Developmental Biology 213, 170–179 (1999)Article ID dbio.1999.9368, available online at http://www.idealibrary.com on

A Role for xGCNF in Midbrain–HindbrainPatterning in Xenopus laevis

Kening Song,1,2 Ken-Ichi Takemaru,1 and Randall T. MoonHoward Hughes Medical Institute, Department of Pharmacology,and Center for Developmental Biology, University of WashingtonSchool of Medicine, Seattle, Washington 98195

Cells in the presumptive neural ectoderm of Xenopus are committed to neural fate through a process called neuralinduction, which may involve proteins that antagonize BMP signaling pathways. To identify genes that are induced by theBMP antagonists and that may be involved in subsequent neural patterning, we used a suppression PCR-based subtractionscreen. Here we investigate the prospective activities and functions of one of the genes, a nuclear orphan receptor previouslydescribed as xGCNF. In animal cap assays, xGCNF synergizes with ectopic chordin to induce the midbrain–hindbrainmarker engrailed-2 (En-2). In Keller explants, which rely on endogenous factors for neural induction, similar increases inEn-2 are observed. Expression in embryos of a dominant interfering form of xGCNF reduces the expression of endogenousEn-2 and Krox-20. These gain-of-function and prospective loss-of-function experiments, taken with the observation thatxGCNF is expressed in the early neural plate and is elevated in the prospective midbrain–hindbrain region, whichsubsequently expresses En-2, suggest that xGCNF may play a role in regulating En-2 and thus midbrain–hindbrainidentity. © 1999 Academic Press

Key Words: neural induction; xGCNF; nuclear orphan receptor; Engrailed-2; pattern formation; midbrain–hindbrain;

suppression PCR.

w1

fte1B(iaan

INTRODUCTION

The initial step of vertebrate neural development is thecommitment of the prospective ectoderm to a neural fate, aprocess known as neural induction. The committed neuraltissue then undergoes regional specification and eventuallydifferentiates into diverse neural structures. The classicexperiments by Spemann and Mangold (1924) demonstratedthis process by transplanting the dorsal lip from gastrulaembryos to the ventral side of host embryos and observingthe formation of secondary axis which included a completeset of neural tissues. The induction is likely to be mediatedby diffusible factors emanating from the gastrula organizer(Saxen, 1961; Gimlich and Cooke, 1983), such as noggin(Smith and Harland, 1992), chordin (Sasai et al., 1994), andfollistatin (Hemmati-Brivanlou et al., 1994). These factorscan directly induce neural tissue in ectoderm explants

1 Contributed equally.

2 Current address: The Genetics Institute, 87 Cambridge Park

Drive, Cambridge, MA 02140.

170

ithout inducing mesoderm (Lamb et al., 1993; Sasai et al.,995).Several lines of evidence link neural induction by these

actors to BMP signaling pathways. BMP4 has been showno inhibit neural fate and to induce epidermis in dissociatedctodermal explants (Wilson and Hemmati-Brivanlou,995). The expression of dominant negative BMP ligands orMP receptors in animal caps promotes neural cell fate

Hawley et al., 1995). Furthermore, the distinct neuralnducers noggin and chordin both bind BMPs with highffinity and prevent activation of their receptors (Piccolo etl., 1996; Zimmerman et al., 1996). This interesting mecha-ism is conserved in Drosophila (Sasai et al., 1995; Schmidt

et al., 1995). Specifically, decapentaplegic (dpp) and shortgastrulation (sog), related to vertebrate BMP and chordin,respectively, have antagonistic effects in dorsoventral pat-terning (Ray et al., 1991; Ferguson and Anderson, 1992;Francois et al., 1994). These findings favor the notion thatvertebrate neural induction is a default pathway that resultsfrom the elimination of BMP signaling by neural inducers,

using an evolutionarily conserved mechanism (reviewed inWilson and Hemmati-Brivanlou, 1997).

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

All rights of reproduction in any form reserved.

MtnftHi11mbeuctrIpircr

T

aeXicfRipae

171xGCNF in Xenopus Midbrain–Hindbrain Patterning

Neural tissues induced in response to noggin or chordinexpress anterior neural molecular markers (Lamb et al.,1993; Hemmati-Brivanlou et al., 1994; Sasai et al., 1995;

cGrew et al., 1995; Bouwmeester et al., 1996), suggestinghat these factors do not act alone to induce the full range ofeural cell fates. Later in development, this neural tissue isurther specified to a more posterior character prior toerminal cellular differentiation (reviewed by Wilson andemmati-Brivanlou, 1997). Both FGFs and Wnts have been

mplicated in posteriorization of neural tissue (Green et al.,992; Kengaku and Okamoto, 1995; Lamb and Harland,995; McGrew et al., 1995, 1997), though the specificechanisms remain largely unknown. To contribute to a

etter understanding of the pathway by which prospectivectoderm becomes committed to a neural cell fate, wendertook a screen using PCR-based subtraction to isolateDNA clones enriched in ectodermal explants in responseo chordin. One clone turned out to be a nuclear orphaneceptor previously described as xGCNF (Joos et al., 1996).n the present study we employ both gain-of-function androspective loss-of-function assays to test whether xGCNFs a candidate for participating in neural patterning. Weeport that it has a pattern of expression and an activityonsistent with a role in patterning the midbrain–hindbrainegion.

