EphA4 Is Required for Cell Adhesion and Rhombomere-Boundary Formation in the Zebrafish

Current Biology, Vol. 15, 536–542, March 29, 2005, ©2005 Elsevier Ltd All rights reserved. DOI 10.1016/j.cub.2005.02.019

EphA4 Is Required for Cell Adhesion and Rhombomere-Boundary Formation in the Zebrafish Julie E. Cooke,2 Hilary A. Kemp,1 and Cecilia B. Moens1,* 1 Howard Hughes Medical Institute and Division of Basic Science Fred Hutchinson Cancer Research Center B2-152 1100 Fairview Avenue N P.O. Box 19024 Seattle, Washington 98109

Summary The formation of boundaries between or within tissues is a fundamental aspect of animal development. In the developing vertebrate hindbrain, boundaries separate molecularly and neuroanatomically distinct segments called rhombomeres. Transplantation studies have suggested that rhombomere boundaries form by the local sorting out of cells with different segmental identities [1–4]. This sorting-out process has been shown to involve repulsive interactions between cells expressing an Eph receptor tyrosine kinase, EphA4, and cells expressing its ephrinB ligands [5–7]. Although a model for rhombomere-boundary formation based on repulsive Eph-ephrin signaling is well established in the literature, the predictions of this model have not been tested in loss-of-function experiments. Here, we eliminate EphA4 and ephrinB2a proteins in zebrafish with antisense morpholinos (MO) and find that rhombomere boundaries are disrupted in EphA4MO embryos, consistent with a requirement for Eph-ephrin signaling in boundary formation. However, in mosaic embryos, we observe that EphA4MO cells and EphA4-expressing cells sort from one another, an observation that is not predicted by the Eph-ephrin repulsion model but instead suggests that EphA4 promotes cell adhesion within the rhombomeres in which it is expressed. Differential cell adhesion is known to be an effective mechanism for cell sorting. We therefore propose that the well-known EphA4-dependent repulsion between rhombomeres operates in parallel with the EphA4-dependent adhesion within rhombomeres described here to drive the cell sorting that underlies rhombomere-boundary formation. Results and Discussion EphA4 and ephrinB2a Are Required for Rhombomere-Boundary Formation In the zebrafish hindbrain, interacting Ephs and ephrins are expressed in complementary, rhombomere-restricted domains such that each rhombomere boundary forms at an interface between one or more receptor-ligand pair *Correspondence: [email protected] 2 Present address: Comparative Genomics Group, MRC Rosalind Franklin Centre for Genomics Research, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SB, UK.

(Figure S1A [7–11]). We knocked down EphA4 (expressed in r3 and r5) and ephrinB2a (expressed in r1, r4, and r7) with antisense morpholinos (MO) and found that they eliminate the corresponding proteins by whole-mount immunostaining (Figure S1B–S1E) and Western blotting (data not shown). In embryos injected with EphA4MO, we observed a subtle but consistent disorganization of the boundaries of krox20 expression at 18 hr postfertilization (hpf) [12], with the outer boundaries of krox20 expression (r2/3 and r5/6) being most strongly affected (Figures 1A and 1B). Analysis of boundary-specific markers sema3Gb, foxb1.2, and rfng showed a corresponding effect (Figures 1E and 1F and data not shown). Although knocking down ephrinB2a alone has very little if any effect on rhombomere boundaries (Figures 1C and 1G), in embryos injected with both EphA4MO and ephrinB2aMO, krox20 expression and boundary-specific markers are disorganized or absent throughout the segmented region of the hindbrain (Figures 1D and 1H). We infer that loss of EphA4 and ephrinB2a results in a failure of cell sorting at the forming boundary, and as a result, a specialized boundary cell population is not specified. Neuronal Patterning Defects in Embryos Lacking Normal Rhombomere Boundaries Segmentation is thought to be an evolutionary mechanism for generating regional diversity within a tissue because segment boundaries allow adjacent populations of cells to follow distinct developmental trajectories and acquire different fates. In the hindbrain, the function of segment (rhombomere) boundaries has only been tested under conditions that disrupt overall anterior-posterior patterning of the hindbrain [13], so the specific functions of rhombomere boundaries in neural patterning has not been addressed. In EphA4MO and EphA4MO; ephrinB2aMO embryos, the specification of regional identities along the anterior-posterior axis based on hox gene expression is normal (data not shown); however, neuronal populations that are normally separated by rhombomere boundaries are fused. At 26 hpf, segment-restricted domains of early neuronal differentiation are disorganized in EphA4MO embryos and EphA4MO; ephrinB2aMO embryos (Figure S2). At 48 hpf in wild-type embryos, motor neurons of the trigeminal nerve (nV) have differentiated in r2 and r3, and motor neurons of the abducens nerve (nVI) have differentiated in r5 and r6 (Figures 1I and 1M) [14–16]. In EphA4MO embryos, these motor neuron clusters are closer together or fused (Figures 1J and 1N), and in EphA4MO; ephrinB2aMO embryos, these phenotypes are even more severe (Figures 1L and 1P). These effects on neuronal patterning are consistent with the effects we observed on krox20 and boundaryspecific markers. We infer that rhombomere boundaries have an important function in separating neuronal populations in the hindbrain. Motor neurons in adjacent rhombomeres normally have distinct functions: for example, trigeminal motor neurons in r2 innervate the ab-

