogenic clusters are defined by the co-expression of Dl, l'sc andactivated MAPK

Careful mapping studies of themesoderm, using a probe for the l'scgene, had revealed nineteen clusters (C1–C19) defined by (transient-ly) high levels of L'sc. (Carmena et al., 1995; Fig. 5A). These sameclusters also co-express pMAPK, a reporter for MAPK signaling. Wecan identify two of these clusters, C14 and C16, as the high-Dl,definitive cardiogenic mesoderm described in the previous section. C2and C15 correspond to the Eve-positive clusters. Thus, in situ andantibody stainings revealed a cluster (C2) of L'sc positive cells in thecenter of the early cardiogenic mesoderm of the stage 11 embryos(Fig. 5B), and three clusters covering the definitive cardiogenicmesoderm. The latter three included the two clusters (C14, C16) andthe late Eve-cluster (C15) of the early stage 12 embryo (Fig. 5C). Usingdouble labeling with an antibody against Dl and pMAPKwe confirmedthat the same clusters also express pMAPK during stage 11 (C2) andstage 12 (C14–16; Fig. 5D, E). These findings suggest that both theMAPK pathway as well as L'sc may act in the cardiogenic mesoderm tomaintain/focus the expression of Dl, and thereby provide necessaryinput for the development of pericardial nephrocytes and bloodprogenitors. To test this hypothesis we carried out loss and gain offunction studies for members of both pathways.

MAPK activity, mediated by the EGF and FGF pathways, is required forthe specification of blood progenitors and pericardial nephrocytes

Past studies had already shown that Ras, a member of the MAPKpathway, is capable of activating Dl expression. Constitutive activationof Ras leads to a persistence of Dl expression in the clusters C2 and C15of L'sc positive cells present in the cardiogenic mesoderm (Carmenaet al. 1998; 2002). We wanted to investigate whether this pathway isalso active in the cardiogenic mesoderm, that is, clusters C14 and C16.Overexpression of a constitutively active form of Ras (RasV12) using aMef2-gal4 driver, expressed globally in the somatic and cardiogenicmesoderm (Taylor et al., 1995; Ranganayakulu et al., 1996) resulted in

Fig. 3. Morphogenesis and cell fate in the cardiogenic mesoderm. Photographic panels show lateral views of embryos at stages indicated to the left (for subdivision of stage 12, seeMaterials and methods). Each panel shows several consecutive segments of the mesoderm, labeled by anti-Dl (green); segments are indicated at bottom of panels (T1–T3: thoracicsegments 1–3; A1–A3: abdominal segments 1–3). In most cases, a lowmagnification view and a highmagnification view of the same specimen are shown in consecutive panels (e.g.,E and E'; F and F'). For stage 12A, panel C presents a view of the cardiogenic mesoderm, and panel D shows a more medial section corresponding to the visceral mesoderm. Similarly,for stage 12B, E/E' and F/F' present the cardiogenic mesoderm, I and J the visceral mesoderm. In all panels of left column, the second marker used besides anti-Dl is anti-Tinman (Tin;red) which labels the cardiogenic mesoderm. Embryos shown in the panels of the second column, as well as panels C and D of the third column, were labeled with anti-Phosphohistone3 (PH3), a marker for mitotic cells, as well as anti-Even-skipped (Eve; blue). In the third column, Seven-up (Svp)-positive cells are labeled by svp-Gal4NUAS-GFP(red). In these embryos, the lumen of the segmental tracheae and the demarcating segmental boundaries, are rendered in blue. Panels on the right (Q–T) show 3D digital models ofthe cardiogenic mesoderm at stages indicated at the top of each panel. Individual cells of cardiogenic mesoderm are shown as colored spheres, with colors representing differenttransient and definitive cell fates: cb cardioblasts (red in S–T); cmd definitive cardiogenic mesoderm (orange in Q; comprises anterior (a) and posterior (p) high-Dl cluster); cme

early cardiogenic mesoderm (light orange/green in Q); cbsvp Svp-positive cardioblasts (pink in T); Dl- low-Dl clusters (pale yellow in R); eve-pc Even skipped-positive pericardialcells (brown in R-T); lg blood progenitors of the lymph gland (magenta in T); pcn pericardial nephrocytes (magenta in T); pre-lg/pcn precursors of blood progenitors and pericardialnephrocytes (yellow in S); vme early visceral mesoderm (blue in Q); vml late visceral mesoderm (cyan and dark blue in R–T, corresponding to Tin-positive and Tin-negative latevisceral mesoderm cells, respectively). Other abbreviations: dt, dorsal tracheal branch; dlt, dorsal longitudinal trunk of the trachea; tp, tracheal pit; Bars: 25 μm (A–P); 10 μm (E', F',H'; K'–P').

110 M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

Fig. 3 (continued).

111M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

an increase in cardioblast number (Fig. 6B). Both blood progenitorsand pericardial nephrocytes were also increased in number. In thisexperiment, high levels of Dl expression were also seen in Mef2positive muscle cells, whereas inWT embryos no such expression wasapparent (data not shown).

Ras is a downstream activator of both the EGF and FGF receptortyrosine kinase pathways. In order to determine which of these wasacting in the cardiogenic mesoderm, we carried out gain and loss offunction studies for both the EGFR [Drosophila homolog: faint little ball(flb)] and FGFR [Drosophila homolog: heartless (htl)] pathways. Mef2-Gal4-driven expression of a constitutively active FGFR construct(Fig. 6C) or EGFR construct (not shown) mimicked the phenotype ofRas expression, in terms of a strongly increased number of blood

progenitors/nephrocytes and cardioblasts. EGFR loss of functionmutants showed a highly reduced level of Dl in the entire mesodermof stage 11/12 embryos (Fig. 6K). Labeling of late mutant embryoswith anti-Odd, anti-Pericardin or anti-Mef2 demonstrated that bloodprogenitors, nephrocytes and cardioblasts were virtually absent inEGFR mutants (Fig. 6E, H). Since in other tissues (Kurada and White,1998; Protzer et al., 2008), EGFR is an important factor for cellsurvival, we also carried out anti-caspase staining to monitor celldeath in the cardiogenic mesoderm. Significant amounts of cell deathwere seen in the cardiogenic mesoderm of stage 11/12 EGFR loss offunction mutants (Fig. 6F). These data support the idea that EGFRactivity is required for the maintenance of Dl expression and survivalof the cardiogenic mesoderm.

Fig. 4. Specification of different cell fates in the cardiogenic mesoderm. Schematic depiction of cardiogenic mesodermal cells (colored spheres) of two adjacent thoracic segments(left) and two abdominal segments (right) at stage 11 (top), mid stage 12 (middle) and stage 13 (bottom). Different cell types are annotated and colored in the same way as in themodels shown in Fig. 3.

112 M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

It had been previously established thatmutations in the FGFR gene,htl, result in a loss of the dorsal vessel and blood progenitors in lateembryos (Beiman et al., 1996; Gisselbrecht et al., 1996; Shishido et al.,1997; Mandal et al., 2004). In order to follow the fate of thecardiogenic mesoderm-derived lineages and Dl expression levels inembryos lacking FGFR, we did a developmental series of FGFRmutantslabeled with anti-Dl and markers for the different cardiogeniclineages. FGFR mutant embryos at stages 11 and 12 showed anabnormally shaped mesoderm; however, metameric clusters of Dl-positive cells were still visible in the cardiogenic mesoderm (Fig. 6L).By stage 14, these cells were no longer present (Fig. 6I), indicating thatFGFR signaling, like EGFR, is required for the maintenance ofcardiogenic lineages, but acts at a later phase than EGFR.