MATERIALS AND METHODS

Embryonic Manipulations

Culture of Xenopus embryos and microinjection of syntheticRNA were performed as described in Moon and Christian (1989).Embryo staging was according to Nieuwkoop and Faber (1967).Animal caps were prepared from stage 8–9 embryos and cultured tothe desired stages as described previously (McGrew et al., 1995).Keller sandwiches were prepared from stage 10.25 embryos asdescribed (Doniach et al., 1992).

PCR-Based Subtractive Hybridization

RNA was isolated by the acidic phenol method (Chomczynskiand Sacchi, 1987), and poly(A) RNA was selected (Sambrook et al.,1992). cDNA synthesis and a PCR-based subtractive hybridizationwere performed by using the PCR-Select cDNA Subtraction kitfrom Clontech (PT1117-1) following the manufacturer’s recom-mendations. The final PCR products were resolved on a polyacryl-amide gel. The DNA bands representing differentially expressedsequences were cut out of the gel and eluted. The eluted cDNAswere amplified by one round of PCR (12 cycles) before being cloneddirectly into Bluescript vector (Stratagene). The cDNAs then wereanalyzed by sequencing, Northern blot, and in situ hybridization.

RT-PCR and Northern Blots

Total RNA was isolated from explants or embryos at differentstages and further treated by RQ1 DNase (Promega). Reverse-transcription PCR (RT-PCR) was carried out as described previ-

ously (Lai et al., 1995). The primers used in this study weredescribed previously in Hemmati-Brivanlou and Melton (1994) and

io

Copyright © 1999 by Academic Press. All right

Lai et al. (1995). The primer sequences for amplifying the xGCNFcoding region (Joos et al., 1996) were 59 CGC ACT GGT ATG AGA

GG ACA C 39 and 59 TTA GCA TCT CTC TCA CTC CTT G 39.Where indicated, RNA samples (10 mg per lane) were resolved on

a 1% agarose–3% formaldehyde gel. Transfer, hybridization, andwashing were performed as described in Sambrook et al. (1989).The cDNA clone of xGCNF, corresponding to nucleotides 325–700of the xGCNF coding region, was labeled using a random primingkit from Ambion to generate a DNA probe.

In Situ Hybridization

Whole-mount in situ hybridization was performed as describedpreviously (Harland, 1991). The digoxigenin probe for xGCNF wasgenerated by T7 RNA polymerase using CS2-xGCNF linearizedwith HindIII as template.

Constructs and Synthetic mRNAs

The coding region of xGCNF amplified by RT-PCR was gelpurified and subcloned into plasmid CS21 at the StuI site. Theresulting plasmid was named CS2-xGCNF and the sequence wasverified by sequencing. Synthetic capped RNA for xGCNF wastranscribed with SP6 RNA polymerase from NotI-linearized CS2-xGCNF using the MessageMachine kit (Ambion). CS2-xGCNF waslinearized with HindIII and transcribed with T7 polymerase togenerate digoxigenin-labeled probe for in situ hybridization (Har-land, 1991). To make a dominant interfering construct of xGCNF,CS2-xGCNF was digested with BglII and XhoI and blunted and the4.3-kb fragment purified; pSp64T EnR was digested with EcoRI andBamHI and blunted and the 1.2-kb fragment purified. The 4.3- andthe 1.2-kb fragments were then ligated to generate DBD-EnR,which contains the DNA binding domain of xGCNF (Joos et al.,1996) fused to the Drosophila engrailed repressor domain (Han andManley, 1993). To make capped RNA, DBD-EnR was linearizedwith XbaI and transcribed with SP6. Capped chordin RNA wasproduced using pSp65T-chd as described in Sasai et al. (1995).

RESULTS

xGCNF Is Expressed in the Neural Plate and ItsExpression Is Elevated by Chordin

While animal caps of Xenopus blastula-stage embryosdifferentiate in vitro into epidermis, they can be diverted toa neural fate by prior injection of RNA encoding noggin(Lamb et al., 1993) or chordin (Sasai et al., 1995). We tookdvantage of these observations to generate neural andpidermal tissue from the same explanted region of theenopus embryo and then used a subtractive strategy to

solate genes induced in response to neural induction byhordin. The PCR-based subtraction employed cDNA poolsrom animal caps of embryos injected with either chordinNA or BMP4 RNA as diagrammed in Fig. 1. The BMP4

njection was used rather than a control RNA since itromotes an ectodermal cell fate in animal caps (Wilsonnd Hemmati-Brivanlou, 1995), thus magnifying any differ-nces between the two pools and increasing the chance of

solating differentially expressed genes. Several clones werebtained and further tested by Northern blot analysis, DNA

s of reproduction in any form reserved.

tstX(

aNta(B2(atiTa

ac

sne3enwpoaftFs(

172 Song, Takemaru, and Moon

sequencing, and preliminary in situ hybridization to ensurehat candidates for further study were expressed in neurula-tage embryos. A search of GenBank revealed that one ofhe clones that was expressed in neurula was also theenopus homologue of murine germ cell nuclear factor

xGCNF), a nuclear orphan receptor (Joos et al., 1996).To test whether xGCNF is differentially expressed in

nimal caps from embryos injected with chordin RNA,orthern blot analysis was performed. When sibling con-

rol embryos have reached stage 12.5, xGCNF is expressedt high levels in animal caps from chordin-injected embryosFig. 2, lane 1), while the expression in animal caps fromMP4-injected embryos is substantially lower (Fig. 2, lane), similar to that in animal caps from uninjected embryosdata not shown). As the animal caps of uninjected embryosre cultured to embryonic stage 14 the levels of xGCNFranscripts increase to levels comparable to those observedn stage 12.5 chordin-injected animal caps (data not shown).