EphA4 in Rhombomere-Boundary Formation 537

Figure 1. EphA4 and ephrinB2a Are Required for Normal Rhombomere Boundaries and Hindbrain Neuronal Patterning (A–H) Whole-mount RNA in situs in lateral views with anterior to the left showing expression of krox20 at 18 hpf (A–D) and sema3Gb at 20 hpf (E–H) in embryos treated as indicated at top. Although r3 and r5 are specified normally in each condition, the boundaries of r3 and r5 are disorganized in EphA4MO and EphA4MO; ephrinB2aMO-injected embryos. Asterisks in (F) indicate missing r2/3 and r5/6 boundaries, dotted lines in (F) and (H) indicate incomplete boundaries. (I–P) Fusion of segmental motor nuclei in embryos lacking rhombomere boundaries. (I–L) Confocal images showing isl1-GFP expression in trigeminal motor neurons in r2 and r3 (arrows) and DM-GRASP immunoreactivity in hindbrain commissures that project along rhombomere boundaries (red staining, indicated by white arrowheads). Treatments are as for (A)–(H). Images show dorsal views of the right side of each embryo with anterior to the left. The dotted line indicates the midline. Green arrowheads indicate the VIIth (facial) motor nerve, which exits the hindbrain in r4. Note that trigeminal motor nuclei lie close together in EphA4MO embryos (J) and are fused in EphA4MO; ephrinB2aMO embryos (L). (M–P) Confocal images showing DM-GRASP immunostaining of abducens motor neurons at 44 hpf (arrows). Embryos are shown in dorsal views with anterior to the left; dotted line indicates the midline. Note that abducens motor nuclei are closer together or fused in EphA4MO embryos (N) and are entirely fused in EphA4MO; ephrinB2aMO embryos (P). Although the cell bodies in (N) are fused on one side, the axons that project ventrally out of the hindbrain are not (arrowheads in inset). In contrast, both cell bodies and axons are fused in (P) (arrowheads in inset). Scale bars, 50 ␮m.

ductor mandibulae muscle, whereas trigeminal motor neurons in r3 innervate four other muscles in the first pharyngeal arch [15]. We predict that this specificity of target innervation may be lost in the absence of the r2/3 boundary and that the autonomy of movement of first arch-derived muscles may be correspondingly compromised. EphA4 Is Required for Normal Cell Adhesion within r3 and r5 Our observations thus far are consistent with an important role for EphA4 and ephrinB2a in rhombomereboundary formation and/or maintenance, as predicted by the Eph-ephrin repulsion model [6, 17–19]. A further prediction of this model is that because cell sorting is the active response to Eph-ephrin signaling, cells lacking Eph expression will be essentially inert, being neither able to induce a repulsive response by activating reverse signaling in adjacent cells nor to initiate a cellautonomous repulsive response to ephrin signals. In or-

der to test this prediction, we made mosaic embryos in which EphA4MO cells were transplanted into the hindbrain of a wild-type host and vice versa. To aid in this analysis, we used a transgenic line, pGFP5.3, that expresses GFP in r3 and r5 as host embryos [20]. The prediction of the repulsion-based model is that cells lacking EphA4 will be able to contribute normally to all rhombomeres. However, this is not what we observed. Although EphA4MO cells contribute normally to even-numbered rhombomeres, they sort robustly to the edges of r3 and r5 of wild-type host embryos (Figure 2B; n = 59/70 mosaics). In occasional cases where EphA4MO cells were transplanted into the dorsal hindbrain, they were observed to accumulate at the dorsalmost surface of r3 and r5 (data not shown). The sorting out of EphA4MO cells, which is not observed in control mosaics (Figure 2A), is rescued by coexpression of fulllength zebrafish EphA4 (Figure S3). EphA4MO cells that accumulate at the edges of r3 and r5 maintain expression of r3 and r5 markers and do not appear to move

