The nieuwkoid/dharma Homeobox Gene Is Essential for bmp2b ... · Based on these results, we propose that gastrula organizer speciﬁcation requires the ... Further evidence supporting

i

wtphmdgzNtf

Developmental Biology 215, 190–207 (1999)Article ID dbio.1999.9479, available online at http://www.idealibrary.com on

The nieuwkoid/dharma Homeobox GeneIs Essential for bmp2b Repressionn the Zebrafish Pregastrula

David S. Koos and Robert K. HoDepartment of Molecular Biology, Princeton University, Princeton, New Jersey 08544

Dorsoventral specification of the zebrafish gastrula is governed by the functions of the dorsal shield, a region of the embryofunctionally analogous to the amphibian Spemann organizer. We report that the bozozok locus encodes the transcriptionfactor nieuwkoid/dharma, a homeobox gene with non-cell-autonomous organizer-inducing activity. The nieuwkoid/dharma gene is expressed prior to the onset of gastrulation in a restricted region of an extraembryonic tissue, the yolksyncytial layer, that directly underlies the presumptive organizer cells. A single base-pair substitution in the nieuwkoid/dharma gene results in a premature stop codon in bozm168 mutants, leading to the generation of a truncated protein product

hich lacks the homeodomain and fails to induce a functional organizer in misexpression assays. Embryos homozygous forhe bozm168 mutation exhibit impaired dorsal shield specification often leading to the loss of shield derivatives, such asrechordal plate in the anterior and notochord in the posterior, along the entire anteroposterior axis. Furthermore, bozomozygotes feature a loss of neural fates anterior to the midbrain/hindbrain boundary. Characterization of homozygousutant embryos using molecular markers indicates that the boz ventralized phenotype may be due, in part, to the

erepression of a secreted antagonizer of dorsal fates, zbmp2b, on the dorsal side of the embryo prior to the onset ofastrulation. Furthermore, ectopic expression of nieuwkoid/dharma RNA is sufficient to lead to the down regulation ofbmp2b expression in the pregastrula. Based on these results, we propose that gastrula organizer specification requires theieuwkoop center-like activity mediated by the nieuwkoid/dharma/bozozok homeobox gene and that this activity reveals

he role of a much earlier than previously suspected inhibition of ventral determinants prior to dorsal shieldormation. © 1999 Academic Press

Key Words: Nieuwkoop center; homeodomain; nieuwkoid/dharma/bozozok; BMP; zebrafish.

nino

INTRODUCTION

The proper specification and regionalization of the verte-brate dorsal embryonic axis rely on the essential functionsof a potent inductive signaling center functionally analo-gous to the amphibian Spemann organizer (for review, seeLamaire and Kodjabachain, 1996; Harland and Gerhart,1997). In the frog Xenopus, the process of dorsal specifica-tion begins shortly after fertilization with the establish-ment of regional asymmetries along the presumptivedorsal–ventral axis of the zygote. Within the first cell cycle,a dramatic cytoplasmic reorganization involving a corticalrotation leads to the localization or local activation ofmaternally derived “determinants” on the dorsal side of theembryo. Recent evidence has suggested that the maternally
deposited b-catenin protein, an effector of the WNT signal-ing pathway, may be sufficient to fulfill the role of a dorsal f
190

determinant (for review, see Moon and Kimmelman, 1998).The nuclear accumulation of b-catenin protein in cells onthe dorsal side of the embryo represents one of the earliestknown molecular asymmetries along the presumptivedorsal–ventral axis. The inheritance of dorsal determinantsby descendents of the dorsal blastomeres is thought toestablish a primary inductive signaling activity. This pri-mary inductive activity is predominantly, but not exclu-sively, restricted to the large yolky dorsal-vegetal cells ofthe blastula stage embryo, a region often referred to as theNieuwkoop center (Gerhart et al., 1989). Subsequently,

on-cell-autonomous activity of the Nieuwkoop centernfluences the establishment of a secondary inductive sig-aling center in the overlying mesoderm, the Spemannrganizer.
The Spemann organizer can induce and pattern neural
ates in the overlying ectoderm, dorsalize lateral mesoderm,

0012-1606/99 $30.00Copyright © 1999 by Academic Press

All rights of reproduction in any form reserved.

itw(cdsdmaagsamos

wecHdsaoctuieY

sBtYltmpaYa

prld

zada

tntmaaYldcp1ritnct

(

smaasmfla

191Zebrafish nieuwkoid/dharma Role in bmp2b Repression

and, when transplanted into the ventral side of the embryo,generate a secondary dorsal axis including neural plate(reviewed in Gerhart et al., 1989). The Spemann organizernductive activity has been proposed to result, in part, fromhe inhibition of instructive ventralizing signals associatedith members of the BMP family of secreted growth factors

reviewed in Graf, 1997). Accordingly, the organizer se-retes proteins, such as chordin and noggin, that canirectly bind to and antagonize BMP signals. In addition toignaling activities, cells of the organizer and their descen-ents contribute directly to tissues such as prechordal plateesoderm, notochord, neural tissue, and gut endoderm,

long the rostral to caudal extent of the dorsal axis (Harlandnd Gerhart, 1997). Although analogous Spemann-type or-anizer regions have been identified in all vertebrate speciestudied, such as the node in the mouse, Henson’s node inmniotes, and the dorsal shield in teleosts, the molecularechanisms underlying and mediating the establishment

f the organizer remain less well understood in vertebratepecies outside amphibians (Eyal-Giladi, 1997).Recent efforts have been applied toward determining tohat degree the amphibian organizer specification pathway

mploying the Nieuwkoop center is conserved in the pro-ess of dorsal axis specification in other vertebrate embryos.owever, differences in embryonic morphology and tissueistributions among different vertebrate embryos makepatial correlations less obvious (Eyal-Giladi, 1997; Tamnd Quinlan, 1996). For example, the Nieuwkoop center orrganizer-inducing activity is maximal in the dorsal-vegetalells in Xenopus; however, corresponding vegetal cells inhe zebrafish are not obviously present. In fact, directlynderlying the presumptive organizer cells in the zebrafishs a specialized region of the yolk cell which forms anxtraembryonic tissue called the yolk syncytial layer orSL (Kimmel and Law, 1985b).Historically, the YSL has been proposed to play an in-

tructive role in the specification of blastoderm cell fates.lastoderm transplantation experiments performed in therout and goldfish have indicated that a small region of theSL possesses organizer-inducing or Nieuwkoop center-

ike activities (Long, 1983; Yamaha et al., 1998). In addition,he animal region of yolk cell, including the YSL, exhibitsesoderm-inducing and patterning activity when trans-

lanted onto the animal region of a host embryo (Mizuno etl., 1996). Further evidence supporting a role of the teleostSL in the process of dorsal specification has come fromnalysis of b-catenin protein localization in the early em-

bryo. In the midblastula-stage zebrafish embryo, b-cateninrotein becomes localized to the nuclei in a restrictedegion of the dorsal YSL; subsequently the nuclear accumu-ation occurs in both the dorsal YSL and the overlyingorsal blastomeres (Schneider et al., 1996).Insight into the process of organizer specification in the

ebrafish has come from the independent isolation andnalysis of the zygotically expressed nieuwkoid and
harma homeobox genes (Koos and Ho, 1998; Yamanaka etl., 1998). Comparison of the sequence indicates that these
Copyright © 1999 by Academic Press. All right

wo genes encode the same paired-like homeobox gene.ieuwkoid/dharma expression is first detected shortly afterhe midblastula transition in the dorsal blastoderm cellargin, directly overlying the region of b-catenin nuclear

ccumulation in the YSL. At the onset of gastrulationpproximately 3 h later, expression is detected only in theSL directly underneath the dorsal shield, the fish equiva-

ent of the Spemann organizer. In addition, the nieuwkoid/harma paired-like homeobox gene exhibits non-ell-autonomous organizer-inducing activities upon misex-ression in blastoderm cells or in the YSL (Koos and Ho,998; Yamanaka et al., 1998). The dynamic and dorsallyestricted expression of nieuwkoid/dharma combined withts dorsalizing and organizer-inducing activity suggestedhat the nieuwkoid/dharma gene is an important compo-ent in the process of shield specification, possibly bothharacterizing and mediating a Nieuwkoop center-like ac-ivity.

Zebrafish embryos with a mutation in the bozozokbozm168, i2) gene fail to form a morphologically distinctshield at the onset of gastrulation, suggesting that thesemutant embryos have reduced or impaired organizer activ-ity (Solnica-Krezel et al., 1996; Blagden et al., 1997). Con-istent with diminished gastrula organizer activity, bozutant embryos show severe defects in the development of

xial mesendoderm along the entire rostral to caudal axisnd also exhibit severe deficiencies in forebrain and ventralpinal cord regionalization. The most severely affected bozutants display a complete loss of both notochord andoorplate in the trunk and tail combined with reduced orbsent eyes and forebrain (Solnica-Krezel et al., 1996;

Fekany et al., 1999).In this paper, we report that the bozozok mutation

represents a defect within the nieuwkoid/dharma ho-meobox gene. A single base-pair substitution in the bozm168

mutant allele of the nieuwkoid/dharma gene results in apremature stop codon which encodes a truncated proteinproduct lacking the putative DNA binding domain. Thistruncated protein product fails to induce a functional orga-nizer in misexpression assays. Morphological and markergene analyses demonstrate that homozygous mutant em-bryos exhibit impaired specification of dorsal mesendodermand neuroectodermal cell fates along the entire anteropos-terior axis. The mutant phenotype can be rescued in anonautonomous manner by misexpression of nieuwkoid/dharma directly in the YSL. Furthermore, the bozozok-associated dorsal deficiencies are accompanied by eitherderepressed or expanded expression of genes associatedwith the specification of ventral fates. In addition, ectopicexpression of nieuwkoid/dharma/bozozok RNA is suffi-cient to lead to a down regulation of the zygotic expressionof the ventralizing factor, zbmp2b, prior to the onset ofgastrulation. These results indicate that dorsal–ventral an-
tagonism may already occur at the level of the Nieuwkoopcenter before the process of organizer specification.
s of reproduction in any form reserved.