FIG. 1. Identification of xGCNF as a gene differentially expressedin response to chordin. Two cell embryos were injected with RNAencoding either chordin (2 ng per embryo) or BMP4 (0.5 ng perembryo) at the animal pole. Animal cap explants were preparedfrom the injected embryos at stage 8.5 and cultured to lategastrula–early neurula stage (around stage 12.5), and poly(A) RNAswere then isolated and used to generate cDNAs. A subtractedhybridization between cDNA from chordin-injected caps (tester)and from animal caps injected with BMP4 (driver) was thenperformed to isolate sequences that are differentially expressed inchordin-injected animal caps, yielding xGCNF.

hus, injection of chordin RNA promotes a more rapidccumulation of xGCNF, but as previously noted (Joos et l

Copyright © 1999 by Academic Press. All right

l., 1996) this transcript normally accumulates in animalap explants.More detailed in situ hybridization with xGCNF anti-

ense RNA probes then revealed that while transcripts areot detected in early gastrula-stage embryos (Fig. 3A),xpression is evident beginning at the neurula stage (Fig.B). Expression increases during neurulation, with greatestxpression anteriorly, decreasing toward the posterior, andegligible expression in the median of the neural plate,hich represents the future floorplate (Fig. 3C). Expressionersists in tailbud embryos (Fig. 3D). While these patternsf expression generally agree with an earlier report (Joos etl., 1996), we also observed a band of high expression at theront of the neural plate, approximately where the presump-ive midbrain–hindbrain boundary will form (arrowheads,ig. 3C). This band of expression can be seen as early astage 13, but becomes less prominent beginning at stage 14data not shown).

FIG. 2. Total RNA (10 mg per lane) was run on an agarose gel,transferred to a nylon membrane, and hybridized with a xGCNFDNA probe. Transcripts of a single size of about 10 kb werestrongly expressed in chordin-injected caps (lane 1, upper blot). ThexGCNF band was substantially lower in BMP4-injected animalcaps (lane 2). 28S ribosomal RNA was used as a loading control

(lower bands), and reprobing with a max cDNA confirmed equaloading (data not shown).

s of reproduction in any form reserved.

li

axytwdd

ilbud

173xGCNF in Xenopus Midbrain–Hindbrain Patterning

xGCNF Synergizes with Chordin to Induce theMidbrain–Hindbrain Marker En-2 in Animal Caps

The relative lack of expression of xGCNF during earlygastrulation suggests that it acts later than initial neuralinduction. Its elevated expression in the anterior neuralplate (Fig. 3C) in the presumptive midbrain–hindbrainregion suggests that it may be involved in regional neuralpatterning. To examine this potential activity of xGCNF,we cloned the remainder of its coding region (Joos et al.,1996) by RT-PCR and prepared an expression construct.

As a transcription factor (Joos et al., 1996), xGCNF maymediate the function of neural inducers by regulating thetranscription of neural genes. We first tested this possibilityin animal cap assays (Fig. 4A). Capped xGCNF RNA wassynthesized and injected into the animal/dorsal side of two-to four-cell embryos. Animal caps were prepared at stage 8and cultured until the sibling embryos reached stage 22 to24. Expression of a number of regionally restricted neural

FIG. 3. The spatial pattern of expression of xGCNF during develostage embryos using xGCNF antisense probe. (A) A stage 10 emindicates the emerging dorsal lip. (B) A stage 12.5 embryo. Transcrof the neural plate. (C) A stage 13.5 embryo. Strong xGCNFanterior-to-posterior gradient of xGCNF signal with the higher exwhich represents the future floorplate. Stripes of strong staining (whapproximately to the midbrain–hindbrain. (D) A stage 23 embryo (taexpression was seen in the somites and in the branchial arches.

markers was measured by RT-PCR. xGCNF on its own (Fig.4A, lane 5, and Fig. 5B, lane 6) does not induce significant

R3

Copyright © 1999 by Academic Press. All right

evels of region-specific neural markers like Xanf-2 (pitu-tary, Lai et al., 1995), Otx-A (forebrain, Lai et al., 1995),

En-2 (midbrain–hindbrain boundary, Hemmati-Brivanlouet al., 1991), Krox-20 (hindbrain, Bradley et al., 1993), andHoxB9 (spinal cord, Sharpe et al., 1987), nor does it inducethe panneural marker NCAM (Kintner and Melton, 1987) atRNA dosages varying from 0.1 to 2 ng (data not shown).Therefore, xGCNF is not sufficient for neural induction.

Expression of chordin in animal caps induces the expres-sion of the panneural marker, NCAM, as well as regional-specific markers of anterior character (Sasai et al., 1995),nd likely other factors involved in neural induction. SinceGCNF is an orphan nuclear receptor induced by chordin,et insufficient to induce neural gene expression, its activi-ies may require ligands or other protein factors. To testhether xGCNF functionally interacts with chordin or itsownstream effectors, we coinjected RNAs encoding chor-in and xGCNF and analyzed neural gene expression by

nt. Whole-mount in situ hybridization was performed on differentZygotic xGCNF transcripts were not detected. Solid arrowhead

or xGCNF were first detected at this stage in the anterior portionssion was observed in the neural plate. There is an apparention to the anterior. The expression is excluded from the midline,rrowheads) were seen at the front of the neural plate corresponding). The expression of xGCNF was decreased at this stage. Persistent

pmebryo.ipts fexprepressite a

T-PCR. Similar to injection of chordin RNA (Fig. 4A, lane. and Fig. 5B, lane 3), coinjection of chordin and xGCNF

s of reproduction in any form reserved.

