a RESEARCH ARTICLE

Gsx Transcription Factors Repress IroquoisGene ExpressionEmily F. Winterbottom, Simon A. Ramsbottom, and Harry V. Isaacs*

We have previously shown that the Gsx family homeobox gene Gsh2 is part of the regulatory networkspecifying dorsoventral pattern of primary neurons in the developing amphibian embryo. Here, we inves-tigate the role of Gsx transcription factors in regulating the transcription of Iroquois family homeoboxgenes in the amphibian neural plate. Iroquois genes are key regulators of neural patterning and theirexpression is coincident with that of the Gsx genes during open neural plate stages. We show that Gsxproteins repress Iroquois expression in the embryo and conversely, inhibition of Gsx activity with eitherantisense morpholino oligos or an anti-morphic Gsx protein up-regulates Iroquois expression. These dataindicate that Gsx factors act as negative regulators of Iroquois gene expression in the amphibian neuralplate and support a model in which the Gsx proteins promote neuronal differentiation by repressing theexpression of known inhibitors of neuronal differentiation such as Iro3. Developmental Dynamics240:1422–1429, 2011. VC 2011 Wiley-Liss, Inc.

Key words: BMP; Gsh1; Gsh2; Iroquois; neurogenesis; Xenopus

Accepted 27 March 2011

INTRODUCTION

In nonamniotic vertebrates, an earlyphase of primary neurogenesis occursbefore neural tube closure, which gen-erates the neurons controlling simplemovement and reflex responses of thefree-swimming larvae (Forehand andFarel, 1982; Hartenstein, 1989, 1993).For many years, the process of pri-mary neurogenesis in the open neuralplate of Xenopus, the African clawedtoed frog, has provided a relativelysimple and accessible model for thestudy of vertebrate central nervoussystem (CNS) development (Kintner,1992; Chitnis and Kintner, 1995;Chitnis, 1999).

We have recently shown that Gsx(‘‘genomic screened homeobox’’) genes

are expressed in the open neural plateof Xenopus during primary neurogene-sis and Gsh2 is part of the network ofrepressive homeodomain transcriptionfactors that establishes the patterns ofgene expression in the open neuralplate of amphibian embryos (Illes et al.,2009; Winterbottom et al., 2010).

The Gsx genes belong to the Para-Hox family of homeobox genes andhave homologues in all major groupsof bilaterians, as well as cnidarians.Expression is seen in developing nerv-ous tissue in all bilaterian speciesstudied, suggesting an ancestral rolefor this class of transcription factorsin neural patterning and/or differen-tiation (Valerius et al., 1995; Weisset al., 1998; Frobius and Seaver, 2006;Galliot et al., 2009; Illes et al., 2009).

In this report, we investigate theactivities of the Xenopus Gsx proteins.Using Gsx domain swap mutant pro-teins, we show that the activities ofwild-type Gsx proteins are similar tothat of fusion proteins containingeither the Gsh1 or Gsh2 peptidesequence fused to the transcriptionalrepression domain from the Drosoph-ila Engrailed protein. Conversely, Gsxfusion proteins containing the domi-nantly acting transcriptional activa-tion domain from the Herpes simplexVP16 protein function as antimorphicGsx proteins. Our data are consistentwith Xenopus Gsx proteins acting astranscriptional repressors.Our previous study showed that

Gsh2 is required for the normal devel-opment of primary interneurons, and

Dev

elop

men

tal D

ynam

ics

Additional Supporting Information may be found in the online version of this article.

Department of Biology, University of York, York, United Kingdom*Correspondence to: Harry V. Isaacs, Area 11, Department of Biology, University of York, York, YO10 5YW, UK.E-mail: hvi1@york.ac.uk

DOI 10.1002/dvdy.22648Published online 2 May 2011 in Wiley Online Library (wileyonlinelibrary.com).

DEVELOPMENTAL DYNAMICS 240:1422–1429, 2011

VC 2011 Wiley-Liss, Inc.

mutual repression between the Gsh2and Dbx1 genes is important forestablishing the limits of Gsh2 andDbx1 expression at the interneuron/progenitor domain boundary (Winter-bottom et al., 2010).

Iroquois genes are key regulators inthe specification and patterning of thevertebrate nervous system (Bellefroidet al., 1998; Gomez-Skarmeta et al.,1998; Cavodeassi et al., 2001). TheIroquois gene, Iro3, is expressed in theamphibian neural plate in a domainthat encompasses the medial motorneuron domain and extends laterallyto the interneuron domain (Bellefroidet al., 1998). Gsh2 is expressed at thelateral boundary of Iro3 expression(Illes et al., 2009). We present dataindicating that, during normal devel-opment, Gsx proteins negatively regu-late expression of Iro3 and the otherIroquois genes, Iro1 and Iro2. Further-more, our experiments show that Gsxproteins are able to regulate Iro3expression in the absence of proteintranslation, suggesting that Iro3 islikely to be a proximal target of Gsx-mediated repression. We propose thatGsx factors play a role in regulatingthe levels and extent of Iroquois geneexpression in the open neural plateand that these interactions are impor-tant for regulating the differentiationof primary interneurons.

