Rhombomere boundaries are Wnt signaling ... - Wiley Online Library

DEVELOPMENTAL DYNAMICS 231:278 –291, 2004

RESEARCH ARTICLE

Rhombomere Boundaries Are Wnt Signaling Centers That Regulate Metameric Patterning in the Zebrafish Hindbrain Bruce B. Riley,* Ming-Yung Chiang, Elly M. Storch, Rebecca Heck, Gerri R. Buckles, and Arne C. Lekven

The vertebrate hindbrain develops from a series of segments (rhombomeres) distributed along the anteroposterior axis. We are studying the roles of Wnt and Delta–Notch signaling in maintaining rhombomere boundaries as organizing centers in the zebrafish hindbrain. Several wnt genes (wnt1, wnt3a, wnt8b, and wnt10b) show elevated expression at rhombomere boundaries, whereas several delta genes (dlA, dlB, and dlD) are expressed in transverse stripes flanking rhombomere boundaries. Partial disruption of Wnt signaling by knockdown of multiple wnt genes, or the Wnt mediator tcf3b, ablates boundaries and associated cell types. Expression of dlA is chaotic, and cell types associated with rhombomere centers are disorganized. Similar patterning defects are observed in segmentation mutants spiel-ohne-grenzen (spg) and valentino (val), which fail to form rhombomere boundaries due to faulty interactions between adjacent rhombomeres. Stripes of wnt expression are variably disrupted, with corresponding disturbances in metameric patterning. Mutations in dlA or mind bomb (mib) disrupt Delta–Notch signaling and cause a wide range of patterning defects in the hindbrain. Stripes of wnt1 are initially normal but subsequently dissipate, and metameric patterning becomes increasingly disorganized. Driving wnt1 expression using a heat-shock construct partially rescues metameric patterning in mib mutants. Thus, rhombomere boundaries act as Wnt signaling centers required for precise metameric patterning, and Delta signals from flanking cells provide feedback to maintain wnt expression at boundaries. Similar feedback mechanisms operate in the Drosophila wing disc and vertebrate limb bud, suggesting coaptation of a conserved signaling module that spatially organizes cells in complex organ systems. Developmental Dynamics 231:278 –291, 2004. © 2004 Wiley-Liss, Inc. Key words: boundary-formation; Wnt; tcf/lef; Delta–Notch; erm; ephA4; pou2; radial glia; reticulospinal; commissural neuron Received 10 November 2003; Revised 23 April 2004; Accepted 25 April 2004

INTRODUCTION The vertebrate hindbrain is overtly segmented during early embryonic development. Hindbrain segments, or rhombomeres, serve to organize the hindbrain along the anterior– posterior axis. Numerous cell types and gene expression patterns are repeated in successive rhombomeres (Trevarrow et al., 1990; Clarke and Lumsden, 1993; Lumsden

and Krumlauf, 1996; Wingate and Lumsden, 1996), yet each segment develops its own unique identity and produces specialized structures accordingly. Several genes have been identified that are required for specification of segment identity. Anterior Hox genes play a prominent role in this process. Misregulation or disruption of Hox gene function leads to loss of specific segments or altered

segment identity (Lufkin et al., 1991; Chisaka et al., 1992; Mark et al., 1993; Alexandre et al., 1996; Studer et al., 1996; Jungbluth et al., 1999; Bell et al., 1999; Rossel and Capecchi, 1999; McClintock et al., 2001; Waskiewicz et al., 2002). Retinoic acid (RA) secreted from posterior mesoderm regulates Hox gene expression in the hindbrain. Treatment of embryos with exogenous RA, or

Biology Department, Texas A&M University, College Station, Texas Grant sponsor: National Institutes of Health, NIDCD; Grant number: R01 DC034806; Grant sponsor: American Heart Association; Grant number: 0365081Y. *Correspondence to: Bruce B. Riley, Biology Department, Texas A&M University, College Station, TX 77843-3258. E-mail: [email protected] DOI 10.1002/dvdy.20133 Published online 28 July 2004 in Wiley InterScience (www.interscience.wiley.com).

© 2004 Wiley-Liss, Inc.

RHOMBOMERE BOUNDARIES, Wnts, AND METAMERIC PATTERNING 279

conditions that disrupt RA signaling, result in changes in Hox gene expression and fundamental changes in hindbrain patterning (Papalopulu et al., 1991; Marshall et al., 1992; Simeone et al., 1995; Alexandre et al., 1996; Maden et al., 1996; Dupe et al., 1999; Niederreither et al., 2000; White et al., 2000). Several other segmentation genes, Krox20, vhnf1, and Kreisler/valentino (mafB), regulate even earlier stages of hindbrain segmentation and are required for normal activation of various Hox genes (Sham et al., 1993; Nonchev et al., 1996; Manzanares et al., 1997, 1999; Prince et al., 1998). Krox20 is expressed in rhombomeres 3 and 5 (r3 and r5), which fail to form in krox20 mutant mice (Swiatek and Gridley, 1993). The mouse Kreisler gene, and its zebrafish ortholog valentino (val), are expressed in r5 and r6. In mutant embryos, this region develops with some attributes of both rhombomeres but fails to establish normal positional identity of either r5 or r6 (Cordes and Barsh, 1994; McKay et al., 1994; Moens et al., 1996, 1998). In zebrafish, vhnf1 acts as an upstream regulator of val, and mutations in vhnf1 cause a phenotype very similar to val (Sun and Hopkins, 2001; Wiellete and Sive, 2003). In addition, expression of Fgf3 and Fgf8 in r4 is required for expression of val (Maves et al., 2002; Walsh et al., 2002). In embryos knocked down for both Fgfs, r5/6 does not develop properly. Rhombomeres are maintained as lineage-restricted compartments due to expression of ephrins and ephrin receptors in alternating segments (Xu et al., 1995, 1999; Bovenkamp and Greer, 1997; Mellitzer et al., 1999; Cooke et al., 2001; reviewed by Lumsden, 1999; Cooke and Moens, 2002). This signaling system establishes zones of mutual exclusion and, thereby, prevents intermixing of cells between even and odd numbered rhombomeres. Successive rhombomeres are also separated by boundary regions showing distinctive cellular morphology and localized expression of boundaryspecific genes (Trevarrow et al., 1990; Guthrie et al., 1991; Heyman et al., 1993, 1995; Mahmood et al., 1995, 1996; Moens et al., 1996; Yoshida and Colman, 2000). Although

it has been argued that rhombomere boundaries serve as physical barriers to interrhombomeric cell migration, environmental factors and mutations that specifically ablate rhombomere boundaries do not lead to increased segmental intermixing (Nittenberg et al., 1997; White et al., 2000). Presumably, the ephrin signaling system is sufficient to prevent intermixing even in the absence of morphological boundaries. What, then, is the role of rhombomere boundaries? One possibility is that they serve as signaling centers that help regulate the reiterated pattern of similar cell types within all rhombomeres. Boundary formation is increasingly recognized as an important regulator of coordinated development of complex morphogenetic fields (Reviewed by Guthrie, 1995; Lumsden and Krumlauf, 1996; Blair, 1997, 2003; Gaunt, 1997; Dahmann and Basler, 1999; Irvine, 1999; Irvine and Rauskolb, 2001; Sanson, 2001). The Drosophila wing margin, for example, forms at the dorsal–ventral boundary of the wing disc and expresses the paracrine factor, Wingless (Wg). The resulting concentration gradient of Wg coordinately regulates outgrowth and differentiation of the wing blade. A similar role is played by the apical ectodermal ridge (AER) in the vertebrate limb bud, although in this case, fibroblast growth factor (Fgf) proteins serve as the organizing signal. Regulation and maintenance of these signaling centers relies on interactions between Delta ligands and Notch receptors on adjacent rows of cells abutting the boundary (Couso et al., 1994, 1995; Kim et al., 1995, 1996; Rulifson and Blair, 1995; de Celis et al., 1996; Rulifson et al., 1996; de Celis and Bray, 1997; Laufer et al., 1997; Micchelli et al., 1997; Rodriguez-Esteban et al., 1997). Thus, boundary formation serves to coordinate patterns of growth and differentiation in adjacent developmental compartments and is often regulated by coherent fronts of Delta–Notch signaling. Here, we examine the role of rhombomere boundaries as Wnt signaling centers. At least four wnt genes, including wnt1, are preferen-

tially up-regulated at rhombomere boundaries soon after the boundaries form. Partial disruption of Wnt signaling leads to loss of fates associated with rhombomere boundaries and disorganization of cell types associated with rhombomere centers. Similar defects are seen in segmentation mutants in which upregulation of wnt gene expression does not occur properly at rhombomere boundaries. Various delta genes, including deltaA (dlA) normally form coherent fronts of expression in cells flanking rhombomere boundaries and are required to maintain boundaries as Wnt signaling centers. In dlAdx2 and mind bomb (mib) mutants, which are disrupted in Delta–Notch signaling (Appel et al., 1999; Itoh et al., 2003), rhombomere boundaries initially form but subsequently break down, and hindbrain structure becomes increasingly disorganized. Forced expression of wnt1 partially rescues hindbrain patterning in mib mutants. These data suggest that elevated expression of wnt genes at rhombomere boundaries helps regulate metameric patterning in the hindbrain and that Delta signals from flanking cells provide feedback to maintain the boundaries as Wnt signaling centers.