OmlliR(Isiatt

f s of KL e xG

174 Song, Takemaru, and Moon

RNAs induces the anterior neural markers Xanf-2 andtx-A but does not induce the more posterior neuralarkers Krox-20 and HoxB9 (Fig. 4A, lane 4, and Fig. 5B,

ane 4). However, in contrast to the absent or very lowevels of the midbrain–hindbrain marker En-2 in chordin-njected animal caps, coinjection of chordin and xGCNFNAs significantly increases the levels of En-2 transcripts

Fig. 4A, lane 4 vs lane 3, and Fig. 5B, lane 4 vs lane 3).mportantly, the effects of xGCNF on neural gene expres-ion in these in vitro assays occur without concomitantnduction of dorsal mesoderm, as monitored by RT-PCRnalysis for muscle actin transcripts (Fig. 5B). We conclude

FIG. 4. RT-PCR analysis of neural gene expression. (A) Lane 1, stmarkers. Lane 2, animal caps from uninjected embryos express nexpress anterior neural markers. Lane 4, animal caps from embryosneural markers. Lane 5, xGCNF alone does not induce neural markKrox-20, HoxB9, and EF-1a which serves as a loading control (see Mrom uninjected embryos serve as controls. Lane 2, RT-PCR analysiane 3, RT-PCR of Keller explants from embryos injected with th

hat xGCNF regulates En-2 expression during neural induc-ion in vitro, in the absence of dorsal mesoderm.

Copyright © 1999 by Academic Press. All right

Expression of xGCNF Increases the Steady-StateLevels of En-2 Expression in Keller SandwichExplants

The results from coexpression of chordin and xGCNF inanimal cap assays suggest that xGCNF may synergize withother factors that are downstream of primary induction. Wefurther tested this hypothesis in Keller sandwich explants,which more closely recapitulate the condition in an embryothan animal cap explants. It is known that neural inductionis mediated by both vertical and planar induction (reviewedin Doniach, 1993). In vertical induction the inducing sig-nals from the invaginating dorsal mesoderm signal verti-

2 whole embryos show a complete anteroposterior range of neuralral markers. Lane 3, animal caps from chordin-injected embryos

cted with chordin and xGCNF express En-2 in addition to anterioranimal caps. PCR primer sets used were for Xanf-2, Otx-A, En-2,

rials and Methods). (B) Lane 1, RT-PCR analysis of Keller explantseller sandwich explants from embryos injected with xGCNF RNA.CNF-EnR RNA.

age 2o neuinje

ers inate

cally to the overlaying ectoderm, while in planar inductionthe signals work horizontally through the plane of the

s of reproduction in any form reserved.

IowctmtaHstE(s

wacRd

175xGCNF in Xenopus Midbrain–Hindbrain Patterning

ectoderm. In Keller sandwiches, which are held in a two-

FIG. 5. (A) A diagram of a dominant interfering construct of xGCNF(xGCNF-EnR). The nucleotides encoding 1–101 amino acids (theDNA binding domain) of xGCNF were cloned in frame to thenucleotides coding for the first 298 amino acids (repressor domain) ofDrosophila engrailed (Han and Manley, 1993). (B) Lane 1, RT-PCRanalysis of embryos reveals detectable levels of neural genes (Xanf-2,Otx-A, En-2) as well as muscle actin (M. actin), a dorsal mesodermalgene. Lane 2, animal cap explants express none of the marker genes.Lane 3, animal caps from embryos injected with chordin RNA (360 pg)express Xanf-2 and Otx-A. Lane 4, animal caps from embryos injected

ith chordin (360 pg) and xGCNF (100 pg) RNAs express En-2 as wells Xanf-2 and Otx-A. Lane 5, animal caps from embryos injected withhordin (360 pg) and xGCNF (100 pg) RNAs, as well as xGCNF-EnRNA (200 pg), express Xanf-2 and Otx-A but not En-2. Lane 6, xGCNFoes not induce any marker in animal cap explants.

dimensional sheet by coverslips, planar signaling is primar-ily responsible for neural induction (Doniach et al., 1992).

Copyright © 1999 by Academic Press. All right

n this assay, xGCNF RNA was injected into the dorsal sidef two- to four-cell-stage embryos and Keller sandwichesere prepared at early gastrula stage (stage 10) and then

ultured until sibling embryos reached stage 16. RT-PCRhen was performed to monitor the levels of different neuralarkers along the anteroposterior axis. Compared to con-

rol explants (Fig. 4B, lane 1), xGCNF injection does notffect levels of Otx-A, Krox-20, or HoxB9 (Fig. 4B, lane 2).owever, the midbrain–hindbrain marker En-2 is increased

ignificantly by xGCNF injection (Fig. 4B, lane 2 comparedo lane 1). This result is consistent with the induction ofn-2 in animal caps coinjected with chordin and xGCNF

Figs. 4A and 5B) and suggests a role for xGCNF in regionalpecification of the midbrain–hindbrain.

