Bone morphogenetic protein (Bmp) signaling is crucial for the formation and patterning of zebrafish ventral and posterior mesoderm. Mutants defective in the Bmp pathway have expanded trunk muscle, abnormal tails and severely impaired development of ventral mesodermal derivatives such as vasculature,blood and pronephros. As Bmps continue to be expressed in the ventral and posterior mesoderm after gastrulation, it is likely that Bmp signaling continues to play an important developmental role during outgrowth of the posterior body. However, because Bmp signaling plays an essential role during the gastrula stages, it has not been possible with mutants or standard disruption techniques to determine the later functions of the Bmp pathway. To study the role of Bmp signaling in the ventral and posterior mesoderm during trunk and tail outgrowth, we generated a transgenic zebrafish line containing a heatshock-inducible dominant-negative Bmp receptor-GFP fusion. Our data show that Bmps are important for tail organizer formation and for patterning the ventral mesoderm during early gastrulation. However, from mid-gastrulation to the early somitogenesis stages, Bmp signaling is important for ventral tail fin development and for preventing secondary tail formation. We conclude that the role of Bmp signaling in the ventral and posterior mesoderm changes as gastrulation proceeds.

The formation of the vertebrate trunk and tail is a complex process involving carefully regulated cell differentiation and morphogenesis (reviewed by Griffith et al., 1992). The tailbud is a specialized domain arising from the ventrolateral regions of gastrula-stage fish and frog embryos, which contain precursors for all the tail tissues except the notochord (Gont et al., 1993; Kanki and Ho,1997; Davis, 2000). A variety of studies in both fish and frogs have suggested that the molecular signaling pathways important for development of the tail could exert their influence beginning at the early gastrula stages(Gont et al., 1993; Mullins et al., 1996). Furthermore, genes that will mark the tailbud during trunk and tail morphogenesis, such as eve1/Xhox3, are already expressed in early gastrula embryos (Joly et al.,1993), suggesting that the patterning of the posterior body begins very early in development, well before a morphologically apparent tailbud has formed. Whether or not the tailbud cells continue to be under the influence of patterning signals during later gastrulation and elongation of the tailbud has been less clear.

One crucial signaling pathway for vertebrate trunk and tail mesoderm formation is the Bone morphogenetic protein (Bmp) pathway (reviewed by Dale and Jones, 1999). Bmps are members of the TGFβ superfamily of secreted proteins, and they bind TypeI/TypeII receptor complexes to initiate a signaling cascade that activates transcription of downstream targets such as the homeobox genes voxand vent (Onichtchouk et al.,1996; Melby et al.,2000; von Bubnoff and Cho,2001; Ramel and Lekven,2004). During gastrulation in Xenopus and zebrafish, Bmp signaling is highest on the ventral side of the embryo, where it patterns the mesoderm to adopt fates including blood, vasculature, pronephros and tail muscle. Loss of Bmp signaling in both animals leads to expansion of dorsal structures, such as trunk muscle, at the expense of ventral structures(Re'em-Kalma et al., 1995; Mullins et al., 1996; Kishimoto et al., 1997). Zebrafish mutants support the model that a gradient of Bmp signaling patterns various tissue types during gastrulation in the developing embryo (reviewed by Dale and Wardle, 1999). For example, the swirl mutant, which lacks bmp2b, has radialized trunk somites, reduced blood, vasculature and pronephros, and no tail tissue(Kishimoto et al., 1997). Other mutants in the Bmp pathway, including snailhouse hypomorphs(Dick et al., 2000),piggytail (Kramer et al.,2002), lost-a-fin(Mintzer et al., 2001) and minifin (Connors et al.,1999), have progressively less severe phenotypes than swirl mutants. For instance, in zygotic lost-a-fin embryos,which are mutant for the Bmp receptor alk8, the patterning of the ventral and posterior mesoderm is normal, but the ventral tail fin is absent(Mintzer et al., 2001). Such data suggest that the highest levels of Bmp signaling are required for formation of the ventral tail fin, while lower levels of Bmp signaling are required for other ventral and posterior tissues.

The role of Bmps in tail development has been further elucidated by experiments involving overexpression of bmp RNA in the early zebrafish embryo. When injected in combination with RNA encoding Wnt8 and the Nodal ligand Cyclops (Cyc), Bmp induces the formation of ectopic tails(Agathon et al., 2003). Moreover, whereas tails can form at lower efficiency when Bmp is expressed with either Wnt8 or Cyc alone, Bmp is indispensable for tail formation. A second set of experiments suggests that Bmp signaling is important for reserving a population of tailbud cells for somitogenesis in the tail. When dorsal determinants, which activate Bmp antagonists such as chordin(Sasai et al., 1994) and noggin (Zimmerman et al.,1996), are removed before gastrulation, embryos do not develop trunk somites and instead only develop tail somites(Ober and Schulte-Merker,1999). If, however, dorsal determinants are removed in a swirl background, which lacks zygotic Bmp signaling, the embryos recover trunk somite formation and lose the tail somites. These experiments indicate that Bmp signaling negatively regulates trunk somitogenesis while promoting tail somitogenesis, potentially by setting aside a population of cells to form tail somites rather than trunk somites.

Such studies show a requirement for Bmp signaling in formation of the zebrafish tail organizer. However, they do not determine whether Bmps exert their influence on tailbud cells only during gastrulation, or if Bmps continue to be required in the post-gastrula tailbud. Bmps are expressed in the tailbud and surrounding tissues throughout somitogenesis(Martinez-Barbera et al.,1997; Dick et al.,2000), making it seem likely that they are playing important roles during trunk and tail formation. As mutants are defective in Bmp signaling from the onset of zygotic transcription, and overexpression, dominant-negative and morpholino experiments have been performed only before the gastrula stages, it is unclear what the role of Bmp is in the ventral and posterior mesoderm after the end of gastrulation. To address this question, we generated a zebrafish transgenic line for conditionally reducing Bmp signaling. We found that Bmp signaling plays an important role in primary tail formation during early gastrulation, and that it prevents the formation of ectopic tails from the mid-gastrula stages through early somitogenesis. High levels of Bmp signaling appear to be important for the development of most ventral mesodermal derivatives during early gastrulation, whereas the ventral fin requires Bmp signaling from early gastrulation through the early somitogenesis stages.

