Otx2 Is Required to Respond to Signals from Anterior Neural Ridge for Forebrain Specification

Developmental Biology 242, 204–223 (2002) doi:10.1006/dbio.2001.0531, available online at http://www.idealibrary.com on Otx2 Is Required to Respond to Signals from Anterior Neural Ridge for Forebrain Speciﬁcation

E Tian,* Chiharu Kimura,*,† Naoki Takeda,‡ Shinichi Aizawa,*,†,1 and Isao Matsuo*,† *Department of Morphogenesis, Institute of Molecular Embryology and Genetics (IMEG), and †Vertebrate Body Plan Group, RIKEN Center for Developmental Biology, and ‡Division of Transgenic Technology, Center for Animal Resources and Development (CARD), Kumamoto University, Honjo 2-2-1, Kumamoto 860-0811, Japan

Previous analysis employing chimeric and transgenic rescue experiments has suggested that Otx2 is required in the neuroectoderm for development of the forebrain region. In order to elucidate the precise role of Otx2 in forebrain development, we attempted to generate an allelic series of Otx2 mutations by Flp- and Cre-mediated recombination for the production of conditional knock-out mice. Unexpectedly, the neo-cassette insertion created a hypomorphic Otx2 allele; consequently, the phenotype of compound mutant embryos carrying both a hypomorphic and a null allele (Otx2frt-neo/Ϫ) was analyzed. Otx2frt-neo/Ϫ mutant mice died at birth, displaying rostral head malformations. Molecular marker analysis demonstrated that Otx2frt-neo/Ϫ mutant embryos appeared to undergo anterior–posterior axis generation and induction of anterior neuroectoderm normally; however, these mutants subsequently failed to correctly specify the forebrain region. As the rostral margin of the neural plate, termed the anterior neural ridge (ANR), plays crucial roles with respect to neural plate speciﬁcation, we examined expression of molecular markers for the ANR and the neural plate; moreover, neural plate explant studies were performed. Analyses revealed that telencephalic gene expression did not occur in mutant embryos due to defects of the neural plate; however, the mutant ANR bore normal induction activity on gene expression. These results further suggest that Otx2 dosage may be crucial in the neural plate with respect to response to inductive signals primarily from the ANR for forebrain speciﬁcation. © 2002 Elsevier Science (USA) Key Words: forebrain; neural plate; anterior neural ridge; regionalization; hypomorphic allele; Otx2.

INTRODUCTION prised of cells at the rostral margin of the neural plate, e.g., the anterior neural ridge (ANR), and the mid/hindbrain junction, also referred to as the isthmus (Joyner, 1996; Anterior–posterior (A-P) patterning of the vertebrate cen- Shimamura and Rubenstein, 1997; Houart et al., 1998; tral nervous system (CNS) begins prior to and during Joyner et al., 2000). The ANR is involved in the de- gastrulation (Beddingtion and Robertson, 1999). The initial velopment of forebrain, whereas the isthmus functions in process is mediated by signals from the anterior visceral the development of midbrain and cerebellar structures, endoderm (AVE) and primitive streak-derived structures respectively. such as the anterior definitive endoderm (ADE) in mouse The molecular mechanisms underlying inductive activ- (Ang and Rossant, 1993; Beddington and Robertson, 1999; ity of the ANR and patterning of the forebrain have been Kimura et al., 2000). Later, at neural plate stages, local shown by tissue explants and genetic studies (reviewed in signaling centers play crucial roles in refining the identity Rubenstein et al., 1998; Wilson and Rubenstein, 2000). The of neuroectoderm at specific A-P positions (Rubenstein et ANR is a morphologically defined structure located at the al., 1998; Wilson and Rubenstein, 2000). Two local signal- junction of the anterior neural plate and the nonneural ing centers have been identified. These regions are com- ectoderm (Couly and Le Douarin, 1988; Eagleson et al., 1995). Notably, removal of the ANR results in loss of 1 To whom correspondence should be addressed. Fax: ϩ81-96- expression of the telencephalic gene BF-1; moreover, addi- 373-6586. E-mail: [email protected]. tional ANR can induce ectopic BF-1 expression in mouse

FIG. 1. Generation of an allelic series of Otx2 mutations. (A) Strategy for production of an allelic series of mutations at the Otx2 locus. Schematic representation of the Otx2 wild-type allele (a), the targeting vector (b), the Otx2frt-neo allele (c), the Otx2ﬂox allele (d), and the Otx2⌬pro-ex1 allele (e), respectively. The Otx2frt-neo allele is obtained by the standard method of homologous recombination (c). Mice heterozygous for the Otx2frt-neo allele are crossed with Flpe transgenic mice to generate animals bearing the Otx2ﬂox allele (d), in which the

© 2002 Elsevier Science (USA). All rights reserved. 206 Tian et al. neural plate explants (Shimamura and Rubenstein, 1997). al., 1996, 1997; Acampora et al., 1997). Furthermore, Otx2 Similarly, in zebrafish, telencephalic gene expression is is essential for diencephalon development in cooperation reduced or absent when cells at the margin of the neural with Emx2 (Suda et al., 2001). However, the molecular plate, referred to as row 1 cells, are ablated at the gastrula mechanism by which Otx2 functions in forebrain develop- stage. Additionally, these cells can induce telencephalic ment remains unclear. gene expression when transplanted to more caudal regions In order to elucidate the precise role of Otx2 in forebrain of the neural plate (Houart et al., 1998). Thus, signals from specification at the neural plate stage, we have attempted to the ANR are crucial for telencephalic development. Fur- generate an allelic series of Otx2 mutations through Flp- thermore, Fgf8 expression in the ANR is suggestive of an and Cre-mediated recombination for the production of important role for Fgf8 in forebrain patterning. Indeed, conditional knock-out mice. In the course of the generation FGF8b-soaked beads can restore BF-1 expression in explants of various alleles, we found that the neo-cassette insertion denuded of the ANR (Shimamura and Rubenstein, 1997), creates a hypomorphic Otx2 allele. The phenotype of the and inhibitors of Fgf function reduce BF-1 expression in mutant embryo carrying both a hypomorphic allele and a neural plate (Ye et al., 1998). Telencephalon formation can null allele was analyzed. This investigation revealed that proceed in the absence of Fgf8 in zebrafish and mice mutant embryos failed to form the forebrain region cor- (Shanmugalingam et al., 2000; Wilson and Rubenstein, rectly. Analyses employing molecular markers and neural 2000); however, Fgf8 is not the sole factor for the induction plate explant studies have indicated that Otx2 may be of telencephalon. This observation implies that other sig- required to respond to inductive signals primarily from the naling molecules may be involved in this inductive process ANR with respect to forebrain specification. (Shanmugalingam et al., 2000; Wilson and Rubenstein, 2000). On the other hand, the genetic mechanism governing competence in the prosencephalic neural plate with respect MATERIALS AND METHODS to signals from the ANR remains to be determined. Otx2 has been shown to be essential in forebrain devel- Construction of the Targeting Vector opment; the entire forebrain and midbrain fail to develop in Construction of the targeting vector was effected via initially Otx2 homozygous mutant embryos (Acampora et al., 1995; subcloning a 3-kb EcoRI–SphI fragment of the mouse Otx2 3Ј Matsuo et al., 1995; Ang et al., 1996). Chimeric and genetic homologous region into modified pBluescriptII with a DT-A frag- studies have indicated that Otx2 is required at an earlier ment (Yagi et al., 1993). Subsequently, the PGKneopA-cassette stage prior to and during gastrulation for the formation of (neo-cassette), flanked by frt sites and including a loxP site inserted AVE (Rhinn et al., 1998; Acampora et al., 1998; Kimura et in 5Ј fashion relative to the frt site, was inserted into 5Ј of the Otx2 al., 2000). Nevertheless, analysis of chimeric embryos, 3Ј region adjacent to the NheI and EcoRI sites. Following neo- Otx1 knock-in mutation, and transgenic rescue experi- cassette insertion, a 2.5-kb BamHI–NheI fragment, containing Otx2 promoter and the first coding exon, was inserted 5Ј of the ments have suggested that Otx2 is also required in the neo-cassette. Finally, a SmaI–BamHI 5.2-kb 5Ј Otx2 region possess- et al., neuroectoderm for forebrain development (Acampora ing a loxP site at the 3Ј end was inserted 5Ј homologous of the Otx2 1998; Rhinn et al., 1998, 1999; Suda et al., 1999; Kimura et promoter. The two frt sequences were generated by annealing the al., 2000). Additionally, Otx2 functions in the formation of two synthetic oligo DNA strands (5Ј-CCGGATCCGAAGTTCCT- mesencephalic territory by determining the position of the ATTCCGAAGTTCCTATTCTCTAGAAAGTATAGGAACTTCA- isthmic organizing center cooperating with Otx1 (Suda et CTGCAGGG-3Ј and 5Ј-CCCTGCAGTGAAGTTCCTATACTTT-

neo-cassette is excised. Heterozygotes for the Otx2flox allele are further crossed with Cre transgenic mice to generate mice carrying the Otx2⌬pro-ex1 allele (e), in which the Otx2 promoter and first coding exon are deleted. Horizontal thick lines represent Otx2 genomic DNA, whereas the thin line represents pBluescript sequences. Filled boxes represent Otx2 coding exons. Filled arrowheads and open arrowheads denote loxP and frt sites, respectively. Neor, neomycin-resistant gene with the PGK promoter and polyadenylation signal; DT-A, diphtheria toxin A fragment gene with MC1 promoter. Restriction enzyme sites: B, BamHI; H, HindIII; E, EcoRI; N, NheI; NI, NotI; S, SphI; SI, SalI; Sm, SmaI; X, XhoI. Probes A, B, C, and D are used for Southern blot identification. P1, P2, F1, N1, and R1 show primers utilized for the identification of the different Otx2 alleles by PCR analysis. (B) Identification and characterization of mutant allele by Southern blot analysis. (a) Genomic DNAs from wild-type and mutant mice were digested with HindIII and hybridized with probe A to examine the correct 3Ј recombination. The 5.6- and 3-kb bands represent wild-type and targeted alleles, respectively. (b) Digested with XhoI and hybridized with probe B for the neo-cassette identification. The 8-kb band represents targeted allele. (c) Digested with BamHI and hybridized with probe C for the correct 5Ј recombination and the presence of the first loxP site. Recombinants generate a 7-kb band in addition to the 20-kb band of a wild-type allele. (d) Digested with NheI and hybridized with probe D for identification of Otx2frt-neo (4 kb), Otx2flox (2 kb), and Otx2⌬pro-ex1 (5 kb) alleles, as well as an 8-kb band of wild-type allele. (C) PCR analysis of genotyping for different alleles. (a) PCR amplification employing a combination of primers F1, N1, and R1 generates a 550- and 360-bp product for the Otx2frt-neo and wild-type alleles, respectively. (b) Primers P1 and P2 amplified a 220- and 160-bp product, indicating the presence of the loxP site located 5Ј upstream in the mutant allele and the absence of the loxP site in the wild-type allele, respectively. (c) In the Otx2flox allele, a 560-bp PCR product is amplified by primers F1 and R1. (d) Primers P1 and R1 are used to detect the Otx2⌬pro-ex1 allele (300 bp).