Bi

emtnTtolathtbcaptfida

192 Koos and Ho

MATERIALS AND METHODS

Isolation of the bozozok allele of the nieuwkoid/dharma gene.A genomic fragment containing the nieuwkoid/dharma gene bodywas isolated using PCR with Pfu high-fidelity polymerase (Strat-agene) templated by genomic DNA isolated from individual ho-mozygous bozozok embryos at 26 hpf, a stage at which the mutantphenotype is obvious. The primers used were based on thenieuwkoid/dharma cDNA sequence: 6.5.P2, 59-GCT AAT CTGATT CCT GAT GAT CC, and 6.5P5, 59-TTT TGG GCA CCC CTACAC. The PCR parameters were hot start, 94°C for 45 s, 54°C for45 s, 72°C for 2 min 30 s, 30 cycles. The PCR products correspond-ing to the nieuwkoid/dharma genomic fragments were modifiedusing the standard A-tailing protocol (Promega) and then ligatedinto pGemTeasy (Promega) and sequenced in both directions by thePrinceton Syn/Seq facility.

Constructs and site-directed mutagenesis. To create morestable transcripts, the entire coding sequence of the nieuwkoid/dharma gene was excised out of pAA1.2 (Koos and Ho, 1998) byBamHI/EcoRI double digestion and ligated into pCS21 (Turner andWeintraub, 1994) to create pDSK68.2. Site-directed mutagenesis toprepare pBOZM1 was performed on pDSK68.2 using the Quick-Change mutagenesis kit (Stratagene) employing the primers M1F,59-CCT TGT CCA TCA TGA CCA GGG ATG TAC TGC, andM1R, 59-GCA GTA CAT CCC TGG TCA TGA TGG ACA AGG.pBOZM2 was prepared using the primers M2F, 59-G GTT ACCGAC TAG CCG ACC CTC GAG ACC, and M2R, 59-GGT CTCGAC GGT CGG CTA GTC GGT AAC C. The changed nucleotideis underlined. To prepare pBOZM3, the amino-terminal half ofnwk/dharma was isolated from pDSK68.2 by double digestion with

amHI and SalI and this fragment containing the sequence encod-ng amino acids 1–85 was ligated into pCS21 previously prepared

by BamHI and XhoI double digestion.Genetic linkage analysis and genotyping. Adult heterozygote

bozm168 fish were generously provided by Dr. Lilianna Solnica-Krezel. Linkage was determined by mapping a BspHI restrictionsite polymorphism between bozm168 carriers and *AB wild-typereference strain. The segregation of this restriction fragment lengthpolymorphism (RFLP) was followed into phenotypically mutantand wild-type sibling pools of 50 F2 embryos derived from a crossbetween a bozm168 heterozygous carrier and the *AB line. The RFLPsegregated with bozm168, demonstrating that nieuwkoid/dharmaand boz are linked. The primer pair used for mapping was 6.5P5 and6.5.P2. PCR was performed with Pfu polymerase (Stratagene)templated by 0.25 embryo DNA equivalent using the cyclingparameters hot start, 94°C for 45 s, 54°C for 45 s, 72°C for 2 min30 s, 33 cycles. Genomic DNA was prepared from individualembryos fresh or post in situ or from adult fin clips as per “TheZebrafish Book” (Westerfield, 1994).

In situ hybridization and histology. Zebrafish embryos wereobtained from natural spawnings and staged as described by Kim-mel et al. (1995). Single-color in situ hybridizations were performedessentially as previously described in Thisse et al. (1993). Antisenseriboprobes for axial, bmp2b, chordin, a-coll2a1, emx1, flh, gsc,krox20, nieuwkoid (pAA1.2), shh, spt/tbx16, and twist were syn-thesized using T7, T3, or SP6 RNA polymerase (Promega) and Diglabeled rNTPs (Boehringer) as appropriate; probes were not hydro-lyzed prior to use. A partial cDNA encoding emx1 was fortuitouslyisolated during a survey of paired-like homeobox genes expressed in
the zebrafish blastula (D.S.K., unpublished) and was used to preparea riboprobe.
rd


Microinjection. Synthetic capped sense nieuwkoid/dharmaRNA was prepared using Ambion Messagemaker in vitro transcrip-tion kit with Sp6 RNA polymerase according to the manufacturer’sdirections. RNAs were resuspended in sterile water and diluted ina 5% solution of fluorescein-conjugated dextran (10K; MolecularProbes) in 0.2 M KCl.

RESULTS

bozozok Encodes the nieuwkoid/dharmaHomeodomain Protein

The zebrafish nieuwkoid/dharma paired-like homeoboxgene exhibits non-cell-autonomous organizer-inducing ac-tivities upon misexpression in blastoderm cells or in theextraembryonic tissue, the YSL (Koos and Ho, 1998; Ya-manaka et al., 1998). To test whether the nonautonomousorganizer-inducing activities of nieuwkoid/dharma indi-cated by the gain of function analyses represent an essentialcomponent in the process of gastrula organizer specifica-tion, we sought to determine if any of the recently isolatedmutants (Solnica-Krezel et al., 1996) which display pheno-types consistent with reduced or impaired organizer speci-fication harbor a mutation in the nieuwkoid/dharma gene.Using this candidate mutant approach we report that thenieuwkoid/dharma homeobox gene encodes the product ofthe bozozok locus.

PCR with nieuwkoid/dharma-specific primers was usedto amplify a genomic fragment encompassing thenieuwkoid/dharma gene body from genomic DNA derivedfrom individual wild-type and homozygous bozm168 mutantmbryos collected at a stage of development at which theutant phenotype is clearly visible. Sequence analysis of

he genomic isolates indicated that the bozm168 allele ofieuwkoid/dharma contains a single base-pair (TGG 3GA) substitution in the coding sequence which changes

he codon for a tryptophan residue at position 70 into anpal nonsense codon (Fig. 1a). The predicted mutant trans-ation product consists of the amino-terminal 69 aminocids of Nieuwkoid/Dharma. This severely truncated mu-ant gene product would include the Goosecoid Engrailedomology (GEH) (Koos and Ho, 1998) at the extreme amino-erminal but would completely lack the putative DNAinding domain (Fig. 1b). In addition to generating a stopodon, the single base-pair substitution within the mutantllele forms a BspHI restriction endonuclease site and thusrovides a polymorphism that is directly associated withhe mutation (Fig. 1c). Using this polymorphism, we con-rmed that only embryos homozygous for the bozm168 lesionisplay the mutant phenotype, thereby confirming that thisllele segregates as a recessive. Based on these results, we
eport that the bozozok locus encodes the nieuwkoid/harma homeobox gene.

oad

bot(

saB

nstt

cL

68 bu


The bozozokm168 Allele Lacks the Ability to Inducea Functional Organizer

Because of the variable severity of the ventralizedphenotype in boz mutant embryos, it was imperative todetermine if the bozm168 lesion resulted in a complete loss

f nieuwkoid/dharma activity. To address this, we ex-mined the dorsalizing activity of a synthetic nieuwkoid/harma RNA carrying the bozm168 lesion (BOZ.M1, W703

opal) in wild-type embryos. Previous studies (Koos andHo, 1998; Yamanaka et al., 1998) have demonstrated thatmisexpression of wild-type nieuwkoid/dharma RNA inearly blastomeres of cleavage stage embryos is sufficientto induce ectopic expression of the organizer-specificgene goosecoid (gsc) (Stachel et al., 1993). As shown inFig. 2b, injection of 20 pg of wild-type nieuwkoid/dharma RNA in both blastomeres of two-cell-stage em-bryos leads to the radially expanded expression of theorganizer gene gsc around the entire blastoderm circum-

FIG. 1. The bozozok locus encodes the Nieuwkoid/Dharma hieuwkoid/dharma from DNA isolated from homozygous wild tubstitution in the nieuwkoid/dharma ORF that changes W70 to anhe mutant base substitution. (b) Schematic depicting the conceptuhe severely truncated 69-aa product derived from the bozm168 allele

box denotes the GEH motif in the amino-terminal. (c) The molecupolymorphism in which the single base-pair substitution TGG 3nieuwkoid/dharma is linked to bozm168, as shown by segregationamplified from genomic DNAs from mapping crosses. A 1.1-kbleaved by BspHI digestion in wild type. The bozm168 mutant versionane 1, homozygous wild-type G0 fish (*AB strain); lane 2, heter

wild-type sibs; lane 5, F2 phenotypically wild-type heterozygousdigested fragment generated by the BspHI site present in the bozm1

RFLP.

ference at the onset of gastrulation (30/30). In contrast,injection of the same concentration of BOZ.M1 RNA was

wo


only weakly able to induce ectopic gsc expression (5/30)and the gsc expression did not extend around the entire

lastoderm circumference (Fig. 2c). Misexpression of an-ther recombinant nieuwkoid/dharma RNA engineeredo contain an amber stop codon at the 25th residueY139 3 amber) of the homeodomain (BOZ.M2) also was

able to weakly induce ectopic gsc expression at the onsetof gastrulation (6/34) (data not shown). The residualgsc-inducing activity of these mutagenized nieuwkoid/dharma RNAs may indicate that the amino-terminal halfof nieuwkoid/dharma possesses dorsalizing activity or itmay represent the activity of a full-length product pro-duced by a low level read through the stop codon. In orderto discern between these two possibilities, we generateda truncated nieuwkoid/dharma construct that containsequence encoding only the amino-terminal 85 aminocids of Nieuwkoid/Dharma (BOZ.M3). Misexpression ofOZ.M3 in both blastomeres of two-cell-stage embryos

odomain protein. (a) Comparison of the genomic sequence of*AB strain) and homozygous bozm168 indicates a single base-pairl stop codon. Asterisk denotes the stop codon. Arrowhead denotestranslated 192-amino-acid product derived from the AB allele andgray-shaded box denotes the homeodomain and the black-shaded

esion in the bozm168 allele generates a restriction fragment lengthdestroys an EaeI recognition site and generates a BspHI site. (d)

spHI restriction polymorphism in nieuwkoid/dharma sequences-amplified product from the nieuwkoid/dharma sequence is notleaved by BspHI digestion, reducing the fragment to 0.8 kb (arrow).ous bozm168 G0 fish; lane 3, size marker; lane 4, F2 homozygouslane 6, F2 mutant sibs (homozygous bozm168). Arrow denotes thet not in the *AB strain. The bozm168 mutation segregates with this

omeype (

opaally. Thelar lTGA

of a BPCR

is cozygsibs;

as unable to induce ectopic expression of gsc at thenset of gastrulation (0/18) (Fig. 2d). These results dem-


fa

rnb

isbrpmpheTqdF

ocrodtdeFi

ane 2trol.

194 Koos and Ho

onstrate that the amino-terminal half of nieuwkoid/dharma is insufficient to mediate the induction of anorganizer and gsc expression. The possibility exists thatlow level readthrough of the bozm168 lesion may accountor the incomplete penetrance of the mutant phenotypemong homozygous mutant embryos.