Dominant-negative Bmp receptor cloning

The Xenopus type Ia Bmp receptor(Graff et al., 1994) was amplified with the primers GAGATCTAAAATGAGAAAACGACTTTTCATTGC and CTCTAGACTGGATCTGCTTGGCTATAGTACG, truncating the receptor just before the kinase domain. This fragment was inserted into the BamHI and XbaI sites of CS XLT GFP (kind gift of J. Miller), fusing GFP in frame to the C-terminus of the truncated receptor downstream of the hsp70 promoter. IsceI meganuclease sites flank the transgene.

Generation of a stable transgenic line

Uncut DNA (1 nl) was injected into 1-cell stage WIK/AB embryos. IsceI meganuclease was co-injected with the DNA to maximize the number of integration events. The injection mix was: 100 pg/nl DNA, 0.5× IsceI Buffer, 10% IsceI enzyme. The F1generation was screened for heatshock-inducible GFP fluorescence, and a stable line was generated.

Heatshock conditions

Embryos from an F1 out-cross were typically heatshocked for 1 hour at desired stages by placing them for 1 hour in a 37°C air incubator,and then screened for GFP fluorescence after 1.5 hours at 28.5°C. Embryos were either fixed at 2 hours post-heatshock for in-situ hybridization analysis or left in a 28.5°C air incubator for later visualization. For PCR machine heatshocks, single embryos were separated into PCR tubes with 20 μl of embryo medium. The PCR block was ramped to 37°C for 1 hour, then held at 28.5°C.

Confocal microscopy

Truncated Bmp receptor-GFP/tbx6-gfp embryos were heatshocked from shield to bud stage in a 37°C air incubator. For late tail growth, embryos with ectopic tails and GFP fluorescence in the tail were identified at the 22-24-somite stage. Embryos were mounted in embryo medium containing 1%low-melt agarose. For early tail growth, 20 embryos were mounted at the 16-17-somite stage in 1% low-melt agarose with their tails exposed to the medium. Images were acquired using a Zeiss LSM Pascal confocal microscope and a 40 × lens. Slices (1.6 μm) were taken once per hour.

Transplants

For wild-type transplants (Moens and Fritz, 1999), 1-cell stage WIK/AB embryos were injected with fluorescein-dextran (Mr 10,000). At the dome stage, cells were transplanted from labeled donor embryos to the ventral side of shield-stage host embryos from a transgenic in-cross. Host embryos were immediately heatshocked in an air incubator at 37°C through bud stage to maximize ectopic tail formation. Embryos with ectopic tails at 24 hours were scored for the presence or absence of fluorescein-labeled cells within the ectopic tails. Control transplants from transgenic donors were performed in the same manner, but donor embryos were collected from in-crosses of transgenic fish. PCR analysis of genomic DNA was used to determine whether donor embryos were wild type or transgenic.

Generation of a zebrafish transgenic line to conditionally attenuate Bmp signaling

Overexpression of truncated Bmp receptors is a powerful method for specifically reducing Bmp signaling during development in both fish and frogs(Graff et al., 1994; Maeno et al., 1994; Suzuki et al., 1994; Schmidt et al., 1995; Neave et al., 1997). The truncated receptor acts as a dominant-negative by associating with endogenous receptor complexes and preventing intracellular phosphorylation of receptor SMADs upon ligand binding (Fig. 1) (Graff et al.,1994). The advantage to using a dominant-negative receptor as an inhibitor rather than a secreted Bmp inhibitor such as Noggin(Sasai et al., 1994) or Chordin (Zimmerman et al.,1996) is that it allows for cell-autonomous inhibition of Bmp signaling, which is very useful in transplant studies.

A previous study showed that a truncated Xenopus Bmp receptor is a very effective inhibitor of Bmp signaling in zebrafish embryos(Neave et al., 1997). This receptor is truncated just after the transmembrane domain, so it lacks the kinase domain and adjacent sequences. In order to trace the cells expressing the dominant-negative Bmp receptor, we generated a construct with GFP fused to the C-terminus of the truncated receptor. However, this fusion protein no longer inhibited Bmp signaling when overexpressed (data not shown). Thus, we generated a second construct with GFP replacing the kinase domain, but we left the sequences between the transmembrane domain and the original kinase domain intact. To test the efficacy of the new construct, we injected 1-cell zebrafish embryos with RNA encoding this truncated receptor-GFP fusion. We found that the new truncated receptor-GFP protein was as effective at inhibiting Bmp signaling as the original truncated Bmp receptor (data not shown).

To express the truncated Bmp receptor-GFP fusion at different times of development, we placed this construct under the control of the hsp70promoter (Halloran, 2000) and generated a zebrafish transgenic line. To determine if the induction of the transgene would recapitulate phenotypes of known Bmp pathway mutants, we heatshocked embryos before the onset of gastrulation and observed their phenotypes during the development of the posterior body. Transgenic embryos were sorted by GFP fluorescence 1 hour after the completion of the heatshock and scored at both the 14-somite stage and 30 hours post-fertilization (hpf) for the degree of dorsalization using the scale developed by Mullins et al.(Mullins et al., 1996). In this scale, C5 represents an embryo completely lacking Bmp signaling(swirl phenotype), whereas C1 is a mildly dorsalized embryo. Heatshocking embryos at 3 hpf produced severely dorsalized phenotypes in transgenic embryos (Fig. 2B,C). By contrast, sibling non-transgenic embryos that received the same heatshock treatment had no discernible defects (Fig. 2A), demonstrating that the effects we saw were not due to the heat treatment. The majority of transgenic embryos were in the C4 dorsalization class, with severe expansion of somites and curling of the trunk and tail (Fig. 2B; Table 1). Embryos in this class resembled hypomorphic snailhouse (bmp7) mutants(Dick et al., 2000). Some embryos were severely dorsalized at the end of gastrulation (10 hpf, data not shown), but they did not survive beyond the end of somitogenesis. These embryos were scored as a C5 phenotype, resembling swirl (bmp2b)mutant embryos (Kishimoto et al.,1997). A third class of embryos had C3 phenotypes, with expansion of posterior somites and curling at the end of the tail(Fig. 2C). Embryos in the C3 class resemble piggytail (smad5) mutants(Kramer et al., 2002). Interestingly, heatshocking embryos at 3 hpf in a PCR machine rather than an air incubator produced a more severe dorsalization, with all the embryos having a C5 phenotype (Table 2). These data demonstrate that induction of the transgene before gastrulation phenocopies zebrafish mutants with severe reductions in Bmp signaling.