The bHLH gene l'sc is required for all cell types derived from thecardiogenic mesoderm

In order to identify a possible role of l'sc in the cardiogenicmesoderm, we visualized the cardiogenic lineages in embryoscarrying the deficiency Df(1)scB57, which removes the entireachaete–scute complex (AS-C) (Fig. 7). Homozygous mutants showeda highly significant decrease in the number of pericardial cells andlymph gland blood progenitors, as well as a small decrease in thenumber of cardioblasts. A similar reduction in blood progenitors/nephrocytes and cardioblasts was observed in embryos where a l'sc-Rnai was driven with aMef2-Gal4 driver. In neither the deficiency northe RNAi knock-down experiment did we see a significant reductionin Dl expression in the cardiogenic mesoderm (data not shown).These findings indicate that l'sc is required for the appropriate numberof cells within all cardiogenic lineages to form, even though it does notnoticeably alter expression levels of the Notch ligand, Dl.

Controlling the spatial domains of activity of Dl, l'sc, EGFR and FGFR

We next addressed the mechanisms operating upstream of Dl, l'scand EGFR/FGFR to set up the spatial domains in which these factors areactive. Previous data had shown that the cardiogenic mesoderm as awhole depends on the combined presence of high levels of Wg and

Dpp (Lockwood and Bodmer, 2002). Removal of either of these genesresults in the complete absence of all cardiogenic lineages (Mandalet al., 2004; Fig. 8A). Similarly, the early cardiogenicmesoderm, whichis strongly positive for Mef2 in wild type (Fig. 6J), does not becomespecified in a wg mutant (Fig. 8E). Both Dpp and Wg act upstream ofDl; loss of dpp (data not shown) orwg causes the failure to upregulateDl or L'sc expression in the mesoderm. Thus, the cardiogenic clustersC14 and C16 which are strongly Dl-positive and L'sc positive in thewild type (Fig. 8D) are not apparent in a wg LOF mutant (Fig. 8E). Weconfirmed that Wg, detected by an antibody and lacZ reporterconstruct, indeed coincides with the cardiogenicmesoderm. Through-out the early stages of mesodermal development (up to stage 10; 5 hAEL), Wg is expressed in a stripe-like domain covering the centralpart of the ectoderm and the underlying mesoderm of each segment.The early cardiogenic mesoderm forms the lateral part of the Wg-positive mesodermal domain (Fig. 8B). At the point when the visceralmesoderm separates from the cardiogenic mesoderm, the cells withinthe latter rearrange, spreading out in the longitudinal axis to coveralmost the entire length of a segment. During early stage 12, the timeof appearance of C14/C16, Wg signal can be detected in these clusters(Fig. 8C).

The expression of Dl, as well as activation of the MAPK pathway,becomes highly polarized in presumptive cardioblasts. This polariza-tion may be important for Dl to exert its proper function, that is, tomodulate Notch activation in the C14/C16 clusters such that theproper balance of cardioblasts vs. pericardial nephrocytes/bloodprogenitors is achieved. We wondered whether a spatially restrictedexpression of the ligands and/or receptors of the EGFR and FGFRpathways may help explain the localized activity of MAPK, which inturn maintains/polarizes Dl expression. The EGF receptor and FGFreceptor (Htl) are widely expressedwithin the cardiogenic mesoderm(Fig. 9A–A', C). Likewise, Spitz (Spi), one of the ligands capable ofactivating the EGFR signal is expressed throughout the cardiogenicmesoderm in stage 11–13 embryos (Fig. 9B). However, the FGFRligands, Thisbe (Ths) and Pyramus (Pyr) are expressed in clusters ofectodermal cells that appear in areas overlying the cardiogenicmesoderm (Fig. 9D–E'; Stathopoulos et al., 2004; Klingseisen et al.,2009; Kadam et al., 2009). This directional signal could be involved in

Fig. 5. Overlapping expression of Dl, L'sc and pMAPK in the cardiogenic mesoderm.A: Map of the pattern of L'sc/pMAPK-positive mesodermal clusters of one segment(adapted from Carmena et al., 1995). B–G: Three pairs of photographs (B, C; D, E; F, G)show a lateral view of the mesoderm of two consecutive trunk segments. Upper panelof each pair (B, D, F) represents early stage 11; lower panel (C, E, G) early to mid stage12 (stage 12B). B and C show in situ hybridizations, with a probe against l'sc. In D and E,a double labeling of anti-Dl (green) with anti-Phosphorylated MAP kinase (pMAPK,red) is shown. Specimens shown in F and G were labeled with anti-Dl (grey). Note thatthe dorsal clusters (C2 at stage 11; C14 and C16 at stage 12) that express l'sc (B, C) andpMAPK (D, E) correspond to the anterior (a) and posterior (p) high-Dl clusters in thecardiogenic mesoderm (F, G). A light blue circle represents the area where cluster C2 isfound (F). Dark blue circles represent areas where clusters C14 and C16 are located (G).The purple circle represents the area where cluster C15 is located (G). Abbreviations: dtdorsal tracheal branch; tp tracheal pit. Yellow bracket in D indicates spatial proximitybetween dorsal tracheal branch and posterior high-Dl cluster (see also Fig. 3). Largenumbers in (A) identify the L'sc-positive mesodermal clusters; small numbersunderneath indicate number of cells per cluster. Bar: 20 μm (B–G).

113M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

triggering signaling asymmetries in the underlying C14/C16 clusters,resulting in the eventual accumulation of higher MAPK activity(followed by Dl expression) in the dorsal-most cells of the clusterswhich will become cardioblasts.

Previous genetic studies had already shown that loss of pyrresulted in a reduction of the Eve-positive cells derived frommesodermal cluster C2 (Klingseisen et al., 2009; Kadam et al.,2009). We confirmed that the same mutation had significantlyreduced numbers of pericardial nephrocytes and blood progenitors,indicating that the Pyr-mediated signal plays a definitive role inactivating the MAPK pathway (Fig. 9E). By contrast, loss of ths, thesecond FGFR ligand, did not result in any abnormalities in thecardiogenic lineages (data not shown).

Discussion

In this study we pursued two goals: to elucidate the preciselocation and cellular composition of the cardiogenicmesoderm, and toanalyze the mechanism by which Notch becomes activated in therestricted subset of these cells that become blood progenitors. Ourfindings show that the cardiogenic mesoderm is comprised ofsegmentally reiterated pairs of clusters (cardiogenic clusters) definedby high expression levels of Dl, L'sc and activated MAPK. The MAPKpathway, activated through both EGFR and FGFR signaling, is requiredfor the specification (EGFR) and maintenance (EGFR and FGFR) of allcardiogenic lineages. As shown previously, the default fate of allcardiogenic cells is cardioblasts (Hartenstein et al., 1992; Mandalet al., 2004). Notch activity triggered by Dl is required for thespecification of blood progenitors (thoracic cardiogenic clusters) andnephrocytes (abdominal cardiogenic clusters), respectively. One ofthe downstream effects of MAPK signaling is tomaintain high levels ofDl in the cardiogenic clusters, and to help localize Dl expressiontoward a dorsal subset of cells within these clusters, which willbecome cardioblasts. Dl stimulates Notch activity in the surroundingcells, which triggers the blood progenitor/nephrocyte fate in thesecells.

