Biochimica et Biophysica Acta 1763 (2006) 1184–1191www.elsevier.com/locate/bbamcr

Review

Calcium transients and calcium signalling during early neurogenesis in theamphibian embryo Xenopus laevis

Catherine Leclerc a,⁎, Isabelle Néant a, Sarah E. Webb b, Andrew L. Miller b, Marc Moreau a

a Centre de Biologie du Développement, UMR 5547 et GDR 2688, Université Paul Sabatier, 118 route de Narbonne 31062 Toulouse, Franceb Calcium Aequorin Imaging Laboratory, Department of Biology, HKUST, Clear Water Bay, Kowloon, Hong Kong, PR China

Received 6 July 2006; accepted 8 August 2006Available online 10 August 2006

Abstract

Development of the vertebrate embryonic nervous system is characterized by a cascade of signalling events. In Xenopus, the initial step in thiscascade results from signals emanating from the dorsal mesoderm that divert the fate of the ectoderm from an epidermal to a neural lineage. Thesesignals include extracellular antagonists of the bone morphogenetic protein (BMP). Experiments performed with isolated ectoderm suggest thatepidermis is induced by BMP,whereas neural fates arise by default following BMP inhibition; however, we show that this mechanism is not sufficientfor neural determination. Ca2+ imaging of intact Xenopus embryos reveals patterns of Ca2+ transients in the dorsal ectoderm but not in the ventralectoderm. These increases in intracellular calcium concentration ([Ca2+]i), which occur via the activation of dihydropyridine (DHP)-sensitive Ca2+

channels, are necessary and sufficient to orientate the ectodermal cells toward a neural fate. On the one hand, the treatments that antagonize theincrease in [Ca2+]i, inhibit neuralization, while on the other hand, an artificial increase in [Ca

2+]i, whatever its origin, neuralizes the ectoderm. Usingthese properties, we have constructed a subtractive cDNA library between untreated ectoderm and caffeine-treated ectoderm. The caffeine stimulatesan increase in [Ca2+]i and thus orientates the cells towards the neural pathway. We have identified early Ca2+ target genes expressed in neuralterritories. One of these genes, an arginine methyl transferase, controls the expression of the early proneural gene, Zic3. Here, we discuss analternativemodel where Ca2+ plays a central regulatory role in early neurogenesis. Thismodel integrates the activation of a Ca2+-dependent signallingpathway due to an influx of Ca2+ throughDHP-Ca2+ channels.While Ca2+ is required for neural determination, epidermal determination occurs whenCa2+-dependent signalling pathways are inactive.© 2006 Elsevier B.V. All rights reserved.

Keywords: Calcium; DHP-channel; Neural determination; Xenopus laevis; Gene expression

1. Introduction

In the amphibian, the formation of the nervous system isinitiated very early in development (9 h post fertilization), duringgastrulation, with a process called neural induction. This is thefirst phase in neurogenesis. Neural induction in vertebrate spe-cies may exhibit a degree of morphological and signalling simi-larities, however, only a few species have been studied in greatdetail. Most current information regarding Ca2+ signallingduring neural induction comes from amphibian development,and this shall be used as amodel to discuss the phenomenon. Theclassical experiments of Spemann and Mangold in 1924 [1],using the urodele amphibian model system, show that neural

⁎ Corresponding author.E-mail address: leclerc@cict.fr (C. Leclerc).

0167-4889/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.bbamcr.2006.08.005

induction involves an interaction between the dorsal mesodermand the ectoderm, which leads to a diversion of the epidermallineage towards the neural fate. During this process, the ecto-derm of the embryo becomes regionalized to form the highlyspecialized and interconnected regions found later in the adultnervous system [2]. The cells develop in defined temporal andspatial patterns as a result of several classes of signalling mole-cules and a precise local control of gene expression. Thus, im-mature ectoderm cells are faced with a series of binary choices,the first of which is to become an epidermal or a neural cell.

In the last 10 years it has been suggested that neurogenesisresults from the opposing action of ventralizing signals (e.g.Bone Morphogenetic Proteins, BMP) coming from the ecto-derm, and dorsalizing signals from the dorsal mesoderm (e.g.noggin, chordin, follistatin, XnR3 and Cerberus) (reviewed by[3]). However, there is increasing evidence to suggest that

Fig. 1. Neuralization by dissociation of ectodermal cells is associated with a Ca2+

signal. (A) Recording of internal Ca2+ concentration during dissociation of Xe-nopus ectodermal cells in medium free of divalent ions (filled circles) and afterpreincubation of the ectodermal cells with the membrane permeant form ofBAPTA (open circles, BAPTA-AM, 0.4 mM in the external medium). Tomeasure Ca2+, the cells were previously loaded with aequorin as previouslydescribed [13]. Arrow indicates substitution of normal saline medium by Ca2+—Mg2+-free medium. (B) Expression of the pan-neural marker NCAM in animalcaps was measured by RT-PCR. Dissociated caps differentiated into neural cellsexpressing NCAM. BAPTA-AM loaded caps before dissociation (dissociated+BAPTA) show a dramatic reduction in NCAM expression. Intact cap: animalcaps not dissociated; sibling control embryo (stage 20; 21 h. post-fertilization)served as positive control (embryo) and PCR on the same RNAwithout reversetranscription was done to check the absence of genomic DNA (−RT). Ornithinedecarboxylase (ODC) gene is used as control.