FIG. 2. The Otx2frt-neo allele produces aberrant Otx2 transcripts by the neo-cassette insertion. (A) (a) RT-PCR analysis with primers E1 and N2 utilizing cDNAs from wild-type (left) and Otx2frt-neo/frt-neo homozygous mutant embryo (right) at 8.5 dpc. The 800-bp products in the mutant indicate that the neo-cassette contains a splice acceptor. (b) RT-PCR analysis with primers N1 and E3 indicates the presence of two types of aberrant mRNA transcripts (see below). (c) RT-PCR analysis with primers E1, N0, and E3. The 272-bp products correspond to wild-type mRNA transcripts, whereas the 200-bp products correspond to Otx2 and neo hybrid mRNA, respectively. In the mutant embryo, the level of wild-type mRNA expression is downregulated. (B) (a) Schematic diagram of the Otx2frt-neo allele. Shaded boxes represent Otx2 coding exons; an open box represents neo-cassette. Filled and open arrowheads denote loxP and frt sites, respectively. The arrows indicate primers used in RT-PCR analysis. (b, c) Schematic diagrams representing exon structures of mRNAs produced from this allele. The exon structures were determined by sequence analysis of RT-PCR products. Asterisks indicate the stop codon of neo. Aberrant Otx2 transcripts were also evident in Otx2frt-neo/frt-neo mutant embryos at 6.5- and 7.5-dpc stages (data not shown).

CTAGAGAATAGGAACTTCGGAATAGGAACTTCGGATCCGG), ampliﬁcation (primers: 5Ј-CGGGATCCTGATATCGGCTGGA- and by PCR ampliﬁcation (primers: 5Ј-GCCAAGCTTGAAGTTC- CGTAAACTCCTCTT-3Ј and 5Ј-AAAGATCTCGATCATATTC- CTATTCCGAAGTTC-3Ј and 5Ј-CCGGTACCGAATTCTGAA- AATAACCCTT-3Ј, and 5Ј-AGAGATCTAATAACTTCGTATA- GTTCCTATACTTTCTAG-3Ј), respectively. The two loxP se- GCATA-3Ј and 5Ј-CGGGATCCTTAATATAACTTCGTATA-3Ј, quences were obtained from pBS246 plasmid (GIBCO BRL) by PCR respectively).

FIG. 3. Morphological features of Otx2frt-neo/Ϫ mutant embryos. (A–D) Lateral view of wild-type (A), Otx2ϩ/Ϫ (B), Otx2frt-neo/ϩ (C), and Otx2frt-neo/Ϫ (D) embryos at 18.5 dpc on the CBA genetic background, respectively. Otx2ϩ/Ϫ and Otx2frt-neo/ϩ show no noticeable abnormalities in the CBA background (B, C). Otx2frt-neo/Ϫ mutants fail to develop dorsal and rostral regions of the head; however, they do undergo appropriate formation of ventral portions of the cranial region (D). Skull morphologies of wild-type (E, G) and Otx2frt-neo/Ϫ mutant embryos (F, H) in the CBA background at 18.5 dpc. Ventral views of the whole-mount skull without mandibles (E, F) and separated mandibles (G, H) are shown. Most of the skull elements are formed, although slightly distorted, in Otx2frt-neo/Ϫ mutant embryos. Lateral view (I, J) and sagittal sections (K, L) of wild-type (I, K) and Otx2frt-neo/Ϫ (J, L) embryos at 10.5 dpc. The telencephalon and diencephalon are not formed in mutant embryos (J, L). The isthmic constriction (arrows in I, K) is not clearly observed in mutants (J, L). The ﬁrst branchial arch (ba1) develops normally in the mutant embryo (J, L). Abbreviations: as, alisphenoid; bs, basisphenoid; bo, basioccipital; d, diencephalon; eo, exoccipital; m, mesencephalon; mt, metencephalon; mx, maxillar; pl, palatine; pm, premaxillar; sy, symphysis; t, telencephalon.

Generation and Genotyping of Mice Carrying injected into morula-stage embryos to generate chimera mice. Male the Different Otx2 Alleles chimeras were mated with CBA females in order to produce heterozygous mutants. These heterozygotes were crossed with Flpe The targeting vector was linearized with SalI and introduced transgenic mice carrying the CAG-Flpe transgene (provided by Dr. into TT2 embryonic stem (ES) cells via electroporation as described F. Stewart) in order to delete the neo-cassette, resulting in the elsewhere (Yagi et al., 1993). G418-resistant ES clones were iden- production of Otx2flox heterozygous mice (Fig. 1A). The flox mice tified by PCR and confirmed by Southern blot analysis as described can be inverted into ⌬pro-ex1 heterozygous mice, which are below. Two independent homologous recombinant clones were characterized by deletion of the Otx2 promoter and first coding

FIG. 4. Molecular analysis of anterior neural markers at 9.5 dpc. BF-1 (A, B), Emx2 (C, D), Six3 (E, F), Nkx2.1 (G, H), Fgf8 (I, J), and En2 (K, L) expression by whole-mount in situ hybridization in wild-type (A, C, E, G, I, K) and Otx2frt-neo/Ϫ mutant embryos (B, D, F, H, J, L) at 9.5 dpc. BF-1 expression occurs in the telencephalic region of the wild type (A); however, BF-1 expression is not detected in this region (B). Normally, Emx2 and Six3 are expressed in the dorsal forebrain and most anterior forebrain, respectively (C, E); however, expression of these genes is not detected in mutant embryos (D, F). Nkx2.1 expression is found in the ventral forebrain of the wild type (arrow in G); however, it is not observed in the mutant embryo (H). Fgf8 expression, seen in the telencephalic commissural plate and isthmic constriction in the wild type (arrows in I), is present throughout the entire rostral region in mutant embryos (arrowheads in J). The expression domain of En2, which covers the mesencephalon and anterior metencephalon in the wild type (arrows in K), is shifted anteriorly in mutant embryos (arrows in L). Arrowheads indicate the presence of an En2-negative domain in the most rostral portion of the mutant neuroectoderm (L). 209 210 Tian et al. exon, through mating with Cre transgenic mice (Fig. 1A; Lewan- Explant Culture and FGF8b Bead Implantation doski et al., 1997). Transgenic mice were generated by microinjec- tion of DNA into fertilized eggs as described (Hogan et al., 1994). Mice were mated and fertilization was assumed to occur at the These mutant phenotypes were analyzed mainly in the CBA midpoint of the dark cycle. Zero- to five-somite embryos were genetic background. Genotypes of mutant mice were performed collected in chilled Hank’s solution. The cephalic region was routinely by PCR (Fig. 1C) and confirmed by Southern blot (Fig. isolated from the embryos and digested with 2.5–3.5% pancreatin 1B). F1 (5Ј-GTATTTTCCTTGCTACCAAACTGCCGAGTG-3Ј), and 0.5–0.7% trypsin in MgCl2- and CaCl2-depleted Hank’s solu- N1 (5Ј-TCGTGCTTTACGGTATCGCCGCTCCCGATT-3Ј), R1 tion, depending on embryonic stages. Germ layer separation was (5Ј-CTGGAGGGAAGCCACACCTCTAAGGATTAA-3Ј), P1 (5Ј-TAG- achieved with tungsten needles (Shimamura and Rubenstein, AGGGTATTTAAATAAGGAACGACTGGG-3Ј), and P2 (5Ј-GTC- 1997). CTTCTAAGAGAACTGGATTGTTAACGA-3Ј) primers were used The neural plate from zero- to two-somite wild-type embryos for PCR identification. Four kinds of probes were employed for was isolated exclusively (see Fig. 7, Exp. 1; Shimamura and Ruben- Southern blot. HindIII-digested genomic DNAs were hybridized stein, 1997). The anterior axial mesendoderm, which consists of with Probe A (SphI–HindIII-digested 230-bp fragment of the Otx2 the prechordal plate and anterior portion of the head process from third coding exon). XhoI-digested genomic DNAs were hybridized wild-type or mutant embryos at the zero- to two-somite stage, was with Probe B (BamHI–PstI-digested 650-bp fragment of the neo transplanted to the wild-type neural plate, respectively (see Fig. 7, gene). BamHI-digested genomic DNAs were hybridized with Probe Exps. 2 and 3). These explants were cultured on a nuclepore filter C(EcoRI–EcoRV-digested 600-bp fragment of the Otx2 5Ј region). (Whatman no. 110614) floating in DMEM supplemented with 20% ϫ NheI-digested genomic DNAs were hybridized with Probe D fetal bovine serum and 1 nonessential amino acids in a CO2 (EcoRI–KpnI-digested 750-bp fragment of the Otx2 first intron). incubator at 37°C for 24 h (Shimamura and Rubenstein, 1997; Kimura et al., 2000). Heparin acrylic beads (Sigma) were rinsed in PBS several times. Approximately 50 beads were soaked in 5 ␮l 0.2 mg/ml FGF8b Reverse Transcription (RT)-PCR Analysis recombinant protein solution (R&D) overnight at 4°C (Liu et al., 1999). The beads were rinsed in PBS prior to use. Control beads Total RNA was isolated from wild-type and Otx2frt-neo/frt-neo mu- were soaked in 50 mg/ml BSA–PBS in an identical manner. The left tant embryos at 6.5, 7.5, and 8.5 dpc stages with TRIZOL reagent side of the ANR was removed from the wild-type neural plate at the (GIBCO BRL), respectively. First-strand cDNA synthesis was per- three- to five-somite stages (see Fig. 9, Exp. 1). The ANR from formed with oligo-dT primers and Superscript kit (GIBCO BRL). wild-type or mutant embryos at the three- to five-somite stage was The cDNA was utilized as a PCR substrate by standard protocols. transplanted to the wild-type neural plate in the region where the The PCR primers E1 (5Ј-TATGGACTTGCTGCATCCCTCCG- left side of the ANR had been ablated (see Fig. 9, Exp. 2). The intact TGGGCTA-3Ј), E3 (5Ј-TGGCAGGCCTCACTTTGTTCTGAC- mutant neural plate at the three- to five-somite stages was isolated CTCCAT-3Ј), N0 (5Ј-TTGACAAAAAGAACCGGGCGCCCCTGC- exclusively (see Fig. 9, Exp. 3). The ANR from wild-type embryos at GCT-3Ј), N1 (5Ј-TCGTGCTTTACGGTATCGCCGCTCCCGATT-3Ј), the three- to five-somite stages was transplanted to the intact and N2 (5Ј-AATCGGGAGCGGCGATACCGTAAAGCACGA-3Ј) mutant neural plate (see Fig. 9, Exp. 4). A single heparin acrylic were used. bead soaked in BSA or FGF8b was implanted to the left side of the wild-type neural plate, where the ANR had been excised unilater- ally (see Fig. 9, Exp. 5), or to the intact mutant neural plate (see Fig. Histological and Skeletal Analysis 9, Exp. 6). These neural plate explants were cultured as described above. Embryos were fixed in Bouin’s fixative, dehydrated, embedded in paraffin, and sectioned at 8 ␮m. Sections were then dewaxed, rehydrated, and stained with hematoxylin–eosin solution. Skeletal RESULTS preparation and staining were conducted as described previously (Kelly et al., 1983; Matsuo et al., 1995). Generation of an Allelic Series of Otx2 Mutants To generate a series of Otx2 mutant alleles, a targeting vector was designed in which the Otx2 promoter and first Whole-Mount RNA in Situ Hybridization coding exon region were flanked with two loxP sites; Embryos were dissected in PBS and fixed in 4% para- additionally, a PGKneopA (neo-) cassette flanked by two frt formaldehyde/PBS at 4°C, followed by gradual dehydration in sites was inserted into the Otx2 first intron (Fig. 1A). We methanol/PBS containing 0.1% Tween 20. Specimens were stored planned to excise the neo-cassette in the first intron by in 100% methanol at Ϫ20°C. Whole-mount in situ hybridization Flp-mediated recombination and in the first coding exon by was performed by digoxigenin-labeled riboprobes (Boehringer Cre-mediated recombination; the latter should eliminate Mannheim) as described (Wilkinson, 1993). The following probes Otx2 function since this exon encodes the translational were employed: Otx1 (Suda et al., 1996), Otx2 (Matsuo et al., 1995), start site (Fig. 1A). Following electroporation in TT2 ES En2 (Davis and Joyner, 1988), Fgf8 (Crossley and Martin, 1995), cells, 8 of 182 G418-resistant ES clones were isolated as Gbx2 (Bulfone et al., 1993), BF-1 (Foxg1; Tao and Lai, 1992), Emx2 (Yoshida et al., 1997), Six3 (Oliver et al., 1995), Pax6 (Stoykova and homologous recombinants by PCR and finally confirmed by Gruss, 1994), Nkx2.1 (Kimura et al., 1996), mdkk-1 (Glinka et al., Southern blot (Fig. 1B). Chimeric mice and their mutant 1998), Cer-l (Belo et al., 1997), Foxa2 (Sasaki and Hogan, 1993), offspring were generated from two independent recombi- Cripto (Ding et al., 1998), Shh (Echelard et al., 1993), Foxd4 nant ES clones (E90 and E92). No phenotypic difference was (Kaestner et al., 1995), and Brachyury/T (Herrmann, 1991). evident between these two mutant lines (see below). Het-