Expression of a Dominant Interfering xGCNFReduces the Levels of Endogenous En-2 Expression

We further investigated the roles of xGCNF in a prospec-tive loss-of-function assay. We constructed a dominantinterfering construct of xGCNF (xGCNF-EnR) by fusing theDNA binding domain of xGCNF to the repressor domain ofDrosophila engrailed (Fig. 5A). When injected into em-bryos, the xGCNF-EnR is expected to compete with theendogenous xGCNF for binding to its downstream targets.The repressor domain is expected to keep the target genestranscriptionally inactive, a strategy previously used toeliminate the activity of brachyury in Xenopus (Conlon etal., 1996). Since injection of chordin induces expression ofXanf-2 and Otx-A in animal caps (Fig. 4A, lane 3, and Fig.5B, lane 3), and coinjection of xGCNF induces En-2 inaddition to these genes (Fig. 4A, lane 4, and Fig. 5B, lane 4),we used the expression of En-2 as a readout to test whetherxGCNF-EnR is an inhibitor of xGCNF activity. Support-ingly, injection of xGCNF-EnR and xGCNF RNAs (2:1ratio) along with chordin RNA specifically blocks theinduction of En-2, without affecting the expression ofXanf-2 and Otx-A that are induced by chordin alone (Fig.5B, lane 5 compared to lane 4). These data indicate thatxGCNF-EnR is a repressor of xGCNF function. However,while xGCNF alone or in the presence of chordin does notinduce Krox-20 (Fig. 4A), xGCNF-EnR reduces the expres-sion of this neural marker as well as En-2 in Keller explants(Fig. 4B, lane 3, compared to control lane 1), suggesting thatEn-2 cannot be the sole target of xGCNF-EnR in vivo.

RNA encoding xGCNF, control prolactin, or xGCNF-EnR was then injected into two-cell-stage embryos todetermine if xGCNF-EnR suppressed the expression of En-2and Krox-20 in intact embryos. As initial experimentsdemonstrated that high doses (over 1 ng) of either xGCNFor xGCNF-EnR caused gastrulation defects, we employedlower doses (under 500 pg per embryo), which had no overteffects on development through gastrulation, and grew theembryos to stage 15 prior to fixation and in situ hybridiza-tion. Control embryos injected with RNA encoding prolac-

tin showed robust expression of Otx-A (Fig. 6A), En-2, andKrox-20 (Fig. 6D, mixed in situ hybridization probes, with

s of reproduction in any form reserved.

mdKn

(eO by wH

176 Song, Takemaru, and Moon

En-2 shown by the arrowhead), and HoxB9 (Fig. 6G, Table1). In embryos injected with xGCNF-EnR RNA, the levelsof the forebrain marker Otx-A (Fig. 6C) and spinal cordmarker HoxB9 (Fig. 6I) are similar to those in control (Figs.6A and 6G) embryos and in embryos injected with wild-

FIG. 6. xGCNF-EnR RNA reduces En-2 expression in whole emb0.6 ng per embryo, B, E, and H), or xGCNF-EnR (0.3 ng per embmbryos. Whole-mount in situ hybridization was performed at netx-A (A, B, and C), En-2/Krox-20 (D, E, and F, with En-2 depictedoxB9 (G, H, and I).

type xGCNF RNA (Figs. 6B and 6H). However, xGCNF-EnRconsistently reduces the level of the midbrain–hindbrain

Copyright © 1999 by Academic Press. All right

arker En-2 to low or undetectable levels, with low butetectable levels of Krox-20 (Fig. 6F, arrow denotes reducedrox-20 signal). These data indicate that xGCNF-EnR doesot perturb the expression of very anterior (Otx-A) or

posterior (HoxB9) neural genes, but reduces the expression

RNA encoding prolactin (2 ng per embryo, A, D, and G), xGCNFC, F, and I) was injected into both blastomeres of two-cell-stagea stage using antisense RNA probes for different neural markers:

hite arrowhead and double bands of Krox-20 by black arrows), or

ryos.ryo,urul

of two neural genes expressed between these anterior andposterior markers.

s of reproduction in any form reserved.

tlift

smoaooidTiahe

aW1fmAb

praiccmtdicopssethmxwu

tpeeEDtleKsw(

177xGCNF in Xenopus Midbrain–Hindbrain Patterning

DISCUSSION

We have cloned xGCNF, a prospective nuclear orphanreceptor, by a subtractive PCR-based screen for genes el-evated in response to expression of chordin in Xenopusanimal cap explants. Consistent with the elevation of levelsof xGCNF by chordin, a prospective neural inducer (Sasai etal., 1995), the highest levels of xGCNF during Xenopusdevelopment were observed by in situ hybridization in theneural plate (see also Joos et al., 1996). However, we alsoobserved a heretofore unreported expression of xGCNFaround the presumptive midbrain–hindbrain region. Theprecise site of the midbrain–hindbrain boundary in earlyneural plate-stage embryos was difficult to determine dueto the lack of appropriate molecular markers at this stage.However, we compared the patterns of expression ofxGCNF and En-2 in stage 14 embryos and found that thedomains of expression were very similar (data not shown),though the xGCNF signal is already diminishing in thisregion while that of En-2 is reaching its peak. As this localelevation of xGCNF precedes En-2, and as xGCNF belongso a novel subfamily of nuclear orphan receptors that areigand-dependent transcription factors (Joos et al., 1996), its expressed in an appropriate time and place for a potentialunction in regulating gene expression during patterning ofhe midbrain.