1185C. Leclerc et al. / Biochimica et Biophysica Acta 1763 (2006) 1184–1191

antagonizing BMP signalling is not sufficient to explain neuralinduction and that FGF signalling is also required [4,5].

In this review, we will outline our hypothesis that anothersignalling pathway, involving transient rises in intracellularcalcium concentration ([Ca2+])i, controls the binary choice (i.e.,epidermis or neural tissue) of determination. We will describe(i) the Ca2+ signals that occur during neural induction and theiressential characteristics; (ii) a possible downstream effector thatfits into the signalling cascades involved; and (iii) a new modelto explain the role of Ca2+ in neural induction and thus re-evaluate the concept of “by default” neural induction.

1.1. Ca2+ is involved in the choice between neural andepidermal fate

The several layers of ectoderm cells above the blastocoel inlate blastula- or early gastrula-stage embryos are called theanimal cap. Animal caps can be dissected and kept for severaldays in physiological medium. These cells are multipotent andexhibit plasticity where, without the addition of inducing factors,they give rise to epidermal cells (i.e., atypical epidermis). Whenappropriate neural inducers such as noggin are added to theculture medium, animal cap cells express a variety of neuralmarkers. This assay is therefore a good model to estimate theneural inducing activity of activators or inhibitors.

As early as 1964, Barth and Barth [6] suggested that Ca2+ isimportant to trigger neuralization inRana pipiens ectoderm.Morerecently, the dissociation of animal caps in Ca2+ and Mg2+-freemedium has been shown to orientate the cells toward a neural fate[7–10]. Frequently, the explanation given for this neuralization bydissociation phenomenon is that the epidermal inductors (BMPs)are being diluted from their receptors. This theory was taken asfact for many years because it was reported that when BMP4 isadded at a high concentration during dissociation, the expressionof neural markers is totally abolished [11]. Recently, it has beenshown that cell dissociation induced a sustained activation of theRas/MAPKpathway,which causes the phosphorylation of Smad1at sites that inhibit the activity of this transcription factor. Thisdemonstrates that BMP ligands continue to signal in dissociatedcells [12]. In addition, we have shown that the dissociation ofanimal caps in Ca2+-free medium triggers an increase in [Ca2+]i(Fig. 1A). This increase is due to an efflux of Ca2+ from internalstores resulting from the inversion of the gradient of concentrationin Ca2+ between intra- and extracellular compartments [13]. Todiscriminate whether during cell dissociation, the increase in[Ca2+]i is a cause or a consequence of neural induction, we haveloaded animal cap cells with the Ca2+ chelator BAPTA. Underthese conditions, the neuralization by dissociation is blocked (i.e.,the neural marker NCAM is not expressed, Fig. 1B and Ref. [13]).This demonstrates that, at least in animal caps, a Ca2+-dependentsignal is necessary to trigger neuralization of the ectoderm and toinhibit epidermal determination.

1.2. DHP-sensitive Ca2+ channels and neural induction

The neuralizing protein noggin has been shown to be abinding partner of BMPs and an antagonist of BMP signalling

[14]. Addition of the neural inducer noggin to animal caps fromamphibians (the urodele, Pleurodeles or the anuran, Xenopus)triggers an increase in [Ca2+]i. This increase has a duration of10–20 min and represents a rise to about 15% above that of theresting level of [Ca2+]i [15,16]. It is completely inhibited byantagonists of dihydropyridine (DHP)-sensitive Ca2+ channels,such as nifedipine or nimodipine, thus indicating an influx ofCa2+ from an external source.

The animal caps directly stimulated by specific agonists ofDHP-sensitive Ca2+ channels, such as S(−)Bay K 8644, present atransient increase in [Ca2+]i of 20 min duration. This increase issufficient, even in an active BMP context, to trigger not only theexpression of neural markers but also the formation of neurons andglial cells [16]. Conversely, the blockade of DHP-sensitive Ca2+

channels inhibits neural induction [17]. In addition, methyl-xanthines such as caffeine or theophyline, which are known toreleaseCa2+ from internal stores, are potent neural inducers [16,18].These experiments suggest the crucial role played by Ca2+ since,whatever its provenance, it triggers neuralization of the ectoderm.