RESULTS

Gsh2 and Gsh2-Repressor

Proteins Ventralize

Development

cDNA constructs coding for Gsxfusion proteins were generated tomanipulate the activity of Gsh1 andGsh2 in vivo (Supp. Fig. S1, which isavailable online). Gsh1- and Gsh2-EnRencode fusions with the dominantlyacting repressor domain from the Dro-sophila Engrailed protein (Conlonet al., 1996; Isaacs et al., 1998). X. tro-picalis embryos were bilaterallyinjected with mRNA encoding the wild-type Gsh2 and Gsh2-EnR proteins.Injection of Gsh2 mRNA caused a ven-tralized phenotype of axis truncation,with a reduction of head structures,including loss of eyes and cementglands (Fig. 1A,B). Injection of mRNAcoding for Gsh2-EnR gave a very simi-lar phenotype (Fig. 1C). These data,and our previous study (Winterbottomet al., 2010), indicate that Gsh2 acts asa transcriptional repressor.

Antimorphic Gsh2 Protein

Induces Partial Secondary

Axes

We investigated further the activity ofGsx proteins using an antimorphic

Gsh2 protein, which consists of theGsh2 coding sequence fused to thedominantly acting VP16 transactiva-tion domain (Supp. Fig. S1; Winter-bottom et al., 2010). Given the ven-tralized phenotype that arises fromGsx overexpression, we investigatedthe effects of ventral injection ofGsh2-VP16 and found that this leadsto inhibition of blastopore closure andthe formation of pigmented tissuemasses resembling partial secondaryaxes (Fig. 1D,E). Histological exami-nation of these embryos reveals thatthe partial secondary axes containprofuse neural tissue organized intoneural tube-like structures (Fig. 1F,G).In keeping with the observed pig-mented tissues being partial secondaxes induced by ectopic dorsal orga-nizer tissue, we find that Gsh2-VP16injection leads to a marked expansionof the expression of organizer markerchordin and an inhibition of the ven-tral marker vent2 at the early gastrulastage (Supp. Fig. S2).

Gsx Proteins Promote Bone

Morphogenetic Protein

Activity

The phenotype arising from Gsh2overexpression resembles the ventral-ized phenotype that results from

Fig. 1. Phenotypic effects of Gsh2 and Gsh2 fusion proteins. A–C: Phenotypes at stage 41 of X. tropicalis control embryos (A), and embryos injectedwith (B) 10 pg Gsh2, or (C) 50 pg Gsh2-EnR mRNA. D,E: Phenotypes at stage 33 of (D) control embryos, and (E) embryos injected into the left ventralblastomere with 15 pg of Gsh2-VP16 mRNA. All embryos are with anterior to the left and dorsal to the top. F,G: Section of an embryo from (E), at thelevel of Gsh2-VP16-induced partial secondary axis, histologically stained with borax carmine and picro blue-black. Nuclei stain red, cells green, andaxons turquoise. nc, notochord; nt, neural tube. White arrows indicate partial secondary axes. Red arrows indicate ectopic neural tissue.

Dev

elop

men

tal D

ynam

ics

GSX REPRESSION OF IROQUOIS GENES 1423

injection of low doses of BMP4 mRNA(Dale et al., 1992). Furthermore, thepartial secondary axes that resultfrom ventral injection of antimorphicGsh2 resemble those produced byinjection of inhibitory bone morphoge-netic protein (BMP) receptors (Graffet al., 1994; Suzuki et al., 1994). Withthis in mind, we investigated if manip-ulating Gsx activity in the embryoincreases BMP activity. Reverse tran-scriptase-polymerase chain reaction(RT-PCR) was used to analyze levelsof BMP4 and BMP7 expression inembryos injected with mRNA encodingwild-type Gsh2 and Gsh2-VP16 (Fig.2A). We find that levels of BMP7 andBMP4 mRNAs are up-regulated bywild-type Gsh2 and down-regulated byGsh2-VP16.

We also used western blot analysisof phosphoSmad 1, 5, and 8 levels as areadout of the activity of the BMP sig-nal transduction pathway. Injection ofeither wild-type Gsh2 or Gsh2-EnRresults in an increase in phosphoS-mad1/5/8 levels relative to controlembryos (Fig. 2B). In contrast, theBMP inhibitor Noggin and Gsh2-VP16 both dramatically inhibit BMPsignal transduction. Our experimentsalso indicate that Gsh1 and Gsh1fusion proteins are able to regulateBMP signaling (Fig. 2C).