As the truncated receptor is fused to GFP, we could readily examine the kinetics of induction. We observed the peak levels of GFP expression (and thus, truncated Bmp receptor expression) in embryos at 2 hours post-heatshock. By 6 hours post-heatshock, GFP fluorescence had begun to fade. We hypothesize that this rapid turnover is due to recycling of the truncated receptors within cells. Such a model would fit with previous reports of signaling receptors being rapidly degraded to allow for transient responses to extracellular signals (Rosenfeld et al.,2002; Strous and van Kerkhof,2002; Teis and Huber,2003). Therefore, our transgenic system provides a rapid,short-lived method for reducing Bmp signaling in the zebrafish embryo. As a result, we could study the role of Bmp in patterning tissues during brief windows of time.

Effects of reducing Bmp signaling on ventral and posterior mesoderm at different stages of development

Having developed a system for inducibly interfering with Bmp signaling, we wished to examine the effects of reducing Bmp signaling at different stages of development. Heatshocking embryos at shield stage caused phenotypes that were far less severe than those resulting from 3 hpf heatshocks. At 30 hpf, all transgenic embryos had the C1 dorsalized phenotype, with normal tails lacking ventral tail fins (Fig. 2E; Table 1). The ventral tail fin is thought to be the tissue that is most sensitive to losses of Bmp signaling,revealed by minifin (tolloid) and lost-a-fin (alk8) mutants(Connors et al., 1999; Mintzer et al., 2001). In addition to the C1 phenotype, typically one-fifth of transgenic embryos formed ectopic tail-like structures at their posterior ends(Fig. 2F; Table 1), although in one clutch as many as 45% of the embryos had these structures, which we henceforth refer to as ectopic tails. Heatshocking at the end of gastrulation (bud stage), and during somitogenesis (3-somite and 6-somite stages) also caused loss of the ventral tail fin (Table 1), indicating that ventral tail fin tissue is dependent upon continued Bmp signaling after gastrulation and through the early somitogenesis stages. Ectopic tail phenotypes were also induced by heatshocking at bud through 6-somite stages, but with lower penetrance than heatshocking at shield stage (Table 1). Heatshocking after the 6-somite stage had no effect on posterior development.

We reasoned that heatshocking individual embryos in a PCR machine might decrease the lag time of transgene expression, since this method caused a C5 phenotype in 100% of embryos heatshocked at 3 hpf. Indeed, this method resulted in GFP fluorescence within half an hour of heatshock. As shown in Table 2, heatshocking at shield(early gastrulation) using the PCR machine caused a more severe effect than similar heatshocks in an air incubator, with 41% of embryos having a C3 phenotype. By 70% epiboly (mid-gastrulation), however, heatshocking in the PCR machine caused only a C1 phenotype in transgenic embryos. Percentages of ectopic tails caused by heatshocking at various stages matched closely with similar heatshocks in the air incubator (compare Tables 1 and 2). Thus, Bmp signaling is important for ventral tail fin formation and suppression of ectopic tails from the mid-gastrula stage through early somitogenesis.

eve1 is a zebrafish homeobox gene expressed in the ventral mesoderm during gastrulation and in the tailbud during somitogenesis, thus serving as a good marker of ventroposterior fates within the embryo(Joly et al., 1993; Mullins et al., 1996). As eve1 expression is reduced in bmp mutants(Joly et al., 1993; Mullins et al., 1996), we wished to examine whether eve1 expression is dependent on Bmp signaling only during gastrulation, or whether it continues to be affected by reduction of Bmp signaling after gastrulation. We heatshocked embryos at shield stage, bud stage and 3-somite stage, then assayed for eve1expression by whole-mount in-situ hybridization. As predicted from mutant analyses, heatshocking at shield stage reduced eve1 expression at the ventral margin in transgenic embryos compared with wild-type siblings (compare Fig. 3B with 3A). Interestingly, heatshocking after the end of gastrulation (bud and 3-somite stages) still caused reduced eve1 levels in the tailbuds of transgenic embryos (compare Fig. 3C,E with 3D,F). The size of the tailbud was not decreased, as the homeobox gene vent1 (Kawahara et al.,2000; Melby et al.,2000) was expressed at normal levels in the tailbuds of transgenic embryos after heatshock (data not shown). These data show that eve1expression is dependent upon Bmp signaling during both gastrulation and early somitogenesis.

High levels of Bmp signaling are necessary for the formation of ventral mesodermal derivates, including blood, vasculature and pronephros. In bmp mutant embryos, markers of these cell types are severely reduced or absent (Mullins et al.,1996). Additionally, Bmp signaling is important for the development of blood precursors after gastrulation in Xenopus embryos(Schmerer and Evans, 2003). We sought to examine whether high levels of Bmp signaling are similarly important for development of ventral mesodermal derivatives after gastrulation in zebrafish. We heatshocked one clutch of embryos at 3 hpf to inhibit Bmp signaling throughout gastrulation, and then heatshocked a second clutch continuously from bud to 12-somites. We examined the expression of flk-1 (kdr – Zebrafish Information Network)(Sumoy et al., 1997), gata1 (Detrich et al.,1995) and pax2.1(Pfeffer et al., 1998) to determine whether vascular, hematopoietic and pronephric cells, respectively,were affected. After heatshocking at 3 hpf, flk-1 expression was dramatically reduced in transgenic embryos compared with controls (compare Fig. 4A and 4C). flk-1-expressing cells were limited to the very anterior and posterior ends of transgenic embryos, while these cells were spread throughout the trunk of control embryos. However, embryos heatshocked from bud to 12-somites had normal flk-1 expression (compare Fig. 4B with 4C). Like flk-1, gata1 expression was decreased in transgenic embryos heatshocked at 3 hpf (compare Fig. 4D with 4F). While control embryos had bilateral stripes of lateral plate mesoderm that strongly expressed gata1, this expression was limited to a few cells in the blood islands of transgenic embryos. Blood progenitors were apparently not affected by reductions in Bmp signaling after gastrulation, however, as gata1 expression was normal in transgenic embryos heatshocked from bud to 12-somites (compare Fig. 4E with 4F). Finally, pax2.1 expression was dramatically reduced in the pronephric ducts and tubules of transgenic embryos heatshocked at 3 hpf compared with controls(compare Fig. 4G with 4I). A few disorganized cells could be seen in the posterior of the transgenic embryo, while pronephric ducts and tubules were easily identified in wild-type siblings. Anteriorly, pax2.1 expression marking the otic placode was wider in transgenic embryos, indicating that these embryos, like some of the bmp mutants (Mullins et al.,1996), had expanded anterior tissues. In embryos heatshocked from bud to 12-somites, there was no difference in either the pronephric or anterior pax2.1 expression (compare Fig. 4H with 4I). Taken together, these data suggest that high levels of Bmp signaling are not required after gastrulation for development of vascular, blood or pronephric cells in zebrafish.

