SEARCH

SEARCH BY CITATION

Keywords:

  • primary motoneurons;
  • Rohon-Beard neurons;
  • Xbra;
  • sox2;
  • n-tubulin;
  • chordin;
  • BMP4

Abstract

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

In Xenopus, localized factors begin to regionalize embryonic fates prior to the inductive interactions that occur during gastrulation. We previously reported that an animal-to-vegetal signal that occurs prior to gastrulation promotes primary spinal neuron fate in vegetal equatorial (C-tier) blastomere lineages. Herein we demonstrate that maternal mRNA encoding noggin is enriched in animal tiers and at low concentrations in the C-tier, suggesting that the neural fates of C-tier blastomeres may be responsive to early signaling from their neighboring cells. In support of this hypothesis, experimental alteration of the levels of Noggin from animal equatorial (B-tier) or BMP4 from vegetal (D-tier) blastomeres significantly affects the numbers of primary spinal neurons derived from their neighboring C-tier blastomeres. These effects are duplicated in blastomere explants isolated at cleavage stages and cultured in the absence of gastrulation interactions. Co-culture with animal blastomeres enhanced the expression of zygotic neural markers in C-tier blastomere explants, whereas co-culture with vegetal blastomeres repressed them. The expression of these markers in C-tier explants was promoted when Noggin was transiently added to the culture during cleavage/morula stages, and repressed with the transient addition of BMP4. Reduction of Noggin translation in B-tier blastomeres by antisense morpholino oligonucleotides significantly reduced the efficacy of neural marker induction in C-tier explants. These experiments indicate that early anti-BMP signaling from the animal hemisphere recruits vegetal equatorial cells into the neural precursor pool prior to interactions that occur during gastrulation. Developmental Dynamics 236:171–183, 2007. © 2006 Wiley-Liss, Inc.


INTRODUCTION

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

A combination of cell–cell signaling and cell type–specific transcriptional programs play important roles in neural fate specification at several steps during development. For example, during gastrulation the Organizer expresses both transcription factors that down-regulate epidermal genes and signaling factors that cause the adjacent ectoderm to express a neural fate (reviewed in Wilson and Edlund, 2001; Moody and Je, 2002; De Robertis and Kuroda, 2004). There have been significant advances in delineating the genes that regulate neuronal fate choices after the Organizer-induced commitment to a neural fate (reviewed in Jessell, 2000; Livesey and Cepko, 2001; Bertrand et al., 2002; Guillemot, 2005). However, very little is known about the mechanisms that repress or promote a neural fate prior to the overt formation of the neural plate. In Xenopus, localized factors (e.g., VegT, Vg1, Ecto) set up signaling cascades that specify the progenitors for germ layers and the dorsal axis (Kessler, 1999; Xanthos et al., 2001; Dupont et al., 2005; Moon, 2005). Although fate maps and several blastomere transplantation, deletion, and culture experiments suggested that the neural fate of cleavage stage cells is biased by localized factors prior to gastrulation when classical neural induction occurs (reviewed in Sullivan et al., 1999; Pandur et al., 2002), the responsible molecules were not well defined. Recently, a signaling center (BCNE center) that predisposes dorsal-animal lineages to an anterior neural fate was shown to arise at the early mid-blastula stages and to involve Chordin signaling (Kuroda et al., 2004).

In fish and frog, primary motoneurons (PMN) and primary sensory neurons called Rohon-Beard neurons (RBN) are among the earliest CNS cells to be born, differentiate, and function (Hughes, 1957, 1959; Lamborghini, 1980; Jacobson and Moody, 1984; Roberts et al., 1981; Roberts and Clarke, 1982). Some of the signaling factors and transcriptional programs regulating the differentiation of these two populations of spinal neurons have been identified (reviewed in Lewis and Eisen, 2003). For example, interactions between Delta and Notch expressing cells in the neural plate regulate the numbers of primary neuron precursors (Chitnis, 1999; Appel et al., 2001; Chalmers et al., 2002). Additionally, during neural tube formation an interaction between hedgehog and BMP signaling is necessary for PMN development (Liem et al., 2000; Lewis and Eisen, 2001) and RBN require BMP signaling (Nguyen et al., 2000). Differential expression of several transcription factors in the neural tube also leads to motoneuron versus sensory neuron fate choices and the establishment of functional subtypes (Bisgrove et al., 1997; Anderson, 1999; Jessell, 2000). However, the very early specification of these two types of spinal neurons during gastrulation stages (Lamborghini, 1980; Jacobson and Moody, 1984) also suggests that they may be influenced by pre-neural plate signaling factors.

As a prelude to testing whether there are pre-gastrulation interactions that impact primary spinal neuronal fate choices, we quantified the numbers of PMN and RBN that descend from each cleavage stage Xenopus blastomere (Moody, 1989). These maps demonstrated that the vegetal equatorial blastomeres (C-tier) constitute a boundary between those embryonic cells that largely contribute to primary neurons (A-tier and B-tier) and those that do not (D-tier; Fig. 1). In order to test whether the numbers of C-tier-derived primary neurons are regulated by interactions with neighboring blastomeres, B-tier neighbors were deleted or transiently separated in situ from C-tier cells during cleavage/morula stages (Bauer et al., 1996). Transient interruption of cell–cell interactions well before the onset of gastrulation movements was sufficient to dramatically reduce the numbers of primary spinal neurons produced by C-tier blastomeres, implicating the involvement of an early signal from the animal hemisphere. Interestingly, Chordin signaling from the BCNE center appears to have little effect on the formation of posterior (i.e., spinal) neural markers (Kuroda et al., 2004). Because Noggin, another anti-BMP factor expressed in the BCNE center, also is expressed as a maternal mRNA in Xenopus animal blastomeres (Sive, 1993; Pandur et al., 2002), we tested whether it might be responsible for the B-tier to C-tier interaction that promotes C-tier primary neuronal fate. We demonstrate that maternal mRNA encoding noggin is at a low concentration in vegetal tiers, suggesting a potential endogenous gradient of neural inducing proteins could modulate C-tier neural fates. We show in embryos that altering Noggin levels in neighboring tiers, by both gain and loss of function, significantly affects C-tier neural fate. These effects are replicated in blastomere explants isolated at cleavage stages and cultured in the absence of gastrulation interactions. Co-culture with animal blastomeres enhanced the expression of zygotic neural markers in C-tier blastomere explants, whereas co-culture with vegetal blastomeres repressed them. The expression of these markers in C-tier explants was promoted when Noggin was transiently added to the culture during cleavage/morula stages, and repressed with the transient addition of BMP4. These experiments demonstrate that local levels of Noggin and BMP4 affect the size of the neural field derived from the equatorial cells and that Noggin is a functionally important constituent of pre-gastrulation signaling for the production of primary spinal neurons.