Gsx Proteins Repress

Expression of Iro3

Our data clearly indicate that overex-pression of Gsh2 wild-type and Gsh2domain swap proteins leads to alteredBMP activity during gastrula stages.However, our previous analysis of Gsxgene expression by in situ hybridiza-tion indicates that they are onlyexpressed from the early neurula stageonward (Illes et al., 2009). This is con-

firmed by our Q-PCR analysis of Gsh2expression during early development(Fig. 3A). These data indicate that Gsxproteins are unlikely to modulate BMPactivity during gastrula stages.

One model to explain the abovedata is that Gsx target genes, whichare expressed both during gastrulaand neurula stages, are repressors ofBMP ligand gene expression. Goodcandidates for such Gsx targets areIroquois family genes, which encodehomeodomain-containing transcrip-tional repressors. It has previouslybeen shown that the Iro1 proteinrepresses BMP4 expression in theSpemann organizer and neural plate(Glavic et al., 2001; Gomez-Skarmetaet al., 2001).

Although the Gsh2 expression pro-file indicates that it is unlikely to reg-ulate BMP activity by means of Iro-quois genes at gastrula stages duringnormal development, the regulationof BMP signaling by Gsx proteins pro-vides a useful assay system for inves-tigating interactions within the puta-tive Gsx-Iroquois regulatory pathway.A prediction from our model is that, ifGsx proteins function upstream of Ir-oquois genes, the downstream effectsof the antimorphic Gsh2-VP16 protein,which inhibits BMP signaling, shouldbe rescued by the antimorphic Iroquoisprotein, Iro1-VP16, which promotesBMP signaling through binding tothe BMP4 promoter (Gomez-Skarmetaet al., 2001). In support of this, we findthat the strong inhibition of BMP sig-naling caused by Gsh2-VP16 is rescuedby co-injection of Iro1-VP16 (Fig. 3B).

To determine whether Gsx proteinsinhibit Iroquois gene expression dur-ing open neural plate stages, weexamined Iro2 and Iro3 expression inembryos which were unilaterallyinjected with mRNA encoding wild-

type Gsh1 or Gsh2. Injection of wild-type Gsh1 or Gsh2 inhibited theexpression of Iro2 (Gsh1: 100%, n ¼30, Fig. 3C; and Gsh2: 100%, n ¼ 19,Fig. 3D) and Iro3 (Gsh1: 100%, n ¼25, Fig. 3E; and Gsh2: 100%, n ¼ 7,Fig. 3F). Supplementary Figure S3shows Iro1 is also down-regulated byGsx proteins.

Antimorphic Gsh2

Up-regulates Iroquois

Expression

If Gsx factors directly repress Iroquoisgene expression, antimorphic Gsh2should up-regulate Iroquois gene tran-scription. We first analyzed Iro2 andIro3 expression in animal hemispheretissue explants from embryos injectedwith Gsh2-VP16 mRNA. Figure 3G–Jshows that there is little Iroquois geneexpression in control explants buthigh levels are induced by Gsh2-VP16.To avoid nonphysiological effects

arising from the activity of antimor-phic Gsx in the very early embryo at atime when Gsx genes are not nor-mally expressed, we made use of adexamethasone-inducible version ofGsh2-VP16 (Gsh2-VP16-GR; Supp.Fig. S1). Unilateral expansion of Iro2expression is seen in the injected sidefollowing DEX treatment (90%; n ¼30; Fig. 3K). Iro3 expression is alsoexpanded in embryos injected withGsh2-VP16-GR mRNA and treatedwith DEX at mid-gastrula stage 11(100%; n ¼ 24; Fig. 3L)

Antisense Knockdown of Gsx

Up-regulates Iroquois

Expression

To investigate whether Gsx proteinsrepress Iroquois gene expression inthe open neural plate during normal

Fig. 2. Effects of Gsx proteins on bone morphogenetic protein (BMP) signaling. A: Reverse transcriptase-polymerase chain reaction (RT-PCR)analysis of expression of BMP4 and BMP7 in stage 14 X. laevis control embryos or embryos injected with mRNAs encoding Gsh2 or Gsh2-VP16as indicated. B,C: Western blot analysis of stage 11 X. laevis embryos injected with mRNAs as indicated, and blotted with antibodies againstPhosphoSmad1/5/8 and GAPDH as a loading control. B: Gsh2 and Gsh2 fusion proteins. C: Gsh1 and Gsh1 fusion proteins. Noggin is used as acontrol for BMP signaling inhibition.