The ectopic tails develop only mesoderm and fin tissue

As loss of Bmp signaling had previously been shown to cause a reduction in tail organizer formation (Mullins et al.,1996), we were intrigued by the formation of ectopic tail-like structures when we reduced Bmp signaling after early gastrulation. We examined the tissue types within the ectopic tails to understand the extent of tail formation in these structures. Fin and pigment were clearly apparent at 3 days post-fertilization (dpf) in the ectopic tails(Fig. 5A). This initially suggested that the ectopic tails could organize epidermal and neural crest cells. To determine whether the ectopic tail, like the primary tail, forms a tail-organizing center, we examined eve1 expression at various time points after heatshock. As discussed earlier, Bmp signaling is crucial for normal levels of eve1 in the tailbud even after gastrulation. However, when we heatshocked at bud stage and examined eve1expression at 21 hpf (before ectopic tails were obvious under a dissecting microscope), we saw a new patch of eve1 expression in the ventral portion of the tail (Fig. 5B). This result suggests that secondary tailbuds form before ectopic tails begin growing. eve1 expression was maintained in the secondary tailbud at 30 hpf, after the ectopic tail grew to a significant length(Fig. 5C). Another tailbud marker, no tail, which is critical for primary tail formation(Schulte-Merker et al., 1994),was also expressed in the ectopic tail(Fig. 5D). Additionally, we saw expression of myod (Weinberg et al., 1996) in cells of the ectopic tail, demonstrating that the tails have developing muscle tissue (Fig. 5E). Hypochord cells also migrated into ectopic tails, as marked by coll2a (Yan et al.,1995) expression (Fig. 5F). We next wished to analyze whether the ectopic tails had neural tissue. We examined caudal(Joly et al., 1992) expression(Fig. 5G) and neurogenin (Korzh et al.,1998) expression (Fig. 5H), which are both expressed in the neural tube, but only caudal was expressed in the ectopic tails. As caudal is also expressed in the primary tailbud at this stage, we believe that the cells expressing this gene in the ectopic tails are actually tailbud cells. We examined other markers of neural cell types, including islet1(Inoue et al., 1994), which marks primary motoneurons, the pan-neuronal marker hu(Marusich et al., 1994) and the neural crest marker crestin(Luo et al., 2001), but we did not observe expression of any of these markers in secondary tails (data not shown). Thus, the ectopic tails we examined lacked neural tissue at 30 hpf,and subsequent pigment formation was probably derived from neural crest cells migrating ventrally out of the primary tail. Surprisingly, when we examined sonic hedgehog (Strahle et al.,1996) expression to assess whether floorplate cells were present,we only saw expression in cells that looked like notochord in some of the ectopic tails (Fig. 5I). These cells were columnar in shape, as opposed to the round floorplate cells, and they branched off the posterior part of the primary notochord. Thus, we conclude that secondary tails can have many of the same tissue types as primary tails, including tailbud cells, somitic tissue, notochord, hypochord,epidermis and pigment (which probably migrates into the secondary tails later than 30 hpf). The ectopic tails are mostly mesodermal, however, and they do not actively develop neural tissue.

Visualizing growth of the ectopic tails in living embryos

To examine the growth of the ectopic tails in a more dynamic fashion, we utilized a transgenic zebrafish line previously developed in our laboratory. This line, which expresses GFP under control of the tbx6 promoter,marks cells of the tailbud, presomitic mesoderm and newly formed somites(Szeto and Kimelman, 2004). As the truncated Bmp receptor-GFP transgenic fish are only fluorescent for about 6 hours after heatshock, we crossed these fish to tbx6-gfp fish,heatshocked at shield stage to induce ectopic tail formation, and visualized tbx6 promoter-driven GFP during late somitogenesis stages. Visualizing ectopic tail growth by confocal microscopy from the 25-somite stage, we easily observed GFP expression in the ventral region of the embryo(Fig. 6A). Interestingly, the GFP fluorescence demonstrated that while only a small portion of the ectopic tail separated from the main body, this structure could run a considerable distance along the body axis (arrows in Fig. 6A). We also observed the maturation of the presomitic mesoderm into small somites (the first ectopic somite seen at the 25-somite stage is marked by an asterisk). This closely mirrored the maturation of somitic cells in the primary tail. Rather than being a simple outgrowth containing various cell types from the primary tail,these data show that the secondary tail is active in elongating and forming mature somites.

We next wished to examine whether the ectopic tails arose from the primary tail splitting into two tails, or from a completely distinct tail-forming region. We examined truncated Bmp receptor-GFP/tbx6-gfp transgenic embryos by confocal microscopy beginning at the 17-18-somite stage, just after primary tail formation begins. At this stage, approximately 20% of the transgenic embryos formed small ectopic patches of GFP ventral to the primary tailbud (Fig. 6Ba,c). These ectopic patches, which later produced ectopic tails(Fig. 6Bb,d), started to grow out just after the primary tail had extended beyond the 17-18-somite position. We also examined embryos from earlier stages (12-13 somites), but we did not detect ectopic tail formation until at least the 16-17-somite stage (data not shown). Based on the time and position where the ectopic tails first began forming, we suggest that a population of mesodermal tail progenitors is left behind as the primary tail grows out, and these progenitors form the ectopic tail.