Cardiogenic and myogenic clusters: variations on a common theme inDrosophila mesodermal development

The cardiogenic clusters form part of a larger population ofmesodermal cells defined by high expression levels of l'sc (Carmena etal., 1995). Based on l'sc in situ hybridization, these authors mapped 19clusters of l'sc expressing cells within the somatic mesoderm. Many ofthese clusters (called “myogenic clusters” in the following) give rise toone or two cells that transiently maintain high levels of l'sc, whereasthe remaining cells within the cluster lose expression of l'sc. The l'sc-positive cells segregate from the mesoderm to a more superficialposition, closer to the ectoderm, undergo one final mitotic division,and differentiate as muscle founder cells (Carmena et al., 1995). Dl/Notch mediated lateral inhibition was shown to act during thesingling-out of muscle founders from the myogenic clusters. Loss ofthis signaling pathway caused high levels of L'sc to persist in all cells ofthe myogenic clusters, with the result that all cells developed asmuscle founders. Interestingly, loss of l'sc had only a mild effect,consisting of a slight reduction in muscle founders. This is similar towhat we find in this paper in l'sc-deficient embryos, which show onlya mild reduction in cardioblasts and other cardiogenic lineages.

The developmental fate of most of the L'sc-positive clusters withinthe dorsal somatic mesoderm is different from that of the ventral andlateral myogenic clusters discussed above, even though severalparallels concerning the morphogenesis, proliferation, and depen-dence on Dl/Notch signaling are evident. The somatic (anterior, Wg-positive) mesoderm is divided into a dorsal and ventral domain basedon the expression of Tin. Initially expressed at high levels in the entiremesoderm, this gene is maintained only in the dorsal mesoderm, as aresult of Dpp signaling from the dorsal ectoderm. (Staehling-Hampton et al., 1994; Frasch, 1995). The dorsal somatic mesoderm,which we here call “early cardiogenic mesoderm,” includes four L'sc-positive clusters, C2 and C14–C16. The development of C2 has beendescribed in detail by Carmena et al. (1998; 2002) and Buff et al.(1998). C2 gives rise to a progenitor that divides twice; two of theprogeny become the Eve-positive pericardial cells. Meanwhile, C15which appears later at the same position as C2, seems to behave like a“normal”myogenic cluster. It produces a progenitor that divides onceand forms the founders of the dorsal muscle DA1. As shown in thispaper, the two remaining dorsal clusters, C14 and C16, give rise to thecardioblasts. We note that the Eve-positive progenitors, as well as thecardioblasts, resemble the muscle founders derived from the typicalmyogenic clusters in three aspects. First, they segregate toward a

Fig. 6. Cardiogenic phenotypes in EGFR and FGFR pathway mutants. A–C: dorsal view of stage 16 embryos. Pericardial nephrocytes (pcn) and blood progenitors in the lymph gland(lg) are labeled by anti-Odd (red) and anti-Pericardin (green); Pericardin labels odd positive pericardial nephrocytes as well as Seven-up (Svp) positive cardioblasts (cb). Sytox(blue) labels all nuclei. In embryo where mesodermally expressedMef2-Gal4 drives an activated Ras construct (B), all cardiogenic cells, and in particular cells of the lymph gland andpericardial nephrocytes, are strongly increased in number. The same phenotype results from driving a constitutively active FGF receptor (FGFR (htl); C). D, E: dorsal views of stage 16embryos. Anti-Tinman (Tin; red) labels cardioblasts and Tin-positive pericardial cells (tpc) in wild type embryos (D). Cardioblasts that appear negative for Tin are the Svp positivecardioblasts (D). In embryos carrying a null allele of the EGF receptor gene (EGFR; E), cardioblasts are absent. F: lateral view of stage 12 EGFR-mutant embryo. Labeling with anti-Dl(green) and anti-Caspase 3 (Casp), a marker for apoptotic cell death. Note clusters of Casp-positive cells in dorsal mesoderm (large arrowheads). Small arrowheads point outartefactual labeling of lumen of tracheae. G–I: lateral view of stage 13 embryos labeled with antibodies against Odd (red) and Dl (green). Note normal pattern of Odd-positivepericardial nephrocytes, lymph gland, and epidermal stripes (ep) in wild-type embryo (G). Nephrocytes and lymph gland cells are strongly reduced in loss of function mutants forboth EGFR (H) and FGFR (Htl; I). J–L': lateral views of stage 11 embryos labeled with anti-Dl (green), anti-Mef2 (red) and Sytox (blue). Upper row (J–L) shows sections of cardiogenicmesoderm, located more superficially in the embryo (right underneath the ectoderm). Lower row (J'–L') represent section of deeper (more medial) mesoderm of same embryos asshown in upper row, including (in wild type; J') the visceral mesoderm (vm), fat body mesoderm and somatic mesoderm (msfb/sm). Note strong expression of Dl and Mef2 in earlycardiogenic mesoderm (cme) of wild type embryo (J, J'). In an EGFR mutant (K, K'), the early cardiogenic mesoderm (based on Dl levels and number of Mef2-positive nuclei) isstrongly reduced or absent. In an FGFR (htl)mutant (L, L'), at this early stage, the cardiogenic mesoderm is still present, even though irregularities in the size and shape of cardiogenicmesoderm clusters are apparent. Bars: 20 μm (A–F); 25 μm (G–L').

Fig. 7. Reduction in cardiogenic cell lineages in embryos deficient for l'sc. A: dorsal viewof a stage 16 Df(1)scB57 (AS-C deficient) embryo labeled with anti-Odd skipped (Odd;red) and anti-Pericardin (green). Labeling of nuclei by Sytox (blue). Number of lymphgland blood progenitors (lg) and pericardial nephrocytes (pcn) is significantly reduced(see histogram in B); cardioblasts are mildly reduced. Arrowhead in A points at widegap in row of pericardial nephrocytes. B: For all of following, two samples were countedper organism. Lg (WT: n=8 lobes, mutant: n=4 lobes; p value: 4.85E-06); Pcn(WT: n=6, mutant: n=7; values shown are for 4 hemisegments; p value: 3.34E-05);Cb (WT: n=4, mutant: n=6; values shown are for 4 hemisegments; p value: 0.0197).p Values calculated via Student t-test (1 tailed, equal variance). Bar: 20 μm.

114 M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

superficial position, close to the ectoderm, relative to the remainder ofthe cells within the clusters. Secondly, they undergo one (in case ofC2: two) rounds of division right after segregation. And third, they arerestricted in number by Dl/Notch signaling: in all cases, they areincreased in number following Dl or Notch loss of function.

Cardiogenic clusters, like myogenic clusters, also depend on theMAPK signaling pathway. Past studies have shown that in the Eve-positive C2 and C15 clusters, Ras, is capable of inducing the formationof additional Eve-positive progenitors (Carmena et al., 2002). Ras is adownstream activator of both the EGFR and FGFR tyrosine kinasepathways, both of which have been seen to be important for theformation of the Eve-positive progenitors. With a loss of the FGFRpathway, Eve-positive progenitors of both the C2 and C15 cluster arelost; by contrast, the EGFR pathway affects only C15. The balancedactivity of MAPK and Notch, which in part depends on reciprocalinteractions between these pathways, determines the correct numberof C2/C15 derived progenitors. Ras-induced MAPK activation upre-gulates the expression of other MAPK signaling pathway members