Dev

elop

men

tal D

ynam

ics

1424 WINTERBOTTOM ET AL.

Fig. 3. Repression of Iroquois gene expression by Gsx proteins. A: Quantitative-polymerase chain reaction(Q-PCR) analysis of Gsh2 expressionduring early development. Levels are normalized to ornithine decarboxylase (ODC) expression and are shown relative to expression at Nieuwkoopand Faber (NF) stage 16. B: Western blot analysis of stage 11 X. laevis embryos injected with mRNAs as indicated, and blotted with antibodiesagainst PhosphoSmad1/5/8 and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) as a loading control. C–F: In situ hybridization analysis of(C,D) Iro2 and (E,F) Iro3 expression in neurula stage 14 X. tropicalis embryos unilaterally injected with (C,E) 20 pg of Gsh1 mRNA and (D,F) 5 pgGsh2,mRNA. Dorsal views, anterior to left, injected side at top and marked with asterisk). Injected region indicated by b-galactosidase co-injectionand red-gal staining (pink). G–J: Expression of (G,H) Iro2 and (I,J) Iro3 in animal cap explants from (G,I) control X. laevis embryos and (H,J) embryosinjected with 40 pg Gsh2-VP16 mRNA. K–N: In situ hybridization analysis of (K,M) Iro2 and (L,N) Iro3 expression in neurula stage 14 X. tropicalisembryos unilaterally injected with (K,L) 15pg Gsh2-VP16-GR mRNA with dexamethasone treatment with stage 11or (M,N) Gsh2- and Gsh1-antisensemorpholinos. O: Reverse transcriptase-PCR (RT-PCR) analysis of expression of Iro3 in animal cap explants taken from uninjected X. laevis controlembryos or embryos injected with 20 pg of Gsh2-VP16-GR, and treated with cycloheximide (CHX) and/or dexamethasone (DEX) as indicated.

development, we designed transla-tion-blocking antisense morpholinooligos (AMOs) targeted against Gsh1and Gsh2. We have previously charac-terized these AMOs and have shownthat they effectively block the transla-tion of target mRNAs coding for epi-tope-tagged versions of the Gsh1 andGsh2 proteins respectively (Winter-bottom et al., 2010).

Expression of Iro2 and Iro3 was an-alyzed by in situ hybridization inembryos unilaterally injected with acombination of Gsh1 and Gsh2 AMOs.Iro2 expression is more intense on theinjected side (89%, n ¼ 19, Fig. 3M)and Iro3 expression is expanded inthe injected side, particularly in theindicated area (red arrows; 66%, n ¼18, Fig. 3N). The anterior limit of Iro3expression is at the midbrain–hind-brain junction, indicating that thisarea of Iro3 up-regulation is a poste-rior region of the presumptive hind-brain (Bellefroid et al., 1998). We notethat the presumptive hindbrainregion of the open neural plate is aprominent domain of Gsh1 and Gsh2expression (Illes et al., 2009).

Antimorphic Gsh2 Activates

Iro3 Expression in the

Absence of Protein Synthesis

The experiments described aboveshow that Iro3 expression in the neu-ral plate is repressed by Gsh2 pro-teins and is up-regulated by an anti-morphic form of Gsh2 in vivo and intissue explants. However, theseexperiments do not indicate whetherGsx directly regulates the transcrip-tion of the Iroquois genes, or affectsthem indirectly by means of the regu-lation of intermediate factors. Toaddress this question, animal capexplants were taken from embryosinjected with inducible antimorphicGsh2, Gsh2-VP16-GR, and cultured inthe presence of dexamethasone (DEX)and/or the protein synthesis inhibitorcycloheximide (CHX). Explants werethen analyzed by RT-PCR for expres-sion of Iro3 (Fig. 3O).

Iro3 expression is not detected inuntreated explants. We find that cy-cloheximide treatment alone activateslow-level Iro3 expression in unin-jected explants. Such effects havebeen previously reported and mightbe due to the loss of labile transcrip-

tional repressors of Gsh2 or increasedstability of Gsh2 mRNA following theloss of labile degradative enzymes(Edwards and Mahadevan, 1992).However, induction of Gsh2-VP16 ac-tivity with DEX in the presence or ab-sence of CHX leads to higher levels ofIro3 expression than with CHX alone.These data indicate that regulation ofIro3 expression by Gsh2 does notrequire the translation of additionalintermediate proteins.