RESULTS Expression of wnt1 is first detected at 8 hours postfertilization (hpf) (80% epiboly) in the future midbrain-hindbrain border (Lekven et al., 2003). Between 11 and 14 hpf (1–10 somites), expression marks dorsal hindbrain and spinal cord, and by 16 hpf (14 somites), expression shows segmental regions of up-regulation (Fig. 1A–C). By 18 hpf (20 somites), the hindbrain domain bifurcates along the midline to form two dorsolateral columns with transverse bands of up-regulation at rhombomere boundaries. Boundary domains are progressively refined to form thin stripes by 27 hpf (Fig. 1D–F) and stripes are maintained through at least 48 hpf (data not shown). Similar patterns of expression are seen for wnt10b and wnt3a (Lekven et al., 2003; Fig. 1I, and data not shown). A different pattern is ob-

280 RILEY ET AL.

Fig. 1. Progressive refinement of wnt expression. All figures show dorsal views of wild-type embryos with anterior to the top. A–F: Expression of wnt1 at 11 hours postfertilization (hpf, A), 14 hpf (B), 16 hpf (C), 18 hpf (D), 22 hpf (E), and 27 hpf (F). Also shown at 27 hpf is expression of krox20 (red), confirming that expression of wnt1 does indeed mark rhombomere boundaries. G,H: Expression of wnt8b at 14 hpf (G) and 24 hpf (H). I: Expression of wnt3a at 27 hpf. J,K: Expression of dlA at 24 hpf (J) and 30 hpf (K). L: Double staining, showing expression of wnt1 (black) and dlA (Fast Red fluorescence) at 30 hpf. Scale bar ⫽ 65 ␮m in L (applies to A–L).

Fig. 3. Hindbrain development in tcf3b morphants. All embryos are tcf3b morphants except in H, which depicts a wildtype embryo. A: Expression of wnt1 at 27 hours postfertilization (hpf). B: Expression of dlA at 30 hpf. C: Expression of erm at 30 hpf. The otic vesicles (ov) and tectum (tec) are indicated. D: Expression of ephA4 at 24 hpf. E: Immunostaining with zn8 antibody at 30 hpf shows that commissural neurons (cn) fail to form. Staining of segmental structures along the lateral edges of the hindbrain marks pharyngeal arches and portions of cranial ganglia just out of the plane of focus. F,G: Immunostaining of reticulospinal neurons (rsn) at 24 hpf with anti-acetylated tubulin shows varying degrees of patterning defects in tcf3b-morphants. H,I: 3A10 antibody staining of Mauthner neurons (Mn) at 30 hpf in a wild-type embryo (H) and a tcf3b morphant (I). J: Immunostaining of radial glia (rg) with zrf1-zrf4. Numbers mark locations of the corresponding rhombomeres. Images show dorsal views with anterior to the top (A–I) or a lateral view with anterior to the left (J). Scale bar in J ⫽ 60 ␮m in E,H,I, 75 ␮m in D,F,G,J, 120 ␮m in A–C.

Fig. 2. Hindbrain development in wnt-deficient embryos. A–D: Expression of dlA at 30 hours postfertilization (hpf) in a Dfw5 mutant (A), a Dfw5 mutant injected with wnt3a-morpholino oligomers (MOs) (B), a wild-type embryo coinjected with wnt3a-MO and wnt8b-MO (C), and a Dfw5 mutant coinjected with wnt3a-MO and wnt8b-MO (D). E,F: Expression of erm at 30 hpf in a wild-type embryo (E) and a Dfw5 mutant coinjected with wnt3a-MO and wnt8b-MO (F). The otic vesicles (ov) and optic tectum (tec) are indicated to serve as landmarks. G,H: Expression of ephA4 at 24 hpf in a wild-type embryo (G) and a Dfw5 mutant coinjected with wnt3a-MO and wnt8b-MO (H). I,J: Expression of hoxb1a at 30 hpf in a wild-type embryo (I) and a Dfw5 mutant coinjected with wnt3a-MO and wnt8b-MO (J). K,L: Anti-acetylated tubulin staining of reticulospinal neurons (rsn) at 24 hpf in a wild-type embryo (K) and a Dfw5 mutant coinjected with wnt3a-MO and wnt8b-MO (L). M,N: Immunostaining with zn8 antibody at 30 hpf in a wild-type embryo (M) and a Dfw5 mutant coinjected with wnt3a-MO and wnt8b-MO (N). The zn8 antibody recognizes a cell adhesion molecule related to DM-Grasp (Fashena and Westerfield, 1999) and normally labels commissural neurons (cn), some cranial ganglia, pharyngeal arches, floor plate, and heart (Trevarrow et al., 1990). Commissural neurons are not produced in the wnt-deficient embryo. O,P: Immunostaining with a combination of four antibodies, zrf1–zrf4 (Trevarrow et al., 1990), was used to label radial glia (rg) at 30 hpf in a wild-type (O) and a Dfw5 mutant coinjected with wnt3a-MO and wnt8b-MO (P). Images show dorsal views with anterior to the top (A–N) or lateral views with anterior to the left (O,P). Numbers mark locations of the corresponding rhombomeres. Scale bar in P ⫽ 60 ␮m in O,P, 70 ␮m in A–D,G–N, 85 ␮m in E,F.


served for wnt8b, which is initially expressed in odd numbered rhombomeres by 14 hpf (Kelly et al., 1995; Fig. 1G). However, wnt8b is later preferentially expressed in cells at rhombomere boundaries and downregulated elsewhere (Fig. 1H). Thus, at least four wnt genes are progressively up-regulated in cells at rhombomere boundaries during hindbrain development. Refinement of wnt expression in the hindbrain is accompanied by progressive refinement of complementary patterns of delta genes. Expression of dlA, for example, marks bilateral clusters of cells in each rhombomere at 18 hpf, reflecting early patterns of neurogenesis (not shown). The clusters elongate into short transverse bands by 24 hpf, and by 30 hpf, expression is further refined to form thin parallel stripes of expression flanking rhombomere boundaries (Fig. 1J–L). Parallel stripes of expression are maintained through at least 48 hpf (data not shown). Similar striped patterns are observed for dlB and dlD, as well as upstream proneural genes zash1a and zash1b (data not shown). These gene expression patterns could indicate simultaneous regulation of neurogenesis and boundary formation, as documented for Delta–Notch signaling in the Drosophila wind disc (Rulifson and Blair, 1995; and see below). The significance of these striped patterns has not been addressed previously. We hypothesized that Wnt signals emanating from rhombomere boundaries serve to regulate segmental patterns of growth and differentiation, a process that is reflected by the complementary fronts of delta gene expression.

Impairment of the Wnt Pathway To examine the role of Wnt signaling in metameric hindbrain patterning, functions of wnt1, wnt3a, wnt8b, and wnt10b were disrupted in various combinations. Embryos homozygous for a deletion that removes both wnt1 and wnt10b (Dfw5, Lekven et al., 2003) develop with nearly normal expression of dlA and hindbrain morphology (Fig. 2A, and data not

shown). Knockdown of wnt3a in Dfw5 mutants causes only modest additional disturbances in patterning, as stripes of dlA are slightly less robust and show intermittent breaks (Fig. 2B). More severe patterning defects are seen in wild-type embryos injected with MOs directed against wnt3a and wnt8b, although some striping of dlA persists (Fig. 2C). However, knocking down wnt3a and wnt8b in Dfw5 mutants results in chaotic expression of dlA, with total loss of coherent stripes (Fig. 2D). This spotty pattern probably indicates that neurogenesis and lateral inhibition continue in wnt-deficient embryos, albeit in a disorganized pattern. Segmental identity is not disturbed in wnt-deficient embryos, as shown by normal expression of segmentation genes ephA4, hoxb1a, and krox20 (Fig. 2G–J, and data not shown), whereas other aspects of metameric patterning are severely altered. Expression of ETS-related gene erm, which normally forms transverse stripes spanning rhombomere centers, is expanded throughout the hindbrain (Fig. 2E,F). Commissural neurons, which normally develop and decussate at rhombomere boundaries, are not produced in wnt-deficient embryos (Fig. 2M,N). Clusters of primary neurons that form in rhombomere centers, including reticulospinal neurons, are disorganized and crowded in wnt-deficient embryos, and axonal trajectories are erratic (Fig. 2K,L). Defects in axonogenesis could reflect disruption of guidance cues normally provided by glial cells. Cellular processes produced by radial glia normally fan out to form parallel sheets that flank rhombomere boundaries (Trevarrow et al., 1990; Fig. 2O). In wnt-deficient embryos, radial glia form dense tangles with little discernible pattern (Fig. 2P). Thus, coexpression of wnt1, wnt3a, wnt8b, and wnt10b provides extensive redundancy in functions required for normal metameric patterning in the hindbrain. Because loss of Wnt ligands throughout the hindbrain does not address boundary-specific functions, we used an alternative approach to specifically ablate rhombomere boundaries: wild-type embryos were injected with morpholino oligomers

(MOs) to knockdown tcf3b, one of several mediators of Wnt signaling in the head (Kim et al., 2000; Dorsky et al., 2002, 2003). Knockdown of tcf3b does not fully disrupt Wnt signaling, because other tcf/lef genes continue to function, but rhombomere boundaries are specifically ablated as judged by loss of morphological, cytological, and molecular markers of boundaries (Dorsky et al., 2003). Accordingly, wild-type embryos knocked down for tcf3b (tcf3b morphants) fail to produce discrete stripes of wnt1 up-regulation in boundary regions, while lower expression persists throughout the rest of the dorsolateral domain (Fig. 3A). The hindbrain shows irregular regions of constriction in which wnt1 expression remains contiguous across the midline. Expression of dlA fails to resolve into discrete parallel stripes (Fig. 3B). In addition, dlA expression marks fewer cells than normal, a phenotype not seen in wntdeficient embryos. This finding could reflect the effects of impairing both Wnt-dependent as well as Wnt-independent functions of Tcf3b (reviewed by Roose and Clevers, 1999). Despite this difference, tcf3b morphants show the same defects in metameric patterning as embryos deficient for the four wnt genes examined above. For example, expression of erm is expanded throughout the hindbrain in tcf3b morphants (Fig. 3C) and commissural neurons are not produced (Fig. 3E). Thus, fates or signaling events associated with rhombomere centers appear to be expanded at the expense of rhombomere boundaries. Segmental identities are not altered since tcf3b morphants show normal expression of ephA4, krox20, hoxb1, hoxb4, and val (Dorsky et al., 2003; Fig. 3D, and data not shown). Loss of rhombomere boundaries does not lead to increased cell-mixing between adjacent segments. Maintenance of alternating domains of ephrins and ephrin receptors is presumably sufficient to prevent intermixing. Reticulospinal neurons are disorganized and crowded, and their axons show erratic trajectories and variable defects in fasciculation (Fig. 3F–I). The pattern of radial glia is correspondingly chaotic (Fig. 3J). Thus, tcf3b morphants develop with a range of patterning defects charac-

282 RILEY ET AL.

terized by loss of rhombomere boundaries and associated cell types, and irregular patterning of cell types associated with rhombomere centers. Development of region-specific cell types such as branchiomotor neurons is relatively normal in tcf3b morphants (Dorsky et al., 2003), a finding consistent with maintenance of normal segmental identity.