Current Biology 538

Figure 2. EphA4 Is Required for Normal Cell Adhesion in r3 and r5 18 hpf mosaic embryos in which donor cells (red) were transplanted into pGFP5.3 transgenic hosts that express GFP in r3 and r5 and in the otic vesicle (ov). The middle and right columns are single channel images of the overlaid images shown in the left column. All embryos are shown in dorsal view with anterior to the left. (A) Uninjected or control MO-injected donor cells contribute evenly throughout the hindbrain of a wild-type host. (B) EphA4MO donor cells are excluded from r3 and r5 of a wild-type pGFP5.3 host and accumulate inside the boundaries. In occasional cases where EphA4MO cells were transplanted into the dorsal hindbrain, they were observed to accumulate at the dorsal-most surface of r3 and r5 (data not shown). (C) Uninjected or control MO-injected donor cells form tight unilateral clusters in the centers of r3 and r5 of an EphA4MO pGFP5.3 host (arrowheads in C#). These clusters are not GFP positive because the donor did not carry the transgene. Brackets indicate r3 and r5; scale bar, 50 ␮m.

into the adjacent rhombomeres (Figures 3A and 3B), although with this analysis we cannot rule out the possibility that some EphA4MO cells move into adjacent rhombomeres and change their identity to match that of their neighbors. In the converse transplantation experiment, in which uninjected or control MO-injected cells were transplanted into an EphA4MO host, donor cells formed large, tightly adherent clumps in r3 and r5, which form sharp boundaries with the surrounding EphA4MO cells (Figure 2C). Like the surrounding host cells, these clusters have r3 and r5 identity (Figure S4C). This unexpected cell sorting behavior could result from upregulation of ephrin expression in EphA4MO cells, which would cause them to sort out of r3 and r5 through the well-established Eph-ephrin repulsion mechanism. However, neither of the two B-type ephrins normally expressed in the hindbrain are upregulated in r3 and r5 of EphA4MO embryos (Figures S5A–S5D). Furthermore, knocking down these ephrins in EphA4MO cells did not restore their ability to contribute to r3 and r5 in mosaics (data not shown). The sorting out of EphA4MO cells is consistent with EphA4 being required for cells within r3 and r5 to adhere normally to one another. Cells lacking EphA4 are at an adhesive disadvantage and sort to the exterior of the EphA4-expressing cells. Differential cell adhesion is an effective mechanism for cell sorting both in tissue recombinates and in vivo [21–26]. We propose that EphA4-dependent adhesion contributes to the cell sorting process that underlies rhombomere-boundary formation (Figure 3). At the forming boundaries of r3

Figure 3. Model for the Role of EphA4 in Rhombomere-Boundary Formation (A) At the forming boundary, where EphA4-expressing cells (blue) are intermingled with ephrinB-expressing cells (pink), contact between EphA4-expressing and ephrinB-expressing cells leads to a repulsive response [6]. At the same time, contacts between EphA4expressing cells leads to an EphA4-dependent adhesive response. The ligand for EphA4 within r3 and r5 is unknown (indicated by a question mark). (B) Cell sorting involving both repulsion between unlike cells and preferential adhesion between like cells is thermodynamically more stable because it minimizes repulsive interactions (red bars) and maximizes adhesive interactions (black joiners).