DISCUSSION

Gsx Proteins Are Conserved

Transcriptional Repressors

In this report, we have used repressor/activator domain-swap mutants toinvestigate the activity of Gsx proteinsin the developing Xenopus embryo.Comparison of the effects of the mu-tant and wild-type proteins on embryophenotype, and developmental signal-ing, indicate that Gsh1 and Gsh2 func-tion as transcriptional repressors. Thisis in keeping with the previous identi-fication of an Engrailed homologytype 1 (Eh1) domain, which mediatestranscriptional repression throughinteraction with members of the TLE/Groucho family of co-repressors, atthe N-terminus of both proteins (Illeset al., 2009). This domain is highlyevolutionarily conserved, being pres-ent in the Gsx proteins of higher ver-tebrates, and in their homologue, Ind,in Drosophila and other insect species(Von Ohlen et al., 2007a; Von Ohlenand Moses, 2009).

Many transcription factors can actas either activators or repressors,depending on the target gene and bio-logical context. Indeed, it has beenshown that Drosophila Ind contains aputative transcriptional activationdomain and can activate its own tran-scription (Von Ohlen et al., 2007b;Von Ohlen and Moses, 2009). How-ever, there have been no reports, thusfar, of vertebrate Gsx proteins actingas direct transcriptional activators,and the sequence encoding the activa-tion domain in Drosophila Ind is notconserved in vertebrates (Weiss et al.,1998; Von Ohlen and Moses, 2009).Consequently, it seems reasonable toview the Gsx proteins as transcrip-tional repressors, and the Gsx-VP16fusion as an antimorphic Gsx protein.

Gsx Proteins Regulate

Iroquois Gene Expression

in the Neural Plate

In the present study, we show thatGsx proteins are repressors of Iro-quois gene expression and that anti-morphic Gsx protein up-regulatesIro3 in the absence of protein synthe-sis, indicating a close relationshipbetween Gsx proteins and Iroquoisgene expression. However, at presentwe have no indication that Gsx pro-teins bind directly to regulatoryregions of Iroquois genes.The Xenopus Iro1, Iro2, and Iro3

genes code for homeodomain tran-scription factors which are expressedin complex and overlapping domainsin the open neural plate (Bellefroidet al., 1998; Gomez-Skarmeta et al.,1998; Cavodeassi et al., 2001). TheXenopus Iroquois proteins have beenshown to promote the formation ofneural tissue at the expense of non-neural ectoderm (Bellefroid et al.,1998; Gomez-Skarmeta et al., 1998).Previous studies support a mecha-

nism by which Iroquois proteins pro-mote the formation of neural tissue,in part at least, by repressing BMPligand gene expression in the pre-sumptive neural plate. In the case ofIro1 this is mediated by direct inter-action with the BMP4 promoter(Gomez-Skarmeta et al., 2001). Suchregulatory interactions are likely toexplain our observation that preco-cious expression of Gsx proteins pro-motes BMP signaling in the earlyembryo. However, Gsx expressiondoes not begin until after BMP ligandexpression has been excluded fromthe open neural plate. It, therefore,seems likely that the Gsx-Iroquoisaxis does not modulate BMP signalingduring early development but ratheroperates in later development to regu-late other processes.We have previously shown that

Gsh1 and Gsh2 expression within thepresumptive hindbrain territory atneurula stages encompasses at leastpart of the differentiating primaryinterneuron domain (Illes et al., 2009).Moreover, at these stages Gsh2 isexpressed at the lateral boundary ofIro3 expression (Illes et al., 2009). Ini-tially, neural plate expression of Iro3 isin a domain posterior to the midbrain/hindbrain junction encompassing the

Dev

elop

men

tal D

ynam

ics

1426 WINTERBOTTOM ET AL.

presumptive floor plate, primarymotor neuron domain and the pri-mary interneuron domain at its lat-eral boundary (Bellefroid et al., 1998).However, Iro3 expression is eventuallydown-regulated in differentiating neu-rons, and highest levels of expressionare detected within the neuronal pro-genitor domain between the primarymotor neuron and interneurondomains (Bellefroid et al., 1998). Inter-estingly, Iro3 overexpression promotesthe expression of early neural markersbut inhibits neuronal differentiation,indicating that high-level Iro3 expres-sion is not compatible with differentia-tion of primary neurons (Bellefroidet al., 1998). In keeping we these obser-vations, we propose that the Gsx genesplay a role in down-regulating expres-sion of Iroquois gene expression in thepresumptive hindbrain region at openneural plate stages and that thisrepression is necessary to allow differ-entiation of primary interneurons. Insupport of this model, we have recentlyshown that expression of the neuronaldifferentiation marker n-tubulin isinhibited in Gsx knockdown embryos(Illes et al., 2009).