Fig. 2. Photometric measurements of fluorescence changes (F/F0) of the Ca2+

indicator fluo-3, revealing an elevation of [Ca2+]i in Pleurodeles ectodermalcells. Neuralization of competent (blastula) animal caps by the lectin Con A(300 μg/mL) is associated with an increase in intracellular Ca2+. (Inset)Competent animal caps treated with succinyl-Con A (S-ConA, 300 μg/mL), alectin with no neural-inducing activity.

Fig. 3. Example of Ca2+ transients that occur in the dorsal ectoderm from arepresentative embryo at stage 10.5 (11 h.p.f.). (A) Lucifer yellow generatedfluorescent image of the embryo showing the relative position of the Ca2+

transient (green circle) with the position of the blastopore lip (dashed line). (B)Sequence taken every 10 s to show the appearance of a Ca2+ transient startingat the anterior dorsal ectoderm. Each panel represents the aequorin-generatedphoton image accumulated for 120 s, data were acquired by using a custom-designed Photon Imaging Microscope (PIM, Sciences Wares, USA, [50]).Colour scale indicates luminescent flux in photons/pixel. Scale bar, 0.3 mm.

1186 C. Leclerc et al. / Biochimica et Biophysica Acta 1763 (2006) 1184–1191

Early work, on neural induction in amphibians, identified thelectin concanavalin A (ConA) as a potent neural inducer whenadded to animal caps of both anurans and urodeles. So far, it isthe only lectin known to have inducing activity [19–21]. Inaddition, ConA has been shown to bind to Ca2+ channels andtrigger their activation [22]. We found that an increase in [Ca2+]ioccurs during ConA-stimulated neural induction in Pleurodeleswaltl animal caps (Fig. 2). In addition, we demonstrated thatspecific DHP-sensitive Ca2+ channel antagonists inhibit thisConA-induced neural induction. Conversely, non-inducing lec-tins did not raise [Ca2+]i [16].

In amphibian embryos, the neural competence, i.e., theability of the ectodermal cells to be induced along the neuralpathway, is acquired shortly before gastrulation and lost duringlate gastrula stages [23]. Both protein kinase C (PKC) andG-protein pathways have been proposed to affect the ectodermalcompetence [24–27]. However, PKC or G-protein alone areunable to induce neural tissue. We have demonstrated in bothPleurodeles and Xenopus, that the appearance of DHP-sensitiveCa2+ channels is correlated with the acquisition of neuralcompetence by ectodermal cells [28,29]. The highest density ofthese channels is reached when competence of the ectoderm isoptimal. Conversely, a decrease in the density of DHP-sensitiveCa2+ channels occurs simultaneously with the normal loss ofcompetence [29]. We show that these channels are functionaland we propose that the molecular basis of the gain or loss ofneural competence is linked to the presence of DHP-sensitiveCa2+ channels in ectodermal cells.

1.3. Imaging Ca2+ transients during neural induction in intactamphibian embryos

Using the photoprotein aequorin and a very sensitive tech-nique of Ca2+ imaging based on positional photon counting

[30], we have visualized Ca2+ movements in Xenopus ecto-dermal cells. The onset of Ca2+ signalling activity occurs at theblastula stage (stage 8), long before the start of gastrulation (i.e.,before mesoderm invagination) and the Ca2+ transients are loca-lized in the most anterior part of the dorsal ectoderm (Fig. 3).Through these observations, we show that neural induction startsearlier than previously thought, and this fact, therefore, needs tobe taken into account when considering the initiation of neuralinduction. If the secreted BMP antagonist, noggin, is alsoresponsible for the Ca2+ transients observed in whole embryos,

1187C. Leclerc et al. / Biochimica et Biophysica Acta 1763 (2006) 1184–1191

then these observations favour the model of planar inductiondescribed by Keller [31]. We have also previously confirmedthat the patterns of Ca2+ transients visualized in equivalentex vivo animal caps are similar to those from whole embryos[34].

A recent study by the De Robertis group, demonstrated thatneural induction requires the combined activity of the Nieu-wkoop Center and the Blastula Chordin and Noggin-expressingcenter (BCNE) located in dorsal animal cells. The BCNEcontains the prospective neuroectoderm and Spemann organizerprecursor cells. By means of different grafting experiments, theauthors demonstrated that the BCNE is required for brainformation [32]. We propose that the Ca2+ transients observed inthe dorsal ectoderm at the blastula stage might well be localizedin the BCNE. Ca2+ transients are therefore the first directlyvisualized event linked to neural induction.

As gastrulation proceeds, the Ca2+ transients increase both innumber and intensity, to reach a peak activity by mid-gastru-lation (stage 11–11.5), a stage where neural determination isthought to have occurred. This activity was found to be strictlyrestricted to the dorsal ectoderm (i.e., the tissue where neuralinduction takes place) and never occurred in the ventral ecto-derm cells, which do not receive neural inductive signals. Theincrease in [Ca2+]i is determinant since neural induction in vivowas totally blocked when the embryos were incubated with thepotent Ca2+ chelator, BAPTA, or treated with specific anta-gonists of DHP-sensitive Ca2+ channels. Under these latter con-ditions, the embryos lacked anterior brain structures [13,16,17].The phenotype observed is similar to the one obtained when theBCNE is removed [32].