Mutants That Fail to Initiate Boundary Formation The defects in tcf3b morphants and wnt-deficient embryos could result from specific loss of organizing signals from rhombomere boundaries or, alternatively, could reflect generalized defects caused by reducing Wnt signaling throughout the hindbrain. To distinguish between these possibilities, we examined hindbrain patterning in spiel-ohne-grenzen (spg) mutants, which are disrupted in pou2 function (Burgess et al., 2002). spg mutants are not impaired in Wnt signal transduction but nevertheless fail to form rhombomere boundaries (Hauptmann et al., 2002; Reim and Brand, 2002), probably due to abnormal interactions between adjacent segments. Like tcf3b morphants, spg mutants fail to show boundary-specific up-

regulation of wnt1, and dlA expression is chaotic (Fig. 4A,B). Commissural neurons are often absent or underproduced, and when present, their axons decussate aberrantly through rhombomere centers (Fig. 4G,H). erm is expressed broadly in the hindbrain, albeit at low levels (Fig. 4C–E). Neurons normally seen in rhombomere centers are crowded together in a disorganized mass, and their axons are poorly fasciculated (Fig. 4F). Radial glia are also poorly organized, forming numerous closely packed sheets with no clear segmental relationship (Fig. 4I,J). Thus, hindbrain patterning in spg mutants resembles that in tcf3b morphants and wnt-deficient embryos. We also examined hindbrain pattering in valentino (val) mutants. In val mutants, defective specification of the r5/6 region results in faulty interactions with r4 and r7 (Moens et al., 1996, 1998). As a consequence, rhombomere boundaries do not form posterior to the r3/4 boundary. Expression of wnt1 in the hindbrain of val mutants often fails to show upregulation in stripes posterior to the r3/r4 boundary, and expression of dlA is correspondingly mispatterned (Fig. 5A,B). However, the degree of disruption of hindbrain patterning is

highly variable. In approximately half of val mutants, wnt1 either shows up-regulation in one broad stripe in the posterior hindbrain (not shown) or two closely spaced posterior stripes (Fig. 5C). There is little correlation between patterning on the right and left sides in individual specimens. When discrete stripes of wnt1 do appear in the posterior hindbrain, expression of dlA is partially rescued, always showing a consistent complementary pattern (Fig. 5C). Commissural neurons are often missing or underproduced in the posterior hindbrain (Fig. 5E), and when present, their axons sometimes decussate aberrantly as seen in spg mutants. Reticulospinal neurons and radial glia are disorganized in the posterior hindbrain (Fig. 5F,G), and erm expression is usually expanded throughout this region (Fig. 5D). In summary, spg and val mutants fail to generate normal rhombomere boundaries, leading to failure of wnt1 to up-regulate in segmental stripes. Metameric patterning is disorganized in a manner similar to wnt-deficient embryos, supporting the notion that high levels of Wnt are required at rhombomere boundaries to regulate development within each segment.

Fig. 4. Hindbrain development in spg mutants. All embryos are spg homozygous mutants except for E and I, which show wild-type controls. A: wnt1 expression at 27 hours postfertilization (hpf). B: dlA expression at 30 hpf. C–E: erm expression at 30 hpf in spg mutants (C,D) and a wild-type embryo (E). Spatial patterns are highly variable between spg mutants, but the level of expression is invariably reduced. For landmarks, the otic vesicles (ov) and tectum (tec) are indicated. F: Reticulospinal neurons (rsn) labeled with anti-acetylated tubulin at 24 hpf. G,H: Commissural neurons (cn) labeled with zn8 antibody at 30 hpf. Commissural neurons are absent in anterior rhombomeres and in posterior rhombomeres produce aberrant commissures (c) through the center of the r5 region. H is an enlargement of the left side of the specimen shown in G. I,J: Radial glia (rg) labeled with zrf1-4 antibodies at 30 hpf in a wild-type embryo (I) and a spg mutant (J). The wild-type rg pattern is an enlargement of the specimen shown in Figure 2O. Numbers indicate the positions of corresponding rhombomeres. Images show dorsal views with anterior to the top (A–C,F–H) or a lateral view with anterior to the left (D,E,I,J). Scale bar in J ⫽ 25 ␮m in H, 50 ␮m in I,J, 65 ␮m in A–C,F,G, 150 ␮m in D,E. Fig. 5. Hindbrain development in val mutants. All embryos are val homozygous mutants. A: wnt1 expression at 27 hours postfertilization (hpf). B: dlA expression at 30 hpf. C: Two-color in situ hybridization, showing wnt1 (black) and dlA (Fast Red fluorescence) on the left side of the hindbrain at 30 hpf. D: erm expression at 30 hpf. The position of the otic vesicles (ov) is indicated. E: Commissural neurons (cn) labeled with zn8 antibody at 30 hpf. The number and positions of commissural axons is highly variable in the posterior hindbrain. Note that the right side of the r5/6 region has only a single commissure (c), which is not in register with the two commissures on the left. F: Reticulospinal neurons (rsn) labeled with anti-acetylated tubulin at 24 hpf. G: Radial glia labeled with zrf1-4 antibodies at 30 hpf. All images show dorsal views with anterior to the top except for G, which shows a lateral view with anterior to the left. Numbers indicate corresponding rhombomeres. Scale bar in G ⫽ 50 ␮m in C,G, 70 ␮m in E, 90 ␮m in F, 120 ␮m in A,B,D. Fig. 6. Hindbrain patterning in Delta–Notch mutants. A–C: Hindbrain morphology at 24 hours postfertilization (hpf) in wild-type (A), dlAdx2 (B), and mibta52b (C) embryos. Positions of rhombomeres 1–7 are indicated. D–F: Dorsal views, showing expression of foxb1.2 at 24 hpf in wild-type (D), dlAdx2 (E), and mibta52b (F) embryos. G: erm expression in mibta52b at 30 hpf. The otic vesicles (ov) are indicated to serve as landmarks. H,I: ephA4 expression at 24 hpf in a wild-type embryo (H) and a mibta52b mutant (I). J,K: Expression of krox20 at 24 hpf in dlAdx2 (J) and mibta52b (K) embryos. L,M: Expression of hoxb1a at 24 hpf in dlAdx2 (L) and mibta52b (M) embryos. N: Reticulospinal neurons (rsn) in a mibta52b mutant labeled with anti-acetylated tubulin at 24 hpf. O: zn8 antibody staining of a mibta52b mutant at 30 hpf shows that commissural neurons (cn) are not produced. P,Q: Mauthner neurons (Mn) labeled with 3A10 antibody at 30 hpf in dlAdx2 (P) and mibta52b (Q) embryos. Abnormal axonal trajectories are indicated (arrows). R: Radial glia (rg) labeled with zrf1-4 antibodies at 30 hpf in a mibta52b mutant. Images show lateral views with anterior to the left (A–C,H,I,R) or dorsal views with anterior to the top (D–G,J–Q). Scale bar in R ⫽ 55 ␮m in A–C, 75 ␮m in O–Q, 100 ␮m in R,125 ␮m in D–N.


Delta–Notch Signaling and Boundary Maintenance

Fig. 4.

Fig. 5.

Fig. 6.