Autoregulation of bozozok Prior to the Onset ofGastrulation

The identification of the defect in the bozozok mutantallowed us to examine if nieuwkoid/dharma function isequired for its own expression. In wild-type embryos,

FIG. 2. The bozozokm168 allele lacks the ability to induce a functiof gastrulation (6 hpf). Dorsal is to the right. Embryos were injectedapped, sense RNA (20 pg per blastomere). (a) gsc expression in unegion of the dorsal margin in the cells of the forming shield. (b) Radf wild-type nieuwkoid/dharma RNA into both blastomeres at thharma containing the bozm168 lesion (W70 to opal), can only weawo-cell stage. Arrowhead highlights ectopic weak expression of gharma that contains coding sequence encoding only the N-termxpression when injected as in (b) and (c).IG. 3. Autoregulation of bozozok prior to the onset of gastrulation homozygous bozm168 embryos (b) as detected by in situ hybridiza

and (b) using PCR and digestion with BspHI. Lane 1, size marker; lhomozygous bozm168 embryo shown in (b); lane 4, no-template con

ieuwkoid/dharma is first expressed shortly after the mid-lastula transition (MBT) (3.3 hpf) in a small group of cells

rt


n the dorsal blastoderm margin. Subsequently, at spheretage (4 hpf), nieuwkoid/dharma expression can be detectedoth in dorsal blastoderm cells and within a restrictedegion of the YSL directly underlying the blastoderm ex-ression (Fig. 3a). At 3.3 hpf only a small fraction of bozutant embryos exhibited reduced nieuwkoid/dharma ex-

ression (data not shown). However, by sphere stage at 4pf, nieuwkoid/dharma expression in most of the mutantmbryos was strongly reduced or absent (16/17) (Fig. 3b).herefore, Nieuwkoid/Dharma activity appears to be re-uired to maintain the expression of the nieuwkoid/harma gene in the dorsal blastoderm margin and YSL.urthermore these results indicate that the amino-terminal

ganizer. All embryos are animal pole facing the viewer at the onsetoth blastomeres at the two-cell stage with the indicated synthetic,ted wild-type embryo (*AB). gsc expression is confined to a smallpression of gsc in the entire margin of embryos following injection-cell stage. (c) BOZ.M1 RNA, an engineered allele of nieuwkoid/

nduce gsc expression when injected into both blastomeres at the) Injection of BOZ.M3 RNA, a recombinant allele of nieuwkoid/5 amino acids of Nieuwkoid/Dharma, fails to induce ectopic gsc

euwkoid/dharma RNA is present in wild-type embryos (a) but notat sphere stage (4 hpf). (c) Post-in situ genotyping of embryos in (a), homozygous wild-type embryo (*AB strain) shown in (a); lane 3,

nal orin b

injecial exe twokly isc. (d

inal 8

n. nition

egion of nieuwkoid/dharma is insufficient for autoregula-ion of its transcription.


mmca

atbccoMgtd

gten

wf1i5gen

a1


Morphological Characterization of the Ventralizedboz Mutant Phenotype

Embryos homozygous for the bozm168 mutation displayorphologies and gene expression, such as loss of dorsalesendoderm and impaired neuroectodermal patterning,

onsistent with a reduced or impaired gastrula organizerctivity. boz mutant embryos cannot be distinguished mor-

phologically until just before the onset of gastrulation. Inthe wild-type embryo, dorsal convergence movements oc-curring in the pregastrula lead to the formation of a con-densation of cells on the dorsal margin that is collectivelyreferred to as the dorsal shield (Warga and Kimmel, 1990),the teleost equivalent of the amphibian Spemann organizer(Fig. 4a). In contrast, boz mutant embryos form a reducedcondensation of cells on the dorsal margin (Fig. 4b), suggest-ing that the process of shield specification is impaired in theboz mutant embryo and may be manifested as a reductionof dorsal convergence movements.

At later stages of development, boz mutant embryosdisplay a severe loss of dorsal mesodermal and ventralneural ectodermal fates. During early somitogenesis in thenormal embryo, the notochord and floorplate occupy thedorsal midline separating bilateral blocks of paraxial somitetissue (Figs. 4c and 4e). In severely affected boz mutantembryos, the notochord and floorplate are absent in thetrunk and the lateral blocks of somites are fused across themidline (Figs. 4d and 4f). Morphological analysis at 24–26hpf revealed that, in addition to the lack of dorsal midlinetissues in the trunk and tail, boz mutant embryos exhibitdefects in rostral patterning of the head often resulting intruncations of the forebrain region and reduction or absenceof eyes (Figs. 4h and 4i). Within a single clutch the expres-sivity of the mutant phenotype varies along the extent ofthe rostral/caudal axis. Mutant phenotypes with defects inthe trunk/tail are more frequent than mutant phenotypeswith trunk/tail and head, possibly indicating differences inthe requirement for nieuwkoid/dharma function in pat-terning the head versus patterning trunk and tail.

Impaired Dorsal Specification in boz MutantGastrulae

To investigate the molecular basis of the morphologicaldefects observed in boz mutant embryos, we first examined

series of genes normally expressed in the shield and alonghe dorsal midline during gastrulation. In the zebrafish latelastula and early gastrula, the expression domain ofhordino (chd), the zebrafish orthologue of the Xenopushordin gene, is thought to correlate with the field ofrganizer activity (Miller-Bertoglio et al., 1997; Schulte-erker et al., 1997). In wild-type embryos at the onset of

astrulation, chd is expressed in a dorsally restricted regionhat encompasses approximately one-quarter of the blasto-erm circumference (Fig. 5a). As gastrulation proceeds, chd

expression is detected in the nascent axis as well as in
bilateral domains flanking the dorsal shield (Fig. 5c). Incontrast, chd expression in boz mutants at the onset of

astrulation is greatly reduced (Fig. 5b). Furthermore,hroughout gastrulation, chd expression in boz mutantmbryos is not expressed in the dorsal mesendoderm in theascent axis (Fig. 5d).In the gastrula embryo, the expression of gsc is associatedith the specification of the anterior mesendodermal cell

ates that will contribute to the prechordal plate (Stachel,993). In the wild-type midgastrula embryo, gsc expressions confined to the anterior end of the developing axis (Fig.e). In severely affected boz mutant embryos, the anteriorsc expression domain is dramatically lost (Fig. 5f). Thexpressions of flh (znot), twist, axial (zebrafish HNF3b), andtl (zebrafish brachyury) in the dorsal midline are associ-

ated with the specification of dorsal mesoendodermal fatessuch as notochord (Talbot et al., 1995; Halpern et al., 1995;Strahle et al., 1995; Schulte-Merker et al., 1994; respec-tively). While these genes are strongly expressed in thedorsal midline tissue of wild-type embryos (Figs. 5g, 5i, and5k) the expression of all four of these genes is reduced orabsent in boz mutant embryos (Figs. 5h, 5j, and 5l). Two ofthese genes are also expressed during gastrulation in non-dorsal midline tissues, such as axial (Strahle et al., 1995) inthe presumptive paraxial endoderm and ntl (Schulte-Merker et al., 1994) in the ventral lateral mesoderm (Figs. 5iand 5k). In contrast to the expression of these genes indorsal midline tissue, the nondorsal midline expression ofthese genes appears unaffected in boz mutant embryos.These results indicate that nieuwkoid/dharma function isrequired specifically for the specification of dorsal midlinecell fates and not ventral or lateral cell fates.

Expanded Ventral Gene Expression in boz MutantGastrulae

The loss of dorsal mesendoderm gene expression in bozmutant gastrula embryos is accompanied by the expandedexpression of genes associated with the specification ofventral fates such as eve1, a zebrafish orthologue of theDrosophila even skipped homeobox gene (Joly et al., 1993),nd the spadetail T box gene (spt/tbx16) (Griffin et al.,998; Ruvinsky et al., 1998). In midgastrula wild-type

embryos, eve1 expression is excluded from dorsal tissue andis expressed only within the ventral posterior epiblast in adomain that encompasses the ventral half of the blastodermcircumference (Fig. 5o) (Joly et al., 1993). Consistent withreduced gastrula organizer activity, eve1 expression in bozmutant embryos is expanded into the dorsal region (Fig. 5p).Interestingly, completely radial eve1 expression was neverobserved in boz mutant embryos (n 5 20).

The specification of paraxial fates, such as trunk somitictissue, requires spt/tbx16 function (Griffin et al., 1998).spt/tbx16 is expressed in wild-type embryos in the ventral–lateral regions of the gastrula and excluded from the poste-rior dorsal midline tissues (e.g., notochord) (Fig. 5q) (Griffinet al., 1998; Ruvinsky et al., 1998). In boz mutant embryos,
spt/tbx16 expression invades the dorsal region often leavingno gap in the expression domain at the dorsal margin (Fig.

oh

taPa

196 Koos and Ho

FIG. 4. Morphological characterization of the ventralized boz mutant phenotype. (a) Animal pole view of wild-type embryo at the onsetf gastrulation (6 hpf). A condensation of cells (arrowhead) denotes the forming shield on the dorsal side of the embryo. (b) Severely affectedomozygous bozm168 mutant embryos fail to form a shield. (c and d) Dorsal views of wild-type and homozygous bozm168 mutant embryos,

respectively, at 12 hpf. Anterior is toward the top. In wild-type embryos the notochord (arrowhead) is present in the midline separating thelateral blocks of somites (arrow) (c). In homozygous bozm168 mutant embryos (d), the notochord is absent, and the somites are fused acrosshe dorsal midline. (e and f) Optical section of the embryos in (c) and (d), respectively, at the level of the first somite. View is from thenterior. The notochord (arrowhead) is present in wild-type embryos (e) and absent in homozygous bozm168 mutant embryos (f). (g and h)
rofiles of wild-type and homozygous bozm168 mutant embryos, respectively, at 26 hpf. Anterior is to the left. Dorsal midline structures suchs notochord and floorplate (arrow) are present in wild-type embryos (g), but lacking in homozygous bozm168 mutant embryos (h). (i and
Copyright © 1999 by Academic Press. All rights of reproduction in any form reserved.

batmhe

bae

mTsgTiEb1cMetidccpr

e

dcbdbhTptah

epfl


5r). spt/tbx16 is also expressed in the anterior of thezebrafish embryo in a domain that is reported to encompasspart of the prechordal plate during gastrulation (Fig. 5q)(Griffin et al., 1998; Ruvinsky et al., 1998). This anteriorexpression of spt/tbx16 is also reduced or absent in severelyaffected boz mutant embryos (Fig. 5r).