thumbnail image

Figure 1. Nomenclature of the blastomeres of the 32-cell Xenopus embryo used in these studies (left side). Blastomeres that produce large numbers of primary neurons are yellow (A- and B-tiers), those that contribute none are blue (D-tier), and those that form the border between these two domains are green (C-tier). The different types of explants that were made are depicted on the right side. Explants were composed of the entire dorsal-to-ventral tier(s) from one side of the embryo (4–8 cells). an, animal pole; veg, vegetal pole, dorsal is to the right. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

RESULTS

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

Altered Noggin Expression in Neighboring Blastomeres Changes the Neural Fate of C-Tier Blastomeres

Transcripts for several signaling proteins and receptors are found in Xenopus cleavage stage embryos, indicating that they may influence cell fate decisions prior to interactions occurring during gastrulation (reviewed in Sullivan et al., 1999). Previous studies reported that noggin transcripts are present in animal blastomeres (Sive 1993; Pandur et al., 2002), suggesting Noggin as a candidate for the previously described animal-to-vegetal signal that promotes a primary neuronal fate in C-tier lineages (Bauer et al., 1996). To characterize this further, we dissected 32-cell embryos into A, B, C, and D-tiers (Fig. 1), isolated mRNA and semi-quantified by PCR the relative levels of maternal noggin and bmp4 mRNAs in each tier. Because cleavage furrows vary between embryos and our samples contained material pooled from about 25 embryos, we verified the accuracy of our dissections by showing that a vegetally localized mRNA (Vg1; Weeks and Melton, 1987) was primarily detected in the vegetal (C-tier and D-tier) samples and an animal-enriched mRNA (Wnt8b; Cui et al., 1995) was excluded from the vegetal-most (D-tier) samples (Fig. 2A). On average, A-tier and B-tier blastomeres contained significantly more noggin mRNA than C-tier or D-tier blastomeres (Fig. 2B) and all tiers contain about the same amounts of bmp4 mRNA (Fig. 2B). These results are consistent with previous in situ hybridization studies performed at blastula and early gastrula stages that show that noggin mRNA is concentrated in the B-tier descendants, with some detected in C- and D-tier descendants (Vodicka and Gerhart, 1995; Kuroda et al., 2004). Together, these data suggest that C-tier blastomeres and their progeny may be subject to Noggin signals from their animal blastomere neighbors, which could modulate their contributions to neural progeny.

thumbnail image

Figure 2. Maternal noggin mRNA is differentially distributed across the 32-cell blastomere tiers. A: A representative PCR experiment demonstrating that mRNA for noggin is abundant in the A-tier and B-tier and present at a low concentration in the C-tier and D-tier. bmp4 mRNA is uniformly distributed (see B). To confirm that tiers were dissected correctly, samples were also processed for detection of an animal-enriched (Wnt8b) and a vegetal-enriched (Vg1) mRNA. B: The density of the bands was quantified and normalized to H4 expression. The results from 4 separate experiments were averaged and expressed relative to whole embryo (WE) expression, which was set at 100%. -, minus reverse transcription. Bars indicate SEM.

Download figure to PowerPoint

If animal-tier Noggin promotes C-tier primary spinal neuron fate, then increased levels of Noggin in animal blastomeres should increase the numbers of primary spinal neurons produced in C-tier lineages, whereas decreased levels should reduce those numbers. To test this, we focused on the two C-tier blastomeres (C2, C3; Fig. 1) that produce the largest numbers of primary spinal neurons (Moody, 1989). First, noggin transcripts were injected into the neighboring B-tier blastomeres (B2, B3; Fig. 1), and the numbers of PMN and RBN produced by their C-tier neighbors were counted. PMN numbers were significantly increased in both the C2 and C3 lineages (Fig. 3A). There were small increases in the number of RBN produced in both lineages that did not reach a significant level (Fig. 3A). To antagonize a potential animal-tier-derived Noggin signal, expression of BMP4, which binds Noggin at high affinity (Zimmerman et al., 1996), was increased by mRNA injection in neighboring blastomeres. When this was done in B-tier neighbors, which are major progenitors of the neural plate (Moody, 1987), neural plate formation (83%, n = 48, sox2 or n-tubulin expression; Fig. 4F) and midline mesoderm (41%, n = 27, chd expression) were so significantly repressed that the resulting patterning defects made it difficult to analyze differentiated primary neurons. Therefore, we injected bmp4 mRNA into a non-neural region that neighbors the C-tier cells, i.e., the D-tier blastomeres (D2, D3; Fig. 1) so that the protein could diffuse into the C-tier environment with minimal direct effects on neural plate formation. In these cases, PMN and RBN numbers were significantly reduced in both C-tier lineages (Fig. 3B). Together, these experiments indicate that the local levels of extracellular Noggin in the equatorial region regulate the numbers of primary spinal neurons produced by the C-tier lineages.

thumbnail image

Figure 3. The numbers of primary neurons produced by C-tier blastomeres are affected by altering levels of BMP signaling in neighboring blastomeres. A: Overexpression of Noggin in B-tier neighbors by mRNA injection significantly increases the numbers of PMN produced by C-tier lineages, but does not significantly alter RBN numbers. B: Overexpression of BMP by mRNA expression in D-tier neighbors significantly reduces the numbers of both PMN and RBN produced by C-tier lineages. Asterisks indicate significant difference from control numbers at the P < 0.05 level.