Fig. 8. Expression and function of Wingless (Wg) in the cardiogenic mesoderm.A: Dorsal view of stage 16 wg mutant embryo, labeled with anti-Odd-skipped (red),anti-Pericardin (green), and Sytox (blue). All cardiogenic lineages are absent. B,C: lateral views of stage 11 (B) and early-mid stage 12 (C) wild-type embryos labeledwith anti-Dl (green), and expressing a wg-lacZ reporter construct (Wg; red). Wg isexpressed in segmentally reiterated stripes that overlap with the early cardiogenicmesoderm (cme in B). The wg reporter is expressed in all cells of the definitivecardiogenic mesoderm (cmd), which, by this stage, has lengthened in the ap-axis, toencompass the anterior and posterior high-Dl clusters (a, p). D, E: Lateral view of earlystage 12 wild-type (D) and wg mutant (E) embryo labeled with anti-Dl (green), anti-L'sc (red), and Sytox (blue). Note anterior and posterior high-Dl clusters (a, p) ofdefinitive cardiogenic mesoderm, which also contain L'sc-positive cells in wild type (D).Strongly Dl-positive cells ventrally adjacent to cardiogenic mesoderm are the emergingdorsolateral trunks and dorsal branches of the tracheae (dlt). In a wg mutant (E), thecardiogenic mesoderm is absent; the smooth epithelial layer dorsal of the Dl-positivetracheal cells (dlt) corresponds to the dorsal ectoderm (dec), which in the absence ofdorsal mesoderm, folds far medially. F, F': Lateral view of early stage 12 wg mutantembryo at a more superficial (F) and a deeper (more medial) level (F'). Labeling iswith anti-Dl (green) and anti-Mef2 (red). Loss of Dl/Mef2 in (F) confirms absenceof cardiogenic mesoderm. Other abbreviations: ep, epidermis; mg, midgut; tptracheal placodes; vec, ventral ectoderm; vm/fb, visceral mesoderm/fat body.Bars: 20 μm (A–C); 20 μm (D–F').

Fig. 9. EGFR and FGFR ligands in the cardiogenic mesoderm. A–C: Lateral view of earlystage 12 embryos labeled with anti-EGFR (red in A; white in A') and anti-Delta antibody(A; green), spitz-lacZ reporter (B; green), and FGFR (htl) reporter (C; green). All of theseare expressed widely in the cardiogenic mesoderm and beyond. Arrows in A and A'point at Dl+cardiogenic clusters that upregulate EGFR. D: lateral view of early stage 11embryo showing Pyramus (Pyr; blue) expression via in situ. Pyr is expressed insegmental ectodermal clusters (Kadam et al., 2009) overlying the cardiogenicmesoderm (white box outline). Figure adapted from Kadam et al., 2009. E, E': lateralview of early stage 12 embryo at superficial level of dorsal ectoderm (E) and levelmedially adjacent to ectoderm, showing cardiogenic mesoderm (E'). Labeling with lacZreporter of Thisbe (Ths; green), the other FGFR (Htl) ligand. Mesoderm cells are labeledby anti-Mef2 (red). Expression of Ths, similar to that of Pyr shown in D, occurs in cellsforming part of the dorsal-most row of the ectoderm, adjacent to underlyingcardiogenic mesoderm. F: Dorsal view of stage 16 embryo mutant for pyr, labeledwith anti-Odd skipped (red) and Sytox (blue). Note significant reduction of lymphgland cells (lg) and pericardial nephrocytes (pcn). Other abbreviations: cm, cardiogenicmesoderm; tp, tracheal placodes. Bars: 20 μm (A–C); 20 μm (D–E).

115M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

(autoregulatory feed-back loop), but also stimulates the antagonistArgos, as well as Dl. Dl-activated Notch, in turn, inhibits MAPKsignaling.

We show here that both Dl/Notch andMAPK signaling are active inthe C14 and C16 clusters, which constitute the definitive cardiogenicmesoderm. MAPK activity is required for the maintenance of alllineages derived from these clusters, as shown most clearly in theEGFR LOF phenotype that entails a lack of cardioblasts, bloodprogenitors, and pericardial nephrocytes. Overexpression of Rasresults in an increased number of all three cell types, which indicatesthat the C14/C16 clusters attain a larger size, possibly by an additionalround of mitosis. The phenotype seen in embryos suffering from loss-or overexpression ofDl/Notch pathwaymembers can be interpreted inthe framework of a classical lateral-inhibition mechanism: Dl isupregulated in the C14/C16 derived cardioblast progenitors (analo-gous to the Eve-progenitors of C2/C15), from where it activates Notchin the remainder of the C14/C16 cells; these cells are thereby inhibitedfrom forming cardioblasts, and instead become nephrocytes/bloodprogenitors. The level of Notch activity affects the expression of tin(low Notch) and the GATA homolog srp (high Notch), which triggersthe fate of cardioblasts and blood progenitors/nephrocytes, respec-tively (Mandal et al., 2004).

MAPK is required for the initial activation of Dl in the cardiogenicclusters (just as in the myogenic clusters). Input from the pathway ismost likely also instrumental in the subsequent restriction of Dl to thecardioblast progenitors. The positive interaction between MAPK and

Notch signaling could occur at several levels. A mechanism shown forthe ommatidial precursors of the eye disc involves Ebi and Strawberrynotch (Sno), which are thought to act downstream of EGFR signalingand lead to an upregulation of Dl through the Su(H) and SMRTERcomplex (Tsuda et al., 2002).

Generally, when one progenitor cell is seen to give rise to twodifferent cell types it is accomplished in one of twoways. One: there isan asymmetric division, where a factor expressed by the progenitor issegregated into only one daughter cell; two: a non-uniformlyexpressed extrinsic signal effects one cell, but not its neighboringsibling. Ward and Skeath (2000) had demonstrated that in theposterior (abdominal) segments of the Drosophila cardiogenicmesoderm, inhibition of Notch by Numb accounts for the asymmetricactivity of Notch in a small set of cardiogenic mesoderm cells, the Svp-positive cells. If Numb function is removed, these cells, whichnormally produce two cardioblasts and two pericardial nephrocytes,instead give rise to four cardioblasts. However, multiple nephrocytesper segment remain in numb loss-of-function mutations; further-more, loss of numb does not cause any defect in the blood progenitors,where asymmetrically dividing Svp-positive cells are absent. Thissuggests that in addition to the numb-mediated mechanism,directional activation of Notch by one of its ligands is required forthe majority of nephrocytes and all of the blood progenitors. Wepropose here that the spatially restricted upregulation/maintenanceof Dl in nascent cardioblasts acts to activate Notch in the remainder ofthe cells within each cardiogenic cluster, which promotes their fate asblood progenitors and nephrocytes.

116 M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

The Notch signaling/bHLH gene cassette in vertebrate blood/vasculardevelopment

The Notch signaling pathway is typically associated with membersof two different types of bHLH transcription factors. One type act asactivators, while the other act as repressors (Fisher and Caudy, 1998).In the context of lateral inhibition, best studied in Drosophilaneurogenesis (Heitzler and Simpson, 1991; Campos-Ortega, 1993),activating bHLH transcriptions factors, including genes of the AS-Clike l'sc, are expressed at an early stage in clusters of ectodermal ormesodermal cells, where they activate genetic programs that promotedifferentiative pathways such as neurogenesis, or myogenesis/cardiogenesis (see above). Subsequently, Notch ligands initiate theNotch pathway in these clusters; cells with high Notch activity turn onmembers of the Hairy/E(spl) (HES) family of bHLH genes which act asrepressors and abrogate the transcriptional programs that had beenset in motion by the activating bHLH factors (Heitzler et al., 1996). Asshown in this paper, the gene cassette consisting of the Notchsignaling pathway, as well as activating and repressing bHLH factors,operates in the cardiogenic mesoderm to determine the balancebetween cardioblasts and blood progenitors/nephrocytes. As dis-cussed in the following, the same cassette also appears to be centrallyinvolved in the specification of vascular endothelial cells andhematopoietic stem cells in vertebrates, which adds to the list ofprofound similarities between Drosophila and vertebrate blood/vascular development.