© 2002 Elsevier Science (USA). All rights reserved. Otx2 Functions in Forebrain Specification 211 erozygous offspring of chimera bearing the Otx2frt-neo allele second exon, which encodes homeodomain, whereas the was viable and fertile; offspring exhibited no noticeable second lacked this exon as a consequence of aberrant phenotype (see Figs. 3A and 3C). However, Otx2frt-neo/frt-neo splicing (Fig. 2B). Therefore, Otx2frt-neo/frt-neo homozygous homozygous mutants died at birth or by weaning (data not embryos exhibited three kinds of transcripts: wild-type shown). Otx2frt-neo/ϩ mice were mated to CAG-Flpe trans- normal transcripts and the two types of Otx2-neo fusion genic animals, which ubiquitously express Flpe, a variant mRNA transcripts (Fig. 2). Nucleotide sequences of two Flp recombinase possessing enhanced thermostability, in aberrant transcripts predicted that neither produced normal order to remove the neo-cassette (Buchholz et al., 1998). Otx2 protein; transcripts contained only 33 amino acids in The mutant progeny carrying the transgene effectively the N-terminal peptides of Otx2 in fusion with all neo inherited the recombined allele (Otx2flox allele) (Figs. 1B and protein products (Fig. 2B). Furthermore, semiquantitative 1C). Subsequently, Otx2⌬pro-ex1/ϩ heterozygous mutants were PCR analysis indicated that the expression level of obtained by mating with ␤-actin-cre mice, which ubiqui- wild-type mRNA transcripts was significantly reduced in tously express Cre under the human ␤-actin promo- Otx2frt-neo/frt-neo mutant embryos (Fig. 2A). Analysis of Otx2 ter (Lewandoski et al., 1997). Otx2flox/ϩ, Otx2flox/flox, and expression by in situ hybridization using the third exon as Otx2⌬pro-ex1/ϩ mutant mice thus produced were fertile and a probe demonstrated that neo-cassette insertion did not showed no noticeable phenotype in the CBA genetic back- direct ectopic Otx2 mRNA expression in Otx2frt-neo/Ϫ mu- ground. Otx2frt-neo/ϩ mice were then crossed with Otx2ϩ/Ϫ tant embryos (data not shown). mutant mice (Matsuo et al., 1995), which carry a null allele, Defects in Otx2frt-neo/Ϫ mutants were more severe than resulting in offspring containing both a frt-neo and a null those in Otx2frt-neo/frt-neo and Otx2ϩ/Ϫ mutants; however, de- allele. Unexpectedly, all of the Otx2frt-neo/Ϫ mutant fetuses fects were milder than those observed in Otx2Ϫ/Ϫ mutants exhibited dramatic head malformations, such as excen- (Fig. 3; and data not shown). This finding demonstrated that cephaly and acephaly, at 18.5 dpc (see Fig. 3D; and data not the frt-neo allele was not functioning normally. Analysis of shown). In contrast, Otx2flox/Ϫ fetuses were viable and dem- mRNA expression and phenotype, taken together, indicated onstrated no apparent abnormalities, suggesting that the that Otx2frt-neo is a hypomorphic allele; moreover, these data Otx2flox allele functions as a wild-type Otx2 allele. indicated that the defects in mutant mice carrying the Otx2frt-neo allele are due to a reduction in the level of functional Otx2 gene expression brought about by the Neo-Cassette Insertion Creates a Hypomorphic presence of the neo-cassette in the first intron of the Otx2 Otx2 Allele gene. As described below, detailed analysis of the defects Neo-cassette was supplied as a selectable marker to caused by the hypomorphic allele has provided new insight isolate homologous recombinant clones; however, it has into Otx2 function regarding the mechanism of forebrain been reported to contain potential cryptic splice sites that development. may interfere with gene expression (Carmeliet et al., 1996; Meyers et al., 1998; Nagy et al., 1998). To examine whether ؊ Forebrain Development Was Arrested in Otx2frt-neo/ the presence of the neo-cassette in the intron affected Otx2 Hypomorphic Mutant Embryos gene expression by aberrant splicing, RT-PCR analysis was performed by using primers from Otx2 and neo sequences Previously, we reported that mice heterozygous for the (Fig. 2A). Indeed, the neo-cassette included both a cryptic Otx2 mutation display a craniofacial malformation which splice acceptor and a splice donor (Fig. 2). Sequence analysis is mainly characterized as the loss of the lower jaw depend- of RT-PCR products indicated that the splice acceptor is ing on the genetic background of the C57BL/6 strain (Mat- located 5 bp upstream of the neo translation start site, suo et al., 1995). The majority of skull elements at the level whereas the splice donor is located 15 bp downstream of the of premandibular and distal portions of the mandibular neo TGA stop codon (Fig. 2B). In addition, two types of neo regions were lost or severely affected by this mutation fusion transcripts were produced: one contained the Otx2 (Matsuo et al., 1995). However, 18.5 dpc Otx2frt-neo/Ϫ mu-

FIG. 5. Molecular analysis of neural markers at 8.5 dpc. Emx2 (A, B), Six3 (C, D), Pax6 (E, F), Otx1 (G, H), En2 (I, J), Gbx2 (K, L), and Shh (M, N) expression by whole-mount in situ hybridization in wild-type (A, C, E, G, I, K, M) and Otx2frt-neo/Ϫ mutant (B, D, F, H, J, L, N) embryos at 8.5 dpc. Emx2 expression is detected in the prospective dorsal forebrain of the wild-type embryo (A); however, it is not observed in the mutant embryo (B). Although Six3 expression, a marker for the most anterior neuroectoderm (C), is present in mutant embryos, the expression domain is signiﬁcantly reduced (D). Pax6 expression, a marker for the prospective diencephalon (E), is not evident in the mutant (F). Otx1 expression, a marker for prospective forebrain and midbrain (G), also occurs in the mutant embryo (H). The En2-positive region is shifted to the anterior side (I, J); however, the En2-negative domain is still present in the anterior neuroectoderm of the mutant embryo (J, arrowheads). Gbx2 expression domain, a marker for anterior metencephalon (K), is expanded and shifted anteriorly in the mutant embryo (L). Shh expression, a marker for the ventral midline of neural tube and axial mesendoderm (M), is unchanged in the Otx2frt-neo/Ϫ mutant embryo (N).