Both gain-of-function and prospective loss-of-functiontudies support a role for xGCNF in patterning theidbrain–hindbrain region. David et al. (1998) report that

verexpression of xGCNF in Xenopus embryos perturbs tailnd somite formation, without overt effects on head devel-pment. In the present study we report that overexpressionf xGCNF in animal cap explants does not appreciablynduce neural markers, while its coexpression with chor-in, but not chordin alone, now induces expression of En-2.his result suggests that xGCNF alone is not sufficient to

nduce En-2 and that its action requires other factors thatre induced by chordin. Considerable research in the mouseas shown that Wnt-1 is required for the maintenance ofxpression of En (McMahon et al., 1992) and that En is a

target of Wnt-1 signaling (Danielian and McMahon, 1996).Further work in mouse has implicated FGF8 and PAX genesin regulation of En (reviewed by Joyner, 1996). Similarly inXenopus, FGF has been shown to induce posterior markers

TABLE 1xGCNF-EnR Reduces Endogenous En-2 Expression

Control embryoprolactin injected (n)

Normal En-2 expression 83% (82)Reduced En-2 expression 17%

such as En-2, Krox-20, and HoxB9 in the presence of neuralinducers (Lamb et al., 1995; Kengaku and Okamoto, 1995),

tc

Copyright © 1999 by Academic Press. All right

nd recent work supports a cooperative interaction betweennt and FGF pathways in neural patterning (McGrew et al.,

997). One mechanism for xGCNF to cooperate with otheractors to regulate En-2 is that xGCNF and other factors

ay bind to different regulatory sites on the En-2 gene.lternatively, xGCNF may regulate En-2 only indirectly,y modulating expression of other genes.The results of the animal cap explants seem to pose a

aradox. We found that levels of xGCNF transcript areelatively low in uninjected animal caps at stage 12.5, butre higher by stage 14, when En-2 expression is establishedn the embryo. Injection of chordin RNA induces preco-ious elevation of xGCNF transcripts in stage 12.5 animalaps, consistent with the observation that chordin is nor-ally expressed earlier in the gastrula (Sasai et al., 1995)

han xGCNF. However, we noticed that chordin by itselfoes not induce significant levels of En-2, although it hasnduced xGCNF, which our data show can work withhordin to induce En-2. This paradoxical lack of inductionf En-2 by chordin, despite the induction of xGCNF, may beartially explained by the following. First, spatial expres-ion patterns of xGCNF show elevated levels in the pro-pective midbrain–hindbrain region compared to other ar-as of the neural plate. Thus, the levels of xGCNFranscripts induced by chordin in animal caps may not beigh enough to induce expression of En-2. Second, xGCNFay be subject to translational controls, since endogenous

GCNF transcripts have very long untranslated regions,hile the injected synthetic xGCNF contains virtually nontranslated region.Analyses based on endogenous neural inducing and pat-

erning signals further support the hypothesis that xGCNFlays a role in midbrain–hindbrain patterning. Ectopicxpression of xGCNF in Keller explants, which rely onndogenous signals for neural induction, elevates levels ofn-2. Conversely, an inhibitory protein consisting of theNA binding domain of xGCNF (Joos et al., 1996) linked to

he Drosophila engrailed repressor domain (Han and Man-ey, 1993), reduced the levels of En-2 and Krox-20 in Kellerxplants and in whole embryos. Since both En-2 androx-20 expression was perturbed in embryos in the presenttudy, and a distinct dominant negative xGCNF interferedith the anterior expression of a retinoic acid receptor

David et al., 1998), it is likely that xGCNF has multiple

xGCNF-injectedembryo (n)

xGCNF-EnR-injectedembryo (n)

80% (80) 13% (93)20% 87%

argets in the early nervous system. Consistent with thisonclusion and our results, David et al. (1998) also report

s of reproduction in any form reserved.

wr

C

C

C

C

D

D

D

D

D

D

F

F

F

G

G

H

H

H

H

H

H

J

J

K

K

L

L

178 Song, Takemaru, and Moon

that their dominant negative xGCNF perturbs head andneural tube formation in Xenopus embryos, as analyzed byhistological studies. Thus, gain-of-function assays, prospec-tive loss-of-function assays, and the pattern of xGCNFexpression are consistent with its being involved inmidbrain–hindbrain patterning. Further work will be re-quired to determine if there are direct xGCNF binding sitesin the regulatory regions of En or Krox-20 genes and

hether genetic reduction of xGCNF function supports aequirement during early development in vertebrates.

ACKNOWLEDGMENTS

We thank Jim Smith, Kathleen Weston, and David Kimelman forproviding engrailed repressor cDNAs and Eddy DeRobertis for thechordin cDNA. We are grateful to the members of our lab fordiscussion and support; Christoph Winkler, Lynn McGrew, andMonica Torres for critically reading the manuscript; and Minh Sumfor preparation of figures. K.S. was an Associate, and R.T.M. is anInvestigator, of the Howard Hughes Medical Institute. K.T. wassupported by a fellowship from the Human Frontiers Foundation.

REFERENCES

Bouwmeester, T., Kim, S.-H., Sasai, Y., Lu, B., and De Robertis,E. M. (1996). Cerberus is a head-inducing secreted factor ex-pressed in the anterior endoderm of Spemann’ organizer. Nature382, 595–601.