Transplantation of cells with reduced Bmp signaling

We next asked whether cells with normal Bmp signaling capabilities could populate ectopic tails. We reasoned that trunk and/or tail cells are mis-specified to form ectopic tails when Bmp signaling levels are reduced at mid-gastrulation or early somitogenesis. To test this hypothesis, we performed transplants of fluorescein-labeled wild-type donor cells into ventral regions of transgenic hosts and heatshocked the hosts at shield stage(Fig. 7A). We then scored for the presence of donor cells in the ectopic tails at 24 hpf. As controls, we performed the same experiment with transgenic donor cells to be sure that cells with reduced Bmp signaling would move into ectopic tails. As shown in Fig. 7, wild-type transplanted cells did not populate ectopic tails (Fig. 7B; 0%, n=12), while transgenic cells did(Fig. 7C; 57%, n=14). The transgenic cells could also populate the primary tail(Fig. 7C). By contrast, the transgenic cells transplanted into wild-type embryos never produced an ectopic tail (data not shown), potentially because we could not get a large enough group of transgenic cells in one location to form an ectopic tail. These data indicate that cells capable of receiving Bmp signals are actively recruited into primary trunk or tail tissue, while cells that have reduced Bmp signaling capacity can go into either the primary trunk and tail or the ectopic tail.

Our results demonstrate the dynamic temporal roles of Bmp signaling during development of ventral and posterior mesoderm in zebrafish. While other groups have revealed the importance of the Bmp pathway in ventroposterior patterning,they could not address the timing of these patterning events during embryogenesis. Data presented here suggest that Bmps are critical for primary tail formation and ventral mesoderm development during the early gastrula stages, but the development of these tissues is not dependent on the same levels of Bmp signaling from the mid-gastrula stages onward. Instead, Bmps are important for preventing the formation of ectopic tail structures and for promoting ventral fin formation through early somitogenesis.

Combinatorial signaling in trunk and tail somite formation

Patterning of trunk and tail mesoderm involves the input of signals as diverse as Wnts (Lekven et al.,2001), Fgfs (Griffin et al.,1995) and Nodals (Thisse et al., 2000) in addition to Bmps. The importance of each of these pathways in forming the posterior body has been revealed by mutant studies,but their temporal-specific roles are less clear. A pharmacological Fgf inhibitor revealed the necessity of Fgf signaling for somite formation during posterior body outgrowth (Griffin and Kimelman, 2003), and we have observed that the Wnt pathway is needed continuously during the somitogenesis stages for trunk and tail development (E. Herbig, U.J.P. and D.K., unpublished). As all four signaling molecules are expressed in the ventral region during gastrulation and in the posterior body during somitogenesis, it is likely that they are not only playing individual roles in forming the trunk and tail, but combinatorial roles as well. Indeed, combinatorial signaling between Bmps and Wnts has been recently shown to be important for maximal expression of the zebrafish tailbud gene tbx6 (Szeto and Kimelman,2004), and the zebrafish organizer-repressing genes voxand vent (Ramel and Lekven,2004). Our data suggest that Bmps are individually important for formation of the tail organizer during early gastrulation, but they are not crucial for trunk or tail formation after this stage. One reason that the role of Bmp in tail development may shift so dramatically is that signaling pathways such as those for Wnt and Fgf can compensate for the loss of Bmp signaling beyond the early gastrula stages. The ventroposterior gene eve1, for example, is regulated by Bmp and Fgf signals(Joly et al., 1993; Griffin et al., 1995). While Bmp is needed for initiation and maintenance of maximal eve1expression, Fgf may be sufficient to maintain enough eve1 expression after early gastrulation for proper tail organizer function. Another possibility is that Bmp signaling initially establishes the ventroposterior domain of mesodermal cells during early gastrulation, but it has no role at later times other than to regulate the formation of the ventral fin. While this is a relatively small domain within the embryo, it is important for normal development and would provide a reason for the embryo to maintain Bmp signals in the tailbud during somitogenesis. With the ability to conditionally reduce Bmp activity in zebrafish embryos it will now be possible to examine the interactions with other signaling pathways to determine if Bmp acts combinatorially with other pathways after gastrulation, or whether Bmp has a much more circumscribed role during this time.

Bmp signaling and ventral mesoderm development

In the gradient model for Bmp signaling, ventral mesodermal derivatives require the highest levels of Bmp activity during gastrulation. Bmps in the most ventral part of the margin specify a population of cells to form multipotent hemangioblast cells (Walmsley et al., 2002) as well as cardiac precursors(Kishimoto et al., 1997) and pronephros cells (Mullins et al.,1996). These cells are subsequently patterned to give rise to different subpopulations of the cardiovascular, blood and pronephric systems. Data in Xenopus using an inducible smad6 Bmp inhibitor construct suggest that Bmp signaling is important for maintenance of blood precursors even after gastrulation(Schmerer and Evans, 2003). However, our data indicate that in zebrafish, neither blood, vascular nor pronephric precursors are affected by reduction of Bmp activity after the gastrula stages. Like the tail mesodermal cells, it seems likely that Bmp induces a group of cells in the margin to adopt the most ventral fates, while other signals maintain and pattern these cells. Such a model is supported by mutant analyses, in which zygotic mutants for somitabun and lost-a-fin have normal ventral mesoderm, but maternal zygotic mutants for these genes have deficient ventral mesoderm(Mintzer et al., 2001; Kramer et al., 2002). Therefore, the early acting maternal supplies of smad5 and alk8 are sufficient for ventral mesoderm formation in the absence of functional zygotic protein.

We cannot exclude a later role for Bmps in regulating downstream patterning or morphogenesis of the ventral mesodermal derivatives. Indeed, the analysis of radar morphants suggests that reducing the post-gastrula activity of this Bmp family member compromises vascular integrity in the trunk and tail(Hall et al., 2002). While we did not observe such a phenotype in later heatshocks (data not shown), it is possible that the dominant-negative Bmp receptor does not inhibit Radar signaling or that we did not heatshock at the appropriate stage. However, we can conclude that Bmps are required for early patterning of ventral mesodermal derivatives during gastrulation, but not afterward.