© 2002 Elsevier Science (USA). All rights reserved. 214 Tian et al. tants properly developed upper and lower jaws (Fig. 3D). (Fig. 4J). Engrailed-2 (En2) expression, which covers the Skeletal analysis also indicated that most of the skull midbrain and anterior hindbrain at 9.5 dpc in wild type (Fig. elements in premandibular and distal components of the 4K; Davis and Joyner, 1988), underwent an anterior shift in mandibular regions were present and developed although Otx2frt-neo/Ϫ hypomorphic embryos (Fig. 4L). the morphology of these structures was distorted (Figs. To determine the initial defects in brain formation, 3E–3H). These structures are derived from mesencephalic expression of neuroectoderm markers at 8.5 dpc was further neural crest cells (Couly et al., 1993); therefore, the neural analyzed (Fig. 5). We found that prosencephalon was not crest at the level of mesencephalon appeared to be formed specified correctly in Otx2frt-neo/Ϫ embryos. At this stage, appropriately in Otx2frt-neo/Ϫ mutant embryos. Emx2 expression, a marker for the prospective telencepha- Otx2frt-neo/Ϫ mutant mice at 18.5 dpc did not develop lon and dorsal diencephalon, was not detected in Otx2frt-neo/Ϫ rostral brain and eyes (Fig. 3D). We next examined earlier mutant embryos (Figs. 5A and 5B). Expression of Six3, the defects in the mutant brain (Figs. 3I–3L). At 10.5 dpc, the earliest anterior ectoderm marker, was observed; however, rostral aspect of the mutant brain appeared to be truncated expression was significantly reduced in mutant embryos with failure in the neural tube closure (Figs. 3I and 3J). (Figs. 5C and 5D). At 8.5 dpc, Pax6 expression, evident in Histological examination revealed the complete absence of the wild-type forebrain (Fig. 5E; Stoykova and Gruss 1994; the structures of the telencephalic vesicles and diencepha- Shimamura et al., 1997), was absent in mutant embryos lon (Figs. 3K and 3L). Additionally, the mesencephalon and (Fig. 5F). Nkx2.1 expression was observed in neither mutant rostral portion of metencephalic regions, including the embryo at this stage (data not shown). Otx1, which is isthmus, also appeared to be malformed in the mutant expressed in the prospective forebrain and midbrain in wild embryos (Figs. 3K and 3L). These morphological data indi- type (Fig. 5G; Simeone et al., 1993), was shifted to the cate that the majority of the rostral brain failed to develop anterior side in mutant embryos (Fig. 5H). En2 expression, a normally in Otx2frt-neo/Ϫ embryos at 10.5 dpc. marker for mesencephalon and anterior metencephalon Expression of neuroectoderm markers by whole-mount in (Fig. 5I), also underwent a shift to the anterior side in situ hybridization at 9.5 dpc was subsequently analyzed in mutant embryos (Fig. 5J). Similarly, the domain of Gbx2 order to define rostral brain abnormalities in Otx2frt-neo/Ϫ expression, a marker for anterior metencephalon (Fig. 5K; embryos (Fig. 4). Expression of molecular markers for fore- Bouillet et al., 1995), was shifted anteriorly in mutant brain and isthmus were absent or severely affected in embryos (Fig. 5L). It is noteworthy that an En2-negative Otx2frt-neo/Ϫ mutant embryos. Normally, BF-1 is expressed in region was present in the most anterior region of the most of the telencephalic region (Fig. 4A; Tao and Lai, mutant neuroectoderm (Fig. 5J), suggesting the occurrence 1992); however, its expression in this region was not de- of anterior structures rostral to mesencephalon in mutant tected in Otx2frt-neo/Ϫ mutant embryos (Fig. 4B). Emx2 is also embryos. In contrast to these neuroectoderm markers along expressed in dorsal telencephalon and diencephalons of the A-P axis, Sonic hedgehog (Shh), a marker for the ventral wild-type 9.5 dpc embryos (Fig. 4C). Emx2 expression was midline of the neural tube and axial mesendoderm (Ech- absent in Otx2frt-neo/Ϫ hypomorphic mutants (Fig. 4D). At elard et al., 1993), was unaffected in mutant embryos (Figs. this stage, Six3 is expressed in the anterior forebrain region 5M and 5N). Analysis of these aforementioned markers and eyes in wild type (Fig. 4E; Oliver et al., 1995); however, suggests that, in Otx2frt-neo/Ϫ hypomorphic mutants, the its expression was undetectable in mutant embryos (Fig. prospective forebrain region marked by Six3 and Otx1 4F). Normally, Nkx2.1 is expressed in ventral forebrain (Fig. expression is formed; however, the forebrain is not specified 4G; Kimura et al., 1996); however, Nkx2.1 expression was correctly as judged by the markers for the prospective not observed in mutant embryos (Fig. 4H). Fgf8 is expressed telencephalon and diencephalon (see Discussion). in the commissural plate of telencephalon and in the dorsal diencephalon and isthmic constriction in wild type (Fig. ؊ Otx2frt-neo/ Hypomorphic Mutants Form AVE and 4I; Heikinheimo et al., 1994; Ohuchi et al., 1994). In frt-neo/Ϫ ADE Normally Otx2 mutant embryos, Fgf8 expression in these regions was not detected; rather, Fgf8 was expressed through- Induction of the anterior neuroectoderm has been pro- out the entire rostral region, including nonneural ectoderm posed to result from signals emanating from AVE prior to

FIG. 6. Normal development of AVE and ADE in Otx2frt-neo/Ϫ mutant embryos. Wild-type (A, C, E, G, I, K, M, O, Q) and Otx2frt-neo/Ϫ mutant embryos (B, D, F, H, J, L, N, P, R). (A–H) 6.5 dpc; (M, N) 7.5 dpc, and (I–L, O–R) 7.8 dpc. The direction of embryo is anterior to the left and posterior to the right (except I and J, which are frontal views). Cer-l and mdkk-1 expression in the mutant AVE at 6.5 dpc are not distinguishable from the wild-type embryo (A–D). Expression of posterior markers Cripto and T is present in the posterior ectoderm or primitive streak (p) in both wild-type and mutant embryos (E–H). Foxa2 expression is evident in the head process (hp) and node (n) appropriately in mutant embryos at 7.8 dpc (I, J). Expression of mdkk-1 and Cer-l in the prechordal plate and ADE, respectively, is not affected in Otx2frt-neo/Ϫ mutants at 7.5–7.8 dpc (K–N). Expression of Six3 (O) and Foxd4 (Q) in the ANE is unchanged in the mutant embryos at 7.8 dpc (P, R). Abbreviation: pp, prechordal plate.

© 2002 Elsevier Science (USA). All rights reserved. Otx2 Functions in Forebrain Specification 215 and in ADE and/or axial mesendoderm after gastrulation at zero- to two-somite stage, could induce Nkx2.1 expres- (Ang and Rossant, 1993; Beddington and Robertston, 1999; sion in neural plate explants from wild-type embryos at the Shawlot et al., 1999; Camus et al., 2000; Kimura et al., identical stage, followed by culture for 24 h (Shimamura 2000; Wilson and Rubenstein 2000; Kiecker and Niehrs, and Rubenstein, 1997; Figs. 7B and 7C). Unexpectedly, the 2001). To examine whether the failure in forebrain specifi- anterior axial mesendoderm from mutant embryos was also cation is a consequence of defects in these tissues, expres- able to induce Nkx2.1 expression transplanted on wild-type sion of molecular markers was analyzed (Fig. 6). We found neural plate explants (Fig. 7D; n ϭ 5). This result indicates that marker expression was unchanged in Otx2frt-neo/Ϫ mu- that the mutant anterior axial mesendoderm is capable of tant embryos (Fig. 6). In Otx2Ϫ/Ϫ mutant embryos, expres- normal induction of Nkx2.1 expression. Furthermore, these sion of AVE markers such as Cer-l and Lim1 occurred; findings suggest that failure in the forebrain specification of however, expression remained in the distal visceral mutant embryos does not result from defects of the axial endoderm even at 6.5 dpc; that is, expression does not shift mesendoderm. Furthermore, additional explant analysis to the anterior side (Acampora et al., 1998; Kimura et al., was conducted in order to establish whether mutant ADE is 2000). In contrast, mdkk-1 expression (Pearce et al., 1999) is able to induce and/or maintain Otx2 expression trans- specifically lost in Otx2Ϫ/Ϫ mutant visceral endoderm planted to the wild-type epiblast (Ang et al., 1994). As we (Zakin et al., 2000; Kimura et al., 2001; Perea-Gomez et al., expected, mutant ADE could induce and/or maintain Otx2 2001). In the epiblast of Otx2Ϫ/Ϫ mutants, the posterior expression in the wild-type epiblast (n ϭ 3; data not shown). markers, T and Cripto (Ding et al., 1998; Herrmann, 1991), These explant studies suggest that failure in forebrain were expanded in the proximal ectoderm (Kimura et al., specification of Otx2frt-neo/Ϫ mutant embryos is not a conse- 2000). Thus, generation of the A-P axis fails in Otx2Ϫ/Ϫ quence of defects in either the ADE or the anterior axial mutant embryos. In Otx2frt-neo/Ϫ mutant embryos, however, mesendoderm; rather, failure is due to later and/or other expression of both Cer-l and mdkk-1 was detected in the developmental events. AVE (Figs. 6A–6D). Moreover, Cripto and T expres- sions were observed normally in the posterior epiblast in frt-neo/Ϫ BF-1 Expression Is Not Initiated in the Otx2 mutants (Figs. 6E–6H). Neuroectoderm of Mutant Embryos At the headfold stage, Foxa2 expression, a marker for axial mesendoderm (Fig. 6I; Sasaki and Hogan, 1993), was Several reports have indicated that the ANR functions as unaffected in mutant embryos (Fig. 6J). At this stage, a local signaling center essential for forebrain specification mdkk-1, a marker for prechordal plate and the lateral region (Shimamura and Rubenstein, 1997; Houart et al., 1998). of ADE (Fig. 6K; Kiecker and Niehrs, 2001), and Cer-l,a Fgf8 is expressed in the ANR and later in the commissural marker for ADE (Fig. 6M; Belo et al., 1997), were also plate of telencephalon and, indeed, can induce and/or main- present in mutant embryos, which is a normal occurrence tain BF-1 expression (Shimamura and Rubenstein, 1997). (Figs. 6L and 6N). In addition, expression of Six3 and Foxd4 In order to investigate the possibility that mutant defects (previously referred to as fkh-2), the earliest anterior neuro- are due to the failure in forebrain specification by the ANR, ectoderm markers (Oliver et al., 1995; Kaestner et al., 1995), expression of markers for the ANR and prosencephalic were also unaffected in mutant embryos at 7.8 dpc (Figs. neural plate was analyzed at early somite stages (Fig. 8). We 6O–6R), supporting that the prospective forebrain region is found that BF-1 and Fgf8 expression of the nonneural once formed appropriately in Otx2frt-neo/Ϫ mutant embryos. ectoderm occurred properly; however, expression of BF-1 These results suggest that Otx2frt-neo/Ϫ hypomorphic mutant and Fgf8 was absent in the neural plate of Otx2frt-neo/Ϫ embryos can establish the A-P axis and form ADE and mutants (Fig. 8). Normally, BF-1 expression is initially anterior axial mesendoderm tissues normally. present in nonneural ectoderm adjacent to the anterior edge of the neural plate around the two- to three-somite stage ؊ (data not shown). Subsequently, BF-1 expression begins in The Anterior Axial Mesendoderm of Otx2frt-neo/ the rostral margin of the ectoderm after the five-somite Hypomorphic Mutants Functions Normally stage in wild type (Fig. 8A; Shimamura and Rubenstein, Notably, the anterior axial mesendoderm has been shown 1997). In mutant embryos, BF-1 expression was detected in to induce and maintain Nkx2.1 expression in the neuroec- the ANR at this stage, as was the case with wild type (Fig. toderm (Ericson et al., 1995; Shimamura and Rubenstein, 8B). At the ensuing eight-somite stage, neuroectodermal 1997; Pabst et al., 2000). Indeed, Nkx2.1 expression was not expression was induced in wild-type embryos (Fig. 8C); evident in Otx2frt-neo/Ϫ hypomorphic mutant embryos at 8.5- however, induction was not evident in Otx2frt-neo/Ϫ mutant and 9.5-dpc stages (Fig. 4F; and data not shown). Thus, in embryos (Fig. 8D). Even at the 12-somite stage, BF-1 expres- order to investigate whether the inductive signals produced sion remained undetectable in the neuroectoderm of mu- in the mutant anterior axial mesendoderm are defective, tant embryos (Figs. 8E and 8F). Similarly, Emx2 expression neural plate explant analysis was performed (Fig. 7). As was initiated in the laterocaudal neural plate at the three- reported previously, the anterior axial mesendoderm frag- somite stage (Shimamura et al., 1995; and data not shown). ment, which consists of the prechordal plate and the However, no Emx2 expression occurred in mutant embryos, anterior portion of the head process from wild-type embryos even at the eight-somite stage (Figs. 8G and 8H).