Delta–Notch signaling appears to serve multiple functions in the hindbrain. Its best known function is in regulating neurogenesis and lateral inhibition, during which delta gene expression is patchy and transitory (reviewed by Lewis, 1996). In contrast, the stripes of delta gene expression shown here are relatively stable features of hindbrain patterning. Stable fronts of delta gene expression are often seen in association with boundary formation, such as in the vertebrate limb bud and the Drosophila wing margin (reviewed by Blair, 1997; Dahman and Basler, 1999). In those systems, Delta signals provide feedback to stabilize and refine signaling centers that emit long-range organizing signals. We hypothesized that the stable fronts of delta expression observed in the hindbrain play a role in maintaining rhombomere boundaries as Wnt signaling centers. Hindbrain patterning, therefore, was examined in mutants disrupted for Delta– Notch signaling. A mutant allele of dlA, dlAdx2, is phenotypically variable (Appel et al., 1999; Riley et al., 1999), but severely affected mutants show defects consistent with loss of rhombomere boundaries. Although rhombomeres appear morphologically normal through 20 hpf in all dlAdx2 mutants (not shown), in severely affected mutants, the hindbrain appears highly disorganized by 24 hpf and boundary regions appear irregular and indistinct (Fig. 6B). Boundary-specific markers are also perturbed. Expression of the forkhead class gene foxb1.2 (formerly mariposa) normally marks the ventral midline of the hindbrain as well as rhombomere boundaries (Moens et al., 1996; Fig. 6D). At 24 hr, severely affected dlAdx2 mutants show strong foxb1.2 expression along the ventral midline but almost no detectable expression at rhombomere boundaries (Fig. 6E). Similar defects are seen in mind bomb (mib) mutants, which are globally disrupted in Delta–Notch signaling (Itoh et al., 2003). The mibta52b phenotype is fully penetrant and is consistently more se-

284 RILEY ET AL.

vere than in dlAdx2 mutants. Morphological boundaries are not discernible in mibta52b mutants at 24 hpf, and foxb1.2 expression is almost totally ablated (Fig. 6C,F). Despite apparent loss of rhombomere boundaries by 24 hpf, markers of segmental identity, including ephA4, krox20, hoxb1a, hoxb4, and val, are maintained in their proper spatial domains, although expression levels appear reduced in mibta52b mutants (Fig. 6H–M, and data not shown). The anterior and posterior edges of these domains are irregular, but there is no apparent intermixing between adjacent segments. Loss of boundary regions is accompanied by expansion of erm expression throughout the hindbrain (Fig. 6G). Although many neurons are overproduced in dlAdx2 and mibta52b mutants due to disruption of lateral inhibition, commissural neurons are not produced (Fig. 6O; and Jiang et al., 1996). Reticulospinal neurons and other r-center neurons are dramatically overproduced and often show axonal pathfinding errors and fasciculation defects (Fig. 6N,P,Q). Radial glia are also produced in large numbers but form highly chaotic patterns of radial fibers (Fig. 6R). Many of the above defects are consistent with the effects of disrupting rhombomere boundaries. The pattern of wnt1 expression in dlAdx2 embryos is essentially normal through 20 hpf, but boundary-specific stripes are strongly disrupted by 27 hpf in severely affected mutants (Fig. 7A–D). Distinct stripes of wnt1 expression are also evident in mibta52b mutants at 16 hr hpf, but expression becomes increasingly disorganized with time (Fig. 7F–I). By 27 hpf, mibta52b mutants lose almost all expression of wnt1. Thus rhombomere boundaries are initially specified in dlAdx2 and mibta52b mutants but are not properly maintained. Accordingly, expression of dlA is nearly normal in dlAdx2 mutants through 18 hpf (not shown) but later becomes increasingly erratic. By 30 hr, severely affected dlAdx2 embryos show chaotic patches of expression with little indication of the parallel stripes (Fig. 7E). In mibta52b mutants, expression of dlA is totally ablated in large portions of the hindbrain by 30 hpf and parallel stripes of

expression are never observed at this time (Fig. 7J).

Rescue of mibta52b Mutants To test the hypothesis that failure to maintain Wnt signaling contributes to the mibta52b phenotype, we examined whether forced expression of wnt1 could bypass some of the defects seen in mibta52b mutants. Misexpression was accomplished using a plasmid construct to drive expression of wnt1 with a zebrafish heat-shock promoter (Shoji et al., 1998). In the construct used here (pHS-wnt1-myc), the wnt1 sequence is fused to a human C-myc epitope to facilitate localization of recombinant protein. Injection of wnt1-myc mRNA into Xenopus embryos efficiently induces axial duplication (A. Lekven, unpublished data), indicating that the recombinant protein retains potent biological activity. Plasmid DNA was injected into blastulae derived from mibta52b/⫹ heterozygous parents, which results in variable mosaic inheritance of the plasmid sequence. To obtain maximal activation of the heat-shock promoter, embryos must be incubated at 39 – 40°C for 1 hr (Shoji et al., 1998). However, examination of dlA expression at 30 hpf showed that a single heat shock delivered at 18 hr or 24 hpf was not sufficient to rescue injected mibta52b mutant embryos (Fig. 7M, and data not shown). We speculated that wnt1 would have to be expressed continuously to maintain expression of dlA. Because prolongation of heat shock for more than 1 hr is lethal, we developed a modified regimen of brief serial heat shocks, administered hourly between 16 and 30 hr, which permits development to proceed relatively normally (see Experimental Procedures section). A different heatshock construct that drives expression of green fluorescent protein (pHS-GFP) was used to optimize the modified heat-shock regimen. With this construct, the modified regimen was found to give lower levels of expression than a single 1-hr heat shock but permitted sustained expression over a prolonged period (data not shown). When mibta52b mutant embryos were injected with

pHS-wnt1-myc and given serial heat shocks beginning at 16 hr, 48% (35/ 73) showed substantial signs of rescue, including markedly improved patterns of dlA expression in the hindbrain at 30 hpf (Fig. 7K; Table 1). Expression of dlA was maintained at relatively high levels and was organized into parallel stripes with relatively straight edges. Rescued specimens also showed more normal hindbrain morphology (not shown). In many cases, organized expression of dlA was maintained only in a portion of the hindbrain, presumably reflecting mosaic expression of the injected construct (see below). Unfortunately, it was not possible to visualize the wnt1-myc protein with anti-myc antibodies. This finding is probably because the Wnt-myc protein is rapidly secreted and does not accumulate to high levels. Nevertheless, we infer that expression of Wnt-myc is responsible for rescue because mibta52b mutant embryos receiving pHS-wnt1-myc injection without heat shock, or heat shock without plasmid injection, failed to maintain dlA expression (Fig. 7L; Table 1). Similarly, mibta52b mutants injected with the pHS-GFP and given serial heat shocks failed to maintain dlA expression (Fig. 7O; Table 1). Thus, forced expression of wnt1 permits mibta52b mutants to maintain better overall organization in the hindbrain. In contrast to the above results, no rescue of hindbrain patterning was detected in plasmid-injected mibta52b mutants that did not begin serial heat shocks until 24 hr (Fig. 7N; Table 1). Presumably, once boundaries are lost and patterning becomes chaotic, forced expression of wnt1 is not sufficient to restore normal patterning. This finding indicates that wnt1 expression must be maintained for normal hindbrain patterning. When wild-type embryos were injected with pHS-wnt1-myc and then given serial heat shocks, hindbrain morphology and dlA expression were relatively normal. In some cases, stripes of dlA expression appeared slightly more diffuse than normal, but the overall pattern was unchanged (Fig. 7P; Table 1). Thus, the level of protein produced by the heat-shock construct does not signif-


Fig. 7. Expression of wnt1 and dlA in Delta–Notch mutants. All images show dorsal views with anterior to the top. A–D: wnt1 expression in dlAdx2 mutants at 16 hours postfertilization (hpf, A), 18 hpf (B), 20 hpf (C), and 27 hpf (D). E: dlA expression in a severely affected dlAdx2 mutant at 30 hpf. F–I: wnt1 expression in mibta52b mutants at 16 hpf (F), 18 hpf (G), 20 hpf (H), and 27 hpf (I). J: Expression of dlA in a mibta52b mutant at 30 hpf. K–N: dlA expression at 30 hpf in mibta52b mutants injected with pHS-wnt1-myc and subjected to serial heat shocks beginning at 16 hpf (K), no heat shock (L), a single heat shock delivered at 24 hpf (M), or serial heat shocks beginning at 24 hpf (N). O: expression of dlA at 30 hpf in a mibta52b mutant injected with pHS-GFP and given serial heat shocks beginning at 16 hpf. P: dlA expression in a wild-type embryo injected with pHS-wnt1-myc and given serial heat shocks beginning at 16 hpf. Q: wnt1 staining in a wild-type embryo injected with pHS-wnt1-myc and given serial heat shocks beginning at 16 hpf. R–T: wnt1 staining in mibta52b mutants injected with pHS-wnt1-myc and given serial heat shocks beginning at 16 hpf. Arrowheads indicate stripes of wnt1 expression at presumptive rhombomere boundaries. Note the close proximity of darkly stained plasmid-containing cells in or near each stripe of endogenous wnt1 (R,S), compared with the absence of endogenous wnt1 stripes when pHS-wnt1myc is absent (T). Red arrows show groups of cells containing pHS-wnt1-myc in the center of rhombomere 3. This rhombomere is particularly easy to locate and assess, because it is abnormally large at this stage in mibta52b mutants, and its position relative to the anterior edge of the otic vesicle is highly reproducible. Note the absence of stripes of endogenous wnt1 expression in r-center cells surrounding plasmid-containing cells. Scale bar in T ⫽ 45 ␮m in S, 90 ␮m in A–R,T.

icantly impact the organizing effects of endogenous Wnt signaling. To better understand the mechanism of mutant rescue, we analyzed wnt1 expression in embryos injected with pHS-wnt1-myc. In principle, the in situ probe should hybridize with both plasmid-derived and endogenous wnt1 mRNA. However, all plasmid-injected embryos showed variable spotty patterns of intense hybridization regardless of whether embryos received heat shocks. These foci of staining usually formed within minutes of starting the colorimetric reaction and presumably reflect hybridization to multiple copies of plasmid DNA inherited mosaically by variable numbers of cells. Widespread plasmid-hybridization obscured the much weaker signal from wnt1 transcription, but endogenous wnt1 expression could still be observed when plasmid was inherited by moderate numbers of cells. In wild-type embryos, presumptive plasmid expression did not disturb the endogenous wnt1 pattern (Fig. 7Q; Table 2), in keeping with the negligible effects of plasmid on dlA gene expression (Fig. 7P). In mibta52b mutants, presumptive plasmid expression appeared to have a dramatic nonautonomous effect on endogenous wnt1 expression. Of 36 mibta52b mutants that were injected and heat shocked, 15 (42%) maintained endogenous wnt1 expression in transverse stripes that appeared to coincide with rhombomere boundaries (Fig. 7R,S; Table 2). In all cases, stripes of endogenous wnt1 expression occurred near clusters of plasmid-containing cells. In contrast, 16 of 36 of the injected mutants contained either few or no plasmid-containing cells within the hindbrain, and none of these specimens maintained endogenous wnt1 expression (Fig. 7T; Table 2). In the remainder of injected mutants (5/36), plasmid hybridization was too intense to clearly visualize endogenous wnt1 expression. The ability to rescue hindbrain patterning with only moderate numbers of presumptive Wnt1-expressing cells is surprising. It is possible that wnt1 regulation at rhombomere boundaries involves an auto-regulatory feedback loop, such that Wnt1 from plasmid-containing cells aug-