Even prior to the appearance of hemangioblasts, the lateralmesoderm of vertebrates is prepatterned by sequentially activatedsignaling pathways and transcriptional regulators similar to thosethat act in flies. The Wnt/Wg pathway, for example, separatessubdomains of the mesoderm in vertebrates and Drosophila, as wellas more ancestral ecdysozoans (Eriksson et al., 2009). Notch signalingplays an essential role in generating boundaries between segmental,as well as intra-segmental, subdomains within the ectoderm andmesoderm (McGrew and Pourquie, 1998; Qiu et al., 2009; Tapanes-Castillo and Baylies, 2004). The FGF signaling pathway predates theappearance of Bilaterians and plays a highly conserved role in earlymesoderm patterning (Vasiliauskas and Stern, 2001; Wilson andLeptin, 2000; Böttcher and Niehrs, 1995; Rebscher et al., 2009).Likewise, specific sets of transcriptional regulators are the targets ofthese signaling pathways (e.g., twist: Furlong, 2004; Vernon andLaBonne, 2004; Yeo et al., 2009); zfh: Su et al., 1999; van Grunsven etal., 2006; Kittelmann et al., 2009; myostatin: Lo and Frasch, 1999;Artaza et al., 2005; Grade et al., 2009) and play a role during theestablishments of cell fate in the mesoderm. It appears, therefore, thatthe bilaterian ancestor featured a mesodermal subdomain, the“cardiogenic/lateral mesoderm,” in which signals of the Wg, BMP,Notch, and FGF pathways and conserved sets of transcriptionalregulators established boundaries and cell fate in the mesoderm.

The vertebrate gene encoding an activating bHLH factor withsequence similarity to the Drosophila AS-C genes is SCL (Begley et al.,1989). SCL expression in the lateral mesoderm marks the firstappearance of hemangioblasts (Gering et al., 1998; Davidson andZon, 2000; note that SCL is also expressed widely in the developingvertebrate CNS; Van Eekelen et al., 2003). In Zebrafish, from their siteof origin in the lateral mesoderm, SCL-positive hemangioblastsmigrate dorso-medially and form the intermediate cell mass (ICM).The ICM is the site of primitive endothelial blood vessel andhematopoietic cell specification. Gain of function studies carried outin zebrafish embryos have shown SCL to be one of the genes importantin specifying the hemangioblast from the posterior lateral platemesoderm. The specification of hemangioblast here comes at theexpense of other mesodermal cell fates, namely the somitic paraxialmesoderm (Gering et al., 2003; Xiong, 2008; Gering and Patient, 2008;Dooley et al. 2005). In mice, lack of SCL affects blood and vasculardevelopment as SCL mutants are bloodless and show angiogenesis

defects in the yolk sac (Xiong 2008; Robb et al 1995; Shivdasani et al1995; Visvader et al 1998).

Vertebrate homologs of the repressive Drosophila Hairy/E(spl)family of bHLH genes are the Hes and Hey (hairy/Enhancer-of-splitrelated with YRPW motif) genes (Fischer and Gessler 2007; Iso et al.2003; Kokubo et al. 1999; Satow et al. 2001; Leimeister et al. 1999,2000; Steidl et al. 2000). A well studied member of the Hey family inzebrafish is gridlock, which is required for the specification ofhematopoietic progenitors from the ICM. Hey 2 mutations in micelead to severe congenital heart defects (Fischer et al. 2007). Inaddition, the Hes protein plays a role in hematopoiesis as it is apositive regulator of Hematopoietic Stem Cell (HSC) expansion(Kunisato et al. 2003).

Genetic studies of the Notch receptors and their ligands invertebrates support the idea that this pathway does indeed play acrucial role in the initial determination of hematopoietic stem cells(Bigas et al., 2010). The yolk sac and the para-aortic splanchnopleura(P-Sp)/AGM (aorta gonad mesonephros) of Notch null mouseembryos lack HSCs (Kumano et al 2003). A similar phenotype isobserved in mutants of Jagged 1, one of the Notch ligands (Robert-Moreno et al., 2005, 2008). Notch is thought to be the deciding factorbetween hematopoietic and endothelial cell fates when the twooriginate from a common precursor or hemangioblast. In murinemutants exhibiting lower Notch1 mRNA levels, a lack of hematopoi-etic precursors is seen and is accompanied by an increase in thenumber of cells expressing endothelial cell markers (Robert-Morenoet al. 2005). Likewise, in Drosophila, loss of Notch is associated with anincrease in cardioblast number and a loss of blood precursor cells(Mandal et al 2004).

Acknowledgments

This research was supported by the Ruth L. Kirschstein NationalResearch Service Award GM007185 and the HFSP Grant RGP0015/2008-C. This work was also supported by a grant from the NIH to bothUtpal Banerjee and Volker Hartenstein. Fig. 5A originates fromCarmena et al., 1995 (doi: 10.1101/gad.9.19.2373) and is used withpermission from both the author and Cold Spring Harbor LaboratoryPress. Fig. 9D originates from Kadam et al., 2009 (doi: 10.1242/dev.027904 ) and is used with permission from the author, TheCompany of Biologists and Development. We would also like to thankM.Muskavitch, B. Shilo, A. Carmena, H. Nguyen, J. Skeath, R. Bodmer, S.Artavanis-Tsakonas, E. Olson, T. Schupbach, A. Stathopoulos, M.Levine, the Bloomington Stock Center, the Developmental StudiesHybridoma Bank, the National Institute of Genetics and the BerkeleyDrosophila Genome Project for fly strains, antibodies and clones.We would also like to thank Utpal Banerjee for discussions andsuggestions regarding this project.

References

Abramoff, M.D., Magelhaes, P.J., Ram, S.J., 2004. Image processing with ImageJ.Biophoton. Int. 11, 36–42.

Al-Adhami, M. A., Kunz, Y. W., Ontogenesis of Haematopoietic Sites in BrachydanioRerio (Hamilto-Buchanan) (Teleostei)*.

Artaza, J.N., Bhasin, S., Magee, T.R., Reisz-Porszasz, S., Shen, R., Groome, N.P.,Meerasahib, M.F., Gonzalez-Cadavid, N.F., 2005. Myostatin inhibits myogenesisand promotes adipogenesis in C3H 10T(1/2) mesenchymal multipotent cells.Endocrinology 146, 3547–3557.

Azpiazu, N., Frasch, M., 1993. tinman and bagpipe: two homeo box genes thatdetermine cell fates in the dorsal mesoderm of Drosophila. Genes Dev. 7,1325–1340.

Baron, M.H., 2003. Embryonic origins of mammalian hematopoiesis. Exp. Hematol. 31,1160–1169.

Baron, M.H., 2005. Early patterning of the mouse embryo: implications forhematopoietic commitment and differentiation. Exp. Hematol. 33, 1015–1020.

Bate, M., 1993. The mesoderm and its derivatives. In: Bate, M., Martinez-Arias, A. (Eds.),The Development of Drosophila melanogaster. Cold Spring Harbor Laboratory Press,Cold Spring Harbor, New York, pp. 1013–1090.

117M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

Begley, C.G., Aplan, P.D., Denning, S.M., Haynes, B.F., Waldmann, T.A., Kirsch, I.R., 1989.The gene SCL is expressed during early hematopoiesis and encodes a differenti-ation-related DNA-binding motif. Proc. Natl Acad. Sci. USA 86, 10128–10132.

Beiman, M., Shilo, B.Z., Volk, T., 1996. Heartless, a Drosophila FGF receptor homolog, isessential for cell migration and establishment of several mesodermal lineages.Genes Dev. 10, 2993–3002.

Bertrand, J.Y., Traver, D., 2009. Hematopoietic cell development in the zebrafishembryo. Curr. Opin. Hematol. 16, 243–248.