Roles of Bmp signaling after early gastrula stages

Our data show that Bmp signaling plays very different roles in posterior development after the early gastrula stages(Fig. 8). Rather than maintaining tail formation during somitogenesis, Bmp is important for preventing the formation of ectopic tails and for promoting the formation of the ventral fin. Heatshocks between shield stage and early somitogenesis stages cause phenotypes closely resembling those of severe minifinmutants (Connors et al., 1999). In minifin mutants, the extracellular metalloprotease Tolloid is mutated, leading to reductions in Bmp signaling levels when zygotic Tolloid has been depleted. As discussed earlier, these mutants do not develop ventral tail fins. Our transgenic analysis allows us to postulate that the zygotic minifin mutant phenotype is due to reduced Bmp signaling after the early gastrula stages, when specification of the primary tail has already occurred. This model is in agreement with data from Connors and Mullins (S. A. Connors and M. C. Mullins, personal communication), in which transgenic rescue of minifin mutant embryos is most efficiently achieved from the late gastrula stages onward. Similar to our results, they find that rescue of the minifin phenotype can only be effectively achieved through the early somitogenesis stages. Using two different approaches, these results demonstrate a role for Bmp signaling during the late gastrula and early somitogenesis stages in ventral tail fin formation.

Lowering Bmp signaling after early gastrulation causes the formation of secondary tails. We can detect, in live embryos or by in-situ hybridization,multiple tissue types, including fin, pigment, somitic cells, hypochord and occasionally notochord in secondary tails. Aside from pigment cells (which probably migrate into the ectopic tails after 30 hpf), we do not detect the presence of neural tissue in ectopic tails. Our results differ from those of Agathon et al. (Agathon et al.,2003), who showed that extra Bmp signaling, in combination with Wnt and/or Nodal, induces ectopic tail formation. The ectopic tails they observed, however, were quite different from the ones we observed,and could have both trunk and tail somites. Whereas their ectopic tails often extended out from the trunk region, we observed ectopic tails forming only within the tail region, and only at the time the primary tail began to grow. Additionally, whereas the ectopic tails they observed had both neural and mesodermal tissue, the ectopic tails formed in our embryos were solely mesodermal (with a surrounding fin). The difference between our results and theirs is likely to be due to the experimental approaches. Whereas Agathon et al. (Agathon et al., 2003)overexpressed a combination of ligands beginning at the 128-256-cell stage, we inhibited only Bmp signaling after the early gastrula stage. These results support the model that the role of Bmp signaling changes during the early period of development.

As bmps are strongly expressed in the ventral portion of the primary tail, it is likely that this signaling normally prevents the organization of secondary tail structures. In line with this, our transplant data indicate that only posterior cells lacking Bmp signaling can contribute to ectopic tail tissues. The formation of ectopic tails is, therefore, a cell-autonomous effect. Wild-type cells in a background of cells with low Bmp signaling behaved like wild-type cells in a wild-type background, whereas only cells with a lowered capacity to respond to Bmp signaling became part of the ectopic tail. Thus, Bmp signaling may be important for limiting the number of tail precursors during the mid-gastrula through early somitogenesis stages.

Two different morphogenetic mechanisms could explain the emergence and growth of secondary tails. In the first, loss of Bmp signaling through the early somitogenesis stages could lead to a split in the tail later in development. In this view, the ventral portion of the primary tailbud splits away from the remainder of the tailbud, pulling away tissue such as the hypochord and notochord from the primary tail during outgrowth. A second mechanism is that the most ventral tail mesoderm could be converted to a new tail organizer when Bmp signaling is disrupted. As this organizer grows out,it forms primarily somites, and it also recruits the hypochord and notochord away from the primary axis. These two models are not mutually exclusive, and we favor a combination of both. Our experiments examining the emergence of the ectopic tail at the 17-18-somite stage suggest that the ventral portion of the primary tailbud dissociates from the main tailbud to form an ectopic tail. This is not, however, a simple split of the tail into equal parts, as is commonly seen in Xenopus embryos with disrupted convergence extension movements (reviewed by Ueno and Greene,2003). Rather, a very specific domain of the embryo gives rise to mesodermal tail tissues independent of the primary tail. We suggest that the ectopic tail forms from a population of tail progenitor cells that detach from the primary tailbud at the onset of tail formation and are left behind in the ventral region of the embryo. Furthermore, we propose that these are the cells that would otherwise form ventral tail mesoderm, but their fate changes due to an early disruption in Bmp signaling. Future studies will be needed to determine why in some cases these cells are left behind to form an ectopic tail, whereas in most cases the cells integrate into the primary tailbud.

New avenues of research

Bmp signaling is used at many times in development to affect a wide variety of cell types. Although its role has been extensively studied during early times in the lower vertebrates, the conditional dominant-negative fish described here will allow investigators to study Bmp signaling throughout development and in adult fish. Because the dominant-negative Bmp receptor acts cell-autonomously, it is well suited for transplant studies. Moreover, as the receptor is GFP-tagged, in crosses of heterozygous fish it is easy to identify the embryos expressing the transgene. Finally, because the dominant-negative Bmp receptor turns over fairly rapidly, it is possible to do timecourse experiments to determine when Bmp signaling is needed for particular developmental events. We expect that these fish will be a useful resource for analysis of Bmp signaling in zebrafish.

We wish to thank Dave White and the UW Zebrafish Facility for fish care and maintenance; Dave Raible and Jared Ragland for helpfuldiscussions; Jessica Lewis, Alex Nechiporuk and Gilbert Weidinger for probes; Daniel Szeto for probes and for help with the tbx6-gfp transgenic fish; Kelly Grant for help with confocal microscopy; Alvaro Sagasti for the IsceIprotocol; and Mary Mullins for communicating results prior to publication. This work was supported by grants from the NIH (GM65469) and March of Dimes(FY2003-120) to D.K. U.P. and A.W. were supported by NIH training grant GM07270.