Download figure to PowerPoint

thumbnail image

Figure 4. Altering Noggin levels in single B-tier or D-tier blastomeres alters the size of the neural plate progenitor pool, indicated by sox2 expression (A–F), and the numbers of nascent primary neurons, indicated by n-tubulin expression (G–T). A: A βgal mRNA-injected (B-tier) embryo demonstrating that the width of the neural plate (purple) on the injected side (white line) is indistinguishable from the width of the neural plate on the uninjected control side (black line). B: A noggin mRNA-injected (B-tier) embryo has an expanded neural plate on the injected side (white line). C: A bmp4 mRNA-injected (D-tier) embryo has a reduced neural plate on the injected side (white line). D: Injection of the control MO (B-tier) on one side does not alter the width of the neural plate (white line). E: Injection of Noggin MO (B-tier) reduces the width of the neural plate (white line) on the side of injection. F: The entire neural plate is grossly reduced when bmp4 mRNA is injected into a B-tier blastomere. G: Control embryo demonstrating the three stripes of nascent primary neurons on each side of the neural plate. Asterisk indicates side injected with control (β-gal) mRNA. H: A noggin mRNA-injected (B-tier) embryo has expanded numbers of cells in the PMN (large arrow) and RBN (small arrow) stripes. I: D-tier expression of BMP4 represses the size of the RBN (small arrow) and interneuron (large arrow) progenitor stripes on the injected side. The PMN stripe is not visible in this neural groove stage embryo. J: Lineage labeling of an A-tier blastomere (red cells) after D-tier bmp4 mRNA injection demonstrates that many A-tier cells continue to express n-tubulin (arrows). J': βgal mRNA-injected A-tier control. In contrast, many fewer B-tier-derived cells (arrows in K) and no C-tier-derived cells (red in L) express n-tubulin following bmp4 mRNA injection of the D-tier neighbor. K': Many βGal-labeled cells (between arrows) in B-tier control; L': Several n-tubulin expressing C-tier cells (arrows) in βgal mRNA-injected control. M: Injection of a control MO on one side (asterisk) does not alter the three primary neuron progenitor stripes in the neural plate. N: Injection of Noggin MO into a dorsal B-tier blastomere reduces the PMN stripe (large arrow) on the side of the injection. O: Injection of Noggin MO into a ventral B-tier blastomere reduces the RBN stripe (small arrow) on the side of the injection. P,Q: Lineage labeling shows that after nMO injection of a B-tier blastomere, many B-tier-derived cells (between arrows) continue to express n-tubulin (P); in contrast, very few C-tier-derived cells (arrows) express n-tubulin after nMO injection of the B-tier neighbor (Q). RT: Lineage labeling shows that after nMO injection of a C-tier blastomere, many A-tier-derived cells (R, between arrows) and B-tier-derived cells (S, between arrows) continue to express n-tubulin; in contrast, very few C-tier-derived cells (arrows) express n-tubulin (T). All examples are dorsal views. A–F are stage 14/15 (neural plate); G,H, J–N, P–T are stage 16/17 (neural fold); I, O are stage 19/20 (neural groove/tube).

Download figure to PowerPoint

To determine whether the changes in primary spinal neuron numbers occurred early in the neural plate when the primary neuron progenitors are first established (Lamborghini, 1980; Jacobson and Moody, 1984; Chitnis, 1999), the extent of the neural plate domain was measured by sox2 expression. In control βgal mRNA-injected embryos, the size of the injected side of the neural plate was not significantly different from the uninjected side (Fig. 4A; mean difference between sides = 5.67% ± 1.3; n = 32; P > 0.05). Increased expression of Noggin in B-tier blastomeres caused the neural plate to expand on the injected side (Fig. 4B; 72.7% of embryos); the width of the neural plate on the Noggin-injected side was significantly larger than that on the uninjected side (mean increase = 24.46% ± 4.24, n = 22, P < 0.001). To lower Noggin signaling in the C-tier environment without directly inhibiting neural plate progenitors, bmp4 mRNA was injected in D-tier blastomeres. In these embryos, the neural plate was significantly smaller on the injected side (Fig. 4C; mean decrease = 25.6% ± 3.1, n = 31, P < 0.001). Thus, modulating the levels of Noggin signaling alters the numbers of cells in the neural plate that ultimately give rise to the primary spinal neurons.

To determine whether the numbers of nascent PMN and RBN were consequently altered, the expression of a neural-specific β-tubulin gene that is expressed by primary neurons near the time of their terminal mitosis (Fig. 4G; Richter et al., 1988; Good et al., 1989; Moody et al., 1996) was monitored after the same manipulations. In those embryos in which Noggin expression was increased in B-tier blastomeres, the n-tubulin expression domain was expanded (Fig. 4H; 72.6%, n = 106), whereas in those embryos in which bmp4 mRNA was injected in D-tier blastomeres, the n-tubulin expression domain was reduced (Fig. 4I; 74.1%, n = 104). Single blastomeres were labeled with lineage tracer (βGal) to determine whether these reductions were confined to the descendants of the C-tier. In A-tier labeled embryos (n = 29), numerous βGal-labeled cells continued to express n-tubulin (Fig. 4J), whereas fewer cells derived from B-tier (n = 23; Fig. 4K) and no cells derived from C-tier (n = 27; Fig. 4L) expressed n-tubulin. Thus, the effects of increased vegetal BMP4 are not confined to the C-tier-derived cells, but are most effective in cells that are in closest proximity to the source of the signal. Together, these experiments demonstrate that altering Noggin/BMP4 levels in the vicinity of C-tier blastomeres affects the production of primary spinal neurons at the earliest stages that they can be distinguished.

To determine whether the promotion of primary spinal neurons derived from the C-tier lineages is specific to Noggin originating from animal tiers, the level of Noggin protein was reduced in B-tier cells in the intact embryo by injection of antisense noggin morpholino oligonucleotides (nMO). No changes were observed in the size of the neural plate (sox2 expression) after unilateral injection of a control morpholino (cMO; Fig. 4D). However, injection of nMO in the B2 or B3 lineage caused a significant reduction (mean decrease = 14.0% ± 1.35, n = 32, P < 0.001) on the injected side (Fig. 4E). Likewise, although there were no changes after injection of the cMO (Fig. 4M), the majority of B2-injected embryos showed a reduction in the expression of n-tubulin in the PMN stripe (74.4%, n = 90; Fig. 4N), and the majority of B3-injected embryos showed a reduction in staining intensity in the RBN stripe (76.8%, n = 69; Fig. 4O) after nMO injection. This differential effect on PMN and RBN is due to the differential population of ventral (B2, C2) versus dorsal (B3, C3) regions of the neural tube (Moody, 1989; Fig. 3). To determine whether these decreases are due to repression of primary neural fate in the B-tier lineage (autonomous effect) or in the C-tier lineage (through signaling), nMO was injected into a B-tier cell and βgal mRNA injected into either the same B-tier cell or its C-tier neighbor. In all B-tier labeled embryos (n = 33), about half of the descendants still expressed n-tubulin (Fig. 4P), whereas in 87.5% (n = 40) of C-tier labeled embryos, nearly no descendants were n-tubulin positive (Fig. 4Q). Thus, the repression of primary neurons resulted from small reductions in the nMO-injected B-tier lineage and large reductions in the neighboring C-tier lineage. Although the cell counts of C-tier lineages (Fig. 3) demonstrate that B-tier Noggin levels affect the numbers of primary spinal neurons produced, C-tier blastomeres and their progeny also contain noggin mRNA at cleavage, blastula, and gastrula stages (Fig. 2; Vodicka and Gerhart, 1995; Kuroda et al., 2004), and thus may promote neural fates in their own and neighboring lineages. To test this, we injected nMO in the C-tier lineage, which resulted in a reduced n-tubulin domain in 84.7% of embryos (n = 92). Lineage labeling demonstrated that in all A-tier labeled (n = 21) and B-tier labeled (n = 24) embryos, large numbers of βGal-labeled cells continued to express n-tubulin (Fig. 4R, S), whereas in 72.5% of C-tier labeled embryos (n = 40) very few βGal-labeled cells expressed n-tubulin (Fig. 4T). These data demonstrate that reducing Noggin in C-tier blastomeres primarily affects only the C-tier lineage.

