SEARCH

SEARCH BY CITATION

Keywords:

  • KLF;
  • Neptune;
  • FGF;
  • posterior;
  • Xenopus;
  • embryo;
  • somite;
  • BMP;
  • tail

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

In order to elucidate the molecular mechanisms underlying the posterior axis and tail formation in embryogenesis, the function of Neptune, a zinc-finger transcription factor, in Xenopus laevis embryos was investigated. Injection of neptune mRNA into the animal pole area of embryos resulted in the formation of an additional tail structure that included a neural tube and muscle tissue. This activity required FGF signaling since coinjection of a dominant-negative FGF receptor RNA (XFD) completely blocked the formation of a tail structure. A loss-of-function experiment using a fusion construct of neptune and Drosophila engrailed (en-neptune) RNA showed that endogenous Neptune is necessary for formation of the posterior trunk and tail. Furthermore, activity of Neptune was necessary for the endogenous expression of brachyury and fgf-8 at the late gastrula stage. These findings demonstrate a novel function of Neptune in the process of anterior-posterior axis formation through the FGF and brachyury signaling cascades. An experiment using a combination explant with ventral and dorsal marginal tissues showed that cooperation of these two distinct tissues is important for the tail formation and that expression of Neptune in prospective ventral cells may be involved in the activation of the process of tail formation. Developmental Dynamics 234:63–73, 2005. © 2005 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Formation of the three germ layers of vertebrate embryos involves highly complex processes. In Xenopus, the ectoderm arises from the animal pole and the endoderm arises from the vegetal pole of the embryo. The mesoderm arises in the marginal zone as a result of inductive interactions between the animal and vegetal tissues of the embryo by signals derived from the endoderm (Nieuwkoop, 1969). In recent molecular studies, VegT, a T-box transcription factor produced maternally and localized vegetally, has been identified as an early endogeneous mesoderm-inducing factor (Zhang et al., 1998). Activity of VegT in mesoderm induction might be mediated through the activation of nodal-related genes (Clements et al., 1999; Kofron et al., 1999; Agius et al., 2000).

Within each layer, molecular signals and morphogenetic movements establish a set of coordinated ventral/posterior fate (Munoz-Sanjuan and Hemmati-Brivanlou, 2001). As for the posteriorizing signal, fibroblast growth factor (FGF) signal is known to play a key role in vertebrates. In Xenopus, addition of FGF induces mesodermal tissues in animal cap explants (Kimelman and Kirschner, 1987; Slack et al., 1987) and expression of dominant-negative FGF receptor leads to loss of the posterior body axis (Amaya et al., 1991). In zebrafish, inhibition of the FGF receptor signal leads to complete loss of both the trunk and tail (Griffin et al., 1995). Furthermore, mutant mice lacking FGF8 are unable to undergo normal gastrulation, leading to loss of mesoderm- and endoderm-derived tissues (Sun et al., 1999a). These findings suggest that the FGF signal plays a role in mesoderm induction and posterior formation.

A well-studied factor that is located downstream of FGF signaling is brachyury. brachyury, a pan-mesodermal T-box gene, was first identified in a mutant mouse lacking a tail and notochord (Herrmann et al., 1990). The zebrafish homolog of mouse brachyury is no tail. Mutation of this gene also leads to lack of a notochord and tail (Halpern et al., 1993; Schulte-Merker et al., 1994). A loss-of-function study in Xenopus embryos showed that brachyury is necessary for differentiation of the notochord and posterior mesoderm (Conlon et al., 1996). Xenopus brachyury has an activity to induce ventroposterior mesoderm in animal cap explants (Cunliffe and Smith, 1992). Furthermore, embryonic FGF (eFGF) and brachyury, both of which are expressed in the marginal zone at the gastrula stage (Issacs et al., 1992; Smith et al., 1991), form a positive-feedback regulatory loop (Issacs et al., 1994; Schulte-Merker and Smith, 1995). FGF and brachyury are known to play key roles in posterior mesoderm specification, but the molecular mechanisms underlying the posterior formation have not been elucidated.

In the present study, we found a novel function of Neptune/biklf (blood island-enriched Krüppel-like factor) in posterior tissue formation. Biklf was first identified as a Krüppel-like factor expressed in the blood islands of the zebrafish embryo (Kawahara and Dawid, 2000). Subsequent studies revealed that biklf was required for erythroid cell differentiation (Kawahara and Dawid, 2001). In Xenopus, Neptune, a factor related to biklf, was isolated and shown to have a function in primitive erythropoiesis (Huber et al., 2001). However, our early studies on the expression and function of Neptune in Xenopus have shown that Neptune is highly expressed in the prospective posterior/ventral area at the gastrula stage and is able to form a tail-like structure if overexpressed in animal pole cells. This activity is very similar to that of brachyury. Therefore, we predicted that Neptune is involved somehow in the process of posterior axis formation. Although the expression of neptune was distinct from the expression of brachyury at the gastrula stage, their expressional domains were very close to each other. Loss-of-function study revealed that Neptune was required for the formation of posterior structures. Findings in the present study suggest that Neptune is a component of the tail organizer that was discovered in the zebrafish embryo (Agathon et al., 2003) and is involved in the formation of posterior structures.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Isolation and Expression Study of neptune in Embryos