Gsx Proteins Are Key

Regulators of Gene

Expression in the Open

Neural Plate

Our previous work indicates thattranscriptional repression mediated byGsx factors is involved in establishingthe lateral expression boundary of theDbx1 homeobox gene (Winterbottomet al., 2010). Dbx1 is expressed in theneuronal progenitor domain between

differentiating primary motor neuronsand interneurons in the intermediateregion of the amphibian neural plate.As is the case with Iro3, Dbx1 hasbeen implicated in the inhibition ofneuronal progenitor differentiation(Gershon et al., 2000; Winterbottomet al., 2010. Our latest findings indi-cate that Gsx proteins have a generalrole of promoting primary interneurondifferentiation through repression ofmultiple genes, including Dbx1 andIro3 (Fig. 4).

EXPERIMENTAL

PROCEDURES

Xenopus Embryological

Methods and Microinjection

Embryos were generated by in vitrofertilization of eggs laid followinginjection with human chorionic go-nadotrophin. X. laevis embryo culturewas in NAM/10 (Slack and Forman,1980) at 14–24�C. X. tropicalisembryo culture was in MRS/9 (Tindallet al., 2007) before gastrulation, andin MRS/20 thereafter, at 21.5–27�C.Jelly coats were removed with 3%L-cysteine (Sigma) in NAM or MRS/9,pH 7.8–8. Embryos were stagedaccording to (Nieuwkoop and Faber,1967).

Embryos were injected in NAM/3 orMRS/9 plus 3–5% Ficoll (Sigma) atthe two- to four-cell stage using aDrummond microinjector. Unilateralinjections were monitored using co-injection of GFP mRNA and visualiza-tion of fluorescence before fixing, orinjection of b-galactosidase mRNA andstaining with 1.5 mg/ml Red-Gal(Sigma). Microinjected embryos weretransferred to NAM/10 or MRS/20before gastrulation. X. laevis animalcap explants were explanted at blastulastage 9. Animal caps were cultured inNAM/2 in agarose-coated wells, andfixed for in situ hybridization or snap-frozen for RT-PCR as required.

Drug Treatments

To activate the Gsh2-VP16-GR fusionprotein, dexamethasone (DEX) wasadded to the culture medium at theappropriate stage to a final concentra-tion of 10 mM. In the experiments todetermine the directness of regulationby Gsh2, cycloheximide (CHX) was

added to the culture medium at afinal concentration of 10 mg/ml. Ani-mal caps were excised from injectedor control embryos and 15–20 capsper group were cultured in the pres-ence of CHX for 30 min. DEX wasthen added to the appropriate wells,and caps were cultured in CHXþDEXfor 2 hr. Caps were then transferredto into appropriate medium withoutCHX for a further 1.5 hr. At stage 10–10.5, caps were fixed for in situhybridization or snap-frozen for RT-PCR. This protocol has previouslybeen shown to result in effective inhi-bition of protein synthesis, as meas-ured by radio-methionine incorpora-tion (Fisher et al., 2002).

Photography and Sectioning

Whole specimens were photographedusing a SPOT 14.2 Color Mosaic cam-era (Diagnostic Instruments Inc.) andSPOT Advanced software, with aLeica MZ FLIII microscope.For histological sections, animal

caps of the desired stage were fixed in4% paraformaldehyde (Sigma), thenstained with borax carmine (10% bo-rax carmine, 35% ethanol). Caps wereembedded in Paraplast paraffin waxand sectioned using a Bright 5040microtome. The sections were coun-terstained with picro blue black(97.5% saturated picric acid, 2.5% of a1% aqueous solution of naphthaleneblue black) and mounted in Histo-mount (National Diagnostics). Sec-tions were photographed using an 18.2Color Mosaic camera (DiagnosticInstruments Inc.) and SPOTAdvancedsoftware with a Leica DM2500 micro-scope. Images were processed usingAdobe Photoshop Elements 4.0.

DNA Constructs and mRNA

Synthesis

Fusion protein constructs were gener-ated using the CS2þ VP16 and CS2þEnR plasmids as described in (Isaacset al., 1998). The Gsh1-VP16 andGsh2-VP16-GR constructs have beenpreviously described (Winterbottomet al., 2010). Briefly, to produce Gsh1-VP16 and Gsh2-VP16, and Gsh1-EnRand Gsh2-EnR, the coding regions ofthe X. tropicalis Gsh1 and Gsh2 geneswere sub-cloned into the CS2þ VP16or CS2þ EnR plasmids at the 30 ends

Fig. 4. A model for promotion of neuronaldifferentiation by Gsx proteins. [Color figurecan be viewed in the online issue, which isavailable at wileyonlinelibrary.com.]

Dev

elop

men

tal D

ynam

ics

GSX REPRESSION OF IROQUOIS GENES 1427

of the VP16 or EnR sequences, at theXhoI/XbaI site. To generate the Gsh2-VP16-GR construct, the ligand bindingdomain from the human glucocorticoidreceptor alpha (GR) gene was sub-cloned from pSP64T-MyoDGR (Kolmand Sive, 1995) into the XhoI site ofGsh2-VP16, between the VP16 andGsh2 sequences. The PCR primersused are given in Supp. Table S1.