Agathon, A., Thisse, C. and Thisse, B. (
2003
). The molecular nature of the zebrafish tail organizer.
Nature
424
,
448
-452.
Connors, S. A., Trout, J., Ekker, M. and Mullins, M. C.(
1999
). The role of tolloid/mini fin in dorsoventral pattern formation of the zebrafish embryo.
Development
126
,
3119
-3130.
Dale, L. and Jones, C. M. (
1999
). BMP signalling in early Xenopus development.
BioEssays
21
,
751
-760.
Dale, L. and Wardle, F. C. (
1999
). A gradient of BMP activity specifies dorsal-ventral fates in early Xenopusembryos.
Semin. Cell. Dev. Biol.
10
,
319
-326.
Davis, R. and Kirschner, M. W. (
2000
). The fate of cells in the tailbud of Xenopus laevis.
Development
127
,
255
-267.
Detrich, H. W., 3rd, Kieran, M. W., Chan, F. Y., Barone, L. M.,Yee, K., Rundstadler, J. A., Pratt, S., Ransom, D. and Zon, L. I.(
1995
). Intraembryonic hematopoietic cell migration during vertebrate development.
Proc. Natl. Acad. Sci. USA
92
,
10713
-10717.
Dick, A., Hild, M., Bauer, H., Imai, Y., Maifeld, H., Schier, A. F., Talbot, W. S., Bouwmeester, T. and Hammerschmidt, M.(
2000
). Essential role of Bmp7 (snailhouse) and its prodomain in dorsoventral patterning of the zebrafish embryo.
Development
127
,
343
-354.
Gont, L. K., Steinbesser, H., Blumberg, B. and DeRobertis, E. M. (
1993
). Tail formation as a continuation of gastrulation:the multiple cell populations of the Xenopus tailbud derive from the late blastopore lip.
Development
119
,
991
-1004.
Graff, J. M., Thies, R. S., Song, J. J., Celeste, A. J. and Melton, D. A. (
1994
). Studies with a Xenopus BMP receptor suggest that ventral mesoderm-inducing signals override dorsal signals in vivo.
Cell
79
,
169
-179.
Griffin, K. J. and Kimelman, D. (
2003
). Interplay between FGF, one-eyed pinhead, and T-box transcription factors during zebrafish posterior development.
Dev. Biol.
264
,
456
-466.
Griffin, K. J. P., Patient, R. K. and Holder, N. H. K.(
1995
). Analysis of FGF function in normal and no tailzebrafish embryos reveals separate mechanisms for formation of the trunk and tail.
Development
121
,
2983
-2994.
Griffith, C. M., Wiley, M. J. and Sanders, E. J.(
1992
). The vertebrate tail bud: three germ layers from one tissue.
Anat. Embryol.
185
,
101
-113.
Hall, C. J., Flores, M. V., Davidson, A. J., Crosier, K. E. and Crosier, P. S. (
2002
). Radar is required for the establishment of vascular integrity in the zebrafish.
Dev. Biol.
251
,
105
-117.
Halloran, M. C., Sato-Maeda, M., Warren, J. T., Su, F., Lele,Z., Krone, P. H., Kuwada, J. Y. and Shoji, W. (
2000
). Laser-induced gene expression in specific cells of transgenic zebrafish.
Development
127
,
1953
-1960.
Inoue, A., Takahashi, M., Hatta, K., Hotta, Y. and Okamoto,H. (
1994
). Developmental regulation of islet-1 mRNA expression during neuronal differentiation in embryonic zebrafish.
Dev. Dyn.
199
,
1
-11.
Joly, J. S., Maury, M., Joly, C., Duprey, P., Boulekbache, H. and Condamine, H. (
1992
). Expression of a zebrafish caudal homeobox gene correlates with the establishment of posterior cell lineages at gastrulation.
Differentiation
50
,
75
-87.
Joly, J. S., Joly, C., Schulte Merker, S., Boulekbache, H. and Condamine, H. (
1993
). The ventral and posterior expression of the zebrafish homeobox gene eve1 is perturbed in dorsalized and mutant embryos.
Development
119
,
1261
-1275.
Kanki, J. P. and Ho, R. K. (
1997
). The development of the posterior body in zebrafish.
Development
124
,
881
-893.
Kawahara, A., Wilm, T., Solnica-Krezel, L. and Dawid, I. B.(
2000
). Functional interaction of vega2 and goosecoid homeobox genes in zebrafish.
Genesis
28
,
58
-67.
Kishimoto, Y., Lee, K. H., Zon, L., Hammerschmidt, M. and Schulte-Merker, S. (
1997
). The molecular nature of zebrafish swirl: BMP2 function is essential during early dorsoventral patterning.
Development
124
,
4457
-4466.
Korzh, V., Sleptsova, I., Liao, J., He, J. and Gong, Z.(
1998
). Expression of zebrafish bHLH genes ngn1 and nrd defines distinct stages of neural differentiation.
Dev. Dyn.
213
,
92
-104.
Kramer, C., Mayr, T., Nowak, M., Schumacher, J., Runke, G.,Bauer, H., Wagner, D. S., Schmid, B., Imai, Y., Talbot, W. S. et al.(
2002
). Maternally supplied Smad5 is required for ventral specification in zebrafish embryos prior to zygotic Bmp signaling.
Dev. Biol.
250
,
263
-279.
Lekven, A. C., Thorpe, C. J., Waxman, J. S. and Moon, R. T.(
2001
). Zebrafish wnt8 encodes two wnt8 proteins on a bicistronic transcript and is required for mesoderm and neurectoderm patterning.
Dev. Cell
.
1
,
103
-114.
Luo, R., An, M., Arduini, B. L. and Henion, P. D.(
2001
). Specific pan-neural crest expression of zebrafish Crestin throughout embryonic development.
Dev. Dyn.
220
,
169
-174.
Maeno, M., Ong, R. C., Suzuki, A., Ueno, N. and Kung, H.(
1994
). A truncated bone morphogenetic protein 4 receptor alters the fate of ventral mesoderm to dorsal mesoderm: roles of animal pole tissue in the development of the ventral mesoderm.
Proc. Natl. Acad. Sci. USA
91
,
10260
-10264.
Martinez-Barbera, J. P., Toresson, H., da Rocha, S. and Krauss,S. (
1997
). Cloning and expression of three members of the zebrafish Bmp family: Bmp2a, Bmp2b and Bmp4.
Gene
198
,
53
-59.
Marusich, M. F., Furneaux, H. M., Henion, P. D. and Weston, J. A. (
1994
). Hu neuronal proteins are expressed in proliferating neurogenic cells.
J. Neurobiol.
25
,
143
-155.