FIG. 7. The anterior axial mesendoderm of Otx2frt-neo/Ϫ embryos functions normally in neural plate explant assays. (A) Schematic representation of the neural plate and anterior axial mesendoderm of wild-type and Otx2frt-neo/Ϫ mutant embryos at the zero- to two-somite stage, respectively, and the explant experiments (Exps. 1–3). (B–D) Whole-mount in situ hybridization analysis of Nkx2.1 expression following a 24-h culture of neural plates. In neural plate explant obtained from wild-type embryos exclusively, Nkx2.1 expression is not evident (B; n ϭ 6). The anterior axial mesendoderm from wild-type (C; n ϭ 8) or mutant embryos (D) can induce Nkx2.1 expression in the wild-type neural plate.

Normally, Fgf8 expression commences in the nonneural the mutant neural plate loses competence to the inductive ectoderm around the four-somite stage and in neuroecto- signals from the ANR, neural plate explant assays were derm following the six-somite stage (Shimamura and conducted (Fig. 9; Shimamura and Rubenstein, 1997). Wild- Rubenstein, 1997; and data not shown). In mutant embryos type anterior neural plate explants lose BF-1 expression at the eight-somite stage, however, Fgf8 expression was not upon excision of the ANR at the three- to ﬁve-somite stages observed in the neural plate; on the other hand, normal Fgf8 followed by culture for 24 h (Fig. 9B; n ϭ 9; Shimamura and expression was displayed in nonneural ectoderm (Figs. Rubenstein, 1997). Initially, we examined whether the 8I–8L). At the subsequent 12-somite stage, Fgf8 expression ANR fragment from mutant embryos (ANRmut) can restore in the prospective commissural plate remained undetect- BF-1 expression in wild-type neural plate explants in a able in mutant embryos (Figs. 8M and 8N). Thus, in mutant manner similar to the ANR fragment from wild type embryos, expression of the telencephalic genes BF-1, Emx2, (ANRWT) (Fig. 9C; n ϭ 10). Indeed, the ANRmut was able to and Fgf8 was not initiated in neuroectoderm, although induce BF-1 expression transplanted on the wild-type ex- expression of these genes began normally in nonneural plants (Fig. 9D; n ϭ 14). In order to investigate the compe- ectoderm. tence of the mutant neural plate, the ANRWT was transplanted on the mutant neural plate (Figs. 9E and 9F). The WT ؊ ANR , however, was unable to induce BF-1 expression Otx2frt-neo/ Mutant Neural Plate Cannot Respond (Fig. 9F; n ϭ 7). Further, a FGF8b bead was transplanted on to the Signals from the ANR the mutant neural plate; a FGF8b-bead is demonstrated to To investigate another possibility that the mutant ANR be sufﬁcient for BF-1 induction in the neural plate (Figs. 9G retains the ability to induce telencephalic markers while and 9H; n ϭ 14; Shimamura and Rubenstein, 1997). A

FIG. 8. Defects of neural plate in the Otx2frt-neo/Ϫ hypomorphic mutant embryos. BF-1 (A–F), Emx2 (G, H), and Fgf8 (I–N) expression by whole-mount in situ hybridization at the 5-somite (A, B), 8-somite (C, D, G–L), and 12-somite (E, F, M, N) stages in wild-type (A, C, E, G, I, K, M) and Otx2frt-neo/Ϫ mutant embryos (B, D, F, H, J, L, N). (K, L) Sagittal sections of embryos shown in (I) and (J), respectively. All whole-mount embryos are lateral views. At the ﬁve-somite stage, BF-1 expression occurs in the nonneural ectoderm (arrowhead) of the wild-type embryo but not in the neuroectoderm (A). BF-1 expression is not changed in the nonneural ectoderm of the mutant embryo at the ﬁve-somite stage (B, arrowhead). BF-1 expression is found in the neuroectoderm (arrow) and nonneural ectoderm (arrowhead) in the wild-type embryo at the eight-somite stage (C). In the mutant embryo, BF-1 expression is detected solely in the nonneural ectoderm (D, arrowhead). At the 12-somite stage, BF-1 expression is evident throughout the prospective telencephalic region of wild type (E, arrow). At this stage, BF-1 expression is not induced in the neural ectoderm; however, it remains apparent in the nonneural ectoderm of mutant embryos (F, arrowhead). At the eight-somite stage, Emx2 expression is present in the neuroectoderm in the wild type (G). However, no Emx2 expression occurs in the mutant embryo at this stage (H). At the eight-somite stage, Fgf8 expression is observed in the neuroectoderm (arrow) and nonneural ectoderm (arrowhead) of the wild-type embryo (I and K); in contrast, Fgf8 expression occurs in the nonneural ectoderm (arrowheads) but not in the neuroectoderm in mutant embryos (J, L). At the 12-somite stage, Fgf8 expression is present in the prospective telencephalic commissural plate in the wild-type embryo (M, arrow). In the mutant embryo, expression of Fgf8 is not initiated in the neuroectoderm; however, it remains evident in the nonneural ectoderm (N, arrowhead). Abbreviations: be, branchial ectoderm; fg, foregut; is, isthmic constriction.

FGF8b bead, however, could not induce BF-1 expression in marked by Six3, Foxd4, and Otx1 expression can be formed the mutant neural plate (Fig. 9I; n ϭ 11). Therefore, failure once in Otx2frt-neo/Ϫ hypomorphic mutant embryos normally in forebrain specification of mutant embryos results from (Figs. 5 and 6). Notably, Six3 demarcates the most anterior the defects of the neural plate, but not of the ANR. This neural plate (Oliver et al., 1995). Overexpression of Six3 finding suggests that Otx2 may be required in the neural leads to enlarged forebrain, whereas the dominant activator plate for proper response to inductive signals produced in form of Six3 results in the reduction of forebrain in ze- the ANR. brafish (Kobayashi et al., 1998, 2001). This observation suggests that Six3 may define the forebrain territory. Recently, we found the cis-regulatory region of Otx2, DISCUSSION which governs transgene expression in AVE and ADE but not in anterior neuroectoderm within endogenous Otx2 Generation of Otx2 Hypomorphic Mutation expression (Kimura et al., 2000). Transgenic rescue studies We were able to generate a hypomorphic allele of Otx2 with this cis-regulatory region have demonstrated that through gene targeting capable of neo-cassette insertion Otx2 expression in the anterior neural plate is not required into the intron. Intriguingly, the Otx2frt-neo/Ϫ hypomorphic for induction of Six3 expression (Kimura et al., 2000). Ϫ/Ϫ mutant embryos displayed a unique phenotype, which Furthermore, the fact that Otx2 cells express Six3 based permitted normal A-P axis generation and gastrulation. on chimeric studies (Rhinn et al., 1999), in concert with the These mutants failed to specify the forebrain region, how- aforementioned findings regarding Otx2, supports the pos- ever. These results provide new insights into the roles of sibility that Otx2 expression in neural plate may not Otx2 with respect to forebrain specification. RT-PCR anal- function in an essential capacity in formation of the pro- ysis demonstrated that normal Otx2 mRNA expression of spective forebrain territory. Consistently, marker expres- the hypomorphic allele was significantly reduced as this sion of AVE, primitive streak, ADE, and anterior axial cassette contains cryptic mRNA splice sites and directs mesendoderm, which have been proposed to play important aberrant splicing patterns (Fig. 2). Predicted protein prod- roles in forebrain development in mouse (Beddington and ucts consist of only 33 amino acids of N-terminal residues Robertson, 1999; Shawlot et al., 1999; Camus et al., 2000), frt-neo/Ϫ of Otx2 fused with an entire neo protein resulting from the was not affected during gastrulation in Otx2 mutant aberrant splicing. As Otx2frt-neo/ϩ hypomorphic mutant mice embryos (Figs. 5N, 6, and 7). Indeed, the ADE from frt-neo/Ϫ develop normally, it is unlikely that neo-fused products Otx2 mutant embryos was able to induce or maintain exert dominant-negative effects. Thus, the forebrain defects Otx2 expression transplanted on the epiblast from wild- observed in the mutant embryos may be due to the subcriti- type embryos at early streak stage. Moreover, the mutant cal thresholds of Otx2 expression in the anterior neuroec- anterior axial mesendoderm could induce Nkx2.1 expres- toderm. Recently, neo-cassette insertion into the intron has sion transplanted on wild-type neural plate explants, sug- been shown to create a hypomorphic mutation by targeting gesting that the mutant anterior axial mesendoderm func- of other genes (Meyers et al., 1998; Nagy et al., 1998; Gage tions normally (Fig. 7). These data indicate that AVE, ADE, et al., 1999; Lowe et al., 2001). Early embryonic lethality and anterior axial mesendoderm form appropriately in mu- resulting from complete loss-of-function mutations renders tant embryos; furthermore, these results support the tran- milder alleles valuable in the elucidation of roles at later sient formation of the prospective prosencephalic neural developmental stages; consequently, this approach may plate in mutant embryos. frt-neo/Ϫ provide an applicable method for extensive genetic analysis In Otx2 mutant embryos, however, subsequent of genes of interest. forebrain region marked by BF-1, Emx2, Nkx2.1, and Pax6 expression was not formed correctly (Figs. 4, 5, and 8). Notably, loss-of-function analysis revealed that BF-1 is Otx2 Functions in the Neuroectoderm to Induce essential in the development of the cerebral hemispheres Forebrain Gene Expression (Xuan et al., 1995). Emx2 functions in the formation of Our detailed analysis of Otx2frt-neo/Ϫ hypomorphic mutant dorsal telencephalon and diencephalon (Yoshida et al., embryos provides evidence that Otx2 is required in the 1997; Suda et al., 2001). Small eye mutation analyses have neural plate to induce telencephalic gene expression for demonstrated that Pax6 is required for the development and forebrain specification. Previous studies have suggested regionalization of diencephalon (Stoykova et al., 1996; that Otx2 expression in AVE can direct the induction of Warren and Price, 1997; Grindley et al., 1997). Furthermore, anterior neuroectoderm but is not sufficient to form the overexpression experiments suggest that Pax6 also defines forebrain region correctly (Acampora et al., 1998; Suda et the dien–mesencephalic boundary via repression of En1 and al., 1999; Kimura et al., 2000). Moreover, it was suggested Pax2 (Matsunaga et al., 2000). Nkx2.1 is critical to ventral that Otx2 expression in the neuroectoderm is essential for forebrain development (Kimura et al., 1996; Sussel et al., forebrain development, based on investigations involving 1999). Therefore, Otx2 may be involved in forebrain devel- mouse chimeras containing Otx2Ϫ/Ϫ and wild-type cells opment through direct or indirect transactivation of these (Rhinn et al., 1998, 1999). Consistent with the aforemen- forebrain genes crucial for subsequent forebrain specifica- tioned reports, it appears that the anterior neural plate tion. On the other hand, expression of the mesencephalic