Studying the interaction between Ca2+ signalling duringneural induction in intact embryos is complicated by the fact thatthe whole embryo is a three-dimensional structure where thedorsal ectoderm receives both planar and vertical signals fromthe mesoderm. In an attempt to simplify the experimental model,Keller open-face explants were used [33] as a two-dimensionalsystem to study neural induction. In these explants, only planarsignals pass from the mesoderm to the ectoderm. These signalshave been shown to be sufficient to reproduce many aspects ofneural induction observed in vivo, such as the expression ofneural marker genes, neuronal differentiation, and the inductionof a regionalized neural plate along the antero-posterior axis. Inthis model, Ca2+ transients start from themost anterior part of theopen-face explant and are not exclusively generated at theboundary between the ectoderm andmesoderm (Fig. 4 and [34]).

1.4. What are the Ca2+-sensitive target genes?

Ca2+ has been shown to regulate the expression of immediateearly genes such as c-fos [35]. We have shown that in animalcaps, Ca2+ controls the expression of c-fos [36] and of two othertranscription factors: XlPou2 and Zic3. While Fos is a ubiqui-tous transcription factor, XlPou2 and Zic3 are not. These twotranscription factors, which belong to the POU domain and odd-paired domain families of transcription factors, respectively, arespecific for neural determination [37,38]. They are primaryneural regulators. Our experiments, performed on isolated Xe-

nopus ectoderms, show that XlPou2 is induced very early afteran increase in [Ca2+]i. The mRNA for XlPou2 is detectable asearly as 30min after the [Ca2+]i increase is observed. In addition,specific antagonists of DHP-sensitive Ca2+ channels block theexpression of XlPou2 in response to noggin in animal caps, anddramatically reduce the expression of Zic3 in the whole embryo[39]. In planar explants, the accumulated Ca2+ pattern correlateswith the expression of the early neural precursor gene, Zic3.Furthermore, when the Ca2+ transients were blocked with theDHP-sensitive Ca2+ channel antagonist, nifedipine, the level ofZic3 expression was dramatically reduced [34]. Altogether, ourresults suggest that the increase in [Ca2+]i that occurs duringneural induction in the dorsal ectoderm can create domains ofelevated [Ca2+]i. These domains could activate genes with pro-neural activity.

The molecular mechanism by which Ca2+ orientates the cellstowards a neural fate remains poorly understood. To identifynew Ca2+ target genes involved in neural induction, we con-structed a subtractive cDNA library between untreated (i.e.ectodermal) and short duration (i.e., 15–45min) caffeine-treated(i.e. neuralized) animal caps, in order to release Ca2+. Caffeinetreatment, which triggers neural induction through an increase in[Ca2+]i [16], allows the differential isolation of the earliest Ca

2+-dependent genes involved in neural determination [40]. Morethan 30 early genes, which were shown to be controlled by Ca2+

and expressed in presumptive neural territories, were selected.The function of one of these genes, the Xenopus homologue ofthe mammalian arginine methyltransferase PRMT1, calledxPRMT1b [15], has been studied in further detail. In animalcaps, the expression of xPRMT1b is an early response to anincrease in Ca2+ that does not require de novo protein synthesis.Its expression is also triggered following the application ofnoggin or by the inhibition of BMP signalling with tBR (a nonfunctional form of BMP4 receptor). These effects are speci-fically blocked by the Ca2+ chelator, BAPTA. In the wholeembryo, xPRMT1b is expressed in neural territories. The earlyexpression of xPRMT1b at the gastrula stage also occurs througha Ca2+-dependent mechanism that is mediated by the activa-tion of DHP-sensitive Ca2+ channels. The overexpression ofxPRMT1b in neural territories activates the expression of theneural precursor gene Zic3. In addition, an antisense approach,with a morpholino oligonucleotide against xPRMT1b, blocksthe expression of neural markers induced by an increase in Ca2+

such as Zic3 in animal caps, and impairs anterior neural deve-lopment in the whole embryo [15]. Identical phenotypes areobtained with antagonists of DHP-sensitive Ca2+ channels [39],or when the BCNE is deleted [32]. These results suggest that theenzyme xPMT1b is a direct link between [Ca2+]i increase anddownstream events involved in neural induction.