Reduced BMP Signaling in C-Tier Blastomeres Rescues the Effects of Cleavage Stage Dissociation

The animal-to-vegetal signaling that promotes C-tier blastomere neural fate was previously shown to require close cellular proximity during cleavage/morula stages (Bauer et al., 1996). Transiently dissociating cells (for 1–2 hr prior to the mid-blastula transition [MBT]) within the vitelline membrane so that cells did not change neighbors caused C-tier blastomeres to produce significantly fewer primary spinal neurons, implicating either contact-dependent signaling or signaling relying on local concentrations of diffusible factors. If Noggin were the responsible signaling molecule, then expressing Noggin in the dissociated C-tier blastomeres to increase local levels should rescue the dissociation phenotype. This was tested by injecting noggin mRNA into a C-tier blastomere, transiently dissociating the embryo in citrate solution, and counting the numbers of primary neurons descended from the injected lineage. As previously described, a 2-hr dissociation prior to MBT caused a significant decrease in the number of PMN and RBN in both the C2 and C3 lineages (Fig. 5). Injection of noggin mRNA in the C-tier cells reversed this effect for PMN (Fig. 5). The effect on RBN was not reversed (Fig. 5), so the experiment was repeated using a truncated BMP receptor construct (tBMPR) that acts in a dominant-negative fashion to inhibit BMP signaling in the recipient cells (Graff et al., 1994). In these cases, the numbers of both PMN and RBN were dramatically increased in the C-tier lineages after dissociation treatment (Fig. 5). Thus, C-tier expression of primary spinal neurons depends upon local reduction of BMP signaling.

thumbnail image

Figure 5. Lowered BMP signaling in C-tier blastomeres rescues the effect of transient dissociation. C-tier blastomeres were injected at the 32-cell stage with: lineage tracer only and allowed to develop (normal), lineage tracer only then dissociated (citrate), lineage tracer plus noggin mRNA and then dissociated (noggin-citrate), or lineage tracer plus tbmpr mRNA then dissociated (tBMPR-citrate). Expression of Noggin in C-tier blastomeres rescues PMN but not RBN numbers after transient pre-MBT dissociation. However, blockade of BMP signaling in the C-tier lineages via tBMPR expression rescues both phenotypes. Asterisks indicate significant difference from control (normal) numbers at the P < 0.05 level.

Download figure to PowerPoint

The Neural Fate-Promoting Signal Occurs in Blastomere Explants

We tested the neural-promoting signaling of B-tier blastomeres in explants that were dissected at cleavage stages and cultured in simple medium without growth factors to determine if the putative interaction occurs in the absence of gastrulation movements and their associated signaling interactions. Explants were assayed for elongation, a morphological indicator of dorsal axial tissue differentiation (Wilson and Keller, 1991), and zygotic markers that are expressed during gastrulation and/or neurulation: Xbra, a pan-mesodermal gene; chd, a dorsal mesoderm gene; sox2, a neural plate gene; and n-tubulin, a primary neuron gene. C-tier explants cultured alone (C-tier; Fig. 1) elongated slightly (Fig. 6A) in 45.7% of cases (n = 127). Nearly all expressed abundant Xbra and about half expressed abundant chd (Fig. 6A; Table 1). sox2 and n-tubulin were expressed at lower levels and frequency (Fig. 6A; Table 1). To test whether any of these genes were expressed as a result of endogenous noggin mRNA (Fig. 2), 8-cell precursors of the C-tier cells were injected with nMO to reduce the translation of noggin transcripts, and C-tier explants were made 1 hr later at the 32–64 cell stage. The frequency and intensity of Xbra and chd expression were not altered, but sox2 expression intensity was reduced and n-tubulin expression was extinguished (Fig. 6B; Table 1). These data indicate that by the time the explants were removed from the cleavage embryo, C-tier cells are autonomously capable of expressing a mesodermal fate, with some bias to express a neural (sox2) program independent of further signaling, in accord with some earlier reports (Cardellini 1988; Gallagher et al., 1991; Pandur et al., 2002). However, primary neuron development (n-tubulin) requires further exposure to Noggin.

thumbnail image

Figure 6. Blastomere explant expression of mesodermal (Xbra, chd) and neural (sox2, n-tubulin) genes. A: C-tier explants express abundant levels of Xbra and chd, albeit the latter at a lower frequency (see Table 1 for numbers and statistical comparisons). About half of the explants express low levels of sox2 (*) and less than half express n-tubulin (arrows). B: When C-tier blastomeres were injected with nMO, sox2 expression (*) is reduced in intensity and n-tubulin expression is extinguished. C: When C-tier blastomeres were co-cultured with B-tier blastomeres, the explants elongate more elaborately than C-tier (A) or B-tier (D) explants. Although mesoderm gene expression is not altered, the frequency and intensity of neural gene expression increases. D: B-tier explants express mesodermal genes at low levels and in fewer explants (*). sox2 expression (*) is less broad and n-tubulin expression (arrow) is less frequent. E,F: Injection of βgal mRNA into a B-tier cell demonstrates that both sox2-expressing (E) and n-tubulin-expressing (F) cells are derived in part from the B-tier lineage. G,H: Injection of βgal mRNA into a C-tier cell demonstrates that both sox2-expressing (G) and n-tubulin-expressing (H) cells are derived in part from the C-tier lineage. I: When C-tier blastomeres are co-cultured with D-tier blastomeres, elongation is reduced, but the expression of mesodermal genes is unaffected. However, sox2 expression is confined to small domains (arrows) and n-tubulin expression is greatly reduced (arrow).