286 RILEY ET AL.

TABLE 1. Pattern of dlA in Plasmid-Injected, Heat-Shocked Embryos Number of embryos Genotype ta52b

mib mibta52b mibta52b mibta52b mibta52b mibta52b ⫹/⫹

Plasmid

Heat shock

Rescue/normala

Non-rescue/abnormal

Total

pHS-wnt1-myc pHS-wnt1-myc None pHS-wnt1-myc pHS-wnt1-myc pHS-GFP pHS-wnt1-myc

1 hour, 24 hpf Serial, 16 hpf Serial, 16 hpf None Serial, 24 hpf Serial, 16 hpf Serial, 16 hpf

0 35 (48%) 0 0 0 0 31 (100%)

13 (100%) 38 (52%) 18 (100%) 17 (100%) 18 (100%) 16 (100%) 0

13 73 18 17 18 16 31

a

Denotes mibta52b embryos showing parallel stripes of dlA expression in the hindbrain/wild-type embryos showing a normal pattern of dlA expression.

TABLE 2. Pattern of wnt1 in Embryos Injected With pHS-wnt1-myc and Given Serial Heat Shock No. of embryos (no. with plasmid-signal in hindbrain) Genotype

Rescue/normala

Non-rescue/abnormal

Indeterminateb

Total

mibta52b ⫹/⫹

15 (15) 28 (20)

16 (2) 0

5 (5) 2 (2)

36 30

a

Denotes mibta52b embryos showing robust stripes of endogenous wnt1 expression in the hindbrain/wild-type embryos showing a normal pattern of endogenous wnt1 expression. b Endogenous wnt1 expression is obscured by widespread plasmid hybridization.

ments endogenous signaling and helps maintain wnt1 stripes. Several features of the wnt1 stripes in rescued mutants are noteworthy. First, unlike the pattern in wild-type embryos, stripes of wnt1 in rescued mutants usually spanned the hindbrain. This finding suggests incomplete separation of dorsolateral regions of the hindbrain, possibly reflecting defects in midline development known to occur in mibta52b mutants (Appel et al., 1999; Bingham et al., 2003). Second, stripes of wnt1 expression at boundaries often extended well beyond the vicinity of plasmid-containing cells, whereas the presence of plasmid-containing cells was not sufficient to induce extensive wnt1 expression in rhombomere centers. These data suggest that cells at rhombomere boundaries but not cells in rhombomere centers are predisposed to propagate or relay Wnt signals. Finally, stripes of wnt1 expression, although much improved over the pattern in nonrescued mibta52b mutants, were less regular than in wild-type embryos. This finding is consistent with the notion that Delta signals from flanking cells are re-

quired to maintain, sharpen, and refine patterning at rhombomere boundaries.

DISCUSSION We have examined the role of rhombomere boundaries as signaling centers that help organize metameric patterning in the hindbrain. At least four wnt genes begin to show up-regulation in stripes at rhombomere boundaries soon after the boundaries appear, and stripes of wnt expression are maintained long after boundaries are no longer morphologically visible. Disruption or knockdown of all four wnt genes disturbs metameric organization of the hindbrain: molecular markers and neurons associated with rhombomere boundaries are ablated, and markers and cell types associated with rhombomere centers are crowded together and disorganized. Knockdown of tcf3b specifically ablates morphological boundaries and causes similar changes in metameric patterning. Although boundary-stripes of wnt1 up-regulation are lost, basal wnt1 expression is

maintained throughout the dorsolateral domain. Thus, patterning defects in tcf3b morphants are relatively specific and do not reflect wholesale loss of Wnt signaling or loss of the roofplate. Similar defects in metameric patterning are observed in spg, val, and Delta–Notch mutants, which show loss of rhombomere boundaries by means of different underlying mechanisms. In spg and val mutants, loss of boundaries arises from faulty interactions between adjacent segments (Moens et al., 1996, 1998; Hauptmann et al., 2002). Of interest, while val mutants invariably fail to form morphological boundaries in the posterior hindbrain, regulation of wnt1 stripes is variable, ranging from near normal to complete loss. Metameric organization is correspondingly variable, and there is strong correlation between altered patterns of wnt1 stripes and flanking stripes of dlA expression. Parallel stripes of delta gene expression appear to play a role in maintaining rhombomere boundaries. Disruption of Delta– Notch signaling in dlAdx2 and mibta52b mutants results in progres-


sive deterioration of boundaries and increasingly chaotic patterning in the hindbrain. Patterning is particularly severe in mibta52b mutants, as wnt1 expression is almost totally ablated by 30 hpf and dlA is limited to small random patches. In agreement with a previous study (Bingham et al., 2003), we find that expression of segmentation genes is reduced in mibta52b, but this need not indicate loss of segment identity. Because cells in the CNS differentiate prematurely in Delta– Notch mutants, down-regulation of segmentation genes might reflect acceleration of the normal time course seen in wild-type embryos. Furthermore, dlAdx2 mutants are less severely compromised in Delta–Notch signaling and do not show appreciably reduced expression of segmentation genes, yet they show the same defects in metameric patterning seen in mibta52b. Importantly, forced expression of wnt1 in mibta52b mutants partially rescues hindbrain patterning, as revealed by persistence of robust parallel stripes of dlA expression, maintenance of stripes of wnt1 expression, and general improvement of hindbrain morphology. Together, these data indicate that boundary-specific stripes of wnt expression are required for normal hindbrain patterning and are maintained by Delta signals from flanking cells. Establishment of signaling centers at rhombomere boundaries marks a relatively late stage of hindbrain patterning. Hindbrain segmentation genes act first during gastrulation to lay down rhombomere territories, after which alternating patterns of ephrins and ephrin receptors maintain rhombomere integrity by establishing zones of mutual exclusion (reviewed by Lumsden, 1999; Cooke and Moens, 2002). Ephrin signaling is also responsible for establishing the morphological boundaries between segments. These early patterning mechanisms are sufficient to generate repeated patterns of some early-forming cell types, such as primary reticulospinal neurons. However, various other cell types are not produced or are mispat-

terned if rhombomere boundaries are not formed or maintained properly. Understanding the role of rhombomere boundaries has been elusive. The data presented here suggest that rhombomere boundaries provide a series of Wnt signaling centers that help induce or organize repeated patterns of similar cell types in successive segments. This chain of regulatory events can be compared with development of body segments in Drosophila (reviewed by Sanson, 2001). Segmentation genes first establish parasegment territories, which are later subdivided into a repeated series of homologous cell types by segment polarity genes, including Wg. A closer analogy can be found in the Drosophila wing disc. Dorsal and ventral compartments are first specified by homeotic selector genes, and the DV interface is then established as a Wg signaling center that organizes subsequent outgrowth and patterning of the wing. Wg expression at the margin promotes expression of Delta and Serrate in flanking cells, which in turn provide feedback to maintain Wg at the margin (Couso et al., 1994, 1995; Kim et al., 1995, 1996; Rulifson and Blair, 1995; de Celis et al., 1996; Rulifson et al., 1996; de Celis and Bray, 1997; Micchelli et al., 1997). A similar mechanism operates in the vertebrate limb bud to induce and maintain the apical ectodermal ridge (AER). While the AER emits Fgf as a long-range organizing signal, it is initially induced by Wnt3/Wnt3a and requires a coherent front of Delta–Notch signaling (Laufer et al., 1997; Rodriguez-Esteban et al., 1997; Kengaku et al, 1998; Kawakami et al., 2001; Barrow et al., 2003). The similarities between these diverse systems could reflect cooption of a conserved developmental module—a self-reinforcing genetic/signaling feedback loop that spatially organizes large fields of cells. It is still unclear how Wnt signaling regulates metameric patterning in the hindbrain. Boundary signals might act at a distance, forming morphogen gradients that coordinate patterns of growth and differentiation throughout each rhom-

bomere. At the very least, boundary signals are required locally for development of commissural neurons, and also appear to provide foci for organizing radial glia (Trevarrow et al., 1990; Yoshida and Colman, 2000, this study), the fibers of which provide important guidance cues for axonal pathfinding.