2. Discussion

2.1. Does the role of Ca2+ fit into the generally accepted ideasabout neural induction?

The acquisition of a neural fate has long been considered to bea permissive event, which only requires the inhibition of BMP

Fig. 4. Representative example of the Ca2+ transients observed in the same open-face explant for approximately 4.5 h starting at early gastrula (9 h.p.f.). Eachcolour panel represents the aequorin-generated photon image accumulated for 240 s. White dashed lines outline the shape of the explant in each photon imageand the yellow circles/ellipses highlight the location of the Ca2+ events. (A) (a) Brightfield image of the explant 1 h after the start of imaging. (B) (a) Brightfieldimage of the explant 4 h after the start of imaging. Ect.=ectoderm, Mesod.=mesoderm. Colour scale indicates luminescent flux in photons/pixel. Scale bar,100 μm.

1188 C. Leclerc et al. / Biochimica et Biophysica Acta 1763 (2006) 1184–1191

signalling. This idea lead to the notion of the “by default me-chanism”, which is prominent in the field today [41]. This “bydefault” model has allowed us to understand some of the pro-cesses at the molecular level, (i.e., for early neurogenesis andepidermal determination), however, a number of importantquestions remain unsolved. Particularly, the “by default” modelconflicts with data from chick and ascidian embryos, whichindicate that neural induction is initiated by FGF signalling in a

partly BMP-independent manner [5,42]. For example, the “bydefault” model has difficulty explaining the inhibition of neu-ralization (triggered by noggin) that occurs in isolated ectodermthat expresses truncated forms of FGF receptors [43]. In addition,in intact Xenopus embryos the group of Kodjabachian [4], hasshowed that BMP inhibition is required but not sufficient forneural induction and that pre-gastrula FGF signalling is requiredin the ectoderm for the emergence of neural fates.

Fig. 5. New model for neural induction in Xenopus laevis embryo. In this model,Ca2+ plays a central role, by directly activating Ca2+ target genes such asxPRMT1b, which in turn control neural gene transcription either directly or viathe activation of a Ca2+ calmodulin kinase type II (see text for more details). TheCa2+ signals may also inhibit the BMP signalling pathway by activatingcalcineurin, which prevents the phosphorylation of Smads. The activation of theDHP-Ca2+ channels is performed by an as yet unidentified mechanism at thelevel of the binding of BMP by neural inducers. In this model, neuralization isnot a permissive process, but an instructive mechanism, as in epidermalinduction.

1189C. Leclerc et al. / Biochimica et Biophysica Acta 1763 (2006) 1184–1191

Finally, a number of observations considered in this reviewindicate that an increase in [Ca2+]i is a necessary and sufficientevent to turn the fate of ectoderm cells from epidermis to neuraltissue. These results imply a permissive role played by Ca2+.The identification and functional characterization of new Ca2+

target genes, such as xPRMT1b, will help us to make the linkbetween Ca2+ influx and neural determination.

2.2. Control of the choice of determination: a new model

One of the important unsolved questions raised by our data,concerns the mechanism by which DHP-sensitive Ca2+ channelsare activated during gastrulation in the dorsal ectoderm. Thisquestion is linked to another one: how can noggin produce aninflux of Ca2+ through DHP-sensitive Ca2+ channels? In thisrespect, it is of interest to further consider the relationshipbetween noggin and the FGF receptor. For example, in severalmodels such as in chick embryonic neurons, it has beendemonstrated that during cell proliferation, activation of the FGFreceptor stimulates the release of arachidonic acid and itsmetabolites, which in turn activates a Ca2+ influx [44,45].

Another question is why Ca2+ signals start in the anterior partof the ectoderm, since the inducing signal has been proposed tobe diffusing molecules secreted by the dorsal mesoderm. Thisproblem is more puzzling in open-face explants, since they arekept flat under a coverslip. Under these conditions, neuralinduction depends only on the planar signals generated by themesoderm; however, we observed that the Ca2+ transients occurat first in the anterior part of the planar explant and notexclusively at the border between mesoderm and ectoderm [34].This anterior region of Ca2+ transients at the blastula stagemight correspond to the Blastula Chordin and Noggin-expressing center. The presence of neural inducers, such asnoggin, in the blastula ectodermal precursor cells [32] and theevidence that noggin activates DHP-sensitive Ca2+ channels inanimal caps [15,36] support this hypothesis.

2.3. How might Ca2+ up-regulate neural genes?

Control of gene expression by Ca2+ very often involveschanges in the transactivating properties of transcription factorsafter the induction of Ca2+-dependent kinases and phosphatases.Activation of Ca2+ calmodulin kinase II (CaMkinase II) has beenreported to occur during the transcriptional activation ofimmediate early genes such as c-fos [35]. In isolated ectodermfrom the newt, Pleurodeles waltl, a Fos-related protein is up-regulated during neural induction. This up-regulation ismediated by the activation of DHP-sensitive Ca2+ channels.The expression of this Fos-related protein is dependent on theactivation of a CaMkinase, blocked by KN62, which in turnphosphorylates the Ca2+/cAMP response element binding pro-tein (CREB) [36].