Download figure to PowerPoint

Table 1. Neighboring Blastomeres and Noggin Levels Affect Explant Expression of Neural Genesa
 Xbra % (n)chordin % (n)sox2 % (n)n-tubulin % (n)
  • a

    ND, not determined.

  • *

    P < 0.05;

  • **

    P < 0.001.

  • ***

    **P < 0.05;

  • P < .001 when data from C-nMO + B explants were compared to data from C + B-nMO explants.

  • C-tier-nMO, C + B-tiers, C + D-tiers, C-tier + Noggin results were compared to those from C-tier alone explants. C + B-tiers + BMP, C + B-nMO, C + B-cMO and C-nMO + B results were compared to those from C + B-tiers explants. C + B-nMO + Noggin results were compared to those from C + B-nMO explants.

C-tier alone91.9 (37)58.1 (43)52.0 (50)38.1 (21)
C-tier-nMO93.3 (30)57.7 (45)43.5 (92)0 (26)**
C + B-tiers100 (35)50.0 (48)84.8 (59)**75.0 (56)**
B-tier alone10.0 (30)30.9 (42)38.8 (67)9.3 (43)
C + D-tiers98.0 (50)54.8 (42)46.3 (54)13.3 (45)**
C-tier + NogginND64.1 (39)68.3 (41)*50.0 (30)*
C + B-tiers + BMPND37.8 (37)*22.0 (50)**19.4 (36)**
C + B-nMO96.4 (28)45.2 (62)*26.1 (46)**23.9 (67)**
C + B-nMO + NogginND64.2 (19)**55.2 (29)**59.3 (27)**
C + B-cMOND62.2 (37)65.9 (41)77.3 (44)
C-nMO − B100 (33)67.9 (28)*,44.1 (34)**,***34.4 (32)**,***

When C-tier blastomeres were cultured in continued association with their B-tier neighbors (C+B-tier; Fig. 1), neural fate was strongly promoted. These explants elongated at a significantly higher frequency (80.2%, n = 106; P < 0.005) and to a greater morphological complexity compared to C-tier explants (Fig. 6C). Xbra and chd expression was comparable in frequency and intensity to C-tier explants, whereas sox2 and n-tubulin were expressed at significantly higher frequencies and levels compared to either C-tier alone or B-tier alone explants (compare Fig. 6C to A,D; Table 1). Because Xbra and chd expression was similar in C+B-tier and C-tier explants, the increase in neural gene expression in C+B-tier explants probably is not due to increased mesoderm formation, consistent with the report that BCNE-derived neural induction is independent of mesoderm signaling (Kuroda et al., 2004). Analysis of B-tier explants cultured alone (B-tier; Fig. 1) demonstrated that their endogenous ability to express mesodermal or neural genes also does not account for the increase in C+B-tier explants; all four markers were expressed at significantly lower frequencies than in C+B-tier explants (Fig. 6D: P < 0.001, Table 1). However, when either B-tier (Fig. 6E,F) or C-tier (Fig. 6G,H) blastomeres were marked with a lineage tracer prior to co-culture, both sox2 (92.5%, n = 80) and n-tubulin (100%, n = 38) expressing cells were derived from both blastomeres. Thus, it is possible that the enhanced neural gene expression derives from mutual blastomere interactions.

Since neighboring blastomeres on the animal pole side promote C-tier neural fate, we asked whether neighboring blastomeres on the vegetal pole side repress it, as suggested by experiments in which BMP4 expression was increased in D-tier cells (Fig. 3B) and studies showing that vegetal pole blastomeres contain maternal molecules that repress a neural fate (Fig. 2; Kessler, 1999; Moore and Moody, 1999). Consistent with this proposal, explants made from C+D tier blastomeres (Fig. 1) elongated less frequently than C-tier controls (37.8%; n = 98) and the extent of elongation was less extensive than in C+B-tier explants (Fig. 6I). Although the frequency of sox2 expression was comparable to that of C-tier explants (Table 1), it was significantly reduced in 2 out of 3 independent assays (8 and 36%), and levels of expression were clearly reduced (compare Fig. 6I to 6A). n-tubulin expression was significantly reduced in frequency and intensity (Fig. 6I; Table 1). Since Xbra and chd expression was comparable to that in C-tier explants (Fig. 6I; Table 1), the effects on neural gene expression are likely independent of mesoderm formation. These data demonstrate that association with D-tier blastomeres represses the neural, including primary spinal neuronal, fate of C-tier blastomeres.

To determine whether these effects are due to Noggin expressed prior to gastrulation, blastomeres were cultured under conditions that either increased or decreased Noggin during cleavage/morula stages. First, C-tier explants were cultured in Noggin-containing medium for 2 hr, washed and then returned to factor-free medium shortly before MBT. These explants expressed sox2 and n-tubulin at significantly higher frequencies and greater staining intensities compared to C-tier explants cultured alone, whereas chd expression did not change (Fig. 7A; Table 1). To test whether the putative blastomere signal could be Noggin, C+B-tier explants were cultured in the presence of BMP4 protein for 2 hr, washed and then returned to factor-free medium shortly before MBT. All three marker genes were significantly repressed in frequency in these explants (Table 1), and the expression levels of the neural genes were markedly reduced (Fig. 7B). Blockade of endogenous Noggin was additionally achieved by injecting the 8-cell precursors of either the B-tier or C-tier with nMO to reduce the translation of noggin transcripts. Injection of a control MO (C+B-cMO) had no effect on any of the monitored genes (Table 1). In those explants in which the nMO was injected into B-tier precursors (C+B-nMO), the frequency of chd expression was reduced (Table 1), but not the intensity of the staining (Fig. 7C), and Xbra expression was unaltered (Table 1). In contrast, both the frequency and intensity of sox2 and n-tubulin expression were significantly reduced (Fig. 7C; Table 1). Lineage labeling showed that in 82.4% of sox2-positive explants (n = 34) and 63.0% of n-tubulin-positive explants (n = 46), the neural cells were derived from both B-tier and C-tiers; in the remaining explants, the cells were derived from the B-tier alone. Adding Noggin to the culture medium of C+B-nMO-injected explants for 2 hr prior to MBT reversed these reductions (Fig. 7D; Table 1). As shown in whole embryos (Fig. 4T), Noggin expression in the C-tier blastomeres also is required for the expression of the neural genes. In C+B-tier explants in which the C-tier was injected with nMO (C-nMO+B), neither mesoderm gene was altered, but both neural markers were reduced in frequency and intensity (Fig. 7E; Table 1). Lineage labeling showed that those cells expressing either Xbra or chd were derived from both blastomeres in every explant; in only 40% of sox2-positive explants (n = 15) and 34.4% of n-tubulin-positive explants (n = 32) were neural cells derived from the C-tier. In fact, comparison of the expression of chd, sox2, and n-tubulin to the C+B-nMO explants demonstrated a significantly higher frequency of expression (Table 1), suggesting that C-tier nMO injection had mostly an autonomous effect. Thus, in C+B-tier explants, blocking endogenous noggin translation in the B-tier lineage affects both B- and C-tier derived neural progeny whereas blocking endogenous noggin translation in the C-tier lineage predominantly affects C-tier derived neural progeny, consistent with the whole embryo data (Fig. 4).