© 2002 Elsevier Science (USA). All rights reserved. Otx2 Functions in Forebrain Specification 219 marker, En2, was not altered; however, the expression Otx2frt-neo/Ϫ mutant embryos. Consequently, mutant neural domain underwent an anterior shift in mutant embryos plate might result in the loss of the entire forebrain region (Figs. 4 and 5). This result is supported by the previous at later stages. More interestingly, this study suggests that report that Otx2 expression in the neural plate may not be Otx2 may be involved downstream of the signals from the required for the initiation, but rather the maintenance, of ANR, i.e., Fgf signaling, to mediate the induction of telen- En2 expression indirectly (Rhinn et al., 1999). It is notewor- cephalic genes for forebrain regionalization. Otx2 is a thy that En2 was not expressed in the most anterior portion transcription activating factor (Simeone et al., 1993); there- of mutant neuroectoderm (Figs. 4L and 5J), suggesting the fore, it is likely that Otx2 transactivates telencephalic presence of the prospective prosencephalic neural plate in genes as a target directly or indirectly in response to Fgf mutant embryos. In conjunction, these observations sug- signals. It is notable that Otx2Ϫ/Ϫ cells in the forebrain gest that A-P axis development and induction of the pro- region subsequently underwent apoptosis in chimeric em- spective prosencephalic neural plate appear to have been bryos (Rhinn et al., 1999). Apoptosis in these cells may be undergone normally once; however, at subsequent stages, induced as a result of the incompetence of mutant neural the forebrain region failed to be specified in Otx2frt-neo/Ϫ plate with respect to growth or differentiation signals (i.e., mutant embryos. Fgfs produced by the ANR for forebrain specification). Additional molecular analysis is required in order to determine which point(s) of the pathway is/are regulated by Otx2 May Define the Competence of Prosencephalic Otx2 in the prosencephalic neural plate. Neural Plate to Inductive Signals from the ANR Recently available evidence suggests the existence of Expression of forebrain markers such as BF-1 and Emx2 differential competence within the neuroectoderm to in- was not induced in Otx2frt-neo/Ϫ mutant embryos from its ductive signals produced by the ANR or isthmic organizer initial manifestation (Fig. 8). Recently, the ANR in mouse (Shimamura and Rubenstein, 1997; Rubenstein et al., 1998; or the first row of cells in zebrafish has been proposed to Rubenstein and Beachy, 1998). For example, Fgf8 induces function as a local signaling center necessary for forebrain BF-1 expression in the prospective telencephalic neural specification (Shimamura and Rubenstein, 1997; Houart et plate, whereas in more caudal regions of the neural plate, al., 1998). Removal of the ANR results in the loss of Fgf8 induces En2 or Gbx2 expression. Thus, such distinct expression of the telencephalic gene, BF-1; additional ANR neuroectodermal gene expression at differing positions can induce ectopic BF-1 expression in mouse neural plate along the A-P axis within the neural plate is due to intrinsic explants (Shimamura and Rubenstein, 1997). To assess the differences in competence to the signals. (Shimamura and cause of the Otx2frt-neo/Ϫ mutant defects, molecular markers Rubenstein, 1997; Rubenstein et al., 1998). Otx2frt-neo/Ϫ mu- were examined for ANR and prosencephalic neural plate tant embryos do not express BF-1; however, expression of during early somite stage in mutant embryos. At this En2 and Gbx2 is normal (Figs. 4, 5, and 8). Indeed, FGF8b juncture, molecular defects of Otx2frt-neo/Ϫ mutant embryos beads induced neither BF-1 nor En-2 expression in the beginning in the anterior neural plate were detected by the prospective prosencephalic region of mutant neural plates eight-somite stage at the time of forebrain regionalization; (Fig. 9; and data not shown), implying that the competence expression of BF-1, Fgf8, and Emx2 was not induced in the of prosencephalic neural plate may not be transformed to neuroectoderm of mutant embryos, whereas expression of that of the mesencephalic neural plate; rather, prosence- Fgf8 and BF-1 was not affected in nonneural ectoderm (Fig. phalic competence may be lost exclusively in this 8). It is open to question as to whether the ANR or the Otx2frt-neo/Ϫ mutant embryo. neural plate is defective for the inductive interaction be- Additional transplant and genetic studies are necessary in tween the ANR and the neural plate in mutant embryos. order to identify the components involved in the differen- In order to address this issue, neural plate explant studies tial competence of the anterior neural plate. Nevertheless, were performed. These experiments demonstrated that the given the arguments stated above, we believe that Otx2 mutant ANR was able to induce BF-1 expression when most likely involves the competence of the prosencephalic transplanted on the wild-type neural plate; in contrast, neural plate with regard to signals produced mainly in the neither wild-type ANR nor a FGF8b bead, which mimics ANR. the properties of the ANR, induce BF-1 expression in mutant neural plate (Fig. 9). This evidence suggests that Otx2 Is Involved in Critical Processes mutant embryos can form ANR bearing normal induction for Forebrain Development activity; however, mutants are unable to respond to signals of induction produced in the ANR. Concomitantly, Otx2 is In mouse, the A-P axis is initially generated in a expressed in the anterior neuroectoderm but not in the proximal–distal (P-D) direction around 5.5 dpc; subse- ANR of wild-type embryos at the five-somite stage (data not quently, a process of axis rotation occurs prior to primitive shown). Thus, it is likely that, due to the incompetence of streak formation (Beddington and Robertson, 1999). As a mutant neural plate, expression of BF-1 and Emx2 genes, result of this process, the expression of genes that mark the which are crucial for telencephalon development (Xuan et distal visceral endoderm shifts to the anterior visceral al., 1995; Yoshida et al., 1997), cannot be initiated in endoderm, whereas markers expressed in the proximal