In addition, direct effectors of Ca2+-induced gene expressionhave been suggested to exist in the nucleus. An important cis-regulatory element, DRE (Downstream Regulatory Element)can be implied in what can be called excitation- or stimulation-transcription coupling mechanisms [46]. This sequence, located

downstream from the TATA box, is the target of the DRE anta-gonist modulator (DREAM), an EF-hand Ca2+-binding proteinof the recoverin subfamily which, in the absence of Ca2+, bindsto the DRE site and represses transcription. To date, DREAM isthe only Ca2+ sensor known to bind specifically to DNA anddirectly regulate transcription in a Ca2+-dependent manner.

2.4. How might Ca2+ down-regulate epidermal orBMP-dependent genes?

So far, the most intensively used approach for inhibitingBMP signalling has been via the action of BMP antagonists(i.e., noggin and chordin), which prevent BMP4 frominteracting with its receptor. An alternative approach is to actdownstream and to block either BMP receptor activation orSmad phosphorylation. The spatial distribution of activatedSmad1 (phosphorylated Smad1) has been reported to change atthe onset of gastrulation. Prior to gastrulation, the pattern ofphosphorylated Smad1 is equally distributed between the dorsaland ventral sides of the embryo, reflecting an even activation ofthe BMP4 signalling pathway. In contrast, at the late blastula

1190 C. Leclerc et al. / Biochimica et Biophysica Acta 1763 (2006) 1184–1191

stage, Smad1 phosphorylation is enriched in the ventral side andby early gastrulation most of the activated Smad1 is localized tothe ventral side [47]. This correlates with the pattern of Ca2+

increase, which starts in the dorsal ectoderm at the blastula stageand is at a maximum at mid-gastrulation [39]. One can hypo-thesize that the dephosphorylation of Smad1 in the dorsal ecto-derm during gastrulation is controlled by calcineurin, a Ca2+/calmodulin dependent phosphatase 2B. Xenopus calcineurin isexpressed throughout early development [48]. Furthermore,injection of constitutively active mouse calcineurin in a ventrallocation, produces a double axis [49].

Finally, we propose a new model of neural induction tomodulate the concept of the “by default”mechanism. This modelintegrates the activation of a Ca2+-dependent signalling pathwaydue to an influx of Ca2+ through DHP-sensitive Ca2+ channels.While Ca2+ is required for the activation of neural specific genes,epidermal determination occurs when the Ca2+-dependentsignalling pathway is inactive (Fig. 5).

Acknowledgments

This workwas supported by: Centre National de la RechercheScientifique (CNRS); a joint PICS grant funded by the CNRS;the PROCORE France/Hong Kong Joint Research Scheme(F-HK98/99.SC06) sponsored by the Research Grants Councilof Hong Kong and the Consulate General of France in HongKong; Association pour la Recherche sur le Cancer (ARC);and the Research Grants Council of Hong Kong awardsHKUST6279/03M and HKUST6241/04M. We also thankDr. Osamu Shimomura for his generous contribution to aequo-rin-based imaging over the years.

References

[1] H. Spemann, H. Mangold, Über die induktion von embryonalanlagendurch implantation artfremder organisatoren, Wihlem Roux's Arch. Entw.Mech. Org. 100 (1924) 599–638.

[2] S.F. Gilbert, L. Saxen, Spemann's organizer: models and molecules, Mech.Dev. 41 (1993) 73–89.

[3] Y. Sasai, E.M. De Robertis, Ectodermal patterning in vertebrate embryos,Dev. Biol. 182 (1997) 5–20.

[4] E. Delaune, P. Lemaire, L. Kodjabachian, Neural induction in Xenopusrequires early FGF signalling in addition to BMP inhibition, Development132 (2005) 299–310.

[5] C.D. Stern, Neural induction: old problem, new findings, yet morequestions, Development 132 (2005) 2007–2021.

[6] L.G. Barth, L.J. Barth, Sequential induction of the presumptive epidermisof the Rana pipiens gastrula. Biol. Bull. 127 (1964) 413–427.

[7] H. Grunz, L. Tacke, Neural differentiation of Xenopus laevis ectodermtakes place after disaggregation and delayed reaggregation withoutinducer, Cell Differ. Dev. 28 (1989) 211–217.

[8] J.P. Saint-Jeannet, F. Foulquier, C. Goridis, A.M. Duprat, Expression ofN-CAM precedes neural induction in Pleurodeles waltl (urodele,amphibian), Development 106 (1989) 675–683.

[9] J.P. Saint-Jeannet, S. Huang, A.M. Duprat, Modulation of neuralcommitment by changes in target cell contacts in Pleurodeles waltl, Dev.Biol. 141 (1990) 93–103.

[10] J.P. Saint-Jeannet, F. Pituello, S. Huang, F. Foulquier, A.M. Duprat,Experimentally provoked neural induction results in an incompleteexpression of neuronal traits, Exp. Cell Res. 207 (1993) 383–387.