For synthesis of functional cappedmRNA, plasmids were linearized usingenzymes indicated in Supp. Table S2,and transcribed using the SP6MegaScript transcription kit (Ambion),except Iro1-VP16 which was tran-scribed with T3 RNA polymerase,according to a modified protocoldescribed in (Isaacs et al., 1998).

Antisense Morpholino Oligos

The translation-blocking AMOs tar-geted against X. tropicalis Gsh1 orGsh2 used in this report have beenpreviously characterized (Winterbot-tom et al., 2010). A standard controlmorpholino was also used. All AMOswere designed and synthesized byGeneTools. Sequences are as follows:Gsh1-AMO: 50-CCACTAGGAAGGATCGAGGCATGAG-30; Gsh2-AMO: 50-AGCTCAGCAGTCAGACAGCTCCTTC-30;Control MO: 50-CCTCTTACCTCAGTTACAATTTATA-30. Bold type on Gsh1-AMO indicates position of initiatingATG on the complementary genesequence. Gsh2-AMO ends 50 of theinitiating ATG.

First-Strand cDNA Synthesis

and RT-PCR

Embryos and animal caps for RT-PCRwere snap-frozen on dry ice and totalRNA extraction was carried out usingTri-Reagent (Sigma). First-strandcDNA was synthesized using theCloned AMV First-Strand cDNA Syn-thesis Kit (Invitrogen), with OligodTprimers and 2–3 mg of RNA per sam-ple. PCR reactions were conductedusing 2 ml first-strand cDNA and 2�PCR Master Mix (Promega). Primerswere designed to amplify a 200- to500-bp mRNA fragment, when possi-ble from a region flanking an intronin the genomic sequence. Sequencesof the primers used are given in Supp.Table S3.

In Situ Hybridization

Analysis

Supplementary Table S4 gives detailsof the linearization and transcriptionof plasmids for generation of in situhybridization probes. Transcriptionswere performed using 10� digoxigenin(DIG) RNA labeling mix (Roche).In situ hybridization was carried outas described in (Harland, 1991), withthe modifications given in (Pownallet al., 1996). Hybridizing probes weredetected using an anti-DIG AP-coupledantibody (Roche) and BM purple(Roche) as the precipitating substrate.

REFERENCES

Bellefroid EJ, Kobbe A, Gruss P, Pieler T,Gurdon JB, Papalopulu N. 1998. Xiro3encodes a Xenopus homolog of the Dro-sophila Iroquois genes and functionsin neural specification. EMBO J 17:191–203.

Cavodeassi F, Modolell J, Gomez-Skar-meta JL. 2001. The Iroquois family ofgenes: from body building to neural pat-terning. Development 128:2847–2855.

Chitnis AB. 1999. Control of neurogene-sis–lessons from frogs, fish and flies.Curr Opin Neurobiol 9:18–25.

Chitnis A, Kintner C. 1995. Neural induc-tion and neurogenesis in amphibianembryos. Perspect Dev Neurobiol 3:3–15.

Conlon FL, Sedgwick SG, Weston KM,Smith JC. 1996. Inhibition of Xbratranscription activation causes defectsin mesodermal patterning and revealsautoregulation of Xbra in dorsal meso-derm. Development 122:2427–2435.

Dale L, Howes G, Price BM, Smith JC.1992. Bone morphogenetic protein 4: aventralizing factor in early Xenopus de-velopment. Development 115:573–585.

Edwards DR, Mahadevan LC. 1992. Proteinsynthesis inhibitors differentially super-induce c-fos and c-jun by three distinctmechanisms: lack of evidence for labilerepressors. EMBO J 11:2415–2424.

Fisher ME, Isaacs HV, Pownall ME. 2002.eFGF is required for activation ofXmyoD expression in the myogenic celllineage of Xenopus laevis. Development129:1307–1315.

Forehand CJ, Farel PB. 1982. Spinal corddevelopment in anuran larvae: I. Pri-mary and secondary neurons. J CompNeurol 209:386–394.

Frobius AC, Seaver EC. 2006. ParaHox geneexpression in the polychaete annelid Capi-tella sp. I. Dev Genes Evol 216:81–88.

Galliot B, Quiquand M, Ghila L, de RosaR, Miljkovic-Licina M, Chera S. 2009.Origins of neurogenesis, a cnidarianview. Dev Biol 332:2–24.

Gershon AA, Rudnick J, Kalam L, Zim-merman K. 2000. The homeodomain-containing gene Xdbx inhibits neuronaldifferentiation in the developingembryo. Development 127:2945–2954.