Studies in Other Vertebrates Mutations in segmentation genes in mouse also cause local disruption of rhombomere boundaries and disorganization of neural architecture (Lufkin et al., 1991; Chisaka et al., 1992; Mark et al., 1993; McKay et al., 1994; Schneider-Maunoury et al., 1997). Similarly, strong perturbations of RA signaling in embryos of many species leads to loss of rhombomere boundaries and disorganization of cranial nerves and ganglia (Holder and Hill, 1991; Dupe et al., 1999; Niederreither et al., 2000). Unfortunately, these studies are inconclusive with regard to the role of rhombomere boundaries, because all of the above perturbations alter segment identity, which is likely to contribute significantly to defects in cranial nerve development. However, local application of RA beads to the segmented chick hindbrain leads to ablation of rhombomere boundaries without loss of segmental integrity (Nittenberg et al., 1997). Expression of Hox genes is not immediately altered and intermixing of cells between compartments does not occur. Nevertheless, organization and axonal trajectories of cranial nerves are altered. In rat embryos with intermediate RA deficiencies, nearly half show normal Hox gene expression, yet all such embryos show neural disorganization and either indistinct or ablated rhombomere boundaries (White et al., 2000). The loss of rhombomere boundaries may be the most salient cause of irregular neural patterning in these cases. It is unclear whether the role of Wnt ligands as organizing signals is conserved in other vertebrates. Some depictions of mouse Wnt1 expression appear to show up-regulation at rhombomere boundaries (Manzanares et al., 2000). Moreover, a Wnt1 promoter element has been

288 RILEY ET AL.

identified that is sufficient to drive expression of a transgenic reporter gene in stripes at rhombomere boundaries (Rowitch et al., 1998). Disruption of mouse Wnt1 does not significantly alter hindbrain patterning, but this finding probably reflects redundancy. Disruption of both Wnt1 and wnt3a, which are normally coexpressed in the dorsal neural tube, leads to a synergistic deficiency of dorsal cell types (Ikeya et al., 1997). Segmental patterning of cranial nerves is also abnormal. In dreher mutant mouse embryos, Wnt1 expression is reduced both in boundary and interboundary regions and is lost entirely in r6 (Manzanares et al., 2000). The dreher mutation disrupts Limx1a, which is normally expressed in the roof plate and is required for its formation (Millonig et al., 2000). Patterning of cranial nerves in dreher mutants is disorganized in r6 and, to a lesser degree, in r5. Neither dreher mutants nor Wnt1-wnt3a double mutants show detectable changes in regulation of segmental identities. Although we have focused on the role of Wnts, other signaling molecules may contribute to the organizing activity of rhombomere boundaries. In mouse and chick, Fgf3 is preferentially expressed at rhombomere boundaries (Mahmood et al., 1995, 1996). Although expression of fgf3 in the zebrafish hindbrain is restricted to r4 (Phillips et al., 2001; Maves et al., 2002; Walsh et al., 2002), other fgf homologs may yet be identified that potentially regulate metameric patterning. In this context it is interesting that erm, which is thought to be induced by Fgf signaling (Raible and Brand, 2001; Roehl and Nu¨sslein-Volhard, 2001), is expressed in rhombomere centers in zebrafish. However, no Fgf genes have yet been identified that can account for this pattern, neither are other Fgf-inducible genes sprouty4 and pea3 expressed in a similar pattern (Fu¨rthauer et al., 2001; our unpublished observations). Thus, erm might be regulated at this stage by some other pathway(s), possibly including negative regulation by Wnts or other boundary signals. Further studies are needed to resolve these issues.

EXPERIMENTAL PROCEDURES Strains and Developmental Conditions The wild-type strain was derived from the AB line (Eugene, OR). The dlAdx2 mutation was induced in the wild-type background with ENU and encodes a dominant-negative protein (Appel et al., 1999; Riley et al., 1999). The mibta52b mutation was induced with ENU in the Tu wild-type strain (Jiang et al., 1996) and is likely to be a strong hypomorph or null allele (Itoh et al., 2003). The spgHi349 mutation was induced by viral insertion and is thought to be a null allele (Burgess et al., 2002). The valb337 mutation was induced in the AB strain with ENU and is thought to be a null allele (Moens et al., 1996). The Df(LG23)wnt1w5 deletion (herein referred to as Dfw5) was induced by gamma-irradiation and identified in a polymerase chain reaction– based screen for loss of wnt1 (Lekven et al., 2000). Subsequent mapping showed that Dfw5 removes 500 kb or less of LG 23, including the closely linked wnt10b locus (Lekven et al., 2003). Reference to mutants in the text indicates homozygous mutants unless otherwise noted. Embryos knocked down for specific gene activities are referred to as morphants. Descriptions of mutant and morphant phenotypes are based on assessment of 25 or more affected embryos per time-point, except where noted. Embryos were developed in an incubator at 28.5°C in water containing 0.008% Instant Ocean Salts. Developmental stages are expressed as hours postfertilization (hpf).

5.0 mM HEPES, pH 7.6; Nasevicius and Ekker, 2000) at a concentration of 5 ␮g/ ␮l, and approximately 1 nl was injected into the yolk of 1–2 cell wild-type blastulae.

Whole-Mount In Situ Hybridization Whole-mount in situ hybridization was performed at 67°C using probes for dlA, dlB, dlD (Appel and Eisen, 1998; Haddon et al., 1998), ephA4 (Xu et al., 1995), erm (Roehl and Nu¨sslein-Volhard, 2001), hoxb1 and hoxb4 (Prince et al., 1998), krox20 (Oxtoby and Jowett, 1993), foxb1.2 (formerly mariposa, Moens et al., 1996), val (Moens et al., 1998), wnt1 (Molven et al., 1991), wnt3a (A. Lekven, unpublished data), wnt8b (Kelly et al., 1995), and zash1a or zash1b (Allende and Weinberg, 1994).

Immunofluorescence Antibody staining was performed as previously described (Riley et al., 1999). Fixed embryos were incubated with anti-acetylated tubulin (Sigma T6793, diluted 1:100) or various antibodies obtained from the Developmental Studies Hybridoma Bank, including zn8 (diluted 1:1,000); zrf 1, 2, and 4 (diluted 1:100); zrf 3 (diluted 1:250); and 3A10 (diluted 1:20). Embryos were then washed and incubated with one of the following secondary antibodies: Alexa 546-conjugated goat anti-rabbit IgG (Molecular Probes A-11010, diluted 1:50) or Alexa 488-conjugated goat anti-mouse IgG (Molecular Probes a-11001, diluted 1:50).

Misexpression of wnt1 MO Injections MOs were synthesized by Gene Tools. tcf3b-MO (Dorsky et al., 2003) was provided as a gift from Richard Dorsky. Additional MO sequences are as follows: wnt3a-MO sequence, 5⬘-GTTAGGCTTAAACTGACACGCACAC-3⬘; wnt8b-MO sequence, 5⬘AGGGAGACTTTTCTTCACCTTTCAC-3⬘. All MOs are designed to block translation of the gene product in question. In each case, MO was dissolved in Danieux solution (58 mM NaCl, 0.7 mM KCl, 0.4 mM MgSO4, 0.6 mM Ca(NO3)2,

pHS-wnt1-myc plasmid encodes zebrafish Wnt1 fused to human myc epitope tag, downstream of a zebrafish heat-shock promoter (Shoji et al., 1998). A second plasmid, pHSGFP, drives GFP from the heat-shock promoter and was used to optimize a heat-shock regimen for our experimental needs. Circular plasmid (25 ␮g/ml in water containing 3% green food color) was injected into one-cell embryos derived from mibta52b/⫹ heterozygous parents. Homozygous mibta52b/ta52b mutants were identi-


fied after 18 hpf by morphological criteria, which were readily discernible even in mutants with partial rescue of hindbrain patterning. Specifically, mibta52b mutants invariably display a complex syndrome including small malformed ears with otolith deficiencies, loss of floor plate, partial fusion of posterior somites, and a characteristic bulge in the midbrain. These features were not altered by our experimental conditions, so identification of mutant embryos was unambiguous. To maximally induce expression of the heat-shock construct, embryos were incubated at 39°C for 1 hr. For sustained moderate gene expression, embryos were given serial heat shocks beginning at 16 hpf: each hour, embryos were incubated at 39°C for 15 min and then down-shifted to 33°C for 45 min; then the cycle was repeated. Prolonged higher temperature accelerated development such that embryos reached the equivalent of 30 hpf after 10 cycles of heat shock. After the final cycle, embryos were fixed and processed to visualize expression patterns of dlA or wnt1. For procedural controls, noninjected embryos were subjected to 10 heatshock cycles, or, alternatively, embryos were injected with pHS-wnt1myc but not heat shocked.

ACKNOWLEDGMENTS We thank John Postlethwait for foxb1.2, Chris Thorpe for wnt3a plasmid and morpholino, and Randall Moon for his support and discussion of the data. We also thank Rich Dorsky for providing tcf3b-morpholino. B.B.R., E.M.S., R.H., and M.Y.C. were funded by National Institutes of Health, NIDCD and G.R.B. and A.C.L. were funded by the American Heart Association, Texas Affiliate Beginning Grants in Aid.

REFERENCES Alexandre D, Clarke JD, Oxtoby E, Yan YL, Jowett T, Holder N. 1996. Ectopic expression of Hoxa-1 in the zebrafish alters the fate of the mandibular arch neural crest and phenocopies a retinoic acid-induced phenotype. Development 122:735–746. Allende ML, Weinberg ES. 1994. The expression of two zebrafish acheatescute homolog (ash) genes is altered in