[11] P.A. Wilson, G. Lagna, A. Suzuki, A. Hemmati-Brivanlou, Concentration-

dependent patterning of the Xenopus ectoderm by BMP4 and its signaltransducer Smad1, Development 124 (1997) 3177–3184.

[12] H. Kuroda, L. Fuentealba, A. Ikeda, B. Reversade, E.M. De Robertis,Default neural induction: neuralization of dissociated Xenopus cells ismediated by Ras/MAPK activation, Genes Dev. 19 (2005) 1022–1027.

[13] C. Leclerc, C. Rizzo, C. Daguzan, I. Neant, J. Batut, B. Auge, M. Moreau,Neural determination in Xenopus laevis embryos: control of early neuralgene expression by calcium, J. Soc. Biol. 195 (2001) 327–337.

[14] L.B. Zimmerman, J.M. De Jesus-Escobar, R.M. Harland, The Spemannorganizer signal noggin binds and inactivates bone morphogenetic protein4, Cell 86 (1996) 599–606.

[15] J. Batut, L. Vandel, C. Leclerc, C. Daguzan, M. Moreau, I. Neant, TheCa2+-induced methyltransferase xPRMT1b controls neural fate in am-phibian embryo, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 15128–15133.

[16] M. Moreau, C. Leclerc, L. Gualandris-Parisot, A.M. Duprat, Increasedinternal Ca2+ mediates neural induction in the amphibian embryo, Proc.Natl. Acad. Sci. U. S. A. 91 (1994) 12639–12643.

[17] C. Leclerc, C. Daguzan, M.T. Nicolas, C. Chabret, A.M. Duprat,M. Moreau, L-type calcium channel activation controls the in vivotransduction of the neuralizing signal in the amphibian embryos, Mech.Dev. 64 (1997) 105–110.

[18] C. Leclerc, M. Moreau, L. Gualandris-Parisot, G. Drean, S. Canaux, A.M.Duprat, An elevation of internal calcium occurring via L-type channelsmediate neural induction in the amphibian embryo, in: N. Zagris, A.M.Duprat, J. Durston (Eds.), Organisation of the Early Vertebrate, PlenumPress, New York, 1995, pp. 209–226.

[19] L. Gualandris, P. Rouge, A.M. Duprat, Membrane changes in neural targetcells studied with fluorescent lectin probes, J. Embryol. Exp. Morphol. 77(1983) 183–200.

[20] L. Gualandris, P. Rouge, A.M. Duprat, Target cell surface glycoconjugatesand neural induction in an amphibian, J. Embryol. Exp. Morphol. 86(1985) 39–51.

[21] K. Takata, K. Yamamoto, R. Ozawa, Use of lectins as probes for analyzingembryonic induction, Roux's Arch. Dev. Biol. 190 (1981) 92–96.

[22] D.A. Greenberg, C.L. Carpenter, R.O. Messing, Lectin-induced enhance-ment of voltage-dependent calcium flux and calcium channel antagonistbinding, J. Neurochem. 48 (1987) 888–894.

[23] P. Nieuwkoop, A. Johnen, B. Albers, The Epigenetic Nature of EarlyChordate Development. Inductive Interaction and Competence,Cambridge Univ. Press, 1985.

[24] A.P. Otte, I.M. Kramer, A.J. Durston, Protein kinase C and regulation ofthe local competence of Xenopus ectoderm, Science 251 (1991) 570–573.

[25] A.P. Otte, L.L. McGrew, J. Olate, N.M. Nathanson, R.T. Moon, Expressionand potential functions of G-protein alpha subunits in embryos of Xenopuslaevis, Development 116 (1992) 141–146.

[26] A.P. Otte, R.T. Moon, Ectopic induction of dorsal mesoderm by over-expression of Xwnt-8 elevates the neural competence of Xenopusectoderm, Dev. Biol. 152 (1992) 184–187.

[27] F. Pituello, V. Homburger, J.P. Saint-Jeannet, Y. Audigier, J. Bockaert,A.M. Duprat, Expression of the guanine nucleotide-binding protein Gocorrelates with the state of neural competence in the amphibian embryo,Dev. Biol. 145 (1991) 311–322.

[28] G. Drean, C. Leclerc, A.M. Duprat, M. Moreau, Expression of L-typeCa2+ channel during early embryogenesis in Xenopus laevis, Int. J. Dev.Biol. 39 (1995) 1027–1032.

[29] C. Leclerc, A.M. Duprat, M. Moreau, In vivo labelling of L-type Ca2+channels by fluorescent dihydropyridine: correlation between ontogenesisof the channels and the acquisition of neural competence in ecotderm cellsfrom Pleurodeles waltl embryos, Cell Calcium 17 (1995) 216–224.

[30] A.L. Miller, E. Karplus, L.F. Jaffe, Imaging [Ca2+]i with aequorin using aphoton imaging detector, Methods Cell Biol. 40 (1994) 305–338.