Bradley, L., Snape, A., Bhatt, S., and Wilkinson, D. (1993). Thestructure and expression of the Xenopus krox-20 gene: Con-served and divergent patterns of expression in rhombomeres andneural crest. Mech. Dev. 40, 73–84.hen, F., Cooney, A. J., Wang, Y., Law, S. W., and O’Malley, B. W.(1994). Cloning of a novel orphan receptor (GCNF) expressedduring germ cell development. Mol. Endocrinol. 8, 1434–1444.hitnis, A., and Kintner, C. (1996). Sensitivity of proneural genes tolateral inhibition affects the pattern of primary neurons inXenopus embryos. Development 122, 2295–2301.homczynski, P., and Sacchi, N. (1987). Single-step method ofRNA isolation by acid guanidinium thiocyanate–phenol–chloroform extraction. Anal. Biochem. 162, 156–159.onlon, F. L., Sedgwick, S. G., Weston, K. M., and Smith, J. C.(1996). Inhibition of Xbra transcription activation causes defectsin mesodermal patterning and reveals autoregulation of Xbra indorsal mesoderm. Development 122, 2427–2435.ale, L., Howes, G., Price, B. M. J., and Smith, J. C. (1992). Bonemorphogenetic protein 4: A ventralizing factor in early Xenopusdevelopment. Development 115, 573–585.anielian, P. S., and McMahon, A. P. (1996). Engrailed-1 as a targetof the Wnt-1 signalling pathway in vertebrate midbrain develop-ment. Nature 383, 332–334.avid, R., Joos, T. O., and Dreyer, C. (1998). Anteroposteriorpatterning and organogenesis of Xenopus laevis require a correctdose of germ cell nuclear factor (xGCNF). Mech. Dev. 79,137–152.iatchenko, L., Lau, Y.-F. C., Campbell, A. P., Chenchik, A.,Moqadam, F., Huang, B., Lukyanov, S., Lukyanov, K., Gurskaya,

N., Sverdlov, E., and Siebert, P. D. (1996). Suppression subtrac-tive hybridization: A method for generating differentially regu-

Copyright © 1999 by Academic Press. All right

lated or tissue-specific cDNA probes and libraries. Proc. Natl.Acad. Sci. USA 93, 6025–6030.oniach, T. (1993). Planar and vertical induction of anteroposteriorpattern during the development of the amphibian central ner-vous system. J. Neurobiol. 24, 1256–1275.oniach, T., Phillips, C. R., and Gerhart, J. C. (1992). Planarinduction of anteroposterior pattern in the developing centralnervous system of Xenopus laevis. Science 257, 542–545.

ainsod, A., Steinbeisser, H., and De Robertis, E. M. (1994). On thefunction of BMP-4 in patterning the marginal zone of theXenopus embryo. EMBO J. 13, 5015–5025.

erguson, E. L., and Anderson, K. V. (1992). decapentaplegic acts asa morphogen to organize dorsal–ventral pattern in the Drosoph-ila embryo. Cell 71, 451–461.

rancois, V., Solloway, M., O’Neil, J. W., Emery, J., and Bier, E.(1994). Dorsal–ventral patterning of the Drosophila embryo de-pends on a putative negative growth factor encoded by the shortgastrulation gene. Genes Dev. 8, 2602–2616.imlich, R. L., and Cooke, J. (1983). Cell lineage and the inductionof second nervous systems in amphibian development. Nature306, 471–473.reen, J. B. A., New, H. V., and Smith, J. C. (1992). Responses ofembryonic Xenopus cells to activin and FGF are separated bymultiple dose thresholds and correspond to distinct axes of themesoderm. Cell 71, 731–739.an, K., and Manley, J. L. (1993). Functional domains of theDrosophila engrailed protein. EMBO J. 12, 2723–2733.arland, R. M. (1991). In situ hybridization: An improved whole-mount method for Xenopus embryos. Methods Cell Biol. 36,685–695.awley, S. H. B., Wunnenber-Stapleton, K., Hashimoto, C., Lau-rent, M. N., Watabe, T., Blumberg, B. W., and Cho, K. W. Y.(1995). Disruption of BMP signals in embryonic Xenopus ecto-derm leads to direct neural induction. Genes Dev. 9, 2923–2935.emmati-Brivanlou, A., de la Torre, J. R., Holt, C., and Harland,R. M. (1991). Cephalic expression and molecular characterizationof Xenopus En-2. Development 111, 715–724.emmati-Brivanlou, A., Kelly, O. G., and Melton, D. A. (1994).Follistatin, an antagonist of activin, is expressed in the Spemannorganizer and displays direct neuralizing activity. Cell 77, 283–295.emmati-Brivanlou, A., and Melton, D. A. (1994). Inhibition ofactivin receptor signaling promotes neuralization in Xenopus.Cell 77, 273–281.

oos, T. O., David, R., and Dreyer, C. (1996). xGCNF, a nuclearorphan receptor is expressed during neurulation in Xenopuslaevis. Mech. Dev. 60, 45–57.

oyner, A. L. (1996). Engrailed, Wnt and Pax genes regulatemidbrain–hindbrain development. Trends Genet. 12, 15–20.engaku, M., and Okamoto, H. (1995). bFGF as a possible morpho-gen for the anteroposterior axis of the central nervous system inXenopus. Development 121, 3221–3230.intner, C. R., and Melton, D. A. (1987). Expression of XenopusN-CAM RNA in ectoderm is an early response to neural induc-tion. Development 99, 311–325.

ai, C.-J., Ekker, S. C., Beachy, P. A., and Moon, R. T. (1995).Patterning of the neural ectoderm of Xenopus laevis by theamino-terminal product of hedgehog autoproteolytic cleavage.Development 121, 2349–2360.

amb, T. M., and Harland, R. M. (1995). Fibroblast growth factor is

a direct neural inducer, which combined with noggin generatesanterior–posterior neural pattern. Development 121, 3627–3636.

s of reproduction in any form reserved.