We isolated a biklf-related cDNA clone in Xenopus from a tail bud stage cDNA library to elucidate its function in embryogenesis. The amino acid sequence showed 48.5 and 91.6% identity to zebrafish biklf and Neptune, respectively (Kawahara and Dawid, 2000; Huber et al., 2001). Results of analysis of the entire sequence of the cDNA suggested that the clone isolated in this study encodes for a factor that is functionally equivalent to Neptune. Northern blot analysis showed that mRNA was first detected at the mid-blastula stage as a single band of approximately 2.4 kb and was maintained at least until stage 40, although the level of expression was very low (Fig. 1A). A peak of expression level of mRNA was observed at the gastrula stage, and the level of expression decreased at the tail bud stage when its expression was localized in the ventral blood islands (data not shown).

thumbnail image

Figure 1. Temporal and spatial expression patterns of Xenopus neptune. A: Northern blot analysis in Xenopus embryos of various stages. neptune transcript was first detected at the late blastula stage (stage 9) and its expression peaked at the gastrula stage (stage 11). Histon 4 (H4) is shown as an RNA loading control. B–K: Whole-mount in situ hybridization showing the ventral localization of neptune at stage 10.5 (B, F), stage 11 (C, G), stage 12 (D, H), and stage 13 (E, I). The top of each panel indicates the animal pole in B, C, D, F, G, H, and J and indicates the anterior in E and I. Asterisks show the dorsal marginal region of the yolk plug. Note that neptune expression is localized in the prospective ventral and posterior regions at the gastrula stage (B, C, D, F, G, H) and is localized in the posterior ventral region at the end of the gastrula stage (arrow in I). In a dissected stage 11 embryo (J, K), neptune mRNA was highly expressed in outer cells of the animal hemisphere but less expressed in the inner cells, in which brachyury was mainly expressed (data not shown). K is a high-magnification view of the enclosed box in J. Scale bar in B = 500 μm.

Download figure to PowerPoint

Therefore, we verified carefully the spatial expression of neptune in gastrula embryos. Whole-mount in situ hybridization indicated that neptune was expressed in the animal hemisphere of blastula embryos (data not shown) and was localized in the prospective ventral and lateral regions of the animal hemisphere (Fig. 1B–D, F–H). Dissection of stained embryos revealed that neptune was mainly expressed in the prospective ectodermal layer (Fig. 1J,K). At the end of the gastrula stage, the message was restricted to the lateral boundary of the neural and non-neural area (Fig.1E). The expression was also intense in the most anterior and posterior regions. The anterior expression corresponded to the region where the cement gland appeared, and the posterior expression corresponded to the region surrounding the proctodeum (arrow in Fig. 1I). These results suggested a role of Neptune in establishing the pattern of the embryonic axis at this particular stage.

Activity of Neptune in Inducing an Additional Tail Structure

In order to determine the potential role of Neptune in axis formation, we investigated the effects of neptune mRNA injection in embryonic cells. Injection of neptune mRNA into the animal pole region of the embryo frequently resulted in the formation of a tail-like structure (Fig. 2B) (42/55). On the basis of histological study, the extra tail had at least a fin, neural tube, and muscle tissue (Fig. 2I). Whole-mount in situ hybridization and immunostaining showed that sox2 (Fig. 2D) (30/35 positive), N-CAM (Fig. 2F) (14/14 positive), and muscle actin (Fig. 2H) (26/28 positive) were all present in the induced extratail, suggesting that neptune activates a main tract of the tail formation program. A linage-tracing experiment carried out by co-injecting β-galactosidaseRNA (β-gal) revealed that neptune-injected cells were mainly present in the epidermis of the ectopic tail-like structure but were not present in the neural tube or the muscle tissue (Fig. 2J). This suggested that neptune induced the formation of the neural tissue and muscle tissue in a non-cell-autonomous manner.

thumbnail image

Figure 2. Injection of neptune RNA induces the formation of an extra tail-like structure. The animal pole region of 2-cell embryos was injected with 0.3 ng β-gal (A, C, E, G) or 1 ng neptune (B, D, F, H) RNA. The embryos were then cultured to appropriate stages. A, B: Morphological phenotypes of β-gal (A)-injected and neptune (B)-injected embryos at stage 40, showing the appearance of an extra tail (arrowhead in B). C, D: Whole-mount in situ hybridization of stage 23 embryos shows that sox2, an early pan-marker of neural tissues, was expressed in the extra axis (arrowhead in D). E–H: Immunostaining of N-CAM (E, F) and muscle actin (G, H) in β-gal-injected (E, G) and neptune-injected (F, H) embryos at stage 32, showing that N-CAM and muscle actin were expressed in the extra tail (arrowheads in F and H). I, J: A transverse section of a neptune-injected embryo at the position of the ectopic tail (I), showing that the tail contains a fin (f), neural tube (nt), and muscle tissue (m). Another section of a neptune- and β-gal-injected embryo (J), showing that the injected cells are distributed mainly in the epidermis in the additional tail structure. K–N: Morphological phenotypes of untreated stage 32 embryos (K) and stage 32 embryos injected with neptuneRNA (1 ng) (L), neptune (1 ng) + XFD (1 ng) RNAs (M), or neptune (1 ng) + tALK-4 (2 ng) RNAs. The activity of Neptune in inducing the additional tail structure was completely inhibited by XFD injection but was not inhibited by tALK-4 injection. Scale bars in A, C, E, and G = 500 μm; bars in I and K = 100 μm and 1 mm, respectively.