EphA4 in Rhombomere-Boundary Formation 539

and r5, where EphA4-expressing and ephrinB-expressing cells are intermingled, contact between EphA4expressing and ephrinB-expressing cells leads to bidirectional repulsion (left side of Figure 3A) [6]. At the same time, contact amongst EphA4-expressing cells leads to increased adhesion that is dependent on EphA4 and an as-yet unidentified ligand (right side of Figure 3A). The combination of EphA4-dependent interrhombomere repulsion and EphA4-dependent intrarhombomere adhesion means that the hypothetical sorting event schematized in Figure 3B not only minimizes repulsive interactions (represented by red bars between unlike cells) but maximizes adhesive interactions (represented by black joiners between like cells). It is somewhat surprising that EphA4MO cells do not appear to pass into even-numbered rhombomeres, given that EphA4 has been specifically implicated in repulsion at rhombomere boundaries [6]. This indicates that EphA4 activity does not alone prevent cells from crossing boundaries. It is possible that, having r3 and r5 identity, EphA4MO cells do not express genes that are required for cells to adhere within even-numbered rhombomeres. Alternatively, other Eph receptors expressed in r3 and r5 may allow EphA4MO cells to respond to repulsive ephrinB signals in even-numbered rhombomeres. Although repulsion is the most common and wellunderstood outcome of Eph-ephrin interactions, a number of examples of Eph-ephrin-dependent adhesion have recently been described, sometimes in the very cells that under different conditions exhibit repulsion (reviewed in [27–29]). Several mechanisms have been proposed for how Eph-ephrin interactions can promote adhesion (reviewed in [28]). Adhesion can be promoted by conditions that prevent the cleavage or internalization of the inherently adhesive ligand-receptor complex [30– 32], or repulsion and adhesion may be alternate responses to different receptor-ligand stoichiometry [33, 34] and/or to trans- (on different cells) versus cis (on the same cell)-expression of receptor-ligand pairs [35– 37]. At rhombomere boundaries, trans-interactions between EphA4-expressing cells and cells strongly expressing ephrinBs could lead to repulsion, whereas within r3 and r5, cis-interactions between EphA4 and weakly expressed A-type or B-type ligands could promote adhesion. Although ephrinB2a and ephrinB3 are not expressed in r3 and r5, we have found that at least two other ephrins, ephrinA2 and ephrinA5 have low-level homogeneous expression throughout the hindbrain during early somite stages and could act as ligands for EphA4-dependent adhesion (J.E.C., unpublished data). Finally, the effects of EphA4 signaling could be modulated by downstream effectors that are differentially expressed within the hindbrain (reviewed in [28]). Cell Sorting in EphA4MO Mosaics Corresponds with the Timing of Boundary Sharpening In normal embryos, boundaries of segmental gene expression are initially ragged and become razor sharp between about 10 hpf (1 somite stage) and 14 hpf (10 somite stage) (Figures 4A–4E), and we infer that this is the period of active cell sorting leading to boundary sharpening. This period corresponds with the formation

of the neural keel, the precursor to the neural tube, during which neuroepithelial cells elongate mediolaterally to form stable apical contacts with the midline of the forming neural keel and then divide across the midline, generating bilateral clones [38–41]. We reasoned that if the sorting out of EphA4MO cells from EphA4-expressing rhombomeres coincided with the timing of normal rhombomere-boundary formation, this would argue that adhesive differences between EphA4-expressing and -nonexpressing cells could contribute to boundary sharpening during normal development. We made confocal time-lapse movies of mosaic embryos in which EphA4MO cells were transplanted to a wild-type pGFP5.3 host (Figure 4). At t = 0, corresponding to the 3 somite stage, donor-derived EphA4MO cells contribute homogeneously to the left side of the hindbrain (Figure 4F). At this stage, some donor-derived neuroepithelial cells are contacting the midline, and a few have divided across it (arrowheads). Over the next 3 hr, as GFP expression becomes detectable in r3 and r5 (square brackets), donor-derived cells in r4 and r6 continue to divide across the midline (arrowheads, Figures 4H and 4I), whereas those in r3 and r5 retract from it. By the 9 somite stage, donor-derived cells in r2, r4, and r6 have contacted the basal surface on the right side (arrowheads in Figure 4I and data not shown), but those in r3 and r5 have lost contact with the midline and are beginning to be excluded in an anterior-posterior direction, first from the center of r3 (asterisk in Figures 4I–4L) and slightly later from r5 (dot in Figures 4K and 4L). This corresponds with the earlier onset of EphA4 expression in r3 [42]. By the end of the time lapse, EphA4MO cells have accumulated at the edges of r3 and r5 on both sides of the midline (Figure 4L, corresponding to the 13–14 somite stage). These time-lapse analyses demonstrate that the sorting of EphA4MO cells from wild-type host cells corresponds precisely with the normal boundary sharpening period and results from the failure of EphA4MO cells to form and maintain normal contacts with neighboring EphA4expressing host cells, either adjacent to them or across the midline. Boundary sharpening coincides with the dramatic morphogenetic movements that change the neural plate into the neural keel, a process that requires single cells to make changing contacts with their neighbors while maintaining the integrity of the neuroepithelium [38]. Although EphA4 is not strictly required for cells in r3 and r5 to undergo these movements, differences in adhesion between EphA4-expressing and -neighboring EphA4MO cells in mosaic embryos prevent EphA4MO cells from participating in these movements when transplanted into a wild-type host embryo. Extrapolating from these cell behaviors in mosaics to the normal boundary formation process, we hypothesize that during neural keel formation, preferential adhesion between EphA4-expressing cells prevents non-EphA4-expressing cells (i.e., cells with even-numbered rhombomere identities) that find themselves within the presumptive r3 and r5 territories from participating in the normal morphogenetic movements associated with neural keel formation. As a result, non-EphA4-expressing cells are excluded from EphA4-expressing territories. The relative degrees to which EphA4-dependent re-