the embryonic brain of the cyclops mutant. Dev Biol 166:509 –530. Appel B, Eisen JS. 1998. Regulation of neuronal specification in the zebrafish spinal cord by Delta function. Development 125:371–380. Appel B, Fritz A, Westerfield M, Grunwald DJ, Eisen J, Riley BB. 1999. Delta-mediated specification of midline cell fates in zebrafish embryos. Curr Biol 9:247– 256. Barrow JR, Thomas KR, Boussadia-Zahui O, Moore R, Kemler R, Capecchi MR and McMahon AP. 2003. Ectodermal wnt3a/␤-catenin signaling is required for the establishment and maintenance of the apical ectodermal ridge. Genes Dev 17:394 –409. Bell E, Richard JT, Lumsden A. 1999. Homeotic transformation of rhombomere identity after localized Hoxb1 misexpression. Science 284:2168 –2171. Bingham S, Chaudhari S, Vanderlaan G, Itoh M, Chitnis A, Chandrasekhar A. 2003. Neurogenic phenotype of mind bomb mutants leads to severe patterning defects in the zebrafish hindbrain. Dev Dyn 228:451–463. Blair S. 1997. Limb development: marginal fringe benefits. Curr Biol 7:R686 – R690. Blair S. 2003. Lineage compartments in Drosophila. Curr Biol 13:R548 –R551. Bovenkamp DE, Greer P. 1997. Novel Eph-family receptor tyrosine kinase is widely expressed in the developing zebrafish nervous system. Dev Dyn 209: 166 –181. Burgess S, Reim G, Chen W, Hopkins N, Brand M. 2002. The zebrafish spielohne-grenzen (spg) gene encodes POU domain protein Pou2 related to mammalian Oct4 and is essential for formation of the midbrain and hindbrain, and for pre-gastrula morphogenesis. Development 129:905–916. Chisaka O, Musci TS, Capecchi MR. 1992. Developmental defects of the ear, cranial nerves and hindbrain resulting from targeted disruption of the mouse homeobox gene Hox-1.6. Nature 335: 516 –520. Clarke JDW, Lumsden A. 1993. Segmental repetition of neuronal phenotype in the chick embryo. Development 118: 151–162. Cooke JE, Moens CB. 2002. Boundary formation in the hindbrain: Eph only it were simple. Trends Neurosci 25:260 – 267. Cooke JE, Moens CB, Roth LWA, Durbin L, Shiomi K, Brennan C, Kimmel CB, Wilson SW, Holder N. 2001. Eph signalling functions downstream of Val to regulate cell sorting and boundary formation in thecaudalhindbrain.Development128: 571–580. Cordes SP, Barsh G. 1994. The mouse segmentation gene kr encodes a novel basic domain-leucine zipper transcription factor. Cell 79:1025–1034. Couso JP, Bishop S, Martinez Arias A. 1994. The wingless signalling pathway and the patterning of the wing margin

in Drosophila. Development 120:621– 636. Couso JP, Knust E, Martinez Arias A. 1995. Serrate and wingless cooperate to induce vestigial gene expression and wing formation in Drosophila. Curr Biol 5:1437–1448. Dahmann C, Basler K. 1999. Compartment boundaries: at the edge of development. Trends Genet 15:320 –326. de Celis JF, Bray S. 1997. Feed-back mechanisms affecting Notch activation at the dorsoventral boundary in the Drosophila wing. Development 124: 3241–3251. de Celis JF, Carcia-Bellido A, Bray S. 1996. Activation and function of Notch at the dorsoventral boundary in the Drosophila wing imaginal disc. Development 122:359 –369. Dorsky RI, Sheldahl LC, Moon RT. 2002. A transgenic Lef1/␤-catenin-dependent reported is expressed in spatially restricted domains throughout zebrafish development. Dev Biol 241:229 –237. Dorsky RI, Itoh M, Moon RT, Chitnis A. 2003. Two tcf3 genes cooperate to pattern the zebrafish brain. Development 130:1937–1947. Dupe V, Ghyselinck NB, Wending O, Chambon P, Mark M. 1999. Key roles of retinoic acid receptors alpha and beta in the pattering of the caudal hindbrain, pharyngeal arches and otocyst in the mouse. Development 126:5051– 5059. Fashena D, Westerfield M. 1999. Secondary motoneuron axons localize DMGRASP on the fasciculated segments. J Comp Neurol 406:415–424. Fu¨rthauer M, Reifers F, Brand M, Thisse B, Thisse C. 2001. sprouty4 acts in vivo as a feedback-induced antagonist of FGF signaling in zebrafish. Development 128: 2175–2186. Gaunt SJ. 1997. Chick limbs, fly wings and homology at the fringe. Nature 386:324 – 325. Guthrie S. 1995. The status of the neural segment. Trends Neurosci 18:74 –79. Guthrie S, Butcher M, Lumsden A. 1991. Patterns of cell division and interkinetic nuclear migration in the chick embryo hindbrain. J Neurobiol 22:742–754. Haddon C, Smithers L, Schneider-Maunoury S, Coche T, Henrique D, Lewis J. 1998. Multiple delta genes and lateral inhibition in zebrafish primary neurogenesis. Development 125:359 –370. Hauptmann G, Belting HG, Wolke U, Lunde K, So ¨ ll I, Abdelilah-Seyfried S, Prince V, Driever W. 2002. spiel ohne grenzen/pou2 is required for zebrafish hindbrain segmentation. Development 129:1645–1655. Heyman I, Kent A, Lumsden A. 1993. Cellular morphology and extracellular space at rhombomere boundaries in the chick embryo hindbrain. Dev Dyn 198:241–253. Heyman I, Faissner A, Lumsden A. 1995. Cell and matrix specialisations of rhombomere boundaries. Dev Dyn 204:301– 315.

290 RILEY ET AL.

Holder N, Hill J. 1991. Retinoic acid modifies development of the midbrainhindbrain border and affects ganglion formation in zebrafish embryos. Development 113:1159 –1170. Ikeya M, Lee SMK, Johnston JE, McMahon AP, Takeda S. 1997. Wnt signalling required for expansion of neural crest and CNS progenitors. Nature 389:966 – 970. Irvine KD. 1999. Fringe, Notch and making developmental boundaries. Curr Opin Genet Dev 9:434 –441. Irvine KD, Rauskolb C. 2001. Boundaries in development: formation and function. Annu Rev Cell Dev Biol 17:189 –214. Itoh M, Kim CH, Palardy G, Oda T, Jiang YJ, Maust D, Yeo SY, Lorick K, Wright SM, ArizaMcNaughton L, Chandrasekharappa SC, Chitnis AB. 2003. Mind Bomb is a ubiquitin ligase that is essential for efficient activation of Notch signaling by Delta. Dev Cell 4:67–82. Jungbluth S, Bell E, Lumsden A. 1999. Specification of distinct motor neuron identities by the singular activities of individual Hox genes. Development 126: 2751–2758. Jiang YL, Brand M, Heisenberg CP, Beuchle D, Furutani-Seiki M, Kelsh RN, Warga RM, Granato M, Haffter P, Hammerschmidt M, Kane DA, Mullins MC, Odenthal J, van Eeden FJM, Nu¨ssleinVolhard C. 1996. Mutations affecting neurogenesis and brain morphology in the zebrafish. Development 123:205– 216. Kawakami Y, Capdevila J, Bu¨scher D, Itoh T, Rodriguez-Esteban C, IzpisuaBelmonte JC. 2001. Wnt signals control FGF-dependent limb initiation and AER induction in the chick embryo. Cell 104: 891–900. Kelly GM, Greenstein P, Erezyilmaz DF, Moon RT. 1995. Zebrafish wnt8 and wnt8b share a common activity but are involved in distinct developmental pathways. Development 121:1787– 1799. Kengaku M, Capdevila J, Rodriquez-Esteban C, De la Pena J, Johnson RL, Izpisua-Belmonte JC, Tabin C. 1998. Distinct WNT pathways regulating AER formation and dorsoventral polarity in the chick limb bud. Science 280:1274 – 1277. Kim J, Irvine KD, Carroll SB. 1995. Cell recognition, signal induction, and symmetrical gene activation at the dorsalventral boundary of the developing Drosophila wing. Cell 82:795–802. Kim J, Sebring A, Esch JJ, Kraus ME, Vorwerk K, Magee J, Carroll SB. 1996. Integration of positional signals and regulation of wing formation by Drosophila vestigial gene. Nature 382:133–138. Kim CH, Oda T, Itoh M, Jiang D, Artinger KB, Chandrasekharappa SC, Driever W, Chitnis AB. 2000. Repressor activity of Headless/Tcf3 is essential for vertebrate head formation. Nature 407:913– 916. Laufer E, Dahn R, Orozco OE, Yeo CY, Pisenti J, Henrique D, Abbott UK, Fallon

JF, Tabin C. 1997. Expression of Radical fringe in the limb-bud ectoderm regulates apical ectodermal ridge formation. Nature 386:366 –373. Lekven AC, Helde KA, Thorpe CJ, Rooke R, Moon RT. 2000. Reverse genetics in zebrafish. Physiol Genom 2:37–48. Lekven AC, Buckles GR, Kostakis N, Moon RT. 2003. wnt1 and wnt10b function redundantly at the zebrafish midbrainhindbrain boundary. Dev Biol 254:172– 187. Lewis J. 1996. Neurogenic genes and vertebrate neurogenesis. Curr Opin Neurobiol 6:3–10. Lufkin T, Dierich A, LeMuer M, Mark M, Chambon P. 1991. Disruption of the Hox-1.6 homeobox gene results in defects in a region corresponding to its rostral domain of expression. Cell 66: 1105–1119. Lumsden A. 1999. Closing in on rhombomere boundaries. Nat Cell Biol 1:E83– E85. Lumsden A, Krumlauf R. 1996. Patterning the vertebrate neuraxis. Science 274: 1109 –1114. Maden M, Gale E, Kosteeskii I, Zile M. 1996. Vitamin A-deficient quail embryos have half a hindbrain and other neural defects. Curr Biol 6:417–426. Mahmood R, Kiefer P, Guthrie S, Dickson C, Mason I. 1995. Multiple roles for FGF-3 during cranial neural development in the chicken. Development 121: 1399 –1410. Mahmood R, Mason IJ, Morrisskay GM. 1996. Expression of Fgf-3 in relation to hindbrain segmentation, otic pit position and pharyngeal arch morphology in normal and retinoic acid-exposed mouseembryos.AnatEmbryol(Berl)194: 13–22. Manzanares M, Cordes S, Kwan C, Sham M, Barsh GS, Krumlauf R. 1997. Segmental regulation of Hoxb-3 by kreisler. Nature 387:191–195. Manzanares M, Cordes S, Ariza-McNaughton L, Sadl V, Maruthainar K, Barsh G, Krumlauf R. 1999. Conserved and distinct roles of kreisler in regulation of the paralogous Hoxa3 and Hoxb3 genes. Development 122:759 – 769. Manzanares M, Trainor PA, Ariza-McNaughton L, Nonchev S, Krumlauf R. 2000. Dorsal patterning defects in the hindbrain, roofplate and skeleton in the dreher (drj) mouse mutant. Mech Dev 94:147–156. Mark M, Lufkin T, Vonesch JL, Ruberte E, Olivo JC, Dolle P, Gorry P, Lumsden A, Chambon P. 1993. Two rhombomeres are altered in Hoxa-1 mutant mice. Development 119:319 –338. Marshall H, Nonchev S, Sham MH, Muchamore I, Lumsden A, Krumlauf R. 1992. Retinoic acid alters hindbrain Hox code and induces transformation of rhombomeres 2/3 into a 4/5 identity. Nature 360:737–741. Maves L, Jackman W, Kimmel CB. 2002. FGF3 and FGF8 mediate a rhombomere 4 signaling activity in the ze-