Download figure to PowerPoint

It has been shown that FGF signaling is required for posterior tissue formation in mesodermal and neuronal derivatives (Amaya et al., 1991). Therefore, we carried out an experiment using a dominant-negative FGF receptor 1 (XFD) to determine whether the tail-inducing activity of Neptune requires FGF signaling. As shown in Figure 2K–N, an extratail was induced in 83% (68/82) of the embryos injected with neptune (Fig. 2L), whereas an extratail appeared only in 4% (3/80) of the embryos co-injected with neptune and XFD (either 0.125 ng or 1 ng/embryo) (Fig. 2M). We also tested the effect of tALK-4, which has been shown to work as a dominant-negative receptor of activin (Chang et al., 1997). Injection of tALK-4 RNA did not inhibit the formation of an additional tail structure even though a high dose of RNA (2 ng/embryo) was injected. (A tail structure was induced in 91% (20/22) of the embryos.) There was no significant effect on axis formation in the embryos injected with XFD RNA (0.125 and 1 ng/embryo) alone or tALK-4 RNA (2 ng/embryo) alone (data not shown). These results indicate that the activity of Neptune for tail induction is mediated by FGF signaling.

Mesoderm-Inducing Activity of Neptune in Animal Cap Explants

We performed an animal cap assay to examine the activity of Neptune in undifferentiated embryonic cells. neptune RNA was injected into the animal pole region of 2- or 4-cell embryos, and the animal caps excised from blastula embryos were allowed to develop until stage 35. The neptune-injected caps showed an elongated and swollen structure, and such explants had a well-differentiated epidermis (data not shown). Histological analysis revealed that neptune-injected caps had muscle-like mesodermal tissue (Fig. 3B). In order to determine whether the mesoderm-inducing activity of Neptune occurs cell-autonomously, we combined a cap that had been injected with β-gal alone or with β-gal and neptune RNA with another untreated cap. Histological observation of the cultured caps at stage 37/38 showed that β-gal-positive cells differentiated into muscle-like cells if neptune RNA was coinjected (Fig. 3D). No obvious staining was detected in the epidermal cells, indicating that Neptune-overexpressing cells differentiate into mesodermal cells cell-autonomously. This is in sharp contrast with the tail-inducing activity of Neptune described above (Fig. 2). RT-PCR analysis of explants at stage 26 indicated that Neptune induced the expression of muscle α-actin but not the expression of N-CAM (Fig. 3E). Besides α-actin, expression of other mesodermal markers, such as derriere (Sun et al., 1999a), brachyury (Cunliffe and Smith, 1992), and myoD (Hopwood et al., 1989), was elevated in the Neptune-injected caps (data not shown). Thus, it was shown that Neptune has an activity to induce mesoderm directly from animal cap cells even though Neptune is a DNA-binding protein and a factor related to erythropoietic differentiation (Kawahara and Dawid, 2000; Huber et al., 2001).

thumbnail image

Figure 3. Overexpression of neptune induces mesoderm formation in animal caps. A, B: Embryos were injected with water (A) or 1 ng neptuneRNA (B) into the animal pole region at the 2- or 4-cell stage. Animal caps were excised at stage 8 or 9 and cultured until the sibling embryos had reached stage 35. Histological analysis showed that muscle cells had differentiated in the center of the explants (B). C, D:β–gal-injected (C) or β–gal- and neptune-injected (D) animal caps were combined with uninjected caps at the blastula stage and cultured until stage 35. X-gal staining indicates that the neptune-injected cells had differentiated into mesoderm cells (D). E: RT-PCR analysis revealed that expression of α-actin (a muscle cell marker) was induced in the cap after injection of neptune, without N-CAM induction (a neural marker). As control experiments, activin (500 pg)- or noggin (200 pg)-injected caps were subjected to RT-PCR analysis. EF-1α served as an RNA-loading control. F–I: Morphology of the cultured animal caps at stage 24. neptune (1 ng)-injected caps (G) show a rugged structure (G), but this phenotype was completely reversed by coinjection either with XFD (1 ng) (H) or tALK-4 (2 ng) (I). J: Results of RT-PCR analysis support the morphological phenotype. The expression of α-actin was induced in neptune-injected caps but was suppressed in caps coinjected with XFD (0.125 ng or 1 ng) or tALK-4 (2 ng) RNA. Scale bars in A and F = 100 μm and 1 mm, respectively.

Download figure to PowerPoint

Since the activity of Neptune in inducing the tail-like extrusion depended on FGF signaling, we also tested the ability of XFD to interfere with Neptune activity in the animal cap assay. Animal caps that had been injected with neptune exhibited a swollen structure at the neurula stage (Fig. 3G), while co-injection of XFD or tALK-4 completely blocked the morphological change (Fig. 3H, 3I). The results of a RT-PCR assay supported the morphological observation. The expression of α-actin was induced in caps that had been injected with neptune and was suppressed when XFD or tALK-4 RNA was co-injected with neptune (Fig. 3J). These results demonstrated that the mesoderm-inducing activity of Neptune requires both FGF signaling and activin signaling. The introduction of tALK-4 blocked the activin-induced mesoderm formation but failed to block the bFGF-induced mesoderm formation (data not shown). Thus, the mesoderm-inducing activity of Neptune is mainly mediated by activin/nodal-derived signaling.