thumbnail image

Figure 7. The expression of a dorsal mesoderm gene (chd) and neural genes (sox2, n-tubulin) are altered by manipulating the levels of BMP signaling in explant cultures. A: Adding Noggin to C-tier cultures for 2 hr prior to MBT significantly increases elongation and intensifies the expression of sox2 (*) and n-tubulin (arrows) (compare Fig. 6A); chd expression is not significantly altered in frequency or intensity. B: Adding BMP4 protein to C+B-tier co-cultures for 2 hr prior to MBT significantly reduces the expression of all three genes (*).C: Reducing Noggin translation by nMO injection in B-tier blastomeres does not alter chd expression in C+B-tier explants, but reduces both sox2 (*) and n-tubulin (arrows) expression. D: The effect of the B-tier nMO injection is rescued by culturing the C+B-tier explants in exogenous Noggin for 2 hr prior to MBT. chd expression is not altered, but sox2 (*) and n-tubulin (arrows) expression is more intense and detected in more explants. E: Reducing Noggin translation by nMO injection in C-tier blastomeres does not alter chd expression in C+B-tier explants, but reduces both sox2 (*) and n-tubulin (arrows) expression.

Download figure to PowerPoint

DISCUSSION

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

The transcription factors and signaling cascades that lead to the generation of specific neuronal phenotypes once the embryonic nervous system forms have been intensely studied because of the importance of understanding how to manipulate undifferentiated cells to desired neuronal phenotypes (e.g., Limke and Rao, 2002; Ostenfeld and Svenden, 2003). For example, manipulating ES cells through sequential developmental steps that recapitulate the normal post-neural plate segment of neural specification can create motoneuron progenitors that innervate appropriate targets after transplantation (Wichterle et al., 2002). This exciting demonstration of applying the information derived from the normal program of neural specification to replacement therapies provides great incentive for understanding all steps in this developmental program. However, there has been much less progress on delineating the interactions that occur prior to the formation of the neural plate that bias embryonic cells to express a neural fate, even though many localized maternal factors that affect cell fate specification have been identified (reviewed in Sullivan et al., 1999; De Robertis and Kuroda, 2004; Dupont et al., 2005). Previously, we identified a cell–cell signal from animal blastomeres, which occurs prior to gastrulation movements, that promotes the primary spinal neuron fate of neighboring C-tier blastomeres (Bauer et al., 1996). Herein, we ask whether Noggin, whose transcripts are enriched in the animal hemisphere, is responsible for this fate regulation and whether it acts prior to the inductive interactions that occur during gastrulation to establish the bona fide neural plate.

Noggin Signaling From Animal Blastomeres Promotes the Neural Fate of Neighboring C-Tier Blastomeres

The C-tier blastomeres of the 32-cell Xenopus embryo sit on a fate border between those cells that normally contribute heavily to the CNS (A-tier and B-tier), and those that do not contribute to the CNS (D-tier; Fig. 1). This provides a unique opportunity to manipulate gene expression levels in neighboring cells to identify whether the neural fate of this border region is affected by cell–cell signaling. Previously, we showed that the numbers of PMN and RBN produced by C-tier blastomeres were decreased when their contacts with B-tier neighbors were transiently disrupted by a variety of means prior to gastrulation (Bauer et al., 1996). This study indicated that there is an early animal-to-vegetal signal that promotes the ability of C-tier blastomeres to produce primary spinal neuronal progeny. Recently, it was reported that a signaling center (the BCNE center) arises in the dorsal-animal marginal zone during early blastula stages upstream of Spemann's Organizer, and it strongly expresses transcripts for chordin, noggin, siamois, and Xnr3 (Kuroda et al., 2004). The BCNE center is required for anterior neural development in whole embryos, and self-differentiates into neural tissues when cultured as explants or transplanted ventrally. These same properties were reported for the dorsal-animal 16- and 32- cell stage blastomeres (Gallagher et al., 1991), which are the precursors of the BCNE center as shown by mapping blastomere clone pre-gastrulation movements (Bauer et al., 1994; Vodicka and Gerhart, 1995). Thus, the dorsal animal region is predisposed to a neural fate. Kuroda et al. (2004) further showed that the effects on anterior neural markers require Chordin signaling, whereas effects on posterior neural markers are Chordin-independent. Previous findings that maternal transcripts for noggin are enriched in the animal hemisphere, our report herein that they are in an animal-to-vegetal concentration gradient, and reports that noggin mRNAs (likely zygotic) are enriched in B-tier derived cells at blastula and early gastrula stages (Vodicka and Gerhart, 1995; Kuroda et al., 2004), spurred us to investigate whether animal-derived Noggin influences C-tier expression of posterior neural fates, i.e., primary spinal neurons. We found evidence that both B-tier derived (via signaling) and C-tier derived (endogenous) Noggin affects the ability of C-tier cells to express primary spinal neuronal fates. This effect is likely direct, because no significant alterations in mesodermal gene expression were detected, consistent with the demonstration that Noggin (Knecht et al., 1995) and the BCNE center (Kuroda et al., 2004) can affect neural fate independent of mesoderm. These results are consistent with the studies that demonstrate that the BCNE center causes an early neural bias prior to gastrulation. However, although Kuroda et al. (2004) propose that BCNE-derived Noggin has a weaker effect on anterior neural fates compared to Chordin, our experiments indicate that Noggin exerts substantial influence on primary spinal fates, particularly in the C-tier lineages. This effect is not confined to BCNE signaling, since reduction of Noggin in the C-tier lineage reduced neural gene expression in both whole embryos and explants, predominantly in the C-tier derived cells. It should be noted that C-tier signaling also may promote B-tier neural fate, as effects of injection of nMO into C-tier blastomeres were not exclusively on the C-tier lineage.