Current Biology 540

Figure 4. The Timing of Cell Sorting in EphA4MO Mosaics Correlates with the Normal Boundary Sharpening Period (A–E) Whole-mount RNA in situs showing the sharpening of boundaries of krox20 expression in r3 and r5 between 10 and 14 hr post fertilization (hpf). Embryos in (A)–(D) are shown in dorsal view with anterior to the top; embryo in (E) is in lateral view. nc, neural crest. (F–L) Confocal time lapse of EphA4MO cells (red) transplanted into a wild-type pGFP5.3 host embryo, showing that EphA4MO cells fail to undergo normal morphogenetic movements in mosaics. Dorsal views, with anterior to the left. (F) At t = 0 (3 somites; 11 hpf), donor-derived cells are homogeneously distributed along the transplanted (left) side of the embryo and are beginning to make contacts with the midline (dotted line). The subsequent time points correspond to approximately 5 somites (G), 7 somites (H), 9 somites (I), 10.5 somites (J), 12 somites (K), and 13.5 somites (L). By the end of the time lapse, EphA4MO cells in even-numbered rhombomeres have executed the cross-midline division, whereas EphA4MO cells in r3 and r5 have retracted from the midline and sorted out of the center of the rhombomere. Brackets indicate r3 and r5; arrowheads indicate cells that have crossed the midline in even-numbered rhombomeres; asterisks and dots indicate the clearing of donor-derived cells from r3 and r5, respectively. Scale bars, 50 ␮m.

pulsion and EphA4-dependent adhesion contribute to rhombomere-boundary formation remains unknown because eliminating EphA4 affects both processes. It is possible that the two mechanisms are redundant with one another, so that in the absence of EphA4-dependent repulsion boundaries could form by EphA4dependent adhesion and vice versa. By elucidating the mechanism of EphA4-dependent adhesion, we will be able to selectively eliminate either EphA4-dependent adhesion or EphA4-dependent repulsion and answer this important question. Experimental Procedures Morpholino Injections Morpholino oligonucleotides from Genetools, Inc. were resuspended in Danieau solution at the appropriate concentrations, and 1 nl was injected into zebrafish embryos at the 1–4 cell stage. Two different EphA4MOs gave identical phenotypes in injected embryos and mosaics, whereas a control MO neither disrupted rhombomere boundaries nor caused cell sorting in mosaics. Morpholinos were as follows: EphA4TB (translation blocking) MO, AACACAAGCGCAGCCATTGGTGTC, 5 ng/nl; EphA4SB (splice blocking) MO, CGTTTGCTCACCAATGCCGATGACC, 5 ng/nl; ephrinB2aTB MO, CGGTCAAATTCCGTTTCGCGGGA, 10 ng/nl; and control MO, TGATAGGCAGCTTCGGTTGCGACAT, 2–10 ng/nl. RNA In Situs and Immunostaining RNA in situs were performed essentially as described [43]. Whole mount immunostaining was performed with the following antibodies: anti-EphA4 [42], anti-ephrinB2a (R&D Biosystems), anti-Hu