Bigas, A., Robert-Moreno, A., Espinosa, L., The Notch pathway in the developinghematopoietic system. Int J Dev Biol. 54, 1175–88

Bodmer, R., 1993. The gene tinman is required for specification of the heart and visceralmuscles in Drosophila. Development 118, 719–729.

Bodmer, R., Golden, K., Lockwood, W.B., Ocorr, K.A., Park, M., Su, M.T., Venkatesh, T.V.,1997. Heart development in Drosophila. In: Wasserman, P. (Ed.), Advances inDevelopmental Biology, 5. JAI Press, Greenwich, CT, pp. 201–236.

Borkowski, O.M., Brown, N.H., Bate, M., 1995. Anterior-posterior subdivision and thediversification of the mesoderm in Drosophila. Development 121, 4183–4193.

Böttcher, R.T., Niehrs, C., 1995. Fibroblast growth factor signaling during earlyvertebrate development. Endocr. Rev. 26, 63–77.

Buff, E., Carmena, A., Gisselbrecht, S., Jimenez, F., Michelson, A.M., 1998. Signalling bythe Drosophila epidermal growth factor receptor is required for the specificationand diversification of embryonic muscle progenitors. Development 125,2075–2086.

Burns, C.E., Traver, D., Mayhall, E., Shepard, J.L., Zon, L.I., 2005. Hematopoietic stem cellfate is established by the Notch-Runx pathway. Genes Dev. 19, 2331–2342.

Campos-Ortega, J.A., 1993. Mechanisms of early neurogenesis in Drosophila melano-gaster. J. Neurobiol. 24, 1305–1327.

Campos-Ortega, J.A., Hartenstein, V., 1985. The Embryonic Development of Drosophilamelanogaster. Springer-Verlag, Berlin.

Cardona, A., Saalfeld, S., Preibisch, S., Schmid, B., Cheng, A., Pulokas, J., Tomancak, P.,Hartenstein, V., 2010. An integrated micro- and macroarchitectural analysis of theDrosophila brain by computer-assisted serial section electron microscopy. PLoSBiol. 8, 1–17.

Carmena, A., Bate, M., Jimenez, F., 1995. Lethal of scute, a proneural gene, participates inthe specification of muscle progenitors during Drosophila embryogenesis. GenesDev. 9, 2373–2383.

Carmena, A., Buff, E., Halfon, M.S., Gisselbrecht, S., Jimenez, F., Baylies, M.K., Michelson,A.M., 2002. Reciprocal regulatory interactions between the Notch and Ras signalingpathways in the Drosophila embryonic mesoderm. Dev. Biol. 244, 226–242.

Carmena, A., Gisselbrecht, S., Harrison, J., Jimenez, F., Michelson, A.M., 1998.Combinatorial signaling codes for the progressive determination of cell fates inthe Drosophila embryonic mesoderm. Genes Dev. 12, 3910–3922.

Cleaver, O., Tonissen, K.F., Saha, M.S., Krieg, P.A., 1997. Neovascularization of theXenopus embryo. Dev. Dyn. 210, 66–77.

Coffin, J.D., Poole, T.J., 1988. Embryonic vascular development: immunohistochemicalidentification of the origin and subsequent morphogenesis of the major vesselprimordia in quail embryos. Development 102, 735–748.

Crosier, P.S., Kalev-Zylinska, M.L., Hall, C.J., Flores, M.V., Horsfield, J.A., Crosier, K.E.,2002. Pathways in blood and vessel development revealed through zebrafishgenetics. Int. J. Dev. Biol. 46, 493–502.

Crossley, A.C., 1985. Nephrocytes and pericardial cells. In: Kerkut, G.A., Gilbert, L.I.(Eds.), Comprehensive Insect Physiology, Biochemistry, and Pharmacology, Vol. 3.Pergamon, Oxford.

Crozatier, M., Meister, M., 2007. Drosophila haematopoiesis. Cell. Microbiol. 9,1117–1126.

Davidson, A.J., Zon, L.I., 2000. Turning mesoderm into blood: the formation ofhematopoietic stem cells during embryogenesis. Curr. Top. Dev. Biol. 50, 45–60.

de Velasco, B., Mandal, L., Mkrtchyan, M., Hartenstein, V., 2006. Subdivision anddevelopmental fate of the head mesoderm in Drosophila melanogaster. Dev. GenesEvol. 216, 39–51.

Dooley, K.A., Davidson, A.J., Zon, L.I., 2005. Zebrafish scl functions independently inhematopoietic and endothelial development. Dev. Biol. 277, 522–536.

Eichmann, A., Corbel, C., Le Douarin, N.M., 1998. Segregation of the embryonic vascularand hemopoietic systems. Biochem. Cell Biol. 76, 939–946.

Eriksson, B.J., Tait, N.N., Budd, G.E., Akam, M., 2009. The involvement of engrailed andwingless during segmentation in the onychophoran Euperipatoides kanangrensis(Peripatopsidae: Onychophora) (Reid 1996). Dev. Genes Evol. 219, 249–264.

Evans, C.J., Banerjee, U., 2003. Transcriptional regulation of hematopoiesis in Drosophila.Blood Cells Mol. Dis. 30, 223–228.

Evans, C.J., Hartenstein, V., Banerjee, U., 2003. Thicker than blood: conservedmechanisms in Drosophila and vertebrate hematopoiesis. Dev. Cell 5, 673–690.

Fischer, A., Gessler, M., 2007. Delta-Notch—and then? Protein interactions andproposed modes of repression by Hes and Hey bHLH factors. Nucleic Acids Res.35, 4583–4596.

Fischer, A., Steidl, C., Wagner, T.U., Lang, E., Jakob, P.M., Friedl, P., Knobeloch, K.P.,Gessler, M., 2007. Combined loss of Hey1 and HeyL causes congenital heart defectsbecause of impaired epithelial to mesenchymal transition. Circ. Res. 100, 856–863.

Fisher, A., Caudy, M., 1998. The function of hairy-related bHLH repressor proteins in cellfate decisions. Bioessays. 20, 298–306.

Frasch, M., 1995. Induction of visceral and cardiac mesoderm by ectodermal Dpp in theearly Drosophila embryo. Nature 374, 464–467.

Furlong, E.E., 2004. Integrating transcriptional and signalling networks during muscledevelopment. Curr. Opin. Genet. Dev. 14, 343–350.

Gajewski, K., Choi, C.Y., Kim, Y., Schulz, R.A., 2000. Genetically distinct cardial cellswithin the Drosophila heart. Genesis 28, 36–43.

Gering, M., Patient, R., 2008. Notch in the niche. Cell Stem Cell 2, 293–294.

Gering, M., Rodaway, A.R., Gottgens, B., Patient, R.K., Green, A.R., 1998. The SCL genespecifieshaemangioblast development fromearlymesoderm. EMBO J. 17, 4029–4045.

Gering, M., Yamada, Y., Rabbitts, T.H., Patient, R.K., 2003. Lmo2 and Scl/Tal1 convertnon-axial mesoderm into haemangioblasts which differentiate into endothelialcells in the absence of Gata1. Development 130, 6187–6199.

Gisselbrecht, S., Skeath, J.B., Doe, C.Q., Michelson, A.M., 1996. heartless encodes afibroblast growth factor receptor (DFR1/DFGF-R2) involved in the directionalmigration of early mesodermal cells in the Drosophila embryo. Genes Dev. 10,3003–3017.

Go, M.J., Eastman, D.S., Artavanis-Tsakonas, S., 1998. Cell proliferation control by Notchsignaling in Drosophila development. Development 125, 2031–2040.