FIG. 9. The mutant neural plate fails to respond to the signals from the ANR. (A) Schematic representation of neural plate of wild-type and Otx2frt-neo/Ϫ mutant embryos at the three- to five-somite stage and explant experiments (Exps. 1–6). (B–I) Whole-mount in situ hybridization analysis of BF-1 expression following a 24-h culture of neural plates. In wild-type explants lacking left-side ANR, BF-1 expression is completely lost on the left side (B). The ANR fragment from frt-neo/Ϫ wild-type embryos can induce BF-1 expression on the transplanted side where the ANR is excised (C, arrowhead). The ANR from Otx2 mutants restores al. et Tian BF-1 expression in the neural plate (D, arrowhead). No BF-1 expression was detected in Otx2frt-neo/Ϫ mutant explants alone (E; n ϭ 6). In Otx2frt-neo/Ϫ mutant neural plate explant, the ANR from wild type cannot induce BF-1 expression (F). Heparin–acrylic beads soaked with BSA (G) or FGF8b (H) implanted to the left side of the neural plate explant in the area of ANR ablation. A BSA bead cannot induce BF-1 expression in the neural plate (G; n ϭ 14); in contrast, an FGF8b bead can induce BF-1 expression (H). BF-1 expression is not induced by a FGF8b bead transplanted on the mutant neural plate (I). Otx2 Functions in Forebrain Specification 221 epiblast shift to the posterior aspect, where the primitive ology and Genetics, Center for Animal Resources and Develop- streak is to be formed (Beddington and Robertson, 1999). It ment, Kumamoto University, for housing the mice. E.T. is an has been suggested that Otx2 possesses a crucial role in A-P international student supported by the Ministry of Education, axis formation in P-D orientation from its initial manifes- Culture, Sports Science and Technology, Japan. C.K. is supported tation in cooperation with Cripto in epiblast (Kimura et al., by a special postdoctoral researchers program. This work was supported in part by Grants-in-Aid from the Ministry of Education, 2001) and the subsequent autonomous anterior movement Culture, Sports Science and Technology, Japan. of distal visceral endoderm cells (Kimura et al., 2000). The AVE thus formed does not induce anterior neuroectoderm directly; rather, AVE mediates forebrain development by REFERENCES repressing posteriorizing signals (Tam and Steiner, 1999; Klingensmith et al., 1999; Kimura et al., 2000; Stern, 2001). Acampora, D., Mazan, S., Lallemand, Y., Avantaggiato, V., Maury, Moreover, Otx2 is also essential to this suppression M., Simeone, A., and Brulet, P. (1995). Forebrain and midbrain (Kimura et al., 2000). regions are deleted in Otx2Ϫ/Ϫ mutants due to a defective At subsequent mid-to-late-streak stages, ADE derived anterior neuroectoderm specification during gastrulation. Devel- from the primitive streak also plays an important role in the opment 121, 3279–3290. correct patterning of neuroectoderm (Ang and Rossant, Acampora, D., Avantaggiato, V., Tuorto, F., and Simeone, A. (1997). 1993; Ang et al., 1994). Requirement of ADE in forebrain Genetic control of brain morphogenesis through Otx gene dosage development is also suggested by germ-layer explant stud- requirement. Development 124, 3639–3650. ies, surgical removal experiments, and chimeric analysis of Acampora, D., Avantaggiato, V., Tuorto, F., Briata, P., Corte, G., Lim1 mutant embryos (Ang et al., 1994; Shawlot et al., and Simeone, A. (1998). Visceral endoderm-restricted translation 1999; Camus et al., 2000). Thus, ADE may function in the of Otx1 mediates recovery of Otx2 requirements for specification of anterior neural plate and normal gastrulation. Develop- establishment or maintenance of anterior territory that is ment 125, 5091–5104. initiated by AVE (Shawlot et al., 1999; Kimura et al., 2000). Ang, S-L., and Rossant, J. (1993). Anterior mesendoderm induces However, the precise role of Otx2 expression in ADE mouse Engrailed genes in explant cultures. Development 118, remains to be determined. 139–149. Within the anterior neuroectoderm, the neural plate is Ang, S-L., Conlon, R. A., Jin, O., and Rossant, J. (1994). Positive and transversely subdivided into forebrain, midbrain, and hind- negative signals from mesoderm regulate the expression of brain regions, each with characteristic morphological and mouse Otx2 in ectoderm explants. Development 120, 2979– molecular properties. Two local signaling centers, ANR and 2989. mid/hindbrain junction, termed the isthmus, play crucial Ang, S-L., Jin, O., Rhinn, M., Daigle, N., Stevenson, L., and roles in the refinement of the identity within neuroecto- Rossant, J. (1996). A targeted mouse Otx2 mutation leads to severe defects in gastrulation and formation of axial mesoderm derm along the A-P axis (Joyner, 1996; Shimamura and and to deletion of rostral brain. Development 122, 243–252. et al., Otx2 Rubenstein, 1997; Rubenstein 1998). interacts Beddington, R. S. P., and Robertson, E. J. (1999). Axis development in a repressive manner with Fgf8 and Gbx2 for the correct and early asymmetry in mammals. Cell 96, 195–209. positioning of the isthmic organizing center (Acampora et Belo, J. A., Bouwmeester, T., Leyns, L., Ketrtesz, N., Gallo, M., al., 1997; Suda et al., 1997; Liu et al., 1999; Martinez et al., Follettie, M., and De Robertis, E. M. (1997). Cerberus-like is a 1999; Garda et al., 2001; Liu and Joyner, 2001). In the secreted factor with neuralizing activity expressed in the anterior present investigation, we found that Otx2 is also essential primitive endoderm of the mouse gastrula. Mech. Dev. 68, 45–57. for the longitudinal subdivision of the forebrain by defining Bouillet, P., Chazaud, C., Oulad-Abdelghani, M., Dolle, P., and the competence of the neural plate to signals from the Chambon, P. (1995). Sequence and expression pattern of the ANR. Stra7 (Gbx-2) homeobox-containing gene induced by retinoic acid in P19 embryonal carcinoma cells. Dev. Dyn. 204, 372–382. Further elucidation of Otx2 function in processes of brain Buchholz, F., Angrand, P-O., and Stewart, A. F. (1998). Improved specification and regionalization will be resolved through properties of FLP recombinase evolved by cycling mutagenesis. conditional knock-out experiments. These studies can be Nat. Biotechnol. 16, 657–662. flox performed via the mating of mice carrying the Otx2 Bulfone, A., Puelles, L., Porteus, M. H., Frohman, M. A., Martin, allele, which we have generated, with mice that express Cre G. R., and Rubenstein, J. L. (1993). Spatially restricted expression under the control of appropriate tissue-specific regulatory of Dlx-1, Dlx-2 (Tes-1), Gbx-2, and Wnt-3 in the embryonic day regions. 12.5 mouse forebrain defines potential transverse and longitudinal segmental boundaries. J. Neurosci. 13, 3155–3172. Camus, A., Davidson, B. P., Billiards, S., Khoo, P-L., Rivera-Perez, ACKNOWLEDGMENTS J. A., Wakamiya, M., Behringer, R. R., and Tam, P. P. L. (2000). The morphogenetic role of midline mesendoderm and ectoderm We are grateful to Drs. R. Beddington, E. M. De Robertis, M. in the development of the forebrain and the midbrain of the Frohman, P. Gruss, B. G. Herrmann, B. Hogan, A. Joyner, K. mouse embryo. Development 127, 1799–1813. Kaestner, S. Kimura, E. Lai, M. Lewandoski, G. Martin, A. McMa- Carmeliet, P., Ferreira, V., Breier, G., Pollefeyt, S., Kieckens, L., hon, C. Niehrs, M. Shen, and F. Stewart for gifts of plasmids; Drs. Gertsenstein, M., Fahrig, M., Vandenhoeck, A., Harpal, K., Eber- K. Shimamura and M. Nakafuku for discussions; and K. Kitajima hardt, C., Declercq, C., Pawling, J., Monns, L., Collen, D., Risau, for her assistance. We are also grateful to the Division of Microbi- W., and Nagy, A. (1996). Abnormal blood vessel development and

lethality in embryos lacking a single VEGF allele. Nature 380, Joyner, A. L., Liu, A., and Millet, S. (2000). Otx2, Gbx2 and Fgf8 435–439. interact to position and maintain a mid-hindbrain organizer. Couly, G., and Le Douarin, N. M. (1988). The fate map of the Curr. Opin. Cell Biol. 12, 736–741. cephalic neural primordium at the presomitic to the 3-somite Kaestner, K. H., Monaghan, A. P., Kern, H., Ang, S-L.,Weitz, S., stage in the avian embryo. Development 103(Suppl.), 101–113. Lichter, P., and Schutz, G. (1995). The mouse fkh-2 gene. J. Biol. Couly, G. F., Coltey, P. M., and Le Douarin, N. M. (1993). The triple Chem. 270, 30029–30035. origin of skull in higher vertebrates: A study in quail–chick Kelly, W. L., and Bryden, M. M. (1983). A modified differential stain chimeras. Development 117, 409–429. for cartilage and bone in whole mount preparations of mamma- Crossley, P. H., and Martin, G. R. (1995). The mouse Fgf8 gene lian fetuses and small vertebrates. Stain Technol. 58, 131–134. encodes a family of polypeptides and is expressed in regions that Kiecker, C., and Niehrs, C. (2001). The role of prechordal mesend- direct outgrowth and patterning in developing embryo. Develop- oderm in neural patterning. Curr. Opin. Neurobiol. 11, 27–33. ment 121, 439–451. Kimura, S., Hara, Y., Fernandez-Salguero, P., Fox, C. H., Ward, Davis, C. A., and Joyner, A. L. (1988). Expression pattern of the J. M., and Gonzalez, F. J. (1996). The T/ebp null mouse: Thyroid- homeobox-containing genes En-1 and En-2 and the proto- specific enhancer-binding protein is essential for the organogen- oncogene int-1 diverge during mouse development. Genes Dev. esis of the thyroid, lung, ventral forebrain, and pituitary. Genes 2, 1736–1744. Dev. 10, 60–69. Ding, J., Yang, L., Yan, Y. T., Chen, A., Desai, N., Wynshaw-Boris, Kimura, C., Yoshinaga, K., Tian, E., Suzuki, M., Aizawa, S., and A., and Shen, M. M. (1998). Cripto is required for correct Matsuo, I. (2000). Visceral endoderm mediates forebrain devel- orientation of the anterior–posterior axis in the mouse embryo. opment by suppressing posteriorizing signals. Dev. Biol. 225, Nature 395, 702–707. 304–321. Eagleson, G., Ferreiro, B., and Harris, W. A. (1995). Fate of the Kimura, C., Shen, M. M., Takeda, N., Aizawa, S., and Matsuo, I. anterior neural ridge and the morphogenesis of the Xenopus (2001). Complementary functions of Otx2 and Cripto in initial forebrain. J. Neurobiol. 28, 146–158. patterning of mouse epiblast. Dev. Biol. 235, 12–32. Echelard, Y., Epstein, D. J., St-Jacques, B., Shen, L., Mohler, J., Klingensmith, J., Ang, S-L., Bachiller, D., and Rossant, J. (1999). McMahon, J. A., and McMahon, A. P. (1993). Sonic hedgehog, a Neural induction and patterning in the mouse in the absence of member of a family of putative signaling molecules, is impli- the node and its derivatives. Dev. Biol. 216, 535–549. cated in the regulation of CNS polarity. Cell 75, 1417–1430. Kobayashi, M., Nishikawa, K., Suzuki, T., and Yamamoto, M. Ericson, J., Muhr, J., Placzek, M., Lints, T., Jessell, T. M., and (2001). The homeobox protein six3 interacts with the groucho Edlund, T. (1995). Sonic hedgehog induces the differentiation of corepressor and acts as a transcriptional repressor in eye and ventral forebrain neurons: A common signal for ventral pattern- forebrain formation. Dev. Biol. 232, 315–326. ing within the neural tube. Cell 81, 747–756. Kobayashi, M., Toyama, R., Takeda, H., Dawid, I. B., and Gage, P. J., Suh, H., and Camper, S., A. (1999). Dosage requirement Kawakami, K. (1998). Overexpression of the forebrain-specific of Pitx2 for development of multiple organs. Development 126, homeobox gene six3 induces rostral forebrain enlargement in 4643–4651. zebrafish. Development 125, 2973–2982. Garda, A-L., Echevarria, D., and Martinez, S. (2001). Neuroepithe- Lewandoski, M., Meyers, E. N., and Martin, G. R. (1997). Analysis lial co-expression of Gbx2 and Otx2 precedes Fgf8 expression in of Fgf8 gene function in vertebrate development. Cold Spring the isthmic organizer. Mech. Dev. 101, 111–118. Harbor Symp. Quant. Biol. 62, 159–168. Glinka, A., Wu, W., Delius, H., Monaghan, A. P., Blumenstock, C., Liu, A., Losos, K., and Joyner, A. L. (1999). FGF8 can active Gbx2 and Niehrs, C. (1998). Dickkopf-1 is a member of a new family of and transform regions of the rostral mouse brain into a hindbrain secreted proteins and functions in head induction. Nature 391, fate. Development 126, 4827–4838. 357–362. Liu, A., and Joyner, A. L. (2001). EN and GBX2 play essential roles Grindley, J. C., Hargett, L. K., Hill, R. E., Ross, A., and Hogan, B. L. downstream of FGF8 in patterning the mouse mid/hindbrain (1997). Disruption of PAX6 function in mice homozygous for the region. Development 128, 181–191. Pax6Sey-1Neu mutation produces abnormalities in the early Lowe, L. A., Yamada, S, and Kuehn, M. R. (2001). Genetic dissec- development and regionalization of the diencephalon. Mech. tion of nodal function in patterning the mouse embryo. Devel- Dev. 64, 111–126. opment 128, 1831–1843. Heikinheimo, M., Lawshe, A., Shackleford, G. M., Wilson, D. B., Martinez, S., Crossley, P. H., Cobos, I., Rubenstein, J. L., and and MacArthur, C. A. (1994). Fgf-8 expression in the post- Martin, G. R. (1999). FGF8 induces formation of an ectopic gastrulation mouse suggests roles in the development of the face, isthmic organizer and isthmocerebellar development via a re- limbs, and central nervous system. Mech. Dev. 48, 129–138. pressive effect on Otx2 expression. Development 126, 1189– Herrmann, B. G. (1991). Expression pattern of brachyury gene in 1200. whole-mount Twis/Twis mutant embryos. Development 113, Matsunaga, E., Araki, I., and Nakamura, H. (2000). Pax6 defines the 913–917. di-mesencephalic boundary by repressing En1 and Pax2. Devel- Hogan, B., Beddington, R., Costantini, F., and Lacy, E. (1994). opment 127, 2357–2365. “Manipulating the Mouse Embryo: A Laboratory Manual,” 2nd Matsuo, I., Kuratani, S., Kimura, C., Takeda, N., and Aizawa, S. ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, (1995). Mouse Otx2 functions in the formation and patterning of NY. rostral head. Genes Dev. 9, 2646–2658. Houart, C., Westerfield, M., and Wilson, S. W. (1998). A small Meyers, E. N., Lewandoski, M., and Martin, G. R. (1998). An Fgf8 population of anterior cells patterns the forebrain during ze- mutant allelic series generated by Cre- and Flp-mediated recom- brafish gastrulation. Nature 391, 788–792. bination. Nat. Genet. 18, 136–141. Joyner, A. L. (1996). Engrailed, Wnt and Pax genes regulate Nagy, A., Moens, C., Ivanyi, E., Pawling, J., Gertsenstein, M., midbrain–hindbrain development. Trends Genet. 12, 15–20. Hadjantonakis, A-K., Pirity, M., and Rossant, J. (1998). Dissecting