Impaired Anterior Neural Ectoderm Patterning inboz Mutants

Following the end of gastrulation, boz mutant embryosdisplay defects in the ectoderm as well as the mesendodermalong the entire anteroposterior axis. In 14-hpf wild-typeembryos, the emx1 homeobox gene is regionally expressedwithin the prosencephalon, specifically in the anlage oftelencephalon (Morita et al., 1995) (Fig. 6a). In boz mutantembryos, emx1 expression is reduced or absent (Fig. 6b). Inorder to further analyze anterior neural development in bozembryos, 12-hpf-stage wild-type and boz mutant embryoswere doubly stained to reveal the expression of the floor-plate and notochord marker sonic hedgehog (shh) and thehindbrain marker krox20. At this stage in wild-type em-ryos shh expression is detected in the anterior floorplatend krox20 is expressed in the hindbrain region in presump-ive rhombomeres 3 and 5 (Fig. 6c). In severely affected bozutant embryos, the expression of shh is reduced or absent;

owever, the development of the hindbrain indicated by thexpression of krox20 appears unaffected.Monitoring the expression of paxb (Krauss et al., 1993) at

22 hpf revealed further patterning defects in the rostralneuroectoderm of boz mutants. paxb expression in therostral neural ectoderm of wild-type embryos is restrictedto the optic stalks and the midbrain–hindbrain region (Fig.6d). As shown in Figs. 6e and 6f, boz mutants with headdefects can be grouped into a graded series of anteriortruncations of paxb expression with the most severe casesexhibiting the loss of expression rostral to and including themidbrain–hindbrain boundary. Midline development wasfurther characterized by the analysis of the expression ofa-coll2a1 (Yan et al., 1995) at 22 hpf. At this stage,a-coll2a1 is expressed in the dorsal midline in the floor-plate, notochord, hypochord, and axial endoderm of wild-type embryos (Fig. 6g). Severely affected boz mutant em-ryos continue to exhibit defective floorplate specificationnd a severe loss of notochord and hypochord along thentire anteroposterior axis (Fig. 6h).

Rescue of the boz Mutant Phenotype by YSLExpression of nieuwkoid/dharma RNA

In wild-type embryos shortly after the midblastula tran-sition, the nieuwkoid/dharma gene is expressed in a few

j) Profile view of the head region of wild-type and homozygous bo
mutant embryos often lack differentiated eyes (arrowhead) and display amidbrain region (arrow).

arginal blastoderm cells on the dorsal side of the embryo.hen as development proceeds, this expression shifts to amall region of the dorsal YSL which, by the onset ofastrulation, directly underlies the developing organizer.he functional relevance of nieuwkoid/dharma expression

n blastoderm cells relative to YSL expression is unclear.arly cleavage-stage blastomeres are cytoplasmicallyridged to the underlying yolk cell (Kimmel and Law,985a). The YSL itself is derived from the blastomeres thatontinue to maintain cytoplasmic bridges to the yolk cell atBT (Kimmel and Law, 1985b). Thus, RNA injected into

arly blastomeres can be distributed both within the blas-oderm cells and to the YSL. However, an injection directlynto the YSL once it has become cytoplasmically distinct isistributed only within the YSL and to a few specializedells that remain associated to the YSL called forerunnerells (Cooper and D’Amico, 1996). We exploited theseotential distribution differences to gain insight into theegional requirements of nieuwkoid/dharma function in

boz mutant embryos, by comparing the rescue potential ofwild-type nieuwkoid/dharma RNA ectopically expressedither in early blastomeres and YSL or in the YSL only.To test for dorsalizing activity, synthetic nieuwkoid/

harma RNA was injected into the early blastomeres ofleavage-stage embryos obtained from an intercross ofozm168 heterozygotes. The misexpression of nieuwkoid/harma RNA in two distant blastomeres of eight-cell-stageoz mutant embryos and their siblings always generated ayperdorsalized phenotype later in development (Table 1).hese hyperdorsalized embryos contain secondary or ex-anded dorsal axes, as evidenced by ectopic expression ofhe axial-mesoderm and floorplate marker shh (Krauss etl., 1993) and the radially expanded expression of theindbrain marker krox20 (Oxtoby and Jowett, 1993) (Figs.

7b and 7c) at 12 hpf. These results indicate that theexpression of nieuwkoid/dharma in both blastoderm cells,as well as in the YSL, is sufficient to rescue the bozventralized phenotype and lead to the formation of dorsalfates.

Next, we sought to determine if YSL expression ofnieuwkoid/dharma alone was sufficient to rescue the de-velopment of dorsal midline tissues in the trunk and tail ofboz mutant embryos. Using embryos obtained from anintercross of bozm168 heterozygotes, synthetic nieuwkoid/dharma RNA was injected directly into one region of theYSL shortly after YSL formation (3.5 hpf). The injectedembryos were allowed to develop until 22 hpf and stainedfor a-coll2a1 (Yan et al., 1995) expression. As describedarlier, at 22 hpf in wild-type embryos, a-coll2a1 is ex-ressed in dorsal midline tissues such as notochord andoorplate in the trunk and tail (Fig. 7d). As summarized in

mutant embryos, respectively, at 26 hpf. (j) Homozygous bozm168
zm168
loss of rostral forebrain neural fates and aberrant patterning of the


h

w

mgt

198 Koos and Ho

FIG. 5. Impaired dorsal specification boz mutant gastrulae. Analysis marker gene expression in wild-type (a, c, e, g, i, k, m, o, q) andomozygous bozm168 mutant (b, d, f, h, j, l, n, p, r) embryos during gastrulation. Embryos in (a) and (b) are staged at the onset of gastrulation

(6 hpf); embryos in (c–n) are midgastrula stage at 80% epiboly (8 hpf). All embryos are oriented with animal pole toward the top. (a), (b) and(g–n) are dorsal views, (c–f) profile with dorsal to the right. (a) chd expression at the onset of gastrulation (6 hpf) on the dorsal side of

ild-type embryo. (b) In homozygous bozm168 mutant embryos chd expression is reduced at the onset of gastrulation. (c) In wild-typeembryos at 80% epiboly, chd expression is detected in the mesendoderm in the nascent axis (arrowhead) as well as the lateral ectoderm (d).In homozygous bozm168 mutant embryos, chd expression is reduced or absent in the nascent axis. (e) gsc expression in the wild-type

idgastrula is restricted to the cells of the anteriorly migrating prechordal plate (arrowhead). (f) In homozygous bozm168 mutant embryos,
sc expression is reduced or absent. (g, h) In wild-type embryos the expression of flh (g) and twist (i) is detected in dorsal mesendoderm cellshat will later give rise to dorsal embryonic structures such as the notochord. In homozygous bozm168 mutant embryos, the dorsal midline

c(oo

ncwfiep(

dc

t

ec


Table 1 and Fig. 7e, injection of increasing amountsnieuwkoid/dharma RNA reduces the penetrance of the bozmutant phenotype, rescuing notochord and floorplate de-velopment in the trunk and tail. To confirm that theinjected RNA was not distributed into the cells of therescued dorsal embryonic organs, the distribution of lineagetracer that was coinjected with the RNA was analyzed. Asshown in Fig. 7f, no fluorescently labeled cells were de-tected in the dorsal midline tissues of rescued boz mutantsand their siblings. However, a small cluster of labeled cellswas detected in the posterior region of the developing tail(Fig. 7f) and they are most likely descendents of the special-ized forerunner cells that remain associated with the YSLfor extended periods of time (Cooper and D’Amico, 1996).Although none of the YSL-injected embryos (boz mutantsor their siblings) exhibited secondary axes or the hyperdor-salization phenotypes observed after YSL injection, theinjection of 100 pg of nieuwkoid/dharma RNA into the YSLproduced some embryos (3/23 boz homozygotes and 8/70sibling embryos) that exhibited an everted tube-like tailcontaining radial somites by 14 hpf. When monitored fora-coll2a1 expression at 22 hpf, these tube-tail embryoscontain notochord and floorplate in a curly tail (data notshown).

zbmp2b Is Ectopically Expressed in Pregastrulaboz Mutants

During gastrulation, boz mutant embryos already exhibitdefects in dorsal specification that will ultimately result ina loss of dorsal midline tissues. To gain insight into thepregastrula basis of boz ventralized mutant phenotype, weexamined the expression of bmp2b, a gene encoding asecreted factor known to function both as an inducer ofventral fates and as an antagonist of dorsal cell fate speci-fication. zbmp2b is normally first expressed in the zebrafishshortly after the midblastula transition in a ubiquitousmanner throughout the blastoderm. Prior to the onset ofgastrulation, zbmp2b transcripts decrease in abundance onthe presumptive dorsal side of the embryo generating a

expression of flh (h) and twist (j) is dramatically reduced. (k) axial ishe paraxial endoderm (arrowhead) in wild-type embryos. (l) In hom

of axial is very reduced; however, the paraxial endoderm expressioexpressed both in the dorsal axial mesoderm and in the involutingthe dorsal midline expression of ntl is reduced but the ventral latexpression in boz mutant gastrulae. (o and p) The domain of evehomozygous bozm168 mutant embryos. Embryos are oriented animmbryos at 7 hpf, eve1 expression is excluded from the dorsal regiircumference. (p) In homozygous bozm168 mutant embryos at 7 h

embryo. (q and r) The expression of tbx16/spt is altered in homozythe embryos oriented animal pole to the top. (q) In wild-type embrymargin and excluded from the trunk dorsal midline. tbx16/spt is
(arrowhead). (r) In homozygous bozm168 mutant embryos, tbx16/spt is detdomain is greatly reduced or absent.

leared marginal zone easily visible by 40% epiboly or 5 hpfFig. 8a). In boz mutant embryos, the dorsal clearing fails toccur and zbmp2b transcripts are detected both in the cellsf the dorsal margin and in the dorsal YSL (Fig. 8c).To determine if nieuwkoid/dharma activity was suffi-

cient to lead to a reduction in bmp2b expression, wild-typeieuwkoid/dharma RNA was injected into blastomeres ofleavage-stage embryos and the resultant bmp2b expressionas examined at late blastula stage (5 hpf) prior to shield

ormation. The misexpression of nieuwkoid/dharma RNAn both blastomeres of two-cell-stage embryos led to thelimination of detectable bmp2b expression throughout theregastrula embryo, even within the animal pole region32/39) (Fig. 8d). When bozozok RNA was injected intoseparate blastomeres of later cleavage stage embryos, ec-topic or expanded regions displaying reduced expression ofbmp2b were evident at the blastoderm margin (23/31) (Fig.8e). These results support the hypothesis that nieuwkoid/dharma activity is sufficient to lead to the down regulationof bmp2b expression during the process of organizer speci-fication in the pregastrula embryo.