(Molecular Probes), and anti-DM-GRASP (zn-5) [16, 44]. Embryos were fixed in 4% paraformaldehyde overnight at 4°C. Embryos were permeablized with 10 ␮g/ml proteinase K (Sigma), washed, and blocked with PBS, 1% DMSO, 0.1% triton (PBDT) containing 2 mg/ml BSA, and 2% goat serum (or horse serum for antiephrinB2a) before incubation with primary antibody overnight at 4°C in 1% goat serum or horse serum. Secondary antibodies were fluorochrome-conjugated Alexa Fluor 488 or 594 goat anti-mouse or goat anti-rabbit IgG (or donkey anti-goat for anti-ephrinB2a; Molecular Probes). Mosaic Analysis and Confocal Time Lapses Mosaics were made by cell transplantation at early gastrula stages as described [45]. Donor embryos were injected at the 1-cell stage with rhodamine dextran plus EphA4MO or control MO, and cells were transplanted from donor embryos into host embryos at the shield stage (6 hpf). For some experiments, either the donor or host embryo carried the pGFP5.3 transgene [20], which expresses GFP in r3 and r5 so that the distribution of donor cells could be accurately determined in live embryos at 18 hpf. Alternatively, the positions of donor-derived cells were determined after RNA in situ hybridization with krox20. For mRNA rescue experiments, donor embryos were coinjected with 5 ng EphA4MO, 600 pg full-length zebrafish EphA4 mRNA, and 50 pg GFP mRNA as a tracer, and the distribution of donor cells was determined by immunostaining with an anti-GFP antibody (Torrey Pines Biolabs) after RNA in situ hybridization with krox20 as described [7]. For confocal time-lapse analysis, host embryos were mounted hindbrain down at the 1–3-somite stage (11 hpf) in a 35 mm glass bottom dish (World Precision Instruments) in a drop of 1.2% lowmelting-point agarose and were flooded with embryo medium containing tricaine anesthetic (Sigma). Embryos were imaged at 25°C on a Zeiss 510 confocal inverted microscope fitted with an xy-

EphA4 in Rhombomere-Boundary Formation 541

motorized stage by using a 20× objective with a 1.3–1.5× zoom. 30 time points were taken at 15 min intervals; each time point included 30 z sections at 1.28 ␮m intervals. The images shown in Figure 4 are projections of six z sections corresponding to a 7.7-␮m-thick volume approximately 10-␮m deep to the dorsal surface of the hindbrain neuroepithelium.

11.

12. Supplemental Data Supplemental Data include five figures and can be found with this article online at http://www.current-biology.com/cgi/content/full/ 15/6/536/DC1/.

Acknowledgments We wish to thank colleagues who generously provided information, reagents, and transgenic lines: Hitoshi Okamoto and Shin-ichi Higasihijima provided the isl1-GFP line, and Michael Brand and Alex Picker provided the pGFP5.3 line. David Wilkinson provided the anti-EphA4 antibody. Richard Poole kindly provided the sequence of the ephrinB2a morpholinos. The sema3Gb probe was from Hung-Hsiang Yu. The zn8 antibody was from the University of Oregon hybridoma facility. We thank Paul Grant and Julio Vasquez for their help in acquiring the confocal time lapses. Finally, we wish to thank the members of the Moens lab for their thoughtful input into this project. The research was supported by a Wellcome Trust post-doctoral International Prize Traveling Research Fellowship (GR059942; J.E.C.) and National Institutes of Health grant HD37909 (C.B.M.). C.B.M. is an investigator with the Howard Hughes Medical Institute. Received: November 14, 2004 Revised: January 11, 2005 Accepted: January 12, 2005 Published: March 29, 2005 References 1. Guthrie, S., and Lumsden, A. (1991). Formation and regeneration of rhombomere boundaries in the developing chick hindbrain. Development 112, 221–229. 2. Guthrie, S., Prince, V., and Lumsden, A. (1993). Selective dispersal of avian rhombomere cells in orthotopic and heterotopic grafts. Development 118, 527–538. 3. Moens, C.B., Yan, Y.L., Appel, B., Force, A.G., and Kimmel, C.B. (1996). valentino: a zebrafish gene required for normal hindbrain segmentation. Development 122, 3981–3990. 4. Waskiewicz, A.J., Rikhof, H.A., and Moens, C.B. (2002). Eliminating zebrafish pbx proteins reveals a hindbrain ground state. Dev. Cell 3, 723–733. 5. Mellitzer, G., Xu, Q., and Wilkinson, D.G. (1999). Eph receptors and ephrins restrict cell intermingling and communication. Nature 400, 77–81. 6. Xu, Q., Mellitzer, G., Robinson, V., and Wilkinson, D.G. (1999). In vivo cell sorting in complementary segmental domains mediated by Eph receptors and ephrins. Nature 399, 267–271. 7. Cooke, J., Moens, C., Roth, L., Durbin, L., Shiomi, K., Brennan, C., Kimmel, C., Wilson, S., and Holder, N. (2001). Eph signalling functions downstream of Val to regulate cell sorting and boundary formation in the caudal hindbrain. Development 128, 571–580. 8. Cooke, J., Xu, Q., Wilson, S., and Holder, N. (1997). Characterisation of five novel zebrafish Eph-related receptor tyrosine kinases suggests roles in patterning the neural plate. Dev. Genes Evol. 206, 515–531. 9. Chan, J., Mably, J.D., Serluca, F.C., Chen, J.N., Goldstein, N.B., Thomas, M.C., Cleary, J.A., Brennan, C., Fishman, M.C., and Roberts, T.M. (2001). Morphogenesis of prechordal plate and notochord requires intact Eph/ephrin B signaling. Dev. Biol. 234, 470–482. 10. Xu, Q., Alldus, G., Holder, N., and Wilkinson, D.G. (1995). Expression of truncated Sek-1 receptor tyrosine kinase disrupts