the role of N-myc in development using a single targeting vector Stern, C. D. (2001). Initial patterning of the central nervous system: to generate a series of alleles. Curr. Biol. 8, 661–664. How many organizers? Nat. Rev. Neurosci. 2, 92–97. Ohuchi, H., Yoshioka, H., Tanaka, A., Kawakami, Y., Nohno, T., Stoykova, A., and Gruss, P. (1994). Roles of Pax-genes in developing and Noji, S. (1994). Involvement of androgen-induced growth and adult brain as suggested by expression patterns. J. Neurosci. factor (FGF-8) gene in mouse embryogenesis and morphogenesis. 14, 1395–1412. Biochem. Biophys. Res. Commun. 204, 882–888. Stoykova, A., Fritsch, R., Walther, C., and Gruss, P. (1996). Fore- Oliver, G., Mailhos, A., Wehr, R., Copeland, N. G., Jenkins, N. A., brain patterning defects in Small eye mutant mice. Development and Gruss, P. (1995). Six3, a murine homologue of the sine oculis 122, 3453–3465. gene, demarcates the most anterior border of the developing Suda, Y., Matsuo, I., Kuratani, S., and Aizawa, S. (1996). Otx1 neural plate and is expressed during eye development. Develop- function overlaps with Otx2 in development of mouse forebrain ment 121, 4045–4055. and midbrain. Genes Cells 1, 1031–1044. Pabst, O., Herbrand, H., Takuma, N., and Arnold, H. H. (2000). Suda, Y., Matsuo, I., and Aizawa, S. (1997). Cooperation between NKX2. gene expression in neuroectoderm but not in mesend- Otx1 and Otx2 genes in developmental patterning of rostral odermally derived structures depends on sonic hedgehog in brain. Mech. Dev. 69, 125–141. mouse embryos. Dev. Genes Evol. 210, 47–50. Suda, Y., Nakabayashi, J., Matsuo, I., and Aizawa, S. (1999). Pearce, J. J. H., Penny, G., and Rossant, J. (1999). A mouse Functional equivalency between Otx2 and Otx1 in development cerberus/Dan-related gene family. Dev. Biol. 209, 98–110. of the rostral head. Development 126, 743–757. Perea-Gomez, A., Lawson, K. A., Rhinn, M., Zakin, L., Brulet, P., Suda, Y., Hossain, Z. M., Kobayashi, C., Hatano, O., Yoshida, M., Mazan, S., and Ang, S-L. (2001). Otx2 is required for visceral Matsuo, I., and Aizawa, S. (2001). Emx2 directs the development endoderm movement and for the restriction of posterior sig- of diencephalon in cooperation with Otx2. Development 128, nals in the epiblast of the mouse embryo. Development 128, 2433–2450. 753–765. Sussel, L., Marin, O., Kimura, S., and Rubenstein, J. L. (1999). Loss Rhinn, M., Dierich, A., Shawlot, W., Behringer, R. R., Le Meur, M., of Nkx2.1 homeobox gene function results in a ventral to dorsal and Ang, S-L. (1998). Sequential roles for Otx2 in visceral molecular respecification within the basal telencephalon: Evi- endoderm and neuroectoderm for forebrain and midbrain induc- dence for a transformation of the pallidum into the striatum. tion and specification. Development 125, 845–856. Development 126, 3359–3370. Rhinn, M., Dierich, A., Le Meur, M., and Ang, S-L. (1999). Cell Tao, W., and Lai, E. (1992). Telencephalon-restricted expression of autonomous and non-cell autonomous functions of Otx2 in BF-1, a new member of the HNF-3/folk head gene family in the patterning the rostral brain. Development 126, 4295–4304. developing rat brain. Neuron 8, 957–966. Rubenstein, J. L. R., Shimamura, K., Martinez, S., and Puelles, L. Tam, P. P., and Steiner, K. A. (1999). Anterior patterning by (1998). Regionalization of the prosencephalic neural plate. Annu. synergistic activity of the early gastrula organizer and the ante- Rev. Neurosci. 21, 445–477. rior germ layer tissue of the mouse embryo. Development 126, Rubenstein, J. L. R., and Beachy, P. A. (1998). Patterning of the 5171–5179. embryonic forebrain. Curr. Opin. Neurobiol. 8, 18–26. Warren, N., and Price, D. J. (1997). Roles of Pax-6 in murine Sasaki, H., and Hogan, B. (1993). HNF-3b as a regulator of floor diencephalic development. Development 124, 1573–1582. plate development. Cell 76, 103–115. Wilkinson, D. G. (1993). In situ hybridization. In “Essential Devel- Shanmugalingam, S., Houart, C., Picker, A., Reifers, F., Mac- opmental Biology: A Practical Approach” (C. D. Stern and donald, R., Barth, A., Griffin, K., Brand, M., and Wilson, S. W. P. W. H. Holland, Eds.), pp. 257–274. IRL Press, Oxford, U.K. (2000). Ace/Fgf8 is required for forebrain commissure forma- Wilson, S. W., and Rubenstein, J. L. R. (2000). Induction and dorso- tion and patterning of the telencephalon. Development 127, ventral patterning of the telencephalon. Neuron 28, 641–651. 2549–2561. Xuan, S., Baptista, C. A., Balas, G., Tao, W., Soares, V. C., and Lai, E. Shawlot, W., Wakamiya, M., Kwan, K. M., Kania, A., Jessell, T. M., (1995). Winged helix transcription factor BF-1 is essential for the and Behringer, R. R. (1999). Lim1 is required in both primitive development of the cerebral hemispheres. Neuron 14, 1141–1152. streak-derived tissues and visceral endoderm for head formation Yagi, T., Nada, S., Watanabe, N., Tamemoto, H., Kohmura, N., in the mouse. Development 126, 4925–4932. Ikawa, Y., and Aizawa, S. (1993). A novel negative selection for Shimamura, K., Hartigan, D. J., Martinez, S., Puelles, L., and Ruben- homologous recombinants using diphtheria toxin A fragment stein, J. L. R. (1995). Longitudinal organization of the anterior neural gene. Anal. Biochem. 214, 77–86. plate and neural tube. Development 121, 3923–3933. Ye, W., Shimamura, K., Rubenstein, J. L. R., Hynes, M. A., and Shimamura, K., and Rubenstein, J. L. R. (1997). Inductive interac- Rosenthal, A. (1998). FGF and Shh signals control dopaminergic tions direct early regionalization of the mouse forebrain. Devel- and serotonergic cell fate in the anterior neural plate. Cell 93, opment 124, 2709–2718. 755–766. Shimamura, K., Martinez, S., Puelles, L., and Rubenstein J. L. R. Yoshida, M., Suda, Y., Matsuo, I., Miyamoto, N., Takeda, N., Kura- (1997). Patterns of gene expression in the neural plate and neural tani, S., and Aizawa, S. (1997). Emx1 and Emx2 functions in tube subdivide the embryonic forebrain into transverse and development of dorsal telencephalon. Development 124, 101–111. longitudinal domains. Dev. Neurosci. 19, 88–96. Zakin, L., Reversade, B., Virlon, B., Rusniok, C., Glaser, P., Elalouf, Simeone, A., Acampora, D., Mallamci, A., Stornaiuolo, A., J. M., and Brulet, P. (2000). Gene expression profiles in normal D’Apice, M. R., Nigro, V., and Boncinelli, E. (1993). A vertebrate and Otx2Ϫ/Ϫ early gastrulating mouse embryos. Proc. Natl. gene related to orthodenticle contains a homeodomain of the Acad. Sci. USA 97, 14388–14393. bicoid class and demarcates anterior neuroectoderm of the gastrulating mouse embryo. EMBO J. 12, 2735–2747. Received for publication August 7, 2001 Simeone, A. (2000). Positioning the isthmic organizer where Otx2 Accepted November 1, 2001 and Gbx2 meet. Trends Genet. 16, 237–240. Published online January 8, 2002