Grade, C.V., Salerno, M.S., Schubert, F.R., Dietrich, S., Alvares, L.E., 2009. Anevolutionarily conserved Myostatin proximal promoter/enhancer confers basallevels of transcription and spatial specificity in vivo. Dev. Genes Evol. 219, 497–508.

Hadland, B.K., Huppert, S.S., Kanungo, J., Xue, Y., Jiang, R., Gridley, T., Conlon, R.A., Cheng,A.M., Kopan, R., Longmore, G.D., 2004. A requirement for Notch1 distinguishes 2phases of definitive hematopoiesis during development. Blood 104, 3097–3105.

Hartenstein, A.Y., Rugendorff, A., Tepass, U., Hartenstein, V., 1992. The function of theneurogenic genes during epithelial development in the Drosophila embryo.Development 116, 1203–1220.

Hartenstein, V., 2006. Blood cells and blood cell development in the animal kingdom.Annu. Rev. Cell Dev. Biol. 22, 677–712.

Hartenstein, V., Mandal, L., 2006. The blood/vascular system in a phylogeneticperspective. BioEssays 28, 1203–1210.

Heitzler, P., Bourouis, M., Ruel, L., Carteret, C., Simpson, P., 1996. Genes of the Enhancerof split and achaete–scute complexes are required for a regulatory loop betweenNotch and Delta during lateral signalling in Drosophila. Development 122, 161–171.

Heitzler, P., Simpson, P., 1991. The choice of cell fate in the epidermis of Drosophila. Cell64, 1083–1092.

Holz, A., Bossinger, B., Strasser, T., Janning, W., Klapper, R., 2003. The two origins ofhemocytes in Drosophila. Development 130, 4955–4962.

Huppert, S.S., Jacobsen, T.L., Muskavitch, M.A., 1997. Feedback regulation is central toDelta-Notch signalling required for Drosophila wing vein morphogenesis. Devel-opment 124, 3283–3291.

Iso, T., Kedes, L., Hamamori, Y., 2003. HES and HERP families: multiple effectors of theNotch signaling pathway. J. Cell. Physiol. 194, 237–255.

Kadam, S., McMahon, A., Tzou, P., Stathopoulos, A., 2009. FGF ligands in Drosophila havedistinct activities required to support cell migration and differentiation. Develop-ment 136, 739–747.

Kittelmann, M., Schinko, J.B., Winkler, M., Bucher, G., Wimmer, E.A., Prpic, N.M., 2009.Insertional mutagenesis screening identifies the zinc finger homeodomain 2 (zfh2)gene as a novel factor required for embryonic leg development in Triboliumcastaneum. Dev. Genes Evol. 219, 399–407.

Klingseisen, A., Clark, I.B., Gryzik, T., Muller, H.A., 2009. Differential and overlappingfunctions of two closely related Drosophila FGF8-like growth factors in mesodermdevelopment. Development 136, 2393–2402.

Knirr, S., Frasch, M., 2001. Molecular integration of inductive and mesoderm-intrinsicinputs governs even-skipped enhancer activity in a subset of pericardial and dorsalmuscle progenitors. Dev. Biol. 238, 13–26.

Kokubo, H., Lun, Y., Johnson, R.L., 1999. Identification and expression of a novel family ofbHLH cDNAs related to Drosophila hairy and enhancer of split. Biochem. Biophys.Res. Commun. 260, 459–465.

Kumano, K., Chiba, S., Kunisato, A., Sata, M., Saito, T., Nakagami-Yamaguchi, E.,Yamaguchi, T., Masuda, S., Shimizu, K., Takahashi, T., Ogawa, S., Hamada, Y., Hirai,H., 2003. Notch1 but not Notch2 is essential for generating hematopoietic stemcells from endothelial cells. Immunity 18, 699–711.

Kunisato, A., Chiba, S., Nakagami-Yamaguchi, E., Kumano, K., Saito, T., Masuda, S.,Yamaguchi, T., Osawa, M., Kageyama, R., Nakauchi, H., Nishikawa, M., Hirai, H.,2003. HES-1 preserves purified hematopoietic stem cells ex vivo and accumulatesside population cells in vivo. Blood 101, 1777–1783.

Kurada, P., White, K., 1998. Ras promotes cell survival in Drosophila by downregulatinghid expression. Cell 95, 319–329.

Lanot, R., Zachary, D., Holder, F., Meister, M., 2001. Postembryonic hematopoiesis inDrosophila. Dev. Biol. 230, 243–257.

Leimeister, C., Externbrink, A., Klamt, B., Gessler, M., 1999. Hey genes: a novel subfamilyof hairy- and Enhancer of split related genes specifically expressed during mouseembryogenesis. Mech. Dev. 85, 173–177.

Leimeister, C., Dale, K., Fischer, A., Klamt, B., Hrabe de Angelis, M., Radtke, F., McGrew,M.J., Pourquie, O., Gessler, M., 2000. Oscillating expression of c-Hey2 in thepresomitic mesoderm suggests that the segmentation clock may use combinatorialsignaling through multiple interacting bHLH factors. Dev. Biol. 227, 91–103.

Lilly, B., Galewsky, S., Firulli, A.B., Schulz, R.A., Olson, E.N., 1994. D-MEF2: a MADS boxtranscription factor expressed in differentiatingmesoderm andmuscle cell lineagesduring Drosophila embryogenesis. Proc. Natl Acad. Sci. USA 91, 5662–5666.

Lo, P.C., Frasch, M., 1999. Sequence and expression of myoglianin, a novel Drosophilagene of the TGF-beta superfamily. Mech. Dev. 86, 171–175.

Lockwood, W.K., Bodmer, R., 2002. The patterns of wingless, decapentaplegic, andtinman position the Drosophila heart. Mech. Dev. 114, 13–26.

Maeno, M., 2003. Regulatory signals and tissue interactions in the early hematopoieticcell differentiation in Xenopus laevis embryo. Zoolog Sci. 20, 939–946.

Mandal, L., Banerjee, U., Hartenstein, V., 2004. Evidence for a fruit fly hemangioblast andsimilarities between lymph-gland hematopoiesis in fruit fly and mammal aorta-gonadal-mesonephros mesoderm. Nat. Genet. 36, 1019–1023.

Martin-Bermudo, M.D., Martinez, C., Rodriguez, A., Jimenez, F., 1991. Distribution andfunction of the lethal of scute gene product during early neurogenesis in Drosophila.Development 113, 445–454.

118 M. Grigorian et al. / Developmental Biology 353 (2011) 105–118

Martinez-Agosto, J.A., Mikkola, H.K., Hartenstein, V., Banerjee, U., 2007. The hemato-poietic stem cell and its niche: a comparative view. Genes Dev. 21, 3044–3060.

Medvinsky, A.L., Samoylina, N.L., Muller, A.M., Dzierzak, E.A., 1993. An early pre-liverintraembryonic source of CFU-S in the developing mouse. Nature 364, 64–67.

McGrew, M.J., Pourquie, O., 1998. Somitogenesis: segmenting a vertebrate. Curr. Opin.Genet. Dev. 8, 487–493.

Nguyen, H.T., Bodmer, R., Abmayr, S.M., McDermott, J.C., Spoerel, N.A., 1994. D-mef2: aDrosophila mesoderm-specific MADS box-containing gene with a biphasicexpression profile during embryogenesis. Proc. Natl Acad. Sci. USA 91, 7520–7524.

Protzer, C.E., Wech, I., Nagel, A.C., 2008. Hairless induces cell death by downregulationof EGFR signalling activity. J. Cell Sci. 121, 3167–3176.