M

M

M

M

R

S

S

S

S

S

S

S

S

S

T

W

W

Z

179xGCNF in Xenopus Midbrain–Hindbrain Patterning

Lamb, T. M., Knecht, A. K., Smith, W. C., Stachel, S. E., Econo-mides, A. N., Stahl, N., Yancopolous, G. D., and Harland, R. M.(1993). Neural induction by the secreted polypeptide noggin.Science 262, 713–718.angelsdorf, D. J., Thummel, C., Beato, M., Herrlich, P., Schutz,G., Umesono, K., Blumberg, B., Kastner, P., Kark, M., Chambon,P., and Evans, R. M. (1995). The nuclear receptor superfamily:The second decade. Cell 83, 835–839.cGrew, L. L., Lai, C.-J., and Moon, R. T. (1995). Specification ofthe anteroposterior neural axis through synergistic interaction ofthe Wnt signaling cascade with noggin and follistatin. Dev. Biol.172, 337–342.cGrew, L. L., Hoppler, S., and Moon, R. T. (1997). Wnt and FGFpathways cooperatively pattern anteroposterior neural ectodermin Xenopus. Mech. Dev., in press.cMahon, A. P., Joyner, A. L., Bradley, A., and McMahon, J. A.(1992). The midbrain–hindbrain phenotype of Wnt-12/Wnt-12

mice results from stepwise deletion of engrailed-expressing cellsby 9.5 days postcoitum. Cell 69, 581–595.

Moon, R. T., and Christian, J. L. (1989). Microinjection and expres-sion of synthetic mRNAs in Xenopus embryos. Technique 1,76–89.

Muskavitch, M. A. (1994). Delta-notch signaling and Drosophilacell fate choice. Dev. Biol. 166, 415–430.

Nieuwkoop, P. D., and Faber, J. (1967). “Normal Table of Xenopuslaevis (Daudin).” North-Holland, Amsterdam.

Padgett, R. W., Savage, C., and Das, P. (1997). Genetic and bio-chemical analysis of TGF beta signal transduction. CytokineGrowth Factor Rev. 8, 1–9.

Piccolo, S., Sasai, Y., Lu, B., and De Robertis, E. M. (1996).Dorsoventral patterning in Xenopus: Inhibition of ventral signalsby direct binding of chordin to BMP-4. Cell 86, 589–598.

ay, R. P., Arora, K., Nusslein-Volhard, C., and Gelbart, W. M.(1991). The control of cell fate along the dorsal–ventral axis of theDrosophila embryo. Development 113, 35–54.

ambrook, J., Fritsch, E. F., and Maniatis, T. (1989). “MolecularCloning—A Laboratory Manual,” 2nd ed. Cold Spring HarborLaboratory Press, Cold Spring Harbor, NY.

asai, Y., Lu, B., Steinbeisser, H., Geissert, D., Gont, L. K., andDeRobertis, E. M. (1994). Xenopus chordin: A novel dorsalizing

Copyright © 1999 by Academic Press. All right

factor activated by organizer-specific homeobox genes. Cell 79,779–790.

asai, Y., Lu, B., Steinbeisser, H., and De Robertis, E. M. (1995).Regulation of neural induction by the chd and Bmp-4 antagonis-tic patterning signals in Xenopus. Nature 376, 333–336.

axen, L. (1961). Transfilter neural induction of amphibian ecto-derm. Dev. Biol. 3, 140–152.

chmidt, J., Francois, V., Bier, E., and Kimelman, D. (1995).Drosophila short gastrulation induces an ectopic axis in Xeno-pus: Evidence for conserved mechanisms of dorsal–ventral pat-terning. Development 121, 4319–4328.

harpe, C. R., Fritz, A., DeRobertis, E. M., and Gurdon, J. B. (1987).A homeobox-containing marker of posterior neural differentia-tion shows the importance of predetermination in neural induc-tion. Cell 50, 749–758.

mith, W. C., and Harland, R. M. (1992). Expression cloning ofnoggin, a new dorsalizing factor localized to the Spemann orga-nizer in Xenopus embryos. Cell 70, 829–840.

pemann, H., and Mangold, H. (1924). Uber inducktion vonembryo-nanlagen durch implantation artfremder organisatoren.Arch. Mikr. Anat. EntwMech. 100, 599–638.

tandart, N., and Jackson, R. J. (1994). Regulation of translation byspecific protein/mRNA interactions. Biochimie 76, 867–879.annahill, D., Isaacs, H. V., Close, M. J., Peters, G., and Slack,M. W. (1992). Developmental expression of the Xenopus int-2(FGF-3) gene: Activation by mesodermal and neural induction.Development 115, 695–702.ilson, P. A., and Hemmati-Brivanlou, A. (1995). Induction ofepidermis and inhibition of neural fate by Bmp-4. Nature 376,331–333.ilson, P. A., and Hemmati-Brivanlou, A. (1997). Vertebrate neuralinduction: Inducers, inhibitors, and new synthesis. Neuron 18,699–710.immerman, L. B. De Jesus-Escobar, J. M., and Harland, R. M.(1996). The Spemann organizer signal noggin binds and inacti-vates bone morphogenetic protein 4. Cell 86, 599–606.

Received for publication May 5, 1999

Revised June 7, 1999

Accepted June 7, 1999

s of reproduction in any form reserved.