Truncation of Posterior Structures by Depletion of Neptune Activity

In order to understand the physiological importance of Neptune in axis formation, we carried out a loss-of-function study using a dominant-negative construct, en-biklf, a fusion protein of the DNA binding domain of zebrafish biklf (over 90% amino acid identity to the corresponding region of Neptune), and the transcriptional repressor motif of Drosophila engrailed (Kawahara and Dawid, 2001). We also made a Xenopus version of the fusion construct (referred to as en-Neptune) as described in the Experimental Procedures section. When en-biklf and neptune were co-injected into the animal pole region of 2-cell embryos, formation of the ectopic tail structure induced by neptune alone (Fig. 4B) (induced in 19/20 embryos) was completely suppressed (Fig. 4C) (induced in 1/18 embryos). Therefore, en-biklf is able to compete with wild-type Neptune not only in globin gene expression (Kawahara and Dawid, 2001) but also in axis formation. Based on a recent fate map (Lane and Smith, 1999), the posterior tissues of Xenopus embryos are derived from the ventral marginal zone at the gastrula stage. Therefore, we injected en-biklf RNA into the ventral marginal zone of 4-cell-stage embryos. Embryos injected with en-biklf (0.5 ng/embryo) showed large truncation of the trunk and tail (Fig. 4E) (truncated in 48/52 embryos). It seemed that the injected embryos failed to develop a posterior structure. Embryos injected with β-gal RNA as a control experiment appeared normal (Fig. 4D) (truncated in 0/51 embryos). The same effect was also observed when a low dose of en-biklf RNA (50–100 pg) was injected (data not shown). Histological studies demonstrated that the embryos injected with en-biklf lacked somites, notochord, neural tube, and mesenchyme at the posterior level (data not shown). Anterior tissues, such as the optic vesicle, brain, and anterior mesenchyme, were unaffected in these embryos. When wild-type neptune RNA was simultaneously injected with en-biklf RNA, the phenotype of posterior truncation was almost completely rescued (Fig. 4F) (truncated in 12/61 embryos). The same result was obtained in the case of injection of en-neptune RNA. As shown in Figure 4G–I, embryos injected with 100 pg en-neptune RNA had a large truncation (Fig. 4H) (truncated in 31/33 embryos), whereas simultaneous injection of neptune RNA restored the effect (Fig. 4I) (truncated in 14/38 embryos). Therefore, we concluded that the phenotype induced by the introduction of en-biklf or en-Neptune in posterior cells occurred specifically by inhibition of the activity of Neptune in either ectodermal or mesodermal cells.

thumbnail image

Figure 4. Endogenous Neptune is necessary for formation of the tail and posterior structures. A–C: Morphological phenotypes of embryos injected with 0.3 ng β–gal RNA (A), embryos injected with 0.5 ng neptuneRNA (B), and embryos co-injected with 0.5 ng neptune RNA and 0.2 ng en-biklf RNA (C), indicating that en-biklf interferes with the wild-type Neptune activity. D–F: Morphological phenotypes of embryos injected with 0.3 ng β–gal RNA (D), embryos injected with 0.1 ng en-biklf RNA (E), and embryos co-injected with 0.1 ng en-biklf RNA and 1 ng neptune RNA (F). G–I: Morphological phenotypes of embryos injected with H2O (G), embryos injected with 0.1 ng en-neptune RNA (H), and embryos co-injected with 0.1 ng en-neptune RNA and 1 ng neptune RNA (I), showing that both zebrafish- and Xenopus-derived fusion constructs work as dominant-negative. J–O: Activity of Neptune is necessary for the expression of brachyury and fgf-8 at the gastrula stage. Expression of brachyury (J, M), fgf-8 (K, N), and vent-2 (L, O) in embryos injected with 0.3 ng β–gal RNA (J, K, L) or embryos injected with 0.3 ng β–gal and 0.1 ng en-neptune RNAs together (M–O) into one lateral blastomere at the 4-cell stage. Embryos were allowed to develop to the gastrula stage (stage 12) and were subjected to whole-mount in situ hybridization analysis. The expression of brachyury, fgf-8 and vent-2 is shown by purple, while the descendants of injected cells are stained by red (Red-gal). Expression of brachyury and fgf-8 disappeared in the area where en-neptune RNA was injected (M, N), but expression of vent-2 was not affected by the injection of en-neptune RNA (O). Scale bars in A and J = 1 and 0.5 mm, respectively.

Download figure to PowerPoint

Expression of Brachyury and Fgf-8 at the Gastrula Stage Depends on the Existence of Neptune

It has been shown previously that Xenopus brachyury has an activity to induce a tail-like structure if its RNA is injected into animal pole cells (Cunliffe and Smith, 1992; Tada et al., 1997). Since the structures induced by brachyury and Neptune are very similar to each other, the relationship between brachyury expressikon and neptune expression was investigated. Double in situ hybridization of brachyury and neptune showed that expression domains of these two markers apparently overlapped at the prospective ventral/posterior side of the marginal zone (data not shown). In contrast, there was a clear gap between expression domains of brachyury and neptune in the prospective dorsal/anterior side. To elucidate the relationship between Neptune activity and expression of the posterior mesoderm markers at the gastrula stage, en-neptune RNA was injected into one of blastomeres at the 4-cell stage and the expression of brachyury and fgf-8 was examined in the gastrulating embryos. At the same time, β-gal RNA was injected for tracing the descendants of injected cells. Whole-mount in situ hybridization analysis revealed that brachyury and fgf-8 expression at the injected side was suppressed (Fig. 4M,N; brachyury expression suppressed in 10/10 embryos and fgf-8 expression suppressed in 11/11 embryos), whereas vent-2 expression was not affected (Fig. 4O; vent-2 expresssion suppressed in 0/9 embryos). Embryos injected with β-gal alone showed a normal expression pattern of brachyury and fgf-8 (Fig. 4J,K; brachyury expression suppressed in 0/9 embryos and fgf-8 expression suppressed in 0/8 embryos). Thus, the expression of brachyury and fgf-8 is dependent on the presence of Neptune at the gastrula stage.