Melby, A. E., Beach, C., Mullins, M. and Kimelman, D.(
2000
). Patterning the early zebrafish by the opposing actions of bozozok and vox/vent.
Dev. Biol.
224
,
275
-285.
Mintzer, K. A., Lee, M. A., Runke, G., Trout, J., Whitman, M. and Mullins, M. C. (
2001
). Lost-a-fin encodes a type I BMP receptor, Alk8, acting maternally and zygotically in dorsoventral pattern formation.
Development
128
,
859
-869.
Moens, C. B. and Fritz, A. (
1999
). Techniques in neural development.
Methods Cell. Biol.
59
,
253
-272.
Mullins, M. C., Hammerschmidt, M., Kane, D. A., Odenthal, J.,Brand, M., van Eeden, F. J., Furutani-Seiki, M., Granato, M., Haffter, P.,Heisenberg, C. P. et al. (
1996
). Genes establishing dorsoventral pattern formation in the zebrafish embryo: the ventral specifying genes.
Development
123
,
81
-93.
Neave, B., Holder, N. and Patient, R. (
1997
). A graded response to BMP-4 spatially coordinates patterning of the mesoderm and ectoderm in the zebrafish.
Mech. Dev.
62
,
183
-195.
Ober, E. A. and Schulte-Merker, S. (
1999
). Signals from the yolk cell induce mesoderm, neuroectoderm, the trunk organizer, and the notochord in zebrafish.
Dev. Biol.
215
,
167
-181.
Onichtchouk, D., Gawantka, V., Dosch, R., Delius, H.,Hirschfeld, K., Blumenstock, C. and Niehrs, C. (
1996
). The Xvent-2 homeobox gene is part of the BMP-4 signalling pathway controlling dorsoventral patterning of Xenopus mesoderm.
Development
122
,
3045
-3053.
Pfeffer, P. L., Gerster, T., Lun, K., Brand, M. and Busslinger,M. (
1998
). Characterization of three novel members of the zebrafish Pax2/5/8 family: dependency of Pax5 and Pax8 expression on the Pax2.1 (noi) function.
Development
125
,
3063
-3074.
Ramel, M. C. and Lekven, A. C. (
2004
). Repression of the vertebrate organizer by Wnt8 is mediated by Vent and Vox.
Development
131
,
3991
-4000.
Re'em-Kalma, Y., Lamb, T. and Frank, D. (
1995
). Competition between noggin and bone morphogenetic protein 4 activities may regulate dorsalization during Xenopus development.
Proc. Natl. Acad. Sci. USA
92
,
12141
-12145.
Rosenfeld, J. L., Knoll, B. J. and Moore, R. H.(
2002
). Regulation of G-protein-coupled receptor activity by rab GTPases.
Recept. Channels
8
,
87
-97.
Sasai, Y., Lu, B., Steinbesser, H., Geissert, D., Gont, L. K. and DeRobertis, E. M. (
1994
). Xenopus chordin: a novel dorsalizing factor activated by organizer-specific homeobox genes.
Cell
79
,
779
-790.
Schmerer, M. and Evans, T. (
2003
). Primitive erythropoiesis is regulated by Smad-dependent signaling in postgastrulation mesoderm.
Blood
102
,
3196
-3205.
Schmidt, J. E., Suzuki, A., Ueno, N. and Kimelman, D.(
1995
). Localized BMP-4 mediates dorsal-ventral patterning in the early Xenopus embryo.
Dev. Biol.
169
,
37
-50.
Schulte-Merker, S., van Eeden, F. J. M., Halpern, M. E., Kimmel,C. B. and Nüsslein-Volhard, C. (
1994
). no tail(ntl) is the zebrafish homologue of the mouse T(Brachyury) gene.
Development
120
,
1009
-1015.
Strahle, U., Blader, P. and Ingham, P. W.(
1996
). Expression of axial and sonic hedgehog in wildtype and midline defective zebrafish embryos.
Int. J. Dev. Biol.
40
,
929
-940.
Strous, G. J. and van Kerkhof, P. (
2002
). The ubiquitin-proteasome pathway and the regulation of growth hormone receptor availability.
Mol. Cell Endocrinol.
197
,
143
-151.
Sumoy, L., Keasey, J. B., Dittman, T. D. and Kimelman, D.(
1997
). A role for notochord in axial vascular development revealed by analysis of phenotype and the expression of VEGR-2 in zebrafish flh and ntl mutant embryos.
Mech. Dev.
63
,
15
-27.
Suzuki, A., Thies, R. S., Yamaji, N., Song, J. J., Wozney, J. M., Murakami, K. and Ueno, N. (
1994
). A truncated bone morphogenetic protein receptor affects dorsal-ventral patterning in the early Xenopus embryo.
Proc. Natl. Acad. Sci. USA
91
,
10255
-10259.
Szeto, D. P. and Kimelman, D. (
2004
). Combinatorial gene regulation by Bmp and Wnt in zebrafish posterior mesoderm formation.
Development
131
,
3751
-3760.
Teis, D. and Huber, L. A. (
2003
). The odd couple: signal transduction and endocytosis.
Cell Mol Life Sci
60
,
2020
-2033.
Thisse, B., Wright, C. V. and Thisse, C.(
2000
). Activin- and Nodal-related factors control antero-posterior patterning of the zebrafish embryo.
Nature
403
,
425
-428.
Ueno, N. and Greene, N. D. (
2003
). Planar cell polarity genes and neural tube closure.
Birth Defects Res. C Embryo Today
69
,
318
-324.
von Bubnoff, A. and Cho, K. W. (
2001
). Intracellular BMP signaling regulation in vertebrates: pathway or network?
Dev. Biol.
239
,
1
-14.
Walmsley, M., Ciau-Uitz, A. and Patient, R.(
2002
). Adult and embryonic blood and endothelium derive from distinct precursor populations which are differentially programmed by BMP in Xenopus.
Development
129
,
5683
-5695.
Weinberg, E. S., Allende, M. L., Kelly, C. S., Abdelhamid, A.,Murakami, T., Andermann, P., Doerre, O. G., Grunwald, D. J. and Riggleman,B. (
1996
). Developmental regulation of zebrafish MyoD in wild-type, no tail and spadetail embryos.
Development
122
,
271
-280.
Yan, Y. L., Hatta, K., Riggleman, B. and Postlethwait, J. H.(
1995
). Expression of a type II collagen gene in the zebrafish embryonic axis.
Dev. Dyn.
203
,
363
-376.
Zimmerman, L. B., de Jesús-Escobar, J. M. and Harland, R. M. (
1996
). The Spemann organizer signal noggin binds and inactivates bone morphogenetic protein 4.
Cell
86
,
599
-606.