A question that remains to be answered is when during development does this signaling arise? Although we report differential concentrations of maternal noggin and bmp4 mRNAs, are their protein products signaling during cleavage stages? Studies that document the temporal and spatial distributions of the activated intracellular mediators of TGFβ signaling pathways indicate that signaling is not likely significant until early MBT (Faure et al., 2000; Schohl and Fagotta, 2002). The C-terminal phosphorylated form of Smad1, which mediates BMP signaling, is not detected by Western blot or immunohistochemistry until stage 8, even when bmp4 mRNA is over-expressed. Thus, the secretion of anti-BMP molecules by animal blastomeres/BCNE center cells may occur earlier than or just as soon as BMP signaling can be transduced. Schohl and Fagotta (2002) propose that there are three phases of embryonic signaling: (1) an early phase that is influenced by both maternal and very early zygotically transcribed components; (2) a zygotic phase that extends from the late blastula (stage 9.5) to neurula stages that sets up the general embryonic body plan; and (3) a late phase that refines the body plan and initiates organogenesis. Our experiments manipulated signaling systems during the maternal/very early zygotic phase and support the notion that influence on neural fates begin well before the zygotic phase (Kuroda et al., 2004).

A Differential Effect of Noggin mRNA Injection on PMN Versus RBN Fates

Our experiments show that over-expression of Noggin increased PMN numbers, but not RBN numbers, at neural tube stages, even though both cell types were significantly repressed by elevated BMP4. This differential effect is most likely due to later patterning changes caused by residual high levels of Noggin. It is well established that after the neural plate is induced from embryonic ectoderm by Noggin and other BMP antagonists, the dorsal-ventral patterning of the neural tube is also accomplished by the later interaction between a BMP gradient, whose source is the overlying epidermis and roof plate of the dorsal neural tube, and an opposing Shh gradient, whose source is the underlying notochord and floor plate of the ventral neural tube (reviewed in Altman and Brivanlou, 2001; Kintner 2002). Dorsal neural phenotypes, including the neural crest and RBN, are formed in response to high BMP/low Shh levels, whereas ventral neural phenotypes, including motoneurons, are formed in response to low BMP/high Shh levels (Nguyen et al., 2000; Liem et al., 2000). Since RBN are derived from the neural crest (Hughes, 1957), the lack of overproduction of RBN in the neural tube after B-tier noggin mRNA injections is consistent with reports that high levels of Noggin at later stages repress neural crest derivatives (Nguyen et al., 2000; Villanueva et al., 2002). Since RBN numbers are altered by manipulating Noggin levels in other ways (e.g., BMP expression, Noggin morpholino treatment, tBMPR expression), the lack of effects on RBN numbers in Figure 3A is due most likely to later dorsal-ventral patterning effects.

What Fate Is Promoted: Neural Bias, Neural Stem Cell, or Primary Neuron Progenitor?

We have taken advantage of a quantitative fate map for two primary neurons, PMN and RBN, to test for possible pre-gastrulation effects on neuronal fate decisions (Moody, 1989; Bauer et al., 1996). Fish and amphibians develop an early reflex arc comprised of primary neurons. These cell types are the most likely to be affected by pre-gastrulation influences because they complete their terminal cell divisions during gastrulation (Lamborghini, 1980; Jacobson and Moody, 1984), are the earliest to express differentiation genes in the neural plate (Chitnis, 1999; Kintner, 2002), and are the earliest to become functional (Roberts et al., 1981). Because escape behavior depends upon the proper functioning of these cells, perhaps evolutionary pressure would use maternal molecules and early zygotic signaling to ensure the proper development of these critical cells. But does pre-gastrulation signaling directly specify primary neuron fate? We postulate that it does not for two reasons. First, there are zygotic genes expressed later in the neural tube that direct the formation of motoneurons (Jessell, 2000; Marquardt and Pfaff, 2001) and sensory neurons (Anderson, 1999). Although these factors have not yet been directly tested in the establishment of PMN or RBN in fish or frogs, one would expect these cell type–specific gene programs to be conserved and thereby necessary for PMN and RBN specification. Second, most of the experimental manipulations presented herein also affected sox2, a gene expressed directly downstream of neural induction by the definitive neural stem cells of the neural plate, and upstream of several primary neuronal fate specifying bHLH factors (Mizuseki et al., 1998; Kishi et al., 2000). Therefore, pre-gastrulation signaling likely affects the size of the domain that is biased to produce the neural stem cell pool comprising the neural plate, and subsequently impacts primary neuron numbers.

Pre-Gastrulation Patterning of Neural Fate

There has been a tendency to view the program of neural specification as beginning during gastrulation with the inhibition of BMP signaling by factors secreted from the Organizer, which subsequently separates the prospective neural ectoderm from the prospective epidermis and establishes the definitive neural stem cells of the neural plate. However, there is mounting evidence that earlier signaling and transcriptional regulation bias different regions of the embryos to be more (or less) responsive to later signaling centers that develop during morphogenesis. For example, endodermal fate is biased by localization of maternal Vg1 and VegT (Kessler, 1999; Xanthos et al., 2001), neural fate is repressed by vegetal factors (Moore and Moody, 1999), dorsal fate is biased by localized maternal β-catenin (Moon, 2005), and ventral fate is biased by localized maternal Wnt (Pandur et al., 2002) and BMP (Mintzer et al., 2001; Kramer et al., 2002) signaling components. Manipulations of dorsal animal blastomeres have demonstrated an endogenous bias to form neural tissue (Gallagher et al., 1991; Pandur et al., 2002), and even in mouse ES cell culture there is evidence for a pre-neural fate bias (Tropepe et al., 2001). Recent studies demonstrate that a signaling center in the dorsal-animal region, involving Chordin, Noggin, and β-catenin, biases a neural fate prior to gastrulation (Wessely et al., 2001; Kuroda et al., 2004). Herein we demonstrate that levels of Noggin from animal blastomeres during pre-gastrulation stages promote primary spinal neuronal fates in neighboring vegetal equatorial cells, and further provide evidence for endogenous C-tier and perhaps reciprocal signaling. Together these studies indicate that embryonic signaling prior to the classical neural inductive interactions that occur during gastrulation is an important mechanism for biasing multipotent cells toward neuronal phenotypes.