Cooperation of Dorsal and Ventral Tissues Is Important for Formation of the Tail Structure

Wild-type Neptune has an activity to induce an extra tail structure, but our experiments showed that the tail structure was only induced when its RNA was injected into the animal pole area. When the RNA was injected into the dorsal marginal zone, severe damage was observed in the head structures (data not shown). No significant effect was observed in the embryos injected with neptune RNA into the ventral marginal zone. Since a tail-like structure was not induced in animal cap explants (Fig. 3B,G), we hypothesized that Neptune cooperates with factors derived from the dorsal tissues. In order to examine this possibility, animal pole cells were injected with 1 ng neptune RNA and this area was excised together with the dorsal marginal zone (DMZ) or with the ventral marginal zone (VMZ) and further cultured. At stage 33/34, as shown in Figure 5, tail-like extrusions were observed in the explants with the DMZ (31/41) (arrows in Fig. 5B) but not in the explants with the VMZ (0/9) (Fig. 5H). These extrusions were positive in staining of N-CAM and muscle actin with antibodies (Fig. 5C–F), indicating that the structure observed in the DMZ explants is essentially the same as the structure of the extratail induced in a whole body (Fig. 2).

thumbnail image

Figure 5. Tail-inducing activity of Neptune requires the existence of dorsal tissues. neptune RNA (1 ng/embryo) was injected into the animal pole area and embryos at the early gastrula stage were dissected for culture of the animal pole area with the dorsal marginal zone (AP+DMZ) (A, B) or the animal pole area with the ventral marginal zone (AP+VM) (G, H). At stage 33/34, tail-like extrusions (arrows) were observed in explants with the DMZ (B) but not in explants with the VMZ (H). Immunostaining of N-CAM (C, D) and muscle actin (E, F) in control (C, E) and neptune-injected (D, F) explants demonstrates the expression of neural tissue and muscle tissue markers in the extrusions (arrows in D and F). Bars in A, C, E and G = 0.5 mm.

Download figure to PowerPoint

In order to determine how the prospective dorsal and ventral tissues of the gastrula embryo contribute to the formation of the tail structure, we made a combination explant between the DMZ and VMZ tissues. Such a combination explant was previously made to demonstrate the existence of a dorsalizing factor that originates from Spemann organizer and has an activity to convert the ventral tissues to dorsal tissues (Dale and Slack, 1987). In the present study, we excised a large piece of tissue including the marginal zone and 50% of the animal pole tissue together (Fig. 6D), and the two pieces of tissues were combined. While the control explants made from DMZ-DMZ and VMZ-VMZ showed large anterior and ventral structures (Fig. 6A,C), the DMZ-VMZ explants showed an elongated mini-embryo-like structure (Fig. 6B,D). These embryos expressed brachyury, which is known as a tail bud marker (Beck and Slack, 1999), at the end of tail-like extrusion (arrows in Fig. 6B). In order to assess the contribution of DMZ- and VMZ-derived cells to formation of the tail structure, we made an interspecific chimera between Xenopus laevis and borealis tissues (Fig. 6D–F). In 7 chimera explants made from laevis DMZ and borealis VMZ, a total of 6,933 cells in the notochord, neural tissue, and muscle tissue were examined to determine whether the cells were derived from the laevis (DMZ) or borealis (VMZ). As shown in Figure 7G, notochord and neural cells were largely of DMZ origin (0.6 and 24.4% from the VMZ, respectively), whereas muscle cells were mainly derived from the VMZ (80.6% from the VMZ). These findings strongly suggest that cooperation of dorsal tissues such as notochord and neural cells with the prospective ventral tissues is important for tail bud formation and further extension of the tail structure.

thumbnail image

Figure 6. Cooperation of dorsal and ventral cells converts the fate of ventral cells to posterior cells. A–C: Morphological phenotypes and brachyury (bra) expression of combination explants of dorsal-dorsal (DMZ-DMZ), dorsal-ventral (DMZ-VMZ), and ventral-ventral (VMZ-VMZ) marginal zone tissues excised from embryos at the early gastrula stage. At stage 33/34, tail-like extrusion that expressed bra was formed only in the explants combined with VMZ and DMZ tissues (B). D–G: Lineage-tracing experiment in laevis-borealis combination explants. As described in the Experimental Procedures section, borealis-derived cells with brightly spotted nuclei (yellow arrows in F) and laevis-derived cells with dull-stained nuclei were counted in randomly selected cross sections after staining with quinacrine (E, F). Muscle cells were mainly derived from the ventral marginal zone, but notochord and neural cells were mostly derived from the dorsal marginal zone (G). Totally 6,933 cells from 7 explants were analyzed. H, I: Inactivation of Neptune activity caused deformation of the tail-like extrusion. DMZ tissues were combined with untreated VMZ tissues (H) or en-neptune-injected VMZ tissues at the early gastrula stage and cultured until stage 33/34. The mean values of body length were measured in two groups of the experiment as described in the text. Scale bars in A and H = 0.5 mm; scale bar in E = 50μm.