Glavic A, Gomez-Skarmeta JL, Mayor R.2001. Xiro-1 controls mesoderm patterningby repressing bmp-4 expression in the Spe-mann organizer. Dev Dyn 222:368–376.

Gomez-Skarmeta JL, Glavic A, de laCalle-Mustienes E, Modolell J, MayorR. 1998. Xiro, a Xenopus homolog of theDrosophila Iroquois complex genes, con-trols development at the neural plate.EMBO J 17:181–190.

Gomez-Skarmeta J, de La Calle-Mus-tienes E, Modolell J. 2001. The Wnt-activated Xiro1 gene encodes arepressor that is essential for neural de-velopment and downregulates Bmp4.Development 128:551–560.

Graff JM, Thies RS, Song JJ, Celeste AJ,Melton DA. 1994. Studies with a Xeno-pus BMP receptor suggest that ventralmesoderm-inducing signals overridedorsal signals in vivo. Cell 79:169–179.

Harland RM. 1991. In situ hybridization:an improved whole-mount method forXenopus embryos. Methods Cell Biol36:685–695.

Hartenstein V. 1989. Early neurogenesis inXenopus: the spatio-temporal pattern ofproliferation and cell lineages in the em-bryonic spinal cord. Neuron 3:399–411.

Hartenstein V. 1993. Early pattern ofneuronal differentiation in the Xenopusembryonic brainstem and spinal cord. JComp Neurol 328:213–231.

Illes JC, Winterbottom E, Isaacs HV. 2009.Cloning and expression analysis of the an-terior parahox genes, Gsh1 and Gsh2 fromXenopus tropicalis. Dev Dyn 238:194–203.

Isaacs HV, Pownall ME, Slack JM. 1998.Regulation of Hox gene expression andposterior development by the Xenopuscaudal homologue Xcad3. EMBO J 17:3413–3427.

Kintner C. 1992. Molecular bases of earlyneural development in Xenopus embryos.Annu Rev Neurosci 15:251–284.

Kolm PJ, Sive HL. 1995. Efficient hor-mone-inducible protein function in Xen-opus laevis. Dev Biol 171:267–272.

Nieuwkoop PD, Faber J. 1967. NormalTable of Xenopus laevis (Daudin). Am-sterdam: North Holland.

Pownall ME, Tucker AS, Slack JM, IsaacsHV. 1996. eFGF, Xcad3 and Hox genesform a molecular pathway that estab-lishes the anteroposterior axis in Xeno-pus. Development 122:3881–3892.

Slack JM, Forman D. 1980. An interac-tion between dorsal and ventral regionsof the marginal zone in early amphib-ian embryos. J Embryol Exp Morphol56:283–299.

Suzuki A, Thies RS, Yamaji N, Song JJ,Wozney JM, Murakami K, Ueno N. 1994.A truncated bone morphogenetic proteinreceptor affects dorsal-ventral patterningin the early Xenopus embryo. Proc NatlAcad Sci U S A 91:10255–10259.

Tindall AJ, Morris ID, Pownall ME, IsaacsHV. 2007. Expression of enzymesinvolved in thyroid hormone metabolismduring the early development of Xeno-pus tropicalis. Biol Cell 99:151–163.

Von Ohlen T, Syu LJ, Mellerick DM.2007a. Conserved properties of the

Dev

elop

men

tal D

ynam

ics

1428 WINTERBOTTOM ET AL.

Drosophila homeodomain protein, Ind.Mech Dev 124:925–934.

Von Ohlen TL, Harvey C, Panda M.2007b. Identification of an upstreamregulatory element reveals a novelrequirement for Ind activity in main-taining ind expression. Mech Dev124:230–236.

Von Ohlen TL, Moses C. 2009. Identifica-tion of Ind transcription activation and

repression domains required for dorso-ventral patterning of the CNS. MechDev 126:552–562.

Valerius MT, Li H, Stock JL, WeinsteinM, Kaur S, Singh G, Potter SS. 1995.Gsh-1: a novel murine homeobox geneexpressed in the central nervous sys-tem. Dev Dyn 203:337–351.

Weiss JB, Von Ohlen T, Mellerick DM,Dressler G, Doe CQ, Scott MP. 1998.

Dorsoventral patterning in the Drosoph-ila central nervous system: the interme-diate neuroblasts defective homeoboxgene specifies intermediate column iden-tity. Genes Dev 12:3591–3602.

Winterbottom EF, Illes JC, Faas L, IsaacsHV. 2010. Conserved and novel rolesfor the Gsh2 transcription factor in pri-mary neurogenesis. Development 137:2623–2631.

Dev

elop

men

tal D

ynam

ics

GSX REPRESSION OF IROQUOIS GENES 1429