brafish hindbrain. Development 129: 3825–3837. McClintock JM, Carlson R, Mann DM, Prince VE. 2001. Consequences of Hox gene duplication in the vertebrates: an investigation of zebrafish Hox paralogue group 1 genes. Development 128: 2471–2484. McKay IJ, Muchamore I, Krumlauf R, Maden M, Lumsden A, Lewis J. 1994. The kreisler mouse: a hindbrain segmentation mutant that lacks two rhombomeres. Development 120:2199 – 2211. Mellitzer G, Xu Q, Wilkinson DG. 1999. Eph receptors and ephrins restrict cell intermingling and communication. Nature 400:77–80. Micchelli CA, Rulifson EJ, Blair SS. 1997. The function and regulation of cut expression on the wing margin of Drosophila: Notch, Wingless and a dominant negative role for Delta and Serrate. Development 124:1485–1495. Millonig JH, Millen KJ, Hattern ME. 2000. The mouse Dreher gene Lmx1a controls formation of the roof plate in the vertebrate CNS. Nature 403:764 –769. Moens C, Yan YL, Appel B, Force A, Kimmel C. 1996. Valentino: a zebrafish gene required for normal hindbrain segmentation. Development 122:3981– 3990. Moens C, Cordes S, Giorgianni M, Barsh G, Kimmel C. 1998. Equivalence in the genetic control of hindbrain segmentation in fish and mouse. Development 125:381–391. Molven A, Njolstad PR, Fjose A. 1991. Genomic structure and restricted neural expression of the zebrafish wnt-1 (int-1) gene. EMBO J 10:799 –807. Nasevicius A, Ekker SC. 2000. Effective targeted gene “knockdown” in zebrafish. Nat Genet 26:216 –220. Niederreither K, Vermot J, Schuhbaur B, Chambon P, Dolle P. 2000. Retinoic acid synthesis and hindbrain patterning in the mouse embryo. Development 127:75–85. Nittenberg R, Patel K, Joshi Y, Krumlauf R, Wilkinson DG, Brickell PM, Tickle C, Clarke JDW. 1997. Cell movements, neuronal organization and gene expression in hindbrains lacking morphological boundaries. Development 124: 2297–2306. Nonchev S, Vesque C, Maconochie M, Seitanidou T, Ariza-McNaughton L, Frain M, Marshall H, Sham MH, Krumlauf R, Charnay P. 1996. Segmental expression of Hoxa-2 in the hindbrain is directly regulated by Krox-20. Development 122:543–554. Oxtoby E, Jowett T. 1993. Cloning of the zebrafish krox-20 gene (krx-20) and its expression during hindbrain development. Nucleic Acids Res 21:1087–1095. Papalopulu N, Clarke JDW, Bradley L, Wilkinson D, Krumlauf R, Holder N. 1991. Retinoic acid causes abnormal development and segmental patterning of the anterior hindbrain in Xenopus embryos. Development 113:1145–1158.


Phillips BT, Bolding K, Riley BB. 2001. Zebrafish fgf3 and fgf8 encode redundant functions required for otic placode induction. Dev Biol 235:351–365. Prince VE, Moens CB, Kimmel CB, Ho RK. 1998. Zebrafish hox genes: expression in the hindbrain region of the wild-type and mutants of the segmentation gene, valentino. Development 125:393– 406. Raible F, Brand M. 2001. Tight transcriptional control of the ETS domain factors Erm and Pea3 by Fgf signaling during early zebrafish development. Mech Dev 107:105–117. Reim G, Brand M. 2002. spiel-ohne-grenzen/pou2 mediates regional competence to respond to Fgf8 during zebrafish early neural development. Development 129:917–933. Riley BB, Chiang MY, Farmer L, Heck R. 1999. The deltaA gene of zebrafish mediates lateral inhibition of hair cells in the inner ear and is regulated by pax2.1. Development 126:5669 –5678. Rodriquez-Esteban C, Schwabe JWR, De la Pena J, Foys B, Eshelman B, Izpisua Belmonte JC. 1997. Radical fringe positions the apical ectodermal ridge at the dorsoventral boundary of the vertebrate limb. Nature 386:360 –366. Roehl H, Nu¨sslein-Volhard C. 2001. Zebrafish pea3 and erm are general targets of FGF8 signaling. Curr Biol 11:503– 507. Roose J, Clevers H. 1999. TCF transcription factors: molecular switches in carcinogenesis. Biochim Biophys Acta 87456: M23–M37. Rossel M, Capecchi MR. 1999. Mice mutant for both Hoxa1 and Hoxb1 show extensive remodeling of the hindbrain and defects in craniofacial development. Development 126:5027–5040. Rowitch DH, Echelard Y, Danielian PS, Geliner K, Brenner S, McMahon AP. 1998. Identification of an evolutionarily conserved 110 base-pair cis-acting regulatory sequence that governs

Wnt-1 expression in the murine neural plate. Development 125:2735–2746. Rulifson EJ, Blair SS. 1995. Notch regulates wingless expression and is not required for reception of the paracrine wingless signals during wing neurogenesis in Drosophila. Development 121:2813– 2824. Rulifson EJ, Micchelli CA, Axelrod JD, Perrimon N, Blair SS. 1996. wingless refines its own expression domain on the Drosophila wing margin. Nature 384:72–74. Sanson B. 2001. Generating patterns from fields of cells: examples from Drosophila segmentation. EMBO Rep 2:1083– 1088. Schneider-Maunoury S, Seitandou T, Charnay P, Lumsden A. 1997. Segmental and neuronal architecture of the hindbrain of Krox-20 mouse mutants. Development 124:1215–1226. Sham MH, Vesque C, Nonchev S, Marshall H, Frain M, DasGupta R, Whiting J, Wilkinson D, Charnay P, Krumlauf R. 1993. The zinc finger gene Krox20 regulates HoxB2 (Hox2.8) during hindbrain segmentation. Cell 72:183–196. Shoji W, Yee CS, Kuwada JY. 1998. Zebrafish semaphorin Z1a collapses specific growth cones and alters their pathwayinvivo.Development125:1275– 1283. Simeone A, Avantaggiato V, Moroni MC, Mavilio F, Arra C, Cotelli F, Migro V, Acampora D. 1995. Retinoic acid induces stage-specific antero-posterior transformation of the rostral central nervous system. Mech Dev 51:83–98. Studer M, Lumsden A, Ariza-McNaughton L, Bradley A, Krumlauf R. 1996. Altered segmental identity and abnormal migration of motor neurons in mice lacking Hoxb-1. Nature 384:630 –634. Sun Z, Hopkins N. 2001. vhnf1, the MODY5 and familial GCKD-associated gene, regulates regional specification of the gut, pronephros, and hindbrain. Genes Dev 15:3217–3229.

Swiatek PJ, Gridley T. 1993. Perinatal lethality and defects in hindbrain development in mice homozygous for a targeted mutation of the zinc finger gene Krox20. Genes Dev 7:2071–2084. Trevarrow B, Marks DL, Kimmel CB. 1990. Organization of hindbrain segments in the zebrafish hindbrain. Neuron 4:669 – 679. Walsh J, Maroon H, McGonnell IM, Dickson C, Mason I. 2002. Establishment of hindbrain segmental identity requires signaling by FGF3 and FGF8. Curr Biol 12:1117–1123. Waskiewicz AJ, Rikhof HA, Moens CB. 2002. Eliminating zebrafish Pbx proteins reveals a hindbrain ground state. Dev Cell 3:723–733. White JC, Highland M, Kaiser M, ClagettDame M. 2000. Vitamin A deficiency results in the dose-dependent acquisition of anterior character and the shortening of the caudal hindbrain of the rat. Dev Biol 220:263–284. Wiellette EL, Sive H. 2003. vhnf1 and Fgf signals synergize to specify rhombomere identity in the zebrafish hindbrain. Development 130:3821–3829. Wingate JT, Lumsden A. 1996. Persistence of rhombomeric organisation in the postsegmentation hindbrain. Development 122:2143–2152. Xu Q, Alldus G, Holder N, Wilkinson DG. 1995. Expression of truncated Sek-1 receptor tyrosine kinase disrupts the segmental restriction of gene expression in the Xenopus and zebrafish hindbrain. Development 121:4005–4016. Xu Q, Mellitzer G, Robinson V, Wilkinson DG. 1999. In vivo cell sorting in complementary segmental domains mediated by Eph receptors and ephrins. Nature 399:267–271. Yoshida M, Colman DR. 2000. Glial-defined rhombomere boundaries in developing Xenopus hindbrain. J Comp Neurol 424:47–57.