Download figure to PowerPoint

thumbnail image

Figure 7. A model of fate determination of ventral marginal zone cells at the gastrula stage to ventral and posterior tissues. The ventral marginal zone of a Xenopus gastrula embryo comprises common precursor cells for ventral and posterior tissues (as indicated by pink). These cells express neptune, which makes the cells competent to induce signals from dorsal cells (purple). Posterior tissues (blue) such as muscle and fin are induced by this signal. Without induction, these cells differentiate to ventral tissues (red).

Download figure to PowerPoint

Finally, we examined the effect of en-Neptune on the formation of tail-like extrusion in the DMZ-VMZ combination. DMZ tissue was excised together and combined with VMZ tissue that had received injection of 0.1 ng en-neptune RNA into the ventral marginal zone. As shown in Figure 6H and I, en-neptune-injected DMZ-VMZ explants exhibited a truncation in posterior structure. The mean body length of the explants without en-neptune injection was 2.54 ± 0.34 mm (n = 11) and that of the explants with en-neptune injection was 1.61 ± 0.20 mm (n = 11). Taken together, the results suggest that Neptune expressed in the prospective ventral ectoderm has an important role in the interaction of a group of cells located in the ventral area and those in the dorsal area to form the posterior structure.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Activity of Neptune for Formation of Posterior Structures

Previous expressional and functional studies on Neptune, a novel biklf-related KLF (Krüppel-like factor), demonstrated that this factor is necessary for primitive erythropoiesis (Huber et al., 2001). The present study, however, focused on the highest expression level of neptune at the gastrula stage, when blood islands have not been formed, and showed that Neptune is involved in the process of axis formation, particularly in the formation of posterior structures. At the early gastrula stage, neptune was expressed in the animal pole region. As the gastrulation proceeded, neptune expression in the prospective dorsal side disappeared and it remained in the prospective ventral/posterior side (Fig. 1B–D). This expression pattern is similar to the spatial expression pattern of BMP-4-driven target genes, such as vent-1, vent-2, msx-1, and msx-2 (Gawantka et al., 1995; Onichtochouk et al., 1996; Maeda et al., 1997; Suzuki et al., 1997; Su et al., 1991). In spite of the similar expression patterns, Neptune and other downstream factors have quite different activities when overexpressed in embryonic cells. As found in previous studies, vent-1/2 and msx-1/2 possess a ventralizing activity that suppresses the expression of organizer genes and disturbs the formation of head structures (Gawantka et al., 1995; Maeda et al., 1997). In contrast, injection of neptune mRNA into embryos resulted in the formation of a tail-like structure containing mesodermal tissues (Fig. 2). Since recent studies in Xenopus and zebrafish have demonstrated that the BMP signal is essential for outgrowth of the tail bud (Beck et al., 2001; Agathon et al., 2003), Neptune may be a component related to a tail formation program that is activated by the BMP signaling pathway.

Since previous studies showed that FGF-related factors such as brachyury and derriere, have similar posterior-inducing activity (Cunliffe and Smith, 1992; Conlon et al., 1996; Sun et al., 1999a), we further investigated the relationship of these factors. Double staining of neptune and brachyury mRNA in gastrulating embryos indicated that these two transcripts are expressed closely to each other (data not shown). However, careful observation in cut embryos showed that neptune was mainly expressed in the prospective ectodermal layer (Fig. 1), while brachyury was mainly expressed in the prospective mesoderm (data not shown). When en-biklf (a dominant negative construct) was introduced in embryonic cells, the expression of brachyury was completely suppressed at the injected site (Fig. 4). Since there was a gap between the injected site and the border of the brachyury expression domain, it was thought that these two factors interact with each other in a non-cell-autonomous manner. Although the initial expression of neptune and that of brachyury were regulated independently, these two transcription factors probably cooperate in the formation of posterior structures.

Interaction Between Ventral and Dorsal Tissues Is Important for Tail Formation

As shown in Figure 2, Neptune can activate the tail-forming program if overexpressed in the animal pole area. However, it was not able to form a tail-like structure in animal cap explants (Fig. 3). Thus, we predicted the existence of factors that are necessary for tail formation in cells other than those in the animal pole area. Further experiments showed that the dorsal marginal zone (DMZ), but not the ventral marginal zone (VMZ), compensates the activity to form a tail-like structure (Fig. 5). These results suggest that Neptune and factors derived from the DMZ cooperate to initiate the tail-forming program. In order to show the importance of interaction between the dorsal and ventral tissues for formation of the tail structure in vivo, we made a DMZ-VMZ combination explant that forms a tail-like extension after culture. In this explant, cells of the notochord and neural tube at the level of the tail were mainly derived from DMZ cells of the gastrula embryo, and cells of muscle tissue were mainly derived from VMZ cells. This indicates that the DMZ-derived cells dorsalize the VMZ-derived cells to form posterior mesoderm tissues such as muscle but, at the same time, the VMZ-derived cells posteriorize the DMZ-derived cells to form tail extrusion. We propose a function of Neptune as a posteriorizing agent localized in VMZ cells. As summarized in Figure 7, Neptune and factors derived from organizer cells cooperate together to make the posterior ectoderm (neural tube at the tail level and fin) and the mesoderm (muscle and other posterior mesoderm tissues). In this respect, the present findings agree with a study that shows the existence of a tail organizer in zebrafish (Agathon et al., 2003), and Neptune could be a component of the tail organizer in the frog embryo.