13.

14.

15.

16.

17.

18. 19. 20.

21. 22.

23.

24.

25.

26.

27. 28.

29. 30.

31.

32.

the segmental restriction of gene expression in the Xenopus and zebrafish hindbrain. Development 121, 4005–4016. Durbin, L., Brennan, C., Shiomi, K., Cooke, J., Barrios, A., Shanmugalingam, S., Guthrie, B., Lindberg, R., and Holder, N. (1998). Eph signaling is required for segmentation and differentiation of the somites. Genes Dev. 12, 3096–3109. Kimmel, C.B., Ballard, W.W., Kimmel, S.R., Ullmann, B., and Schilling, T.F. (1995). Stages of embryonic development of the zebrafish. Dev. Dyn. 203, 253–310. Nittenberg, R., Patel, K., Joshi, Y., Krumlauf, R., Wilkinson, D.G., Brickell, P.M., Tickle, C., and Clarke, J.D. (1997). Cell movements, neuronal organisation and gene expression in hindbrains lacking morphological boundaries. Development 124, 2297–2306. Chandrasekhar, A., Moens, C.B., Warren, J.T., Jr., Kimmel, C.B., and Kuwada, J.Y. (1997). Development of branchiomotor neurons in zebrafish. Development 124, 2633–2644. Higashijima, S., Hotta, Y., and Okamoto, H. (2000). Visualization of cranial motor neurons in live transgenic zebrafish expressing green fluorescent protein under the control of the islet-1 promoter/enhancer. J. Neurosci. 20, 206–218. Trevarrow, B., Marks, D.L., and Kimmel, C.B. (1990). Organization of hindbrain segments in the zebrafish embryo. Neuron 4, 669–679. Cooke, J.E., and Moens, C.B. (2002). Boundary formation in the hindbrain: Eph only it were simple. Trends Neurosci. 25, 260–267. Lumsden, A. (1999). Closing in on rhombomere boundaries. Nat. Cell Biol. 1, E83–E85. Klein, R. (1999). Bidirectional signals establish boundaries. Curr. Biol. 9, R691–R694. Picker, A., Scholpp, S., Bohli, H., Takeda, H., and Brand, M. (2002). A novel positive transcriptional feedback loop in midbrain-hindbrain boundary development is revealed through analysis of the zebrafish pax2.1 promoter in transgenic lines. Development 129, 3227–3239. Steinberg, M.S. (1963). Reconstruction of tissues by dissociated cells. Science 141, 401–408. Steinberg, M.S., and Takeichi, M. (1994). Experimental specification of cell sorting, tissue spreading, and specific spatial patterning by quantitative differences in cadherin expression. Proc. Natl. Acad. Sci. USA 91, 206–209. Nose, A., Nagafuchi, A., and Takeichi, M. (1988). Expressed recombinant cadherins mediate cell sorting in model systems. Cell 54, 993–1001. Inoue, T., Tanaka, T., Takeichi, M., Chisaka, O., Nakamura, S., and Osumi, N. (2001). Role of cadherins in maintaining the compartment boundary between the cortex and striatum during development. Development 128, 561–569. Lele, Z., Folchert, A., Concha, M., Rauch, G.J., Geisler, R., Rosa, F., Wilson, S.W., Hammerschmidt, M., and Bally-Cuif, L. (2002). parachute/n-cadherin is required for morphogenesis and maintained integrity of the zebrafish neural tube. Development 129, 3281–3294. Masai, I., Lele, Z., Yamaguchi, M., Komori, A., Nakata, A., Nishiwaki, Y., Wada, H., Tanaka, H., Nojima, Y., Hammerschmidt, M., et al. (2003). N-cadherin mediates retinal lamination, maintenance of forebrain compartments and patterning of retinal neurites. Development 130, 2479–2494. Wilkinson, D.G. (2003). How attraction turns to repulsion. Nat. Cell Biol. 5, 851–853. Poliakov, A., Cotrina, M., and Wilkinson, D.G. (2004). Diverse roles of eph receptors and ephrins in the regulation of cell migration and tissue assembly. Dev. Cell 7, 465–480. Holmberg, J., and Frisen, J. (2002). Ephrins are not only unattractive. Trends Neurosci. 25, 239–243. Holmberg, J., Clarke, D.L., and Frisen, J. (2000). Regulation of repulsion versus adhesion by different splice forms of an Eph receptor. Nature 408, 203–206. Zimmer, M., Palmer, A., Kohler, J., and Klein, R. (2003). EphBephrinB bi-directional endocytosis terminates adhesion allowing contact mediated repulsion. Nat. Cell Biol. 5, 869–878. Marston, D.J., Dickinson, S., and Nobes, C.D. (2003). Racdependent trans-endocytosis of ephrinBs regulates Ephephrin contact repulsion. Nat. Cell Biol. 5, 879–888.