[31] R. Keller, J. Shih, A.K. Sater, C. Moreno, Planar induction of convergenceand extension of the neural plate by the organizer of Xenopus, Dev. Dyn.193 (1992) 218–234.

[32] H. Kuroda, O. Wessely, E.M. De Robertis, Neural induction in Xenopus:requirement for ectodermal and endomesodermal signals via Chordin,Noggin, beta-Catenin, and Cerberus, PLoS Biol. 2 (2004) E92.

[33] R. Keller, M. Danilchik, Regional expression, pattern and timing of

1191C. Leclerc et al. / Biochimica et Biophysica Acta 1763 (2006) 1184–1191

convergence and extension during gastrulation of Xenopus laevis,Development 103 (1988) 193–209.

[34] C. Leclerc, M. Lee, S.E. Webb, M. Moreau, A.L. Miller, Calciumtransients triggered by planar signals induce the expression of ZIC3 geneduring neural induction in Xenopus, Dev. Biol. 261 (2003) 381–390.

[35] M.E. Greenberg, M.A. Thompson, M. Sheng, Calcium regulation ofimmediate early gene transcription, J. Physiol. (Paris) 86 (1992) 99–108.

[36] C. Leclerc, A.M. Duprat, M. Moreau, Noggin upregulates Fos expressionby a calcium-mediated pathway in amphibian embryos, Dev. GrowthDiffer. 41 (1999) 227–238.

[37] K. Nakata, T. Nagai, J. Aruga, K. Mikoshiba, Xenopus Zic3, a primaryregulator both in neural and neural crest developement, Proc. Natl. Acad.Sci. U. S. A. 94 (1997) 11980–11985.

[38] S.E. Witta, V.R. Agarwal, S.M. Sato, XIPOU 2, a noggin-inducible gene,has direct neuralizing activity, Development 121 (1995) 721–730.

[39] C. Leclerc, S.E. Webb, C. Daguzan, M. Moreau, A.L. Miller, Imagingpatterns of calcium transients during neural induction in Xenopus laevisembryos, J. Cell. Sci. 113 (Pt 19) (2000) 3519–3529.

[40] J. Batut, I. Neant, C. Leclerc, M. Moreau, xMLP is an early responsecalcium target gene in neural determination in Xenopus laevis, J. Soc. Biol.197 (2003) 283–289.

[41] A. Hemmati-Brivanlou, D.A. Melton, Inhibition of activin receptorsignaling promotes neuralization in Xenopus, Cell 77 (1994) 273–281.

[42] V. Bertrand, C. Hudson, D. Caillol, C. Popovici, P. Lemaire, Neural tissuein ascidian embryos is induced by FGF9/16/20, acting via a combination ofmaternal GATA and Ets transcription factors, Cell 115 (2003) 615–627.

[43] C. Launay, V. Fromentoux, D.L. Shi, J.C. Boucaut, A truncated FGFreceptor blocks neural induction by endogenous Xenopus inducers,Development 122 (1996) 869–880.

[44] S. Antoniotti, A. Fiorio Pla, S. Pregnolato, A. Mottola, D. Lovisolo,L. Munaron, Control of endothelial cell proliferation by calcium influxand arachidonic acid metabolism: a pharmacological approach, J. Cell.Physiol. 197 (2003) 370–378.

[45] C. Distasi, L. Munaron, F. Laezza, D. Lovisolo, Basic fibroblast growthfactor opens calcium-permeable channels in quail mesencephalic neuralcrest neurons, Eur. J. Neurosci. 7 (1995) 516–520.

[46] A.M. Carrion, W.A. Link, F. Ledo, B. Mellstrom, J.R. Naranjo, DREAM isa Ca2+-regulated transcriptional repressor, Nature 398 (1999) 80–84.

[47] S. Faure, M.A. Lee, T. Keller, P. ten Dijke, M. Whitman, Endogenouspatterns of TGFbeta superfamily signaling during early Xenopusdevelopment, Development 127 (2000) 2917–2931.

[48] T. Saneyoshi, S. Kume, T. Natsume, K. Mikoshiba, Molecular cloning andexpression profile of Xenopus calcineurin A subunit(1), Biochim. Biophys.Acta 1499 (2000) 164–170.

[49] R. Nishinakamura, Y. Matsumoto, T. Uochi, M. Asashima, T. Yokota,Xenopus FK 506-binding protein homolog induces a secondary axis infrog embryos, which is inhibited by coexisting BMP 4 signaling, Biochem.Biophys. Res. Commun. 239 (1997) 585–591.

[50] S.E. Webb, K.W. Lee, E. Karplus, A.L. Miller, Localized calciumtransients accompany furrow positioning, propagation, and deepeningduring the early cleavage period of zebrafish embryos, Dev. Biol. 192(1997) 78–92.