DISCUSSION

The bozozok Locus Encodes the nieuwkoid/dharmaHomeobox Gene

We have demonstrated that the locus previously identi-fied by the bozozok mutation encodes the nieuwkoid/

harma gene. These results corroborate a similar identifi-ation of the defect in the bozozok mutant by Fekany et al.

(1999). Previous studies (Koos and Ho, 1998; Yamanaka etal., 1998) have shown that the nieuwkoid/dharma ho-meobox gene has potent dorsalizing activity and, in thispaper, we demonstrate that mutation of this gene leads to asevere ventralization of zebrafish embryos. Sequence anal-ysis of the bozozokm168 allele indicated a single base-pairsubstitution in the nieuwkoid/dharma open reading framethat generates a premature stop codon at position 70.Conceptual translation of this allele suggests that it en-

ressed during gastrulation in the dorsal mesendoderm as well as inous bozm168 mutant embryos, the dorsal mesendoderm expressions not appear affected (arrowhead). (m) In wild-type embryos, ntl isral lateral mesoderm. (n) In homozygous bozm168 mutant embryos,expression is unaffected (arrowhead). (o–r) Expanded ventral genepression is expanded into the dorsal region of the blastoderm inole toward the viewer, with dorsal to the right. (o) In wild-typed detected in cells along the ventral two-thirds of the blastoderm

he eve1 expression domain extends into the dorsal region of thebozm168 mutant embryos at 80% epiboly (8 hpf). Dorsal view withx16/spt is expressed in the ventral lateral region of the blastodermexpressed in the dorsal anterior prechordal plate mesendoderm

expozyg

n doeventeral1 exal p

on anpf, tgousos, tb

also
ected in the trunk dorsal midline and the dorsal anterior expression

I

200 Koos and Ho

FIG. 6. Impaired anterior neural ectoderm patterning in boz mutants. (a and b) Profile views of the developing head region in postgastrulaembryos (14 hpf); the expression of emx1 is detected in the developing forebrain region in wild-type embryos (a) and absent in severelyaffected boz mutant embryos (b). (c and d) The expression of the floorplate and notochord marker, shh, and hindbrain marker, krox20, inthe anterior of wild-type and boz mutant embryos, respectively, at 12 hpf. The embryos have been removed from their yolk cell andmounted flat between coverslips. Anterior is toward the top. At this stage in wild-type embryos (c), shh expression (arrowhead) is detectedin the dorsal midline and krox20 is expressed in the hindbrain region in presumptive rhombomeres 3 and 5 (arrows). In severely affectedboz mutant embryos (d), the expression of shh is reduced or absent, but the development of the hindbrain indicated by the expression ofkrox20 appears unaffected. (e–g) Expression of paxb in wild-type and boz mutant embryos at 22 hpf. Embryos have been removed from theiryolk cell and mounted flat with anterior toward the top. Embryos are aligned with the anterior extent of the pronepherous at the bottom.(e) In the anterior of wild-type embryos, paxb expression is detected in the developing eye stalks (arrowheads), the midbrain–hindbrainboundary (arrow), the otic placodes (asterisks), the neurons, and the pronepherous. (f, g) boz mutant embryos can be grouped into a seriesof truncations of anterior neuronal fates. For example, mildly affected boz mutants (f) exhibit a loss of eye stalk expression and a reducedmidbrain–hindbrain boundary expression of paxb. In severely affected boz embryos (g), these anterior neural deficiencies are enhanced toinclude a loss of midbrain–hindbrain expression of paxb. Note that the otic placodes in these severely affected mutants appear to beenlarged (asterisks). (h and i) Expression of a-coll2a in the trunk and tail of wild-type and boz mutant embryos, respectively, at 22 hpf. Atthis stage in wild-type embryos (h), a-coll2a is detected in the floorplate (fl), notochord (n), hypochord (hy), and axial endoderm (arrowhead).
n boz mutant embryos (i), the floorplate, notochord, and hypochord are reduced or absent. However, the expression of a-coll2a in the axial
endoderm is still detected ventral to the dorsal midline (arrowhead).


tanmobaeic

o

gmFt

poIrsieotsm

b(b

b

b( 70


codes a severely truncated peptide product that wouldcompletely lack the C-terminal portion including the ho-meodomain. Furthermore, this base-pair substitution leadsto a dramatic loss of organizer-inducing activity in misex-pression assays. In addition, the base-pair substitution un-derlying the mutation generates a restriction endonucleasepolymorphism allowing the demonstration that only em-bryos homozygous for the bozm168 lesion display the ven-tralized mutant phenotype. Finally, the ventralized pheno-type of boz mutant embryos was rescued by injection ofwild-type nieuwkoid/dharma RNA either into early blas-tomeres or directly into the YSL. Our findings show thatdorsal specification of the zebrafish embryo requiresnieuwkoid/dharma/bozozok function and that one impor-tant component of this function is to repress the expressionof genes specifying ventral fates such as zbmp2b.

Reduced Dorsal Specification in boz Mutants

Transplantation experiments (Shih and Fraser, 1996) andspecification assays (Sagerstrom et al., 1996) have indicatedhat the dorsal shield is the zebrafish equivalent of themphibian Spemann organizer. Our morphological exami-ation and molecular analysis of gene expression in bozutant embryos has demonstrated that the establishment

f a functional dorsal shield requires nieuwkoid/dharma/ozozok function. The dorsal shield, which normally formss a result of dorsal convergence movements during thearly gastrulation (Warga and Kimmel, 1990), fails to formn boz mutant embryos and represents the first morphologi-al distinction between boz mutant embryos and their

unaffected siblings. The failure to form a dorsal shield at theonset of gastrulation indicates that dorsal specificationrequires nieuwkoid/dharma/bozozok function prior to thenset of gastrulation.The functions of a dorsal shield or organizer during

astrulation are required to specify and contribute to axialesoendodermal fates and to induce ventral neural fates.

TABLE 1Expression of nieuwkoid/dharma Can Rescue the boz Ventralized

Embryo genotypeInjectionlocation

Quantity of RNAinjected (pg) em

ozm618 (2/2) — 01/1), bozm168 (1/2) — 0ozm168 (2/2) Blastomeres 20 per blastomere

(1/1), bozm168 (1/2) Blastomeres 20 per blastomereozm168 (2/2) YSL 60

(1/1), bozm168 (1/2) YSL 60ozm168 (2/2) YSL 100

1/1), bozm168 (1/2) YSL 100

ate-mapping analyses at the onset of gastrulation in wild-ype embryos have indicated that the dorsal shield contains


rogenitors of dorsal midline mesodermal tissues and cellsf the floorplate (Melby et al., 1995; Shih and Fraser, 1995).n fact, surgical deletion or extirpation of the shield canesult in the loss of dorsal midline mesoendodermal tissues,uch as the prechordal plate in the anterior and notochordn the posterior, along the entire anteroposterior axis of thembryo (Shih and Fraser, 1996). Consistent with impairedr reduced organizer activity during gastrulation, boz mu-ant embryos exhibit a severe reduction or loss of expres-ion of genes associated with the specification of dorsalesoendodermal fates such as gsc (Stachel et al., 1993), flh

(Talbot et al., 1995), axial (Strahle et al., 1993), ntl (Schulte-Merker et al., 1994), and twist (Halpern et al., 1995) in thedorsal midline. The nonmidline expression of axial (Strahleet al., 1993) in the paraxial endoderm and ntl (Schulte-Merker et al., 1994) in ventral lateral mesoderm appearsunaffected in boz mutants, indicating that nieuwkoid/dharma/bozozok function is specifically required for thegeneration of dorsal midline fates.

The nonautonomous activities of the dorsal shield, such asthe dorsalization of lateral mesoderm and the induction ofneural ectoderm, are thought to be mediated by diffusion ofpeptide signals released from the organizer tissue (reviewed inHarland and Gerhart, 1997). One of the known peptidessecreted by the organizer is the product of the chd gene. Chdis essential for dorsal specification and has been shown toantagonize BMP signaling by directly binding to BMP pro-teins. The expression of chd, which delimits the entire do-main of organizer activity (Miller-Bertoglio et al., 1997), isalready reduced at the onset and remains affected throughoutgastrulation in boz mutants. The continued decreased expres-sion of chd in the nascent axis further supports the proposalthat organizer activity is strongly reduced in boz mutants andalso provides insight into why the boz mutant phenotypeappears very similar to the phenotype of embryo misexpress-ing BMP family members (see below).

The loss of dorsal fates in boz mutant gastrulae is accom-panied by an expansion of ventral gene expression into the

otype

sWild-typeembryos

Hyperdorsalizedembryos

bozembryos

bozphenotype

(%)

2 0 15 8851 0 0 0

0 13 0 00 39 0 09 0 7 44

46 0 0 017 0 6 2670 0 0 0

Phen

Totalbryo

17511339164623

dorsal region. During gastrulation, the expression of the ven-tral posterior marker eve1 (Joly et al., 1993) is expanded


cismttate

eso1

eosflofIa

abeleof th

202 Koos and Ho

dorsally in boz mutants; however, this expression does notompletely encircle the margin. The dorsal gap in expressionndicates that a dorsal axis is generated, although it appearsignificantly reduced. Normally excluded from the dorsalidline, the ventral lateral expression of tbx16/spt in wild-

ype gastrula-stage embryos is associated with the specifica-ion of ventral lateral fates such as somitic tissue (Griffin etl., 1998). In boz mutant embryos, tbx16/spt expression ex-

FIG. 7. Rescue of the boz ventralized phenotype by expression ofarly blastomeres of either wild-type or boz mutant embryos led toriented anterior to the top. (a) Uninjected wild-type postgastrula ehh and the hindbrain marker krox20. At this stage, shh expressioorplate and notochord. krox20 expression is detected only in thef wild-type nieuwkoid/dharma RNA in distant blastomeres of eormation of expanded or ectopic dorsal fates as indicated by thenjection of nieuwkoid/dharma RNA in boz homozygotes as in (b)nd boz mutant embryos following injection of nieuwkoid/dharm

f) embryos at 22-hpf stage following injection of nieuwkoid/dharmaare in profile and oriented anterior to the top. YSL injection of nieuin boz homozygotes (e) but does not lead to the formation of hypwild-type (d) and heterozygote siblings. (f) Fluorescence image of rtracer that was coinjected with the nieuwkoid/dharma RNA. No la small group of labeled cells was detected in the posterior region

ends into the dorsal midline, often leaving no dorsal gap. Thisxpanded tbx16 expression may be due, in part, to the loss of


xpression of the flh homeobox gene, which has been demon-trated by mutant analyses to function as a negative regulatorf tbx16 activity in the dorsal midline (Amacher and Kimmel,998; Yamamoto et al., 1998).