Current Biology 542

33. Huynh-Do, U., Stein, E., Lane, A.A., Liu, H., Cerretti, D.P., and Daniel, T.O. (1999). Surface densities of ephrin-B1 determine EphB1-coupled activation of cell attachment through alphavbeta3 and alpha5beta1 integrins. EMBO J. 18, 2165–2173. 34. Hansen, M.J., Dallal, G.E., and Flanagan, J.G. (2004). Retinal axon response to ephrin-as shows a graded, concentrationdependent transition from growth promotion to inhibition. Neuron 42, 717–730. 35. Yin, Y., Yamashita, Y., Noda, H., Okafuji, T., Go, M.J., and Tanaka, H. (2004). EphA receptor tyrosine kinases interact with co-expressed ephrin-A ligands in cis. Neurosci. Res. 48, 285– 296. 36. Dravis, C., Yokoyama, N., Chumley, M.J., Cowan, C.A., Silvany, R.E., Shay, J., Baker, L.A., and Henkemeyer, M. (2004). Bidirectional signaling mediated by ephrin-B2 and EphB2 controls urorectal development. Dev. Biol. 271, 272–290. 37. Hornberger, M.R., Dutting, D., Ciossek, T., Yamada, T., Handwerker, C., Lang, S., Weth, F., Huf, J., Wessel, R., Logan, C., et al. (1999). Modulation of EphA receptor function by coexpressed ephrinA ligands on retinal ganglion cell axons. Neuron 22, 731–742. 38. Papan, C., and Campos-Ortega, J. (1994). On the formation of the neural keel and neural tube in the zebrafish Danio (Brachydanio) rerio. Rouxs Arch. Dev. Biol. 203, 178–186. 39. Concha, M.L., and Adams, R.J. (1998). Oriented cell divisions and cellular morphogenesis in the zebrafish gastrula and neurula: a time-lapse analysis. Development 125, 983–994. 40. Geldmacher-Voss, B., Reugels, A.M., Pauls, S., and CamposOrtega, J.A. (2003). A 90-degree rotation of the mitotic spindle changes the orientation of mitoses of zebrafish neuroepithelial cells. Development 130, 3767–3780. 41. Kimmel, C.B., Warga, R.M., and Kane, D.A. (1994). Cell cycles and clonal strings during formation of the zebrafish central nervous system. Development 120, 265–276. 42. Irving, C., Nieto, M.A., DasGupta, R., Charnay, P., and Wilkinson, D.G. (1996). Progressive spatial restriction of Sek-1 and Krox-20 gene expression during hindbrain segmentation. Dev. Biol. 173, 26–38. 43. Prince, V.E., Moens, C.B., Kimmel, C.B., and Ho, R.K. (1998). Zebrafish hox genes: expression in the hindbrain region of wild-type and mutants of the segmentation gene, valentino. Development 125, 393–406. 44. Fashena, D., and Westerfield, M. (1999). Secondary motoneuron axons localize DM-GRASP on their fasciculated segments. J. Comp. Neurol. 406, 415–424. 45. Moens, C.B., and Fritz, A. (1999). Techniques in neural development. Methods Cell Biol. 59, 253–272.