Involvement of Neptune in Posterior Formation Pathway

As described above, Neptune has an activity similar to that of FGF-related factors. Thus, attempts have been made to determine whether Neptune regulates and activates the FGF signal. Since the activity of Neptune in inducing a posterior structure was completely blocked by injection of XFD RNA but not by tALK-4 RNA (Fig. 2), we conclude that Neptune mediates the FGF signaling pathway in the formation of posterior structures. However, we found that the mesoderm-inducing activity of neptune could be inhibited by injection of either XFD or tALK-4 RNA (Fig. 3). This suggests that the mesoderm-inducing activity of Neptune is mediated by activin/nodal signaling. In fact, the phenotype of cultured animal caps injected with Neptune was apparently different from that of caps treated with bFGF (data not shown). Also, lineage-tracing experiments using β-gal showed that neptune-injected cells differentiated into the epidermis of the secondary tail, whereas they differentiated into muscle cells in animal cap explants (Figs. 2J and 3D). Thus, we conclude that the activities of Neptune in inducing the tail and the mesoderm are different in molecular nature. Although the details of molecular basis have not been elucidated yet, it seems likely that Neptune activates multiple pathways in the embryonic cells. Neptune perhaps activates both FGF signaling and nodal/activin signaling in the ectodermal cells but the activity of nodal/activin signaling may be suppressed by the action raised by the adjacent mesodermal cells. For further understanding, analyses should be done to determine whether the intracellular components for signal transduction, such as MAP kinase and/or Smad-2, are phosphorylated after Neptune treatment. Also, elucidation of the direct target genes activated by Neptune in the embryonic cells will help us to dissect the complex activity of Neptune.

Many studies on FGF signaling and posterior axis formation have been carried out. All of the components responsible for posterior axis formation are mesoderm-derived factors. Members of the FGF family are thought to be important in trunk and tail formation (Amaya et al., 1991). Brachyury is also required for formation of the posterior mesoderm (Conlon et al., 1996). Brachyury and eFGF in the prospective mesoderm are activated by each other and these factors maintain a positive-feedback loop (Issacs et al., 1994; Schulte-Merker and Smith, 1995). In addition to these two essential factors, derriere and Gli2 are also shown to have posteriorizing activity (Sun et al., 1999b; Brewster et al., 2000). Over-expression of derriere or Gli2 results in the formation of a tail-like structure, and inhibition of the activity of these factors leads to posterior truncation. Derriere and Gli2 also form a regulatory loop between FGF. Furthermore, the activity of Gli2 for formation of posterior structure is mediated by Xhox3 (Brewster et al., 2000), a homeodomain protein expressed in the posterior tissues at a much later stage (Ruiz i Altaba and Melton, 1989a,b; Ruiz i Altaba et al., 1991; Beck and Slack, 1999). How can Neptune, which is mainly expressed in presumptive ectodermal cells at the gastrula stage, be involved in the posteriorizing signal?

Possible Interaction of Ectoderm and Mesoderm Cells in the Formation of Posterior Structures

Many studies have been carried out to elucidate the tissue interaction in tail bud formation (Tucker and Slack, 1995; Beck and Slack, 1998, 1999). These studies have shown that interaction between the neural ectoderm and the mesoderm is important for development of the tail. The present study suggests that Neptune is mainly localized in the ectodermal side and that it affects the fate of mesoderm derivatives in a non-cell-autonomous manner. Therefore, in the future, it will be important to identify the downstream gene of Neptune that is expressed in the ectoderm and encodes for a secretory factor.

The dominant-negative strategy in the present study has a potential problem that the conserved DNA-binding motif in KLF may bind other target sequences and inactivate such genes non-specifically. For an alternative strategy to avoid this problem, we made two kinds of Morpholino to knock down the expression of Neptune in embryonic cells. The embryos injected with one Morpholino showed a severe tail truncation defect as was observed in embryos injected with en-neptune RNA. We also observed a severe abnormality in the neural tissue and the neural crest formation when en-neptune mRNA or Morpholino was injected into the prospective neural ectoderm (data not shown). This kind of effort will lead to the elucidation of the mechanism of interaction between ectoderm and mesoderm derivatives in tail formation and neural tissue specification.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Isolation of a Xenopus neptune Variant

A Xenopus tail bud (stage 34) cDNA library was screened with a DNA fragment containing the Krüppel-like zinc finger region (aa 270–409) of zebrafish blood island-enriched Krüppel-like factor cDNA (Kawahara and Dawid, 2000, 2001). Positive plaques from 2 x 105 clones were selected and converted to plasmids to synthesize anti-sense RNA probes. Whole-mount in situ hybridization analysis described below was performed to further select clones showing an interesting expression pattern. One clone that exhibited specific expression in ventral blood islands of the tail bud embryo was further characterized. The entire cDNA sequence was determined by the dideoxy-termination method, and the obtained sequence was analyzed by comparison with the GenBank/EMBL/DDBJ database using the DDBJ-FASTA program (accession no. AB067693). On the basis of sequence information, it was assumed that the cDNA clone is derived from a variant gene of Xenopus neptune (AF353715).