Defective Anterior Neural Ectoderm Patterning inboz Mutants

zok RNA. (a–c) Injection of wild-type nieuwkoid/dharma RNA inperdorsalized phenotype. Embryos are at 12-hpf stage. Embryos areo stained simultaneously for the notochord and floorplate markerdetected along the entire anteroposterior axis in the cells of the

brain region and is restricted to rhombomeres 3 and 5. (b) Injectionell-stage stage wild-type or boz heterozygote embryos led to theic expression of shh and radially expanded krox20 expression. (c)o the formation of ectopic dorsal fates. (d–f) Analysis of wild-typeA directly into the YSL. Wild-type (d) and bozozok mutant (e and

into the YSL and staining for the expression of a-coll2a. Embryosd/dharma RNA can rescue the formation of dorsal midline tissuessalized postgastrula phenotype in either boz mutants (e) or theired boz mutant embryo in (e) revealing the distribution of lineaged cells were detected in the midline tissues. However, frequentlye tail (white arrowhead).

bozoa hymbryon ishindight-cectopled t

a RNRNA

wkoierdorescu

Numerous embryonic manipulations performed withvertebrate embryos have indicated that during gastrulation


fbdfafrNdtnart

e

oizi res ar


the organizer and the dorsal mesendoderm tissues derivedfrom it, such as the prechordal plate and posterior noto-chord, can pattern the neural tube along the entire antero-posterior axis (reviewed in Sasai and DeRobertis, 1997;Bronner-Fraser and Fraser, 1997). boz mutant embryos ex-hibit severe truncations in anterior neural tissues (forebraintype) but posterior neural tissues (hindbrain and spinal cordtype) are relatively normal. While these results indicatethat neural tissue expressing forebrain markers is reducedor absent, it is unclear whether these defects represent anelimination of the forebrain or a transformation to a moreposterior fate. Furthermore, it is not clear to what extentthe anterior neural defects in boz mutants represent asecondary consequence of the reduction and loss of pre-

FIG. 8. zbmp2b is ectopically expressed in pregastrula boz mutanmutant embryos prior to the onset of gastrulation. All embryos arewild-type embryos prior to the onset of gastrulation at 40% epibolmbryo. The region free of zbmp2b message is complementary to t

for reference in a slightly younger wild-type embryo (c). (b) In contrin homozygous bozm168 mutant embryos and the persistent zbmp2bf nieuwkoid/dharma RNA is sufficient for the down regulationnjected with nieuwkoid/dharma RNA in both blastomeres at thebmp2b expression is strongly reduced throughout the entire blanjected with nieuwkoid/dharma RNA in two distant blastomeepression are observed (arrows).

chordal plate mesendoderm during gastrulation.It is tempting to speculate that the loss of anterior neural

te


ates is a consequence of reduced dorsal shield activity inoz mutant embryos. Transplantation experiments haveemonstrated that the entire functional shield is requiredor the proper specification of the anterior neural ectodermnd that prechordal plate mesendoderm alone was unable toulfill this role (Grinblat et al., 1998). However, it cannot beuled out that in addition to specifying the shield, theieuwkoop center activity mediated by nieuwkoid/

harma/bozozok in the blastula may specify other cell fateshat have organizing activities and can pattern the anterioreural ectoderm. Recently, a population of cells in thenimal region of the midgastrula have been shown to beequired for the formation of dorsal anterior tissues such ashe forebrain (Houart et al., 1998). However, the origin of

alysis of zbmp2b expression in wild-type and homozygous bozm168

nted with animal pole toward the top and dorsal to the right. (a) Inge, zbmp2b transcripts are not detectable on the dorsal side of theorsal blastoderm region that expresses nieuwkoid/dharma, shownbmp2b transcripts do not clear completely from the dorsal marginression is also detected in the dorsal YSL. (d and e) Misexpressionbmp2b expression in wild-type pregastrula embryos. (d) Embryoell stage and assayed for zbmp2b expression at 40% epiboly stage.rm. (e) zbmp2b expression at 40% epiboly in wild-type embryo

t the eight-cell stage. Ectopic and expanded regions of zbmp2b

ts. Anorie

y stahe dast, z

expof z

two-cstode

hese anterior inductive cells and their relationship to thearly gastrula organizer is unknown.


d

rowarfotdToIfyionitpdNre

204 Koos and Ho

Misexpression of Wild-Type nieuwkoid/dharmaRescues the boz Ventralized Phenotype

To provide further support that reduced nieuwkoid/dharma activity underlies the defects observed in bozmutant embryos, misexpression of wild-type nieuwkoid/dharma RNA in either early cleavage-stage blastomeres ordirectly into the YSL by microinjection was used to rescueboz mutant embryos. The distribution of injected RNA andlineage tracer within the embryo depended on when andwhere the sample was injected. The syncytial nature of theearly embryo combined with the blastoderm cell origin ofthe YSL allows samples injected into early blastomeres tobe widely distributed both among blastomeres and in theYSL. Whereas samples injected into the YSL, after it hasbecome cytoplasmically distinct from the blastoderm, re-main mainly confined to the YSL (Kimmel and Law,1985a,b). Therefore, our injections were designed to com-pare the rescuing activities of these two regions of theembryo. Injection of nieuwkoid/dharma RNA into lateblastomeres often led to the formation of duplicated, sec-ondary dorsal axes. Whereas direct injection of higherconcentrations of nieuwkoid/dharma RNA into the YSLalso significantly rescued the loss of dorsal midline tissuefates in boz mutants, injection into the YSL failed to induceectopic dorsal tissues.

Our misexpression and rescue assays demonstrate that inboth wild-type and boz mutant embryos, misexpression ofnieuwkoid/dharma in the YSL alone is much less effectiveat inducing expanded dorsal fates and secondary axes thaninjection of RNA into early blastomeres, indicating thatnieuwkoid/dharma blastoderm activity is not equivalent tothe YSL activity. One interpretation of results suggests thatexpression of nieuwkoid/dharma in the YSL is only suffi-cient to refine and rescue the incipient embryonic axis. Incontrast, nieuwkoid/dharma expression in the blastodermis sufficient to both influence the incipient embryonicdorsal axis and induce an ectopic dorsal axis.

Is nieuwkoid/dharma Nieuwkoop Center ActivityRequired in the Blastoderm or in the YSL or Both?

In the wild-type midblastula-stage embryo, nieuwkoid/dharma is first expressed in a small group of blastodermcells at the dorsal margin. Subsequently, nieuwkoid/dharma expression becomes progressively restricted to theorsal YSL. nieuwkoid/dharma expression is detected in

the blastoderm cells during most of the time period thatnieuwkoid/dharma is expressed; however, the functionalsignificance of the blastoderm expression relative to theYSL expression of nieuwkoid/dharma has not been deter-mined. Evidence that blastoderm cells expressingnieuwkoid/dharma have Nieuwkoop center-like/organizer-inducing activity has come from transplantation experi-ments (Koos and Ho, 1998) in which it was shown that cellsmisexpressing this RNA could ectopically induce organizer
gene expression in host cells surrounding the transplantedcells. Furthermore, blastoderm rotation experiments in a

elated teleost, the goldfish Carassius auratus, suggest thatrganizer-inducing activity is present and functional bothithin the blastoderm cells and within the YSL (Yamaha etl., 1998). In these experiments, entire blastoderms wereemoved from their yolk cells before the MBT, and there-ore presumably before the expression of any Spemannrganizer-specific genes. These blastoderms were then reat-ached to the yolk cell after a rotation which reversed theorsoventral axis of the blastoderm relative to the yolk cell.hese manipulated embryos later produce two regions ofrganizer-specific gene expression and ectopic dorsal axes.n addition, blastoderms that were physically separatedrom their YSL at MBT and cultured in the absence of theirolk cells were capable of generating mesodermal fatesncluding notochord, a dorsal mesoderm derivative of therganizer (Bozhkova et al., 1994). These results support theotion that a Nieuwkoop center/organizer-inducing activ-ty may exist within the dorsal blastoderm cells as well ashe dorsal YSL. Consistent with this possibility, researcherformed in Xenopus has demonstrated that pre-MBTorsal blastomeres located just animal to the dorsal-vegetalieuwkoop center cells also posses substantial axis-

escuing activity when transplanted into UV axis-deficientmbryos (Gimlich, 1986; Kageura, 1990).

nieuwkoid/dharma Activity Is Required for DownRegulation of zbmp2b in the Pregastrula

The ventralized bozozok mutant phenotype appears verysimilar to the phenotype of embryo-misexpressing BMPfamily members. The misexpression of bmp4 during earlycleavage stages leads to a progressive loss of dorsal midlinemesodermal cell fates (Kishimoto et al., 1997; Neave et al.,1997; Nikado et al., 1997). These data prompted us toexamine the expression of BMPs in bozozok mutant em-bryos. Currently in the zebrafish, two BMP family membersthat are expressed prior to the onset of gastrulation havebeen reported, zbmp2b and zbmp4 (Nikado et al., 1997). Inwild-type embryos, zbmp2b is initially expressed shortlyafter MBT in a ubiquitous manner throughout the blasto-derm. Subsequently, as epiboly proceeds prior to gastrula-tion, zbmp2b expression becomes down regulated in thedorsal side of the embryo, producing a region free of zbmp2transcripts. Visual comparison suggests that this clearedregion of zbmp2 expression is complementary to the smallrestricted expression domain of nieuwkoid/dharma. Anal-ysis of zbmp2b expression in boz mutant embryos prior tothe onset of gastrulation revealed that zbmp2b expressionfails to clear from the presumptive dorsal side of the embryoand this ectopic expression in the most severe cases extendsinto the dorsal YSL.