Qiu, X., Lim, C.H., Ho, S.H., Lee, K.H., Jiang, Y.J., 2009. Temporal Notch activation throughNotch1a and Notch3 is required for maintaining zebrafish rhombomere bound-aries. Dev. Genes Evol. 219, 339–351.

Queenan, A.M., Ghabrial, A., Schupbach, T., 1997. Ectopic activation of torpedo/Egfr, aDrosophila receptor tyrosine kinase, dorsalizes both the eggshell and the embryo.Development 124, 3871–3880.

Ranganayakulu, G., Schulz, R.A., Olson, E.N., 1996. Wingless signaling induces nautilusexpression in the ventral mesoderm of the Drosophila embryo. Dev. Biol. 176,143–148.

Rebscher, N., Deichmann, C., Sudhop, S., Fritzenwanker, J.H., Green, S., Hassel, M., 2009.Conserved intron positions in FGFR genes reflect themodular structure of FGFR andreveal stepwise addition of domains to an already complex ancestral FGFR. Dev.Genes Evol. 219, 455–468.

Riechmann, V., Irion, U., Wilson, R., Grosskortenhaus, R., Leptin, M., 1997. Control of cellfates and segmentation in the Drosophila mesoderm. Development 124, 2915–2922.

Robb, L., Lyons, I., Li, R., Hartley, L., Kontgen, F., Harvey, R.P., Metcalf, D., Begley, C.G.,1995. Absence of yolk sac hematopoiesis from mice with a targeted disruption ofthe scl gene. Proc. Natl Acad. Sci. USA 92, 7075–7079.

Robert-Moreno, A., Espinosa, L., de la Pompa, J.L., Bigas, A., 2005. RBPjkappa-dependentNotch function regulates Gata2 and is essential for the formation of intra-embryonic hematopoietic cells. Development 132, 1117–1126.

Robert-Moreno, A., Guiu, J., Ruiz-Herguido, C., Lopez, M.E., Ingles-Esteve, J., Riera, L.,Tipping, A., Enver, T., Dzierzak, E., Gridley, T., Espinosa, L., Bigas, A., 2008. Impairedembryonic haematopoiesis yet normal arterial development in the absence of theNotch ligand Jagged1. EMBO J. 27, 1886–1895.

Rugendorff, A., Younossi-Hartenstein, A., Hartenstein, V., 1994. Embryonic origin anddifferentiation of the drosophila heart. Dev. Genes Evol. 203, 266–280.

Satow, T., Bae, S.K., Inoue, T., Inoue, C., Miyoshi, G., Tomita, K., Bessho, Y., Hashimoto, N.,Kageyama, R., 2001. The basic helix–loop–helix gene hesr2 promotes gliogenesis inmouse retina. J. Neurosci. 21, 1265–1273.

Shishido, E., Ono, N., Kojima, T., Saigo, K., 1997. Requirements of DFR1/Heartless, amesoderm-specificDrosophila FGF-receptor, for the formation of heart, visceral andsomatic muscles, and ensheathing of longitudinal axon tracts in CNS. Development124, 2119–2128.

Shivdasani, R.A., Mayer, E.L., Orkin, S.H., 1995. Absence of blood formation in micelacking the T-cell leukaemia oncoprotein tal-1/SCL. Nature 373, 432–434.

Staehling-Hampton, K., Hoffmann, F.M., Baylies, M.K., Rushton, E., Bate, M., 1994. dppinduces mesodermal gene expression in Drosophila. Nature 372, 783–786.

Stathopoulos, A., Tam, B., Ronshaugen, M., Frasch, M., Levine, M., 2004. pyramus andthisbe: FGF genes that pattern themesoderm of Drosophila embryos. Genes Dev. 18,687–699.

Steidl, C., Leimeister, C., Klamt, B., Maier, M., Nanda, I., Dixon, M., Clarke, R., Schmid, M.,Gessler, M., 2000. Characterization of the human andmouse HEY1, HEY2, and HEYLgenes: cloning, mapping, and mutation screening of a new bHLH gene family.Genomics 66, 195–203.

Su, M.T., Fujioka, M., Goto, T., Bodmer, R., 1999. The Drosophila homeobox genes zfh-1and even-skipped are required for cardiac-specific differentiation of a numb-dependent lineage decision. Development 126, 3241–3251.

Tapanes-Castillo, A., Baylies, M.K., 2004. Notch signaling patterns Drosophila meso-dermal segments by regulating the bHLH transcription factor twist. Development131, 2359–2372.

Taylor, M.V., Beatty, K.E., Hunter, H.K., Baylies, M.K., 1995. DrosophilaMEF2 is regulatedby twist and is expressed in both the primordia and differentiated cells of theembryonic somatic, visceral and heart musculature. Mech. Dev. 50, 29–41.

Tepass, U., Fessler, L.I., Aziz, A., Hartenstein, V., 1994. Embryonic origin of hemocytesand their relationship to cell death in Drosophila. Development 120, 1829–1837.

Tsuda, L., Nagaraj, R., Zipursky, S.L., Banerjee, U., 2002. An EGFR/Ebi/Sno pathwaypromotes delta expression by inactivating Su(H)/SMRTER repression duringinductive notch signaling. Cell 110, 625–637.

van Eekelen, J.A., Bradley, C.K., Gothert, J.R., Robb, L., Elefanty, A.G., Begley, C.G., Harvey,A.R., 2003. Expression pattern of the stem cell leukaemia gene in the CNS of theembryonic and adult mouse. Neuroscience 122, 421–436.

van Grunsven, L.A., Taelman, V., Michiels, C., Opdecamp, K., Huylebroeck, D., Bellefroid,E.J., 2006. deltaEF1 and SIP1 are differentially expressed and have overlappingactivities during Xenopus embryogenesis. Dev. Dyn. 235, 1491–1500.

Vasiliauskas, D., Stern, C.D., 2001. Patterning the embryonic axis: FGF signaling and howvertebrate embryos measure time. Cell 106, 133–136.

Venkatesh, T.V., Park, M., Ocorr, K., Nemaceck, J., Golden, K., Wemple, M., Bodmer, R.,2000. Cardiac enhancer activity of the homeobox gene tinman depends on CREBconsensus binding sites in Drosophila. Genesis 26, 55–66.

Vernon, A.E., LaBonne, C., 2004. Tumor metastasis: a new twist on epithelial–mesenchymal transitions. Curr. Biol. 14, R719–R721.

Visvader, J.E., Fujiwara, Y., Orkin, S.H., 1998. Unsuspected role for the T-cell leukemiaprotein SCL/tal-1 in vascular development. Genes Dev. 12, 473–479.

Ward, E.J., Skeath, J.B., 2000. Characterization of a novel subset of cardiac cells and theirprogenitors in the Drosophila embryo. Development 127, 4959–4969.

Wilson, R., Leptin, M., 2000. Fibroblast growth factor receptor-dependent morphogen-esis of the Drosophila mesoderm. Philos. Trans. R. Soc. Lond. B Biol. Sci. 355,891–895.

Xiong, J.W., 2008. Molecular and developmental biology of the hemangioblast. Dev.Dyn. 237, 1218–1231.

Yeo, G.H., Cheah, F.S., Winkler, C., Jabs, E.W., Venkatesh, B., Chong, S.S., 2009.Phylogenetic and evolutionary relationships and developmental expressionpatterns of the zebrafish twist gene family. Dev. Genes Evol. 219, 289–300.

Zaffran, S., Xu, X., Lo, P.C., Lee, H.H., Frasch, M., 2002. Cardiogenesis in the Drosophilamodel: control mechanisms during early induction and diversification of cardiacprogenitors. Cold Spring Harb. Symp. Quant. Biol. 67, 1–12.