EXPERIMENTAL PROCEDURES

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

Embryo Collection and Blastomere Injections

Fertilized Xenopus laevis eggs were obtained by natural matings using standard methods (Moody, 1999). mRNAs were synthesized in vitro (Maxiscript; Ambion) and microinjected at the indicated concentrations per cell: βgal (100 pg), gfp (200 pg), noggin (50 pg) (Smith and Harland, 1992), bmp4 (50 pg) (Dale et al., 1992; Nishimatsu et al., 1992). and tbmpr (100 pg) (Graff et al., 1994). noggin, bmp4, and tbmpr transcripts were mixed with lineage tracer transcripts (βgal or gfp) and microinjected into identified blastomeres of the 32-cell embryo (Fig. 1) as described (Moody, 1999, 2000). Concentrations of mRNAs used were based on published effective doses. A noggin antisense morpholino oligonucleotide (nMO: 5′-TCACAAGGCACTGGGAATGATCCAT-3′) and a standard control morpholino (cMO) were synthesized (Gene Tools) and injected into 8-cell blastomere precursors of either B-tier or C-tier cells (16 ng/cell).

RT-PCR.

Embryos were fixed overnight in 98% ethanol/2% acetic acid at −20°C (Sive et al., 2000). Individual tiers of cells were dissected with microsurgical blades and stored in the same solution at −20°C. RNA was purified (RNeasy Protect, QIAGEN), treated with DNase I, and reverse transcription (RT) was performed with 1.0 μg of RNA (SuperScriptTM, Invitrogen). One-tenth of the RT product was used as template for standard PCR amplification (PCR Supermix, Invitrogen). The primers for H4 were 5′CGGGATAACATTCAGGGTATCACT-3′ (forward) and 5′-ATCCATGGCGGTAACTGTCTTCCT-3′ (reverse), 189 bp, 25 cycles (Niehrs et al, 1994). The primers for noggin were 5′-AGTTGCAGATGTGGCTCT-3′ (forward) and 5′-AGTCCAAGAGTCTCAGCA-3′ (reverse), 281 bp, 30 cycles (Agius et al, 2000). The primers for bmp4 were 5′-TCATTCACCTCAACCAAACC-3′ (forward) and 5′-AGTCATTCCAGCCCACATC-3′ (reverse), 252 bp, 25 cycles (this study). The primers for Vg1 were 5′-CCCTCAATCCTTTGCGGTG-3′ (forward) and 5′-CAGAATTGCATGGTTGGACCC-3′ (reverse) (Weeks and Melton, 1987). The primers for Xwnt8b were 5′-TGACTTGAACATCCATTCT-3′ (forward) and 5′-TGGAGAAAGGAATCTGTA (reverse) (Cui et al., 1995). Bands were visualized with a Storm 860 phosphorImager (Molecular Dynamics), and levels of bmp4 and noggin were normalized relative to H4 using ImageQuant software. Each sample consisted of blastomere tiers dissected from 25 different embryos; each tier sample was analyzed for presence of animal-enriched (Wnt8b) and vegetal-enriched (Vg1) mRNAs to confirm accuracy of the dissection. Whole embryo (WE) RNA was extracted from a pool of 3 siblings of the dissected tiers. Four independent experiments were performed, levels of RNA were expressed relative to WE levels that were set at 100%, and these values expressed as a mean ± SEM.

Blastomere dissections for culture.

Embryos were placed on agar-coated dishes in Steinberg's solution and blastomeres were dissected from embryos as described (Moody, 2000). For culture experiments, the entire dorsal-to-ventral tier(s) of cells from one side of the embryo (either B-tier cells alone, C-tier cells alone, C+B-tiers, or C+D-tiers; Fig. 1) were cultured in a simple salt medium (NAM) as described (Pandur et al., 2002). In some cases, either Noggin (50 ng/ml; recombinant mouse/Fc chimera; R&D Systems) or BMP4 (500 ng/ml; recombinant human; R&D Systems) was added to the medium at the start of the culture period. After 2 hr, explants were washed extensively and transferred to new dishes and factor-free NAM for the remainder of the culture period. Concentrations used were based on published effective doses for neural induction (Noggin) and inhibition of neural induction (BMP4) (Knecht et al., 1995; Zimmerman et al. 1996; Wilson et al., 1997).

Transient embryo dissociation.

After blastomere injections at the 32-cell stage, embryos still within their vitelline membranes were dissociated in situ by incubation in calcium-magnesium-free Stern's solution for 2 hr, as described (Bauer et al., 1996). This treatment separates cells from contact with their neighbors, but keeps them in the same spatial arrangement as in controls. After 2 hr, embryos were re-associated by incubating in medium containing normal calcium and magnesium levels and allowed to continue development.

Cell Fate Analysis

Embryos were raised to stages 32–34, fixed in 4% paraformaldehyde, and serially sectioned. Sections were analyzed for the presence of lineage-labeled primary neurons and the total numbers of PMN and RBN were counted along the entire length of the spinal cord as previously described (Moody, 1989; Bauer et al., 1996). Cell counts were compared between experimental and control groups by the t-test.

β-Galactosidase assay, whole mount in situ hybridization.

Embryos and explants were fixed at stages 11–12 (Xbra), 14/15 (chd, sox2), or 17/18 (n-tubulin), assayed for β-Galactosidase activity, then processed for whole-mount in situ hybridization using standard methods (Sive et al., 2000). Full-length, digoxygenin-labeled antisense probes for Xbra (Dale et al., 1992), chd (Sasai et al., 1994), sox2 (Mizuseki et al., 1998; gift of Rob Grainger, University of Virginia), and n-tubulin (Good et al., 1989) were synthesized (Maxiscript; Ambion). In some embryos, the width of the neural plate, as defined by the domain of endogenous sox2 mRNA expression, was measured at 40× magnification with an eyepiece micrometer. The difference in width of the neural plate on the mRNA-injected side versus uninjected side was expressed as a percentage, and differences between experimental and control embryos compared by the t-test.

Because there is variability in the distribution of maternal molecules, in cleavage patterns, and in our ability to cleanly separate tiers of blastomeres during dissections, gene expression in explants was analyzed by in situ hybridization rather than PCR. To normalize for expression levels across samples, probes, and in situ runs, control whole embryos were included in each explant sample during the in situ hybridization procedure. During the color reaction, the staining intensity of the embryos was monitored so that different explant samples were exposed to the chromogenic reaction for equivalent periods, thus allowing a relative comparison of expression between explant samples. Differences in the frequency of expression of markers in explant cultures were compared by the Chi-squared test.

Acknowledgements

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

We thank Drs. Kathryn Moore and Samantha Brugmann for their comments on the manuscript, Lianhua Yang for histological preparations, and Himani Datta Majumdar for in situ hybridization preparations.

REFERENCES

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