The persistent expression of zbmp2b in the dorsal regionof the boz mutant pregastrula embryos strongly suggeststhat nieuwkoid/dharma activity is required to down regu-late bmp2b expression in the zebrafish embryo. Further-
more, misexpression of nieuwkoid/dharma in early em-bryos is sufficient to lead to a down regulation of bmp2b

Rtteeiibptevtvan

psiamr1Gsr1DtNw

mtAmassfeptliew

dtaltnsnppsgTa11prbdvfet

e


transcription throughout the pregastrula embryo, includingthe animal pole region. These results suggest that dorsalantagonism of ventral signaling cascades is required prior tothe onset of gastrulation at the level of the Nieuwkoopcenter for the proper establishment of a functional shield orgastrula organizer. Much of the current understanding ofBMP ventralizing signaling activity has come from workfocused on the dorsal–ventral specification of mesendodermduring gastrulation in Xenopus (reviewed in Graff, 1997).

ecently, BMP signaling has been implicated in regulatinghe dorsoventral nature of the vegetal mesendoderm induc-ion centers during early Xenopus development (Nachalielt al., 1998). In these experiments, the ectopic expression ofither bmp4 or one of its intracellular transducers, smad5,n dorsal-vegetal cells severely impaired the organizer-nducing activity of the vegetal Nieuwkoop center cells inoth embryos and recombinant explants. Furthermore, ex-ression in embryos of a dominant negative BMP2/4 recep-or in the vegetal mesendoderm-inducing cells led to annhancement of dorsal mesoderm induction and inhibitedentral mesoderm induction in the overlying marginal zoneissue. Taken together, these findings suggest that dorso-entral polarity of the pregastrula embryo requires thentagonism of ventralizing activity mediated by BMP sig-aling even at the level of the Nieuwkoop center.It is tempting to speculate that nieuwkoid/dharma sup-

resses expression of zbmp2b directly at the level of tran-cription. Conceptual translation of nieuwkoid/dharmandicates that its protein product shares, in its N-terminal,short amino acid sequence that exhibits homology with aotif present in both fly and vertebrate Goosecoid proteins

eferred to as the GEH (Goriely et al., 1996; Koos and Ho,998). Experiments using chimeric and truncated forms ofoosecoid have demonstrated that this region of the tran-

cription factor is required to mediate a transcriptionalepressive activity (Mailhos et al., 1998; Ferreiro et al.,998). Although the N-terminal region of Nieuwkoid/harma has not been directly demonstrated to possess

ranscriptional repressor activity, by analogy with Gsc,ieuwkoid/Dharma would also have repressor activityhich may directly regulate the transcription of zbmp2b.Another mechanism by which Nieuwkoid/Dharmaight suppress zbmp2b mRNA expression is by affecting

he synthesis or transduction of BMP signaling cascades.nalyses of zebrafish harboring mutations in zbmp2 (Kishi-oto et al., 1997; Nguyen et al., 1998) and smad5 (Hild et

l., 1999), an essential component for transduction for BMPignaling activity, have revealed that zygotic zbmp2b tran-cription requires functional BMP signaling activity. There-ore, Nieuwkoid/Dharma may negatively regulate zbmp2bxpression indirectly by influencing the expression of com-onents downstream of BMP receptor activation andhereby interfere with the BMP autoregulatory feedbackoop. A third possibility is that Nieuwkoid/Dharma actsndirectly to down regulate BMP signaling by promoting the
xpression of dorsal genes, such as chd, which then in turnould antagonize BMP signaling activity.

Fekany et al. (1999) hypothesized that the nieuwkoid/harma gene was functionally similar to the siamois andwin genes of Xenopus. Despite similarities in the timingnd location of expression, our data show that the molecu-ar function of nieuwkoid/dharma appears to differ fromhe functions ascribed to siamois and twin. The zebrafishieuwkoid/dharma gene exhibits only slight sequenceimilarity to the Xenopus siamois gene. In addition,ieuwkoid/dharma appears to functionally repress the ex-ression of ventral-specific genes in the early zebrafishregastrula, whereas it has been demonstrated that bothiamois and twin activate the expression of dorsal-specificenes in Xenopus (Lemaire et al., 1995; Laurent et al., 1997).herefore, rather than activating dorsal-specific genes suchs goosecoid or the nodal-related genes squint (Erter et al.,998; Feldman et al., 1998) and cyclops (Regeblatti et al.,998; Sampath et al., 1998), we hypothesize that one of therimary functions of the nieuwkoid/dharma gene is toepress the expression of ventral-specific genes such asmp, wnt, and vent, thereby allowing the expression oforsal-specific genes at a later time in development. Theery early expression of nieuwkoid/dharma immediatelyollowing the onset of MBT during blastula stages and thexpansion of zbmp2b expression in boz mutants is consis-ent with this model.

nieuwkoid/dharma/bozozok Autoregulation

In order to determine if nieuwkoid/dharma has autoregu-lative activities, we examined its expression in boz mutantmbryos. In boz mutant embryos, nieuwkoid/dharma ex-

pression is already down regulated or absent by the lateblastula stage. These results clearly demonstrate thatnieuwkoid/dharma transcription requires Nieuwkoid/Dharma activity in the pregastrula.

The spatial and temporal juxtaposition of the blastodermexpression domain and the YSL expression of nieuwkoid/dharma are highly suggestive that nieuwkoid/dharma ex-pression in the blastoderm cells influences its expression inthe dorsal YSL. Transplantation experiments will need tobe performed to determine if nieuwkoid/dharma autoregu-lation can function in a cell-nonautonomous manner and ifnieuwkoid/dharma-mediated signals do in fact spread fromthe blastoderm cells to the underlying YSL.

Cell-autonomous activities may also function to inducenieuwkoid/dharma expression in the YSL. Just prior to andduring MBT, the last blastoderm cells that maintain cyto-plasmic connections to the underlying yolk cell fuse to-gether to form the YSL nuclear syncytium in the animalregion of the yolk cell (Kimmel and Law, 1985b). It isconceivable that regional asymmetries inherited or estab-lished within the basal region of the blastoderm prior to
MBT will also be cell-autonomously imprinted into theYSL when it forms at MBT.

206 Koos and Ho

ACKNOWLEDGMENTS

We thank A. Fjose, D. Grunwald, M. Halpern, P. W. Ingham, T.Jowett, J. H Postlethwait, I. Ruvinsky, U. Strahle, and N. Ueno forprobes and also Dr. L. Solnica-Krezel for providing a pair of adultbozozok fish. We thank Tracy Roskoph for technical assistance andfish care. This work was supported by NIH Grant RO1 HD3449 toR.K.H. who is a Rita Allen Foundation Scholar.

REFERENCES

Amacher, S. L., and Kimmel, C. B. (1998). Promoting notochord fateand repressing muscle development in zebrafish axial mesoderm.Development 125, 1397–1406.

Blagden, C. S., Currie, P. D., Ingham, P. W., and Hughes, S. M.(1997). Notochord induction of zebrafish slow muscle mediatedby Sonic hedgehog. Genes Dev. 11, 2163–2175.

Bozhkova, V., Kronnie, G. t., and Timmermans, L. P. M. (1994).Mesoderm differentiation in explants of carp embryos. Roux’sArch. Dev. Biol. 204, 20–29.

Bronner-Fraser, M., and Fraser, S. E. (1997). Differentiation of thevertebrate neural tube. Curr. Biol. 9, 885–891.

Cooper, M. S., and D’Amico, L. A. (1996). A cluster of noninvolut-ing endocytic cells at the margin of the zebrafish blastodermmarks the site of embryonic shield formation. Dev. Biol. 180,184–198.

Erter, C. E., Solnica-Krezel, L., and Wright, C. V. E. (1998). Ze-brafish nodal-related 2 encodes an early mesoendodermal in-ducer signaling from the extraembryonic yolk syncytial layer.Dev. Biol. 204, 361–372.

Eyal-Giladi, H. (1997). Establishment of the axis in chordates: Factsand speculations. Development 124, 2285–2296.

Fekany, K., Yamanaka, Y., Leung. T., Sirotkin, H. I., Topczewski, J.,Gates, M. A., Hibi, M., Renucci, A., Stemple, D., Radbill, A.,Schier, A. F., Driever, W., Hirano, T., Talbot, W. S., and Solnica-Krezel, L. (1999). The zebrafish bozozok locus encodes Dharma,a homeodomain protein essential for induction of the gastrulaorganizer and dorsoanterior embryonic structures. Development126, 1427–1438.

Feldman, B., Gates, M. A., Egan, E. S., Dougan, S. T., Rennebeck,G., Sirotkin, H. I., Schier, A. F., and Talbot, W. S. (1998).Zebrafish organizer development and germ-layer formation re-quire nodal-related signals. Nature 395, 181–185.

Ferreiro, B., Artinger, M., Cho, K. W. Y., and Niehrs, C. (1998).Antimorphic goosecoids. Development 125, 1347–1359.

Gerhart, J., Danilchik, M., Doniach, T., Roberts, T., Rowning, B.,and Stewart, R. (1989). Cortical rotation of the Xenopus egg:Consequences for the anteroposterior pattern of embryonic dor-sal development. Development 107(Suppl.), 37–51.

Gimlich, R. L. (1986). Acquisition of developmental autonomy inthe equatorial region of the Xenopus embryo. Dev. Biol. 115,340–352.

Goriely, A., Stella, M., Coffiner, C., Kessler, D., Mailhos, C.,Dessain, S., and Desplan, C. (1996). A functional homologue ofgoosecoid in Drosophila. Development 122, 1641–1650.

Graff, J. M. (1997). Embryonic patterning: To BMP or not to BMP,that is the question. Cell 89, 171–174.

Graff, J. M., Theis, R. S., Song, J. J., Celeste, A. J., and Melton, D. A.(1994). Studies with a Xenopus BMP receptor suggest that ventral
mesoderm-inducing signals override dorsal signals in vivo. Cell79, 169–179.

Griffin, K. J. P., Amacher, S. L., Kimmel, C. B., and Kimmelman, D.(1998). Molecular identification of spadetail: Regulation of thezebrafish trunk and tail mesoderm formation by T-box genes.Development 125, 3379–3388.

Grinblat, Y., Gamse, J., Patel, M., and Sive, H. (1998). Determina-tion of the zebrafish forebrain: Induction and patterning. Devel-opment 125, 4403–4416.

Halpern, M. E., Thisse, C., Ho, R. K., Riggleman, B., Trevarrow, B.,Weinberg, E. S., Postlewait, J. H., and Kimmel, C. B. (1995).Cell-autonomous shift from axial to paraxial mesoderm develop-ment in zebrafish floating head mutants. Development 121,4257–4264.

Harland, R. M., and Gerhart, J. C. (1997). Formation and function ofSpemann’s organizer. Annu. Rev. Cell. Dev. Biol. 13, 611–667.

Hild, M., Dick, A., Rauch, G.-J., Meier, A., Bouwmeester, T.,Hafter, P., and Hammerschmidt, M. (1999). The smad5 mutationsomitabun blocks Bmp2b signaling during early dorsoventralpatterning of the zebrafish embryo. Development 126, 2149–2159.

Houart, C., Westerfield, M., and Wilson, S. (1998). A small popu-lation of anterior cells patterns the forebrain during zebrafishgastrulation. Nature 391, 788–792.