Manipulations of the Embryo and Microinjection

Xenopus laevis embryos were obtained by artificial fertilization. Females were injected with 250 U of human chorionic gonadotropin and kept at 23°C overnight to induce ovulation. Fertilization was performed by addition of a sperm suspension prepared by mincing testis in 100% MMR (Marc's modified Ringer) solution. After 15 min, eggs were dejellied in 2.5% thioglycolic acid (pH 8.3) and washed several times in 100% Steinberg's solution. Developmental stages were determined according to Nieuwkoop and Faber (1967). For RNA injection, capped mRNAs were synthesized using a MEGAscript kit (Ambion). Xenopuswild-type neptune in pCS2+ was linearized with Not I and RNA was transcribed with Sp6 RNA polymerase. en-neptune/biklf (referred to as en-biklf in this study) derived from zebrafish cDNA was used as a dominant-negative construct (a gift from Dr. A. Kawahara; Kawahara and Dawid, 2001). en-neptune was made by fusing a DNA fragment corresponding to the 208–452–amino acid sequence of Xenopus Neptune and a DNA fragment corresponding to the 1–298–amino acid sequence of Drosophilla engrailed protein. These plasmids were linearized with NotI and RNA was transcribed with Sp6 RNA polymerase. RNA injection was performed according to the method described previously (Takeda et al., 2000). Injected embryos were kept in 50% Steinberg's solution until stages 32–34 for subsequent analyses. For animal cap assays, two blastomeres of 2-cell embryos were injected with a designated amount of mRNA. The animal pole region was dissected at stages 8–9 and further cultured in 100% Steinberg's solution containing antibiotics until appropriate stages.

Northern Blot Analysis

Total RNA from whole embryos or explants was extracted by the AGPC method (Chomczynski and Sacchi, 1987) and then loaded in denatured 1% agarose gel and transferred to Biodyne B nylon membranes (Pall) as described previously (Kikkawa et al., 2001). Probes for Neptune and histon H4 were made from a 1.8-kb EcoRI fragment and a PCR-amplified 0.2-kb fragment by using a random priming kit (Takara). Hybridization was performed in a solution (1 M NaCl, 50 mM Tris-HCl pH 7.4, 40% formamide, 10% dextran sulfate, and 1% SDS) at 42°C overnight. Membranes were washed three times with 0.2% SDS and 0.2x SSC at 50°C for 10 min each time and were exposed to Fuji RX-U film.

Whole-Mount In Situ Hybridization and Cell-Lineage Trace

Whole-mount in situ hybridization was performed as described previously (Shain and Züber, 1996). Embryos were fixed in MEMFA (0.1 M MOPS, 2 mM EDTA, 1 mM MgSO4, 3.7% formaldehyde) for 2 hr at room temperature and stored in 100% methanol at −20°C. Digoxygenin- or fluorescein-labeled anti-sense RNA probes were synthesized as follows: neptune in pBluscript SK+, linearized by Not I and transcribed with T7 RNA polymerase; brachyury in pSP73, linearized by EcoRV and transcribed with T7; α-actin in pSP64, linearized by EcoR I and transcribed with Sp6 polymerase; vent-2 in pBluescript (KS), linearized by Sal I and transcribed with T7 RNA polymerase. The positive signal was visualized using BM purple as a substrate. For the lineage-tracing experiment, β-galactosidase RNA (0.3 ng/embryo) was injected together with en-biklf RNA and descendant cells were visualized by Red-Gal (Research Organics).

borealis-laevis Interspecific Chimera Assay

For the lineage-tracing experiment using laevis-borealis embryos, the dorsal marginal zone (DMZ) tissue from a laevis embryo was combined with the ventral marginal zone (VMZ) tissue from a borealis embryo and the chimeric explants were cultured until stage 33/34. Explants were cut into 7-μm-thick sections and stained with quinacrine as described previously (Maeno et al., 1985). borealis-derived cells with brightly spotted nuclei and laevis-derived cells with dull-stained nuclei were distinguishable (Fig. 7F). Approximately 1,000 cells in each explant and totally 6,933 cells from 7 explants were counted. The mean composition of borealis cells in the notochord, neural tube, and muscle tissues is shown (Fig. 7G).

Whole-Mount Antibody Staining

Whole-mount immunostaining was performed as described previously (Maeda et al., 1997). Monoclonal antibodies against N-CAM (4d, from Developmental Studies Hybridoma Bank) and skeletal actin (anti-human muscle actin, HHF35, Dako) were utilized. Embryos were fixed in Dent solution (20% dimethyl sulfoxide in methanol) and incubated with diluted first antibodies as described above and then with alkaline phosphatase-conjugated secondary antibody. Staining was visualized with NBT/BCIP (Roche).

RT-PCR Analysis

Complementary DNA was synthesized from 500 ng of total RNA extracted from explants or whole embryos at appropriate stages using a Superscript first-strand synthesis system for RT-PCR (Gibco/BRL), and a 1/10 volume of cDNA was subsequently used for PCR reaction. One cycle of the PCR program consisted of 94°C for 1 min, 55°C for 1.5 min, and 72°C for 1 min, and the products were run on 5% polyacrylamide gel and stained by ethidium bromide. The primer sequences and the cycle numbers for genes investigated in this study are as follows: α-actin, 5'-GCT-GAC-AGA-ATG-CAG-AAG-3' and 5'-TTG-CTT-GGA-GGA-GTG-TGT-3', 30 cycles; N-CAM, 5'-CAC-AGT-TCC-ACC-AAA-TGC-3' and 5'-GGA-ATC-AAG-CGG-TAC-AGA-3'; EF-1α, 5'-CCT-GAA-AGG-CCA-TCA-CCC-GAT-TGG-TG-3' and 5'-GAG-GGT-AGT-CTG-AGA-AGC-TCT-CCA-CG-3', 22 cycles.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We thank Drs. A. Kawahara, M. Taira, Y. Sasai and M. Tada for supplying plasmids.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES