SEARCH

SEARCH BY CITATION

Keywords:

  • neogenin;
  • neural development;
  • neural crest;
  • repulsive guidance molecule A (RGM A);
  • Xenopus

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. Funding
  9. References

Background information. RGM A (repulsive guidance molecule A) is a GPI (glycosylphosphatidylinositol)-anchored glycoprotein which has repulsive properties on axons due to the interaction with its receptor neogenin. In addition, RGM A has been demonstrated to function as a BMP (bone morphogenetic protein) co-receptor.

Results. In the present study, we provide the first analysis of early RGM A and neogenin expression and function in Xenopus laevis neural development. Tissue-specific RGM A expression starts at stage 12.5 in the anterior neural plate. Loss-of-function analyses suggest a function of RGM A and neogenin in regulating anterior neural marker genes, as well as eye development and neural crest cell migration. Furthermore, overexpression of RGM A leads to ectopic expression of neural crest cell marker genes.

Conclusions. These data indicate that RGM A and neogenin have important functions during early neural development, in addition to their role during axonal guidance and synapse formation.


Abbreviations used:
BMP

bone morphogenetic protein

BrdU

bromodeoxyuridine

CNS

central nervous system

E8.5 etc.

embryonic day 8.5 etc.

GFP

green fluorescent protein

GPI

glycosylphosphatidylinositol

MO

morpholino oligonucleotide

N

number of analysed embryos in total

n

number of independent experiments

NC

neural crest

NCC

NC cell

ORF

open reading frame

RGM

repulsive guidance molecule

RT

reverse transcription

TdT

terminal deoxynucleotidyl transferase

TUNEL

terminal deoxynucleotidyl transferase-mediated dUTP nick-end labelling

UTR

untranslated region

WMISH

whole-mount in situ hybridization

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. Funding
  9. References

In Xenopus laevis, neural induction starts at gastrulation by signals emanating from the Spemann organiser, dividing the ectodermal germ layer into neural and non-neural tissues. Neural inducers are antagonists of BMPs (bone morphogenetic proteins), such as noggin and chordin (Weinstein and Hemmati-Brivanlou, 1999). Patterning of the neural tissue along the anterior—posterior and dorsal—ventral axes is mediated by different growth factors, such as FGFs (fibroblast growth factors), BMPs or Wnts. At the border between neural tissue and epidermis, NCCs (neural crest cells) are induced that later migrate from the dorsal side of the embryo ventrally and give rise to different tissues and cell types, such as cranial and trunk sensory neurons, facial cartilage and pigmented cells of the epidermis. Signals involved in NC (neural crest) induction are intermediate levels of BMPs and members of the Wnt gene family which signal through the canonical Wnt/β-catenin pathway (Mayor and Aybar, 2001; Wu et al., 2003), whereas non-canonical Wnt signalling has been shown to be involved in regulating NCC migration (De Calisto et al., 2005; Gessert et al., 2007).

Later, the nervous system develops; large and complex neuronal circuits form in which neurons make numerous connections with each other for proper functioning. This network is established by neuronal migration, axonal outgrowth and synapse formation. For these processes, molecular attractive and repulsive guidance cues are required, such as RGM (repulsive guidance molecule) (Brinks et al., 2004; Kyoto et al., 2007; Monnier et al., 2002). RGM is a membrane-bound glycoprotein with repulsive properties on axons (Matsunaga et al., 2006; Stahl et al., 1990; Wilson and Key, 2006). In mouse, three homologues of RGM have been identified, named RGM A, RGM B (also known as dragon) and RGM C (also known as hemojuvelin/HJV) (Niederkofler et al., 2004; Oldekamp et al., 2004; Schmidtmer and Engelkamp, 2004). In Xenopus, only RGM A and B have been described so far (Klein et al., 2002; Samad et al., 2005; Wilson and Key, 2006). In chicken, only one RGM homologue is known, which has the highest homology with mouse RGM A (Schmidtmer and Engelkamp, 2004; Yamashita et al., 2007). In mice, RGM A has a role during neural tube closure (Niederkofler et al., 2004). Further studies have demonstrated that RGM A has a negative effect on the healing of injured tissue of the CNS (central nervous system) and prevents regeneration of axons in this tissue (Doya et al., 2006; Hata et al., 2006; Schwab et al., 2005a, 2005b).

Two molecular mechanisms are known for RGM function. First, RGM A can interact with the single-span transmembrane protein neogenin (Stahl et al., 1990; Matsunaga and Chedotal, 2004; Yamashita et al., 2007). Originally, neogenin was described as a receptor that mediates the attraction of growth cones by netrin-1 (Livesey, 1999). Upon interaction with RGM, however, repulsion of retinal axons is mediated (Rajagopalan et al., 2004). In the neural tube, the overexpression of neogenin or down-regulation of RGM A leads to increased cell apoptosis, and co-overexpression of neogenin and RGM A inhibits the negative effect of neogenin on cell survival (Matsunaga et al., 2004). RhoA, PKC (protein kinase C) and Rho kinase are essential factors of the RGM A-mediated neogenin signal transduction for axonal pathfinding (Conrad et al., 2007). Secondly, and independent of neogenin, RGM A, RGM B and RGM C can function as BMP co-receptors, binding BMP-2, BMP-4 and BMP-12 (Babitt et al., 2005, 2006; Samad et al., 2005; Halbrooks et al., 2007; Xia et al., 2007). In this context, members of the RGM family have been shown to alter receptor specificity and sensitivity of BMP serine/threonine kinase receptors type I and II (Xia et al., 2007).

In the present study, we provide a first functional description of RGM A during early neural and NC development in X. laevis. Loss- and gain-of-function results show that RGM A is required for normal eye development, expression of anterior neural marker genes and the development of NCCs. These novel data indicate that RGM A has additional important functions during early neural development, which are independent of its function in axonal pathfinding and synapse formation during later stages of embryogenesis.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. Funding
  9. References

RGM A is expressed in anterior neural tissue of X. laevis

A X. laevis RGM A sequence was deposited in the GenBank® database (named RGM A1, accession number BC045008), but probably represents an incomplete ORF (open reading frame) for two reasons. First, at the 3′ end, a potential cleavage site for the addition of a C-terminal GPI (glycosylphosphatidylinositol) anchor found in other organisms (Stahl et al., 1990; Monnier et al., 2002) is missing (Supplementary Figure S1 at http:www.biolcell. orgboc100boc1000659add.htm). Secondly, at the 5′ end, the N-terminal signal peptide is shorter in comparison with the Xenopus tropicalis RGM A sequence (accession number NM_203725). Therefore, we aimed to clone a full-length RGM A sequence from X. laevis. Using a primer pair derived from the X. tropicalis sequence and cDNA from X. laevis embryos of stage 29/30, we successfully isolated the probable full-length RGM A, from now on named RGM A2 (accession number EU447275). By comparing the RGM A1 with the RGM A2 sequence, we detected a missing base in the published RGM A1 at position 1293, leading to a premature stop codon at position 1296 before the cleavage site for the GPI anchor. In the RGM A2 sequence, an additional adenine leads to a stop codon at position 1366, resulting in a protein with a GPI anchor cleavage site at position 420 in the amino-acid sequence. In RGM A2, the signal peptide is as long as in the X. tropicalis sequence. Further characteristic domains are the RGD (arginine/glycine/aspartate) site, the partial von Willebrand factor type D domain and a hydrophobic region (Figure 1A). On the basis of these results, we concluded that the newly isolated cDNA clone of RGM A (RGM A2, accession number EU447275) represents a full-length X. laevis homologue.

image

Figure 1. Structure and expression profile of X. laevis RGM A2

(A) Protein structure of RGM A2. The signal peptide is indicated in black. An RGD (arginine/glycine/aspartate) domain, a von Willebrand factor type D domain (vWF) and a hydrophobic domain are indicated in grey. The C-terminal part that is removed for the addition of a GPI anchor is in the range of nucleotide 420–448. (BN) Spatial expression pattern of RGM A2, as determined by WMISH. (B) RGM A2 is expressed in anterior neural tissue at stage (St.) 12.5 (arrowhead). Dorsal is at the top. (C) At stage 13, RGM A transcripts were detected in the anterior neural plate (arrow). The broken line indicates the section shown in (D). Posterior is at the top. (D) Transverse section of the embryo shown in (C). RGM A2 transcripts are found in the sensorial layer of the neuroectoderm (arrow). Posterior is at the top. (E) At stage 18, RGM A2 is expressed in the whole neural tube (arrow). A faint expression can also be detected in the migrating NCCs (arrowhead). Dorsal is at the top. (F) Lateral view at stage 22. RGM A2 is expressed in the neural tube. The strongest expression can be detected in the forebrain (arrowhead). (G) Transverse section at stage 28. Expression of RGM A2 can be detected throughout the diencephalon (arrow) with the exception of the dorsal- and ventral-most parts, as well as the branchial arches (arrowhead). RGM A is not expressed in the eye. (H) Transverse section at stage 28. RGM A2 is expressed in the ventral half of the neural tube (arrow). (I) Lateral view at stage 31. RGM A2 transcripts can be observed in the somites (arrowhead). Broken lines indicate sections shown in (K) and (L). (J) Anterior close up view of the embryo shown in (I). RGM A2 is strongly expressed in the hindbrain (arrowhead) and the branchial arches (arrow). (K) Horizontal section. Expression in the forebrain is highlighted by an arrow. (L) Horizontal section. RGM A2 expression in the NC-derived structures of the branchial arches is highlighted by an arrow. (M) Lateral view at stage 38. (N) Anterior close-up view of the embryo shown in (M). RGM A2 is highly expressed in the branchial arches (arrow). e, eye; nt, neural tube; st, stomatodeum.

Download figure to PowerPoint

Next, we aimed to characterize the temporal and spatial expression pattern of RGM A during Xenopus embryonic development. RT (reverse transcription)-PCR experiments demonstrated that RGM A is expressed during whole embryogenesis of X. laevis (Supplementary Figure S2 at http:www.biolcell.orgboc100boc1000659add.htm). In adult tissue, RGM A can be detected in the CNS, testis, eye, heart, lung, gut and liver (Supplementary Figure S2). The first tissue-specific expression of RGM A was observed at stage 12.5 in the anterior neural plate (Figure 1B) by WMISH (whole-mount in situ hybridization). A horizontal section shows that RGM A is only expressed in the sensorial layer of the neuroectoderm (Figure 1D). At stage 18 and 22, RGM A is expressed in the whole neural tube, with strongest expression in forebrain tissue and faintly in the migrating NCCs (Figures 1E and 1F). Sections at stage 28 reveal that RGM A is expressed in the branchial arches (Figure 1G) and in the ventral and lateral part of the neural tube (Figure 1H). At stage 31, RGM A was additionally detected in skeletal muscles (Figure 1I). Horizontal sections at stage 31 revealed that RGM A is expressed in the forebrain (Figure 1K) and NC-derived structures of the branchial arches (Figure 1L).

Knockdown of RGM A affects eye development

To examine the function of RGM A during Xenopus development, we made use of the antisense MO (morpholino oligonucleotide) strategy for loss-of-function experiments. As the real start codon is not known, we designed two different MOs (RGM A1 MO and RGM A2 MO) that covered both possible ATG start codons (Figure 2A). To test the binding specificity and affinity of both MOs, we used an in vitro-coupled transcription and translation assay. Indeed, both MOs are able to inhibit the translation of the corresponding construct (Figure 2B). In addition, RGM A1 MO inhibited translation of RGM A2. In contrast, translation of deletion constructs that lack either 12 bp [Δ5′ UTR (untranslated region) RGM A1] or 10 bp (Δ5′ UTR RGM A2) of the MO-binding sites was not altered. The control MO also had no effect on translation of RGM A1, as well as RGM A2.

image

Figure 2. RGM A is required for normal eye development

(A) Nucleotide sequences (5′ ends) of the published RGM A1 (accession number BC045008) and the newly isolated RGM A2 (accession number EU447275), which probably represents the full-length ORF. Putative ATG start codons are indicated in green and red letters. The binding sites for two different MOs are shown. Note that RGM A2 MO is specific for the full-length RGM A2, whereas RGM A1 MO targets both putative RGM A transcripts. (B) The in vitro-coupled transcription and translation assay using 35S-labelled methionine revealed that RGM A1 MO was able to block the translation of RGM A1, whereas a control MO did not interfere with protein synthesis. The translation of the Δ5′ UTR RGM A1 construct was not inhibited by the RGM A1 MO. RGM A2 MO specifically inhibits the translation of RGM A2. The Δ5′ UTR RGM A2 contruct was not recognized by the RGM A2 MO. The RGM A1 MO also affects the translation of RGM A2. (C) The unilateral injection of RGM A1 MO (20–30 ng) and RGM A2 MO (30 ng) led to abnormal developed or absent eyes (arrows) at stage 41, whereas a control MO had no effect. Dorsal (top row) and lateral (middle row) views are shown. Transverse vibratome sections (bottom row) confirmed these results. (D) Quantitative presentation of the eye phenotype. Results are the means±S.E.M. (E) Vibratome-produced sections of MO-injected embryos at stage 23. Upper row, whole embryos; lower row, close-up view. In control MO-injected embryos, the optic cups evaginate normally. In RGM A2- and neogenin-depleted embryos, the optic cup does not evaginate properly on the injected side (arrows

Download figure to PowerPoint

Due to the specific expression of RGM A in neural tissue, we were interested in its endogenous function during early neural development of X. laevis, which has not been investigated to date. Therefore, we injected the RGM A1 MO (20–30 ng), as well as the RGM A2 (30 ng) MO, unilaterally into one dorso—animal blastomere of eight-cell-stage embryos to target neural tissue (Moody, 1987). In a first set of experiments, we were interested in general morphological changes upon RGM A knockdown. At stage 41, the pigmented epithelium of the retina is clearly visible in control embryos. The injection of the RGM A1 MO, as well as the RGM A2 MO, resulted in severe morphological defects of the eye, encompassing abnormal (smaller and deformed) and absence of eyes in 72% (RGM A1 MO) and 80% (RGM A2 MO) of the embryos (Figures 2C and 2D). The control MO injection had no effect on eye development. Histological sections confirmed these observations (Figure 2C). As this phenotype can be observed with two independent, different MOs, we concluded that this MO-mediated knockdown is specific and that RGM A is required for proper eye development.

In addition, we performed histological analysis of MO-injected embryos at an earlier stage of development. At stage 23, the evaginating eye cups are clearly visible with the control MO-injected embryos. In contrast, evagination of the eye cup is severely impaired in RGM A-depleted embryos (Figure 2E), indicating that eye development is already disturbed at this stage.

RGM A is required for anterior neural development and NCC migration

Next, we aimed to investigate RGM A function in early neural development on a molecular level. For this purpose, we injected RGM A1/RGM A2 MOs unilaterally into eight-cell-stage embryos. At stage 13, down-regulation of RGM A1 had no influence on the expression of Rx1, Pax-6 and Sox-3 (Figure 3A, and data not shown). Furthermore, the knockdown of RGM A2 did not affect the induction of NCCs, as indicated by the expression of the marker genes Slug and FoxD3. These data indicate that RGM A function is not required for early segregation into the initial neuroectodermal domains.

image

Figure 3. RGM A is required for anterior neural and NC development

(A) Knockdown of RGM A did not have an effect on the induction of the eye-specific marker gene Rx1 and the pan-neural marker gene Sox-3 (RGM A1 MO) at stage (St.) 13, or on the induction of the NCCs, as indicated by the expression of marker genes Slug and FoxD3 (RGM A2 MO) at stage 17. (B) Unilateral depletion of RGM A2 resulted in a decreased expression of Rx1, Emx1, En2 and Krox20 at stage 23 (arrows). Unilateral injection of the control MO had no effect. (C) Unilateral loss of RGM A function led to defects in NCC migration, as shown by the expression of Slug and FoxD3 at stage 20, as well as Twist and Krox20 at stage 23 (arrows), whereas a control MO had no effect. (D) Quantitative presentation of the results shown in (B). (E) Quantitative presentation of the results shown in (C). Results are the means±S.E.M.

Download figure to PowerPoint

At stage 23, however, we observed a reduced expression of Rx1 (eye marker gene), Emx1 (forebrain), En2 (midbrain—hindbrain boundary) and Krox20 (rhombomeres 3 and 5) on unilateral injection of RGM A2 MO (Figure 3B). This phenotype was observed with different efficiencies, with the strongest effect on markers of the forebrain, whereas the effects were more modest on midbrain and hindbrain markers (Figure 3D), which is in agreement with the above-described expression of RGM A in these tissues. The same phenotypes were also observed after injection of the RGM A1 MO (data not shown).

In addition, reduction of RGM A2 function interfered with the migration of NCCs at stage 20, as well as at stage 23. Expression of NCC marker genes Slug, FoxD3, Twist and Krox20 (Figures 3C and 3E) is still detectable; however, these cells reside at the dorsal side and do not migrate ventrally. These data indicate that RGM A function is required for anterior neural development and migration of NCCs.

Down-regulation of RGM A affects cell proliferation

The observed down-regulation of anterior neural marker genes might either reflect an effect of RGM A on differentiation or be due to defects in proliferation or induced apoptosis. Indeed, depletion of RGM A2 resulted in slightly increased apoptosis in approx. 23% of the embryos on the injected side at stage 23, as indicated by TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick-end labelling) assay (Figures 4A and 4B). The down-regulation of RGM A2, in addition, was accompanied by a strong reduction of proliferating cells in the whole anterior neural tissue, as shown by BrdU (bromodeoxyuridine) assays (approx. 50%; Figures 4A and 4B). These data indicate that RGM A may function as a survival factor in anterior neural tissues.

image

Figure 4. RGM A2 is required for cell proliferation in the anterior neural region

(A) Cell apoptosis and proliferation were analysed using TUNEL and BrdU assays. Unilateral RGM A2 depletion resulted in an increase of cell apoptosis. In most embryos affected, the phenotype was modest, with only some apoptotic cells on the injected side (apoptosis, middle panel, arrow). The phenotype was strong with many apoptotic cells in only a few cases (apoptosis, right-hand panel, arrow). Loss of RGM A2 led to a strong decrease of cell proliferation in the anterior neural region on the injected side (proliferation, arrow). Control MO-injected embryos showed no phenotype. MO-injected sides were selected by GFP RNA co-injection (0.5 ng). (B) Quantitative presentation of results shown in (A). Results are the means±S.E.M.

Download figure to PowerPoint

Overexpression of RGM A affects anterior neural and NCC development

In a next set of complementary experiments, we investigated the influence of RGM A2 overexpression on anterior neural development. For this purpose, we unilaterally injected 2 ng of the RGM A2 mRNA and cultured the embryos until stage 41 and observed a strong eye defect in up to 80% (Figure 5A and Supplementary Figure S4 at http:www. biolcell.orgboc100boc1000659add.htm). In contrast, the injection of 2 ng of GFP (green fluorescent protein) had no effect on eye development. Surprisingly, these observations are similar to the loss-of-function phenotype.

image

Figure 5. RGM A2 overexpression results in eye defects

(A) Embryos that were unilaterally injected with RGM A2 RNA exhibited abnormally developed or absent eyes (white and black arrows). Dorsal (upper row) and lateral (lower row) views are given. Injection of GFP RNA served as control. (B) RGM A2 overexpression (2 ng) had no effect on Rx1 expression, whereas the expression domain of Sox-3 was expanded on the injected side (arrow) at stage 13. (C) The unilateral injection of 2 ng of RGM A2 RNA resulted in a decreased expression of the anterior neural marker genes Rx1, Emx1, En2 and Krox20 (arrows) in comparison with GFP RNA injection. (D) Quantification of the phenotype shown in (C). Results are the means±S.E.M. (E) Unilaterally RGM A2 RNA-injected embryos exhibited ectopic patches of Slug and FoxD3 at stage (St.) 17. At stage 20 and 23, the overexpression of RGM A2 led to a defect in NCC migration, as indicated by the expression of Slug, FoxD3, Krox20 and Twist (arrows), as well as ectopic expression of the same marker genes (arrowheads). GFP RNA injections served as controls.

Download figure to PowerPoint

Next, we focused on the influence of RGM A2 mRNA injection on the expression of neural and NCC marker genes. The injection of 2 ng of RGM A2 did not affect Rx1 expression at stage 13 (n=3, N=57, where n is the number of independent experiments and N is the number of analysed embryos in total; Figure 5B). However, we observed a dramatic expansion of the pan-neural marker gene Sox-3 in 54% of injected embryos (n=3, N=69). We then analysed the expression of neural marker genes at later stages and also tested different amounts of RGM A2 RNA to exclude toxic side effects. In all cases, overexpression of RGM A2 (0.1–2 ng) was accompanied by a strong decrease of the expression of Rx1, Emx1, En2 and Krox20 (Figures 5C and 5D, and Supplementary Figure S5 at http:www.biolcell. orgboc100boc1000659add.htm), which is similar to the loss of RGM A function phenotype. To investigate the effect of RGM A2 overexpression on NC development, we unilaterally injected 0.1–2 ng of RGM A2 mRNA and fixed the embryos at stages 17 (NCC induction), 20 and 23 (NCC migration). At stage 17, the overexpression of RGM A2 led to ectopical patches of Slug- and FoxD3-positive cells (Figure 5E and Supplementary Figure S5). At later stages, RGM A2-injected embryos showed defects in NCC migration, as well as ectopic expression of Slug and FoxD3 (stage 20) and Krox20 and Twist (stage 23; Figure 5E). In summary, these data suggest that overexpression of RGM A2 interferes with neural patterning, NCC induction and migration.

Neogenin is required for neural development and NCC migration

Previous studies demonstrated a functional interaction between RGM A and neogenin (Stahl et al., 1990; Matsanuda and Chedotal, 2004; Yamashita et al., 2007). We therefore investigated the spatial expression profile of neogenin and performed corresponding loss-of-function experiments. Neogenin can be detected in the neural tube, the migrating NCCs, the eyes and the somites (Figure 6A). This expression is consistent with the above-mentioned hypothesis. Next, we injected 10 ng of a well-characterized neogenin MO (Wilson and Key, 2006) unilaterally in eight-cell-stage embryos. At stage 41, neogenin-depleted embryos exhibited abnormal or absent eyes on the injected side (Figure 6B). At stage 13, down-regulation of neogenin expression did not interfere with establishment of the early eye field, as indicated by Rx1 expression or, more generally, neural induction, as indicated by Sox-3 expression (Figure 6C). Induction of NCCs was not affected as indicated by Slug and FoxD3 expression. However, the down-regulation of neogenin resulted in a decreased expression of Rx1 (57%; n=3, N=76), Emx1 (30%; n=3, N=70), En2 (20%; n=3, N=84) and Krox20 (12%; n=3, N=63) at stage 23 (Figure 6D). As in the case of RGM A loss-of-function, the strongest reduction was observed in forebrain marker genes. Furthermore, neogenin MO injection led to defects in NCC migration at stage 20 (Slug and FoxD3) and 23 (Krox20 and Twist) (Figure 6E). In addition, neogenin-depletion resulted in defects in eye-cup evagination (Figure 2E). These data indicate that a loss of the potential RGM A receptor neogenin reveals similar or identical phenotypes to the loss of its ligand RGM A.

image

Figure 6. Neogenin is required for eye, neural and NC development

(A) Neogenin expression during Xenopus development. At stage (St.) 20, neogenin is expressed in the neural tube (arrow). At stage 23, neogenin is additionally expressed in the eye (arrowhead) and migrating NCCs (arrow). At stage 31, neogenin expression is seen in the somites (arrow). (B) Unilateral injection of 10 ng of neogenin MO led to defects in eye development (arrow). Quantitative presentation of the results is given. Results are the means±S.E.M. (C) Loss of neogenin function did not interfere with the induction of Rx1 and Sox-3 (stage 13) and Slug and FoxD3 (stage 17). (D) Neogenin-depleted embryos exhibited a decreased expression of Rx1, Emx1, En2 and Krox20 (arrows) at stage 23. (E) At stage 20, depletion of neogenin resulted in an accumulation of Slug- and FoxD3-positive non-migrating NCCs at the dorsal side of the embryos (arrows). Unilateral injection of the neogenin MO led to a failure of NCC migration at stage 23, as shown by the expression of Krox20 and Twist (arrows).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. Funding
  9. References

Isolation of X. laevis full-length RGM A

In the present study, we isolated a X. laevis RGM A clone that differs in two aspects to RGM A sequences published previously (see Supplementary Figure S1). For members of the RGM family, it had previously been shown that a GPI anchor is added to the C-terminus. GPI-anchor sequences are poorly defined by sequence similarity, but rather consist of two amino-acid stretches, a hydrophobic element and a N-terminal element, which are thought to be the site of cleavage and GPI-anchor addition (Moran et al., 1991). In contrast with other organisms, the sequence published previously lacks this site for the addition of a GPI anchor (see Supplementary Figure S1). We were able to isolate a RGM A cDNA clone from X. laevis that contains this anchor site at the 3′ end. This is in agreement with the Xenopus borealis sequence, as found recently (Shin and Wilson, 2008). Furthermore, the isolated RGM A clone in the present study extends at the 5′ end in comparison with the sequence published previously, with a second putative start codon further upstream (see Supplementary Figure S1). Interestingly, this short sequence element is similar to the published X. tropicalis sequence (Supplementary Figure S1). We therefore concluded that sequence described in the present study represents a full-length RGM A clone. It needs to be mentioned, however, that the shorter RGM A1 sequence has a predicted start codon that is also found in other organisms (Supplementary Figure S1), which raises the question of which of the two start codons is used. Protein biochemical approaches, such as nanoelectrospray MS (Monnier et al., 2002) or Edman degradation (Niederkofler et al., 2004), are only able to define the N-terminus of the mature RGM A protein after removal of the signal peptide. As we observed, the same phenotype occurred with both MOs injected, therefore this suggests that the first start codon found in RGM A2 is used. As both versions differ at their 5′ end, this might reflect splice variants with no consequences in the mature protein.

Conserved expression of RGM A and its receptor neogenin in neural tissues

For functional analysis, we performed a detailed description of the RGM A expression. First, we showed a tissue-specific expression of RGM A in the anterior neural tissue at stage 12.5. RT-PCR analysis even indicated a maternal expression. These data are in agreement with previous findings showing RGM A expression that starts at the four-cell stage (Samad et al., 2005). In extension of this previous analysis, we showed that expression of RGM A at stage 18 is lower at the ventral midline and higher in prospective medial and dorsal aspects of the neural plate. Later in development, however, RGM A expression is confined to the ventral half of the neural tube. The dorsal- most part of the neural tube is devoid of RGM A transcripts. In mouse embryos, E8.5 (embryonic day 8.5) is the earliest examined stage in which RGM A is strongly expressed specifically at the tips of the neural folds (Niederkofler et al., 2004). Later on, RGM A is also expressed in the ventral region of the neural tube. Because of the strong expression at E8.5, however, we would expect that RGM A expression also starts earlier in mouse development, which requires further investigation. Also, during chicken embryogenesis at E9.5 and E10.5, RGM A is expressed in the ventral neural tube (Schmidtmer and Engelkamp, 2004). Early expression data for RGM A in chicken are not available. The early neural expression in X. laevis is the first indication that RGM A may possess additional functions during Xenopus embryogenesis, in addition to its function during axonal guidance in the forebrain, as shown previously (Wilson and Key, 2006).

Furthermore, we have examined the spatio-temporal expression of X. laevis neogenin, which is a potential RGM A receptor (Matsunaga and Chedotal, 2004). Neogenin transcripts are observed in the neural tube, the developing eye and migrating NCCs. This expression is partially conserved in zebrafish (Shen et al., 2002), whereas in mice and chicken the early expression is not well described (Gad et al., 1997; Keeling et al., 1997).

The results in the present study are in agreement with a potential function of neogenin and RGM A during early neural development and patterning, as discussed below.

RGM A function during neural development

The well-known functions of RGM A in a neural context include neural tube closure (Niederkofler et al., 2004), axonal pathfinding (Monnier et al., 2002) and synapse formation (Kyoto et al., 2007), as well as regulation of apoptosis (Shin and Wilson, 2008). In the present study, we focused on the function of this molecule during early anterior neural development. We were able to show that a loss of RGM A function results in eye defects that are probably due to earlier deficits in differentiation and morphogenetic movements. In addition, we detected defects in NCC migration. Unfortunately, we could not rescue the loss of RGM A function eye phenotype, as well as marker gene expression, by co-injecting different amounts of RGM A2 mRNA. For this purpose, we tested various amounts of the RGM A2 MO (15 and 30 ng), as well as RGM A2 mRNA (0.05–2 ng, data not shown). Since we observed similar phenotypes due to RGM A2 overexpression that might result in problems with finding the right balance between RGM A2 MO and RNA for rescue experiments. Alternatively, RGM A2 overexpression might induce an unrelated phenotype that prevents the rescue of the MO-induced phenotype, as discussed below. Nevertheless, we consider that the observed RGM A loss-of-function phenotype to be specific, as we observed identical phenotypes (morphological eye phenotype, early neural differentiation) for both RGM A MOs used. Similar, Reggiani et al. (2007) observed identical phenotypes for two different MOs for kidney-specific marker genes.

The observed eye phenotype could be explained by different mechanisms, such as disturbance of neural induction, neural differentiation, cell movements or cell proliferation/apoptosis. First, we investigated whether RGM A might play a role during neural induction or differentiation. Down-regulation of RGM A resulted in a decreased expression of neural marker genes at stage 23, whereas the expression of early neural markers at stage 13 was not affected, indicating that the molecular basis of the defect occurs between these stages. Second, we observed defects in morphogenetic movements during the stages of eye-cup formation, which also point towards a role of RGM A during these early stages and is consistent with the fact that expression of RGM A RNA in anterior neural tissue can be detected at stage 13 and then decreases until stage 18 in the eye region. It needs to be mentioned, however, that RGM A protein might still be present at these stages on the cell surface of the corresponding cells. Furthermore, it needs to be noted that the defects in eye-cup formation might be a secondary consequence of an earlier loss of neural marker gene expression or more general defects in forebrain development. Third, we examined whether a loss of RGM A function alters proliferation or results in induced apoptosis. RGM A inhibition leads to a slight increase in cell apoptosis. In addition, cell proliferation is strongly affected upon RGM A2 depletion. In agreement with the present study, in chicken embryos, knockdown studies showed that RGM A is required for cell survival in the neural tube (Matsunaga et al., 2004). In contrast, RGM A gain- and loss-of-function did not affect cell apoptosis and proliferation in the mouse brain (Niederkofler et al., 2004). And more recently, it has been shown that RGM A overexpression results in increased apoptosis during early development of X. borealis (Shin and Wilson, 2008). These data indicate that the influence of RGM A on proliferation and cell survival/apoptosis might be context-dependent and regulated through interaction partner molecules present in the cell type investigated. Whether RGM A has a more direct effect on differentiation awaits further investigation.

Potential molecular mechanisms of RGM A function

Our data raise the question regarding the molecular mechanism and intracellular pathway through which RGM A function is mediated. Previous studies have shown that RGM A can either interact with neogenin or can function as a BMP co-receptor (Supplementary Figure S6 at http:www.biolcell.orgboc100boc1000659add.htm) (Stahl et al., 1990; Matsanuda and Chedotal, 2004; Babitte et al., 2005, 2006; Samad et al., 2005; Halbrooks et al., 2007; Yamashita et al., 2007; Xia et al., 2007). Furthermore, in light of these findings, it is of interest that we observed similar phenotypes in the gain- and loss-of-function analyses for RGM A, and that RGM A overexpression results in ectopic expression of NCC marker genes. The similar phenotype was observed due to RGM B overexpression, as shown previously (Samad et al., 2005). Thus RGM A seems to have multiple roles during Xenopus development. We would like to propose that under physiological conditions RGM A interacts with neogenin, as loss of RGM A and neogenin function resulted in identical phenotypes. When RGM A2 was overexpressed, however, RGM A interfered with BMP signalling, as demonstrated by the expansion of Sox-3 expression and the observation of ectopic NCCs. BMP signalling has been shown previously to be involved in neural induction and the regulation of NC formation (Liu and Niswander, 2005). Finally, BMP signalling is involved in regulating NCC migration (Correia et al., 2007). The defects in NCC migration observed in the present study are in agreement with this hypothesis. Also, a role for regulating NCC migration has been shown for RGM B (Samad et al., 2005). The hypothesis that RGM A affects different intracellular signal transduction pathways under different experimental conditions (loss and gain of function) would satisfactorily explain the fact that the gain- and loss-of-function studies resulted in comparable phenotypes. Future experiments will uncover these pleiotropic effects of RGM A.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. Funding
  9. References

Embryo cultures

X. laevis embryos were generated and cultured by standard methods (Sive et al., 2000) and the stage was determined according to Nieuwkoop and Faber (1975).

Cloning of RGM A1 and RGM A2

Total RNA was isolated from stage 29/30 X. laevis embryos using the Gentra Purescript™ RNA isolation kit (Biozym) according to the manufacturer's instructions. RT-PCR was performed with Superscript II RNase H reverse transcriptase (Invitrogen) using random primers. Pfu polymerase (Promega) or Advantage DNA polymerase (Clontech), with proofreading activity, were used. All constructs were cloned into the pCR4Blunt-TOPO vector (Invitrogen). For further experiments, the constructs were subcloned into the pCS2+ vector using the EcoRI cleavage site. To clone the constructs, including the 5′ UTR and ORF of the two RGM A (RGM A1, accession number BC045008; RGM A2, accession number EU447275) constructs, the following primers were used: 5′ UTR RGMA1_l, 5′-CGC GGA TCC AGA AGC TGC TTC TTA GTA CAA GTT-3′; 5′ UTR RGMA1_r, 5′-GGG GAA TTC CTA TGG CCT TTC AAA CAA ATG GCA-3′; 5′ UTR RGMA2_l, 5′-CAC TGA CCC CCG TGT GAG CGC TGG GCT GGA-3′; 5′UTR RGMA2_r, 5′-GTT GGG AAT TTA CAC AGT AAA CAT AGA-3′. To generate constructs lacking the MO-binding site (Supplementary Figure S3 at http:www.biolcell.orgboc100boc1000659add.htm), Δ5′ UTR constructs were cloned with the following primers: Δ5′ UTR RGMA1_l, 5′-GGG GAA TCC ATG GGT ATG GGG AGA GGG GCA-5′; 5′ UTR RGMA1_r, 5′-GGG GAA TTC CTA TGG CCT TTC AAA CAA ATG GCA-3′; Δ5′ UTR RGM2_l, 5′-TTT AAA ATG CAG CCG ATA AGG GCG AAG GTT-3′; Δ5′ UTR RGM2_r, 5′-TTA CAG CAA CAT TGC ACA GCA CTG-3′. All constructs were verified by sequencing.

Characterization of RGM A structure

The signal peptide was predicted using the following program: http:www.cbs.dtu.dkservicesSignalP (Nielsen et al., 1999). The cleavage sites for the addition of a GPI anchor were calculated by using the big-PI Predictor program: http:mendel.imp.ac.atgpigpi_server.html (Eisenhaber et al., 1999).

MOs and RNA injections

MOs were obtained by Gene Tools (Philomath, OR, U.S.A.). For loss-of-function experiments we obtained two different MOs, which covered both possible start codons, with the following sequences: RGM A1 MO, 5′-CAT CCA TCC AGC TTG GGC TTT AAC C-3′; RGM A2 MO, 5′-TTA TCG GCT GCA TGA GCC CCT TCC C-3′. The neogenin MO was used as described previously (Wilson and Key, 2006). As a control MO the standard control MO from Gene Tools was utilized. MOs were resuspended in DEPC (diethyl pyrocarbonate)-treated water and injected unilaterally into a single dorso—animal blastomere at the eight-cell stage at amounts of 10 ng (neogenin MO), 20–30 ng (RGM A1 MO) and 30 ng (RGM A2 MO). The binding specificity of the RGM A MOs was tested using the in vitro-coupled transcription and translation (TNT) assay (Promega). For rescue experiments, 15–30 ng of RGM A2 MO and 0.05–2 ng RGM A2 mRNA were injected. For gain of function analysis, 0.1-2 ng of RGM A2 mRNA were injected unilaterally into one dorso—animal blastomere at the eight-cell stage.

RT-PCR

Total RNAs were isolated from X. laevis embryos at different stages and adult organs using the Purescript™ RNA isolation kit (Gentra) according to the manufacturer's instructions. Reverse transcription was performed with Superscript II RNase H reverse transcriptase (Invitrogen) using random primers. Primer sequences for RGM A were: RGMA_RT_l, 5′-GGT TCG AAA AAT GGT GGA GA-3′; RGMA_RT_r, 5′-AGG AGG TCA AAG ACG CAA GA-3′; with an annealing temperature of 55°C, which gave products of 399 bp for RGM A1 and RGM A2. Primers for the histone H4 positive control were: H4_l, 5′-CGG GAT AAC ATT CAG GGT ATC ACT-3′; H4_r, 5′-ATC CAT GGC GGT AAC TGT CTT CTT-3′, resulting in an amplification product of 175 bp. Annealing was carried out at 55°C.

WMISH

Spatial expression pattern of marker genes was analysed by WMISH using standard procedures (Hemmati-Brivanlou et al., 1990). For the spatial expression pattern of RGM A2, wild-type embryos at different stages were fixed with MEMFA [0.1 M MPOS (pH 7.4), 2 mM EGTA, 1 mM MgSO4 and 4% formaldehyde] 2 h at room temperature (21°C), then dehydrated in 100% methanol and stored at −20°C until further processing. Probes were generated using the full-length RGM A2 construct in pCS2+ with T7 RNA polymerase. For analysing the spatial expression of neogenin, a 1.5 kb fragment, including the 5′ UTR and a part of ORF derived from the published X. borealis sequence (accession number DQ173198), was cloned into the pSC-B vector (Stratagene) with following primers: Neo_WM_l, 5′-TTC CAT TAC TGA AGG TTT GCG-3′; Neo_WM_R, 5′-CCG GCA TAC GCC AAG TCA GCT-3′. Probes were generated with Hind III and T7 RNA polymerase. For analysing the expression of marker genes, embryos were injected unilaterally into a single animal—dorsal blastomere at the eight-cell stage and cultured until they reached indicated stages, and then fixed. After staining, the embryos were bleached in 30% H2O2 overnight at room temperature.

Histology

For histological analyses, the embryos were fixed in MEMFA and embedded in gelatin/albumin. Embryos were sectioned on a vibratome at thickness of 20 μm.

BrdU staining

Cell proliferation was detected with BrdU (Hardcastle and Papalopulu, 2000). At stage 23, 10 nl of BrdU (Roche) was injected bilaterally into the region next to the eyes. Over the following 2 h, BrdU was incorporated into newly synthesized DNA. Then the embryos were fixed in MEMFA and the incorporated BrdU was detected with the BrdU labelling and detection kit II (Roche), according to manufacturer's instructions. After fixation, embryos were bleached in 30% H2O2.

TUNEL staining

Whole-mount TUNEL staining of Xenopus embryos was carried out as described previously (Hensey and Gautier, 1999) using BM Purple as chromogenic substrate. This technique detects apoptotic cells using the enzyme TdT (terminal deoxynucleotidyl transferase) (Invitrogen) to directly label the ends of broken DNA strands with dioxygenin—dUTP (Roche). Embryos were fixed in MEMFA for 2 h at room temperature and dehydrated in 100% methanol. After rehydration in PBST [130 mM NaCl, 7 mM Na2HO4, 3 mM NaH2PO4 and 0.1% Tween (pH 7)], embryos were treated with proteinase K (Roche) in PBST, similar to the WMISH protocol. Embryos were incubated in 1× TdT buffer (Invitrogen) for 1 h, followed by the overnight TdT reaction at room temperature (0.5 μM dioxygenin—dUTP and 150 units/ml TdT). Next day, the TdT reaction was stopped by treatment by two washes with PBS/1 mM EDTA at 65°C for 30 min. Subsequently, the embryos were washed with PBST and treated with anti-digoxigenin—AP–Fab fragments (Roche), similar to the WMISH protocol. The following day, the chromogenic reaction was performed with BM Purple. To better visualize the TUNEL staining, embryos were bleached in 30% H2O2. Embryos were considered to have an apoptotic phenotype when a comparison between both sides revealed at least a 3-fold difference in TUNEL-positives cells.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. Funding
  9. References

We thank Doris Weber for technical assistance, and T. Pieler (Institut für Biochemie und Molekulare Zellbiologie, Georg-August-Universität Göttingen, Göttingen, Germany), T. Hollemann (Institut für Physiologische Chemie, University of Halle-Wittenberg, Halle/Saale, Germany), R. Harland (Department of Molecular and Cell Biology, University of California, Berkeley, CA, U.S.A.), M. Sargent (National Institute for Medical Research, London, U.K.) and D. Wedlich (Institut für Zoologie II, Universität Karlsruhe, Karlsruhe, Germany) for providing plasmids.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. Funding
  9. References

This work was supported by the Deutsche Forschungsgemeinschaft (SFB 497, Tp A6).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgments
  8. Funding
  9. References