Joly, J.-S., Joly, C., Schulte-Merker, S., Boulkebache, H., and Con-damine, H. (1993). The ventral and posterior expression of thehomeobox gene eve1 is perturbed in dorsalized and mutantembryos. Development 119, 1261–1275.

Kangeura, H. (1990). Spatial distribution of the capacity to initiatea secondary embryo in the 32-cell embryo of Xenopus laevis.Dev. Biol. 142, 432–438.

Kimmel, C. B., and Law, R. D. (1985a). Cell lineage of zebrafishblastomeres. I. Cleavage pattern and cytoplasmic bridges be-tween cells. Dev. Biol. 108, 78–85.

Kimmel, C. B., and Law, R. D. (1985b). Cell lineage of the zebrafishblastomeres. II. Formation of the yolk syncytial layer. Dev. Biol.108, 86–93.

Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ulmann, B., andSchilling, T. F. (1995). Stages of embryonic development of thezebrafish. Dev. Dyn. 203, 253–310.

Kishimoto, Y., Lee, K.-H., Zon, L., Hammerschmidt, M., andSchulte-Merker, S. (1997). The molecular nature of swirl: BMP2function is essential during early dorsoventral patterning. Devel-opment 124, 4457–4466.

Koos, D. S., and Ho, R. K. (1998). The nieuwkoid gene characterizesand mediates a Nieuwkoop-center-like activity in the zebrafish.Curr. Biol. 8, 1199–1206.

Krauss, S., Johansen, T., Khorz, V., and Fjose, A. (1991). Expressionof the zebrafish paired box gene pax[zf-b] during early neurogen-esis. Development 113, 1193–1206.

Krauss, S., Concordet, J.-P., and Ingham, P. W. (1993). A function-ally conserved homolog of the Drosophila segment polarity genehedgehog is expressed in tissues with polarizing activity. Cell 75,1431–1444.

Laurent, M. N., Blitz, I. L., Hashimoto, C., Rothbacher, U., andCho, K. W. (1997). The Xenopus homeobox gene twin mediatesWnt induction of goosecoid in establishment of Spemann’sorganizer. Development 124, 4905–4916.

Lemaire, P., Garrett, N., and Gurdon, J. B. (1995). Expressioncloning of siamois, a Xenopus homeobox gene expressed indorsal-vegetal cells of blastulae and able to induce a completesecondary axis. Cell 81, 85–94.

Lemaire, P., and Kodjabachain, L. (1996). The vertebrate organizer:Structure and molecules. Trends Genet. 12, 525–531.


S

S

S


Long, W. L. (1983). The role of the yolk syncytial layer in thedetermination of the plane of bilateral symmetry in the rainbowtrout, Salmo gairdneri Richardson. J. Exp. Zool. 228, 91–97.

Mailhos, C., Andre, S., Mollereau, B., Goriely, A., Hemmati-Brivanlou, A., and Desplan, C. (1998). Drosophila Goosecoidrequires a conserved heptapeptide for repression of Paired-classhomeoprotein activators. Development 125, 937–947.

Melby, A. E., Warga, R. M., and Kimmel, C. B. (1995). Specificationof cell fates at the dorsal margin of the zebrafish gastrula.Development 122, 2225–2237.

Miller-Bertoglio, V. E., Fisher, S., Sanchez, A., Mullins, M. C., andHalpern, M. E. (1997). Differential regulation of chordin expres-sion domains in mutant zebrafish. Dev. Biol. 192, 537–550.

Minzuno, T., Yamaha, E., Wakahara, M., Kuroiwa, A., and Takeda,H. (1996). Mesoderm induction in zebrafish. Nature 383, 131–132.

Moon, R. T., and Kimmelman, D. (1998). From cortical rotation toorganizer gene expression: Toward a molecular explanation ofaxis specification in Xenopus. BioEssays 20, 536–545.

Morita, T., Nitta., H., Kiyama, Y., Mori, H., and Mishina, M. (1995).Differential expression of two zebrafish emx homeoproteinmRNAs in the developing brain. Neurosci. Lett. 198, 131–134.

Nachaliel, N., Re’em-Kalma, Y., Eshed, O., Elias, S., and Frank, D.(1998). BMP regulates vegetal pole induction centers in earlyXenopus development. Genes Cells 3, 649–659.

Neave, B., Holder, N., and Patient, R. (1997). A graded response toBMP signaling coordinates patterning of the mesoderm andectoderm in the zebrafish. Mech. Dev. 62, 183–195.

Nikaido, M., Tada, M., Saji, T., and Ueno, N. (1997). Conservationof BMP signaling in zebrafish mesoderm patterning. Mech. Dev.61, 75–88.

Nguyen, V. H., Schmid, B., Trout, J., Conners, S. A., Ekker, M., andMullins, M. C. (1998). Ventral and lateral regions of the zebrafishgastrula, including the neural crest progenitors, are establishedby a bmp2b/swirl pathway of genes. Dev. Biol. 199, 93–110.

Oxtoby, E., and Jowett, T. (1993). Cloning of the zebrafish Krox-20gene (krx-20) and its expression during hindbrain development.Nucleic Acids Res. 21, 1087–1095.

Rebagliati, M. R., Toyama, R., Haffter, P., and Dawid, I. B. (1998).cyclops encodes a nodal-related factor involved in midline sig-naling. Proc. Natl. Acad. Sci. USA 95, 9932–9937.

Ruvinsky, I., Silver, L. M., and Ho. R. K. (1998). Characterization ofthe zebrafish tbx16 gene and evolution of the vertebrate T-boxfamily. Dev. Genes Evol. 298, 94–99.

Sagerstrom, C. G., Grinblat, Y., and Sive, H. (1996). Anteroposte-rior patterning in the zebrafish, Danio rerio: An explant assayreveals inductive and suppressive cell interactions. Development122, 1873–1883.

Sampath, K., Rubinstein, A. L., Cheng, A. M. S., Liang, J. O.,Fekany, K., Solnica-Krezel, L., Korzh, V., Halpern, M. E., andWright, C. V. E. (1998). Induction of the zebrafish ventral brainand floorplate requires cyclops/nodal signalling. Nature 395,185–189.

Sasai, Y., and DeRobertis, E. M. (1997). Ectodermal patterning invertebrate embryos. Dev. Biol. 182, 5–20.

Schneider, S., Steinbeisser, H., Warga, R. M., and Hausen, P. (1996).b-Catenin translocation into nuclei demarcates the dorsalizingcenters in the frog and fish. Mech. Dev. 57, 191–198.

chulte-Merker, S., van Eeden, F., Halpern, M. E., Kimmel, C. B.,and Nusslein-Volhard, C. (1994). no tail (ntl) is the zebrafish


homolog of the mouse T (brachyury) gene. Development 120,1009–1015.

chulte-Merker, S., Lee, K. J., McMahon, A. P., and Hammer-schmidt, M. (1997). The zebrafish organizer requires chordino.Nature 387, 862–863.

hih, J., and Fraser, S. E. (1995). The distribution of tissue progeni-tors within the shield region of the zebrafish gastrula. Develop-ment 121, 2755–2765.

Shih, J., and Fraser, S. E. (1996). Characterization of the zebrafishorganizer: Microsurgical analysis at early-shield stage. Develop-ment 122, 1313–1322.

Solnica-Krezel, L., Stemple, D. L., Mountcastle-Shah, E., Ragini, Z.,Neuhauss, S. C. F., Malicki, J., Scheir, A. F., Stachel, S. E.,Grunwald, D. J., and Meyers, P. Z. (1993). Lithium perturbationand goosecoid expression identify a dorsal specification pathwayin the pregastrula zebrafish. Development 117, 1261–1274.

Stainier, D. Y. R., Zwarkruis, F., Abdelilah, S., and Dreiver, W.(1996). Mutations affecting cell fates and cellular rearrangementsduring gastrulation in the zebrafish. Development 123, 67–80.

Strahle, U., Blader, P., Henrique, D., and Ingham, P. W. (1993).Axial, a zebrafish gene expressed along the developing body axis,shows altered expression in cyclops mutant embryos. GenesDev. 7, 1436–1446.

Talbot, W. S., Trevarrow, B., Halpern, M. E., Melby, A. E., Farr, G.,Postlewait, J. H., Jowett, T., Kimmel, C. B., and Kimmelman, D.(1995). A homeobox gene essential for zebrafish notochord de-velopment. Nature 378, 150–157.

Tam, P. P. L., and Quinlan, G. A. (1996). Mapping vertebrateembryos. Curr. Biol. 6, 104–106.

Thisse, C., Thisse, B., Schilling, T. F., and Postlewait, J. H. (1993).Structure of the zebrafish snail1 gene and its expression in wildtype, spadetail and no tail embryos. Development 119, 1203–1215.

Turner, D. L., and Weintraub, H. (1994). Expression of achaete-scute homolog 3 in Xenopus embryos converts ectodermal cellsto a neural fate. Genes Dev. 8, 1434–1447.

Waga, R. M., and Kimmel, C. B. (1990).Cell movements duringepiboly and gastrulation in zebrafish. Development 108, 569–580.

Westerfield, M. (1994). “The Zebrafish Book: A Guide for theLaboratory Use of Zebrafish.” Univ. of Oregon Press, Eugene.

Yamaha, E., Mizuno, T., Hasebe, Y., Takeda, H., and Yamazaki, F.(1998). Dorsal specification in blastoderm at the blastula stage ingoldfish, Carassius auratus. Dev. Growth Differ. 40, 267–276.

Yamamoto, A., Amacher, S. L., Kim, S.-H., Geissert, D., Kimmel,C. B., and De Rebertis, E. M. (1998). Zebrafish paraxial protcad-herin is a downstream target of spadetail involved in morpho-genesis of gastrula mesoderm. Development 125, 3389–3397.

Yamanaka, Y., Mizuno, T., Sasai, Y., Kishi, M., Takeda, H., Kim,C. H., Hibi, M., and Hirano, T. (1998). A novel homeobox gene,dharma, can induce the organizer in a non-cell-autonomousmanner. Genes Dev. 12, 2345–2353.

Yan, Y.-L., Hatta, K., Riggleman, B., and Postlethwait, J. H. (1995).Expression of a type II collagen gene in the zebrafish embryonicaxis. Dev. Dyn. 203, 363–376.

Received for publication June 2, 1999
Revised September 1, 1999
Accepted September 1, 1999


Documents

The nieuwkoid/dharma Homeobox Gene Is Essential for bmp2b ... · Based on these results, we propose that gastrula organizer speciﬁcation requires the ... Further evidence supporting