Developmental Biology 369 (2012) 1–18

Contents lists available at SciVerse ScienceDirect

Developmental Biology

0012-16

http://d

n Corr

E-m

junya@k

journal homepage: www.elsevier.com/locate/developmentalbiology

Review

Gastrulation and pre-gastrulation morphogenesis, inductions, and geneexpression: Similarities and dissimilarities between urodeleanand anuran embryos

Teruo Kaneda n, Jun-ya Doi Motoki

Advanced Course for Biological Systems Engineering, and Department of Biological and Chemical Systems Engineering, Kumamoto National College of Technology,

2627 Hirayama Shin-Machi, Yatsushiro 866-8501, Japan

a r t i c l e i n f o

Article history:

Received 12 November 2011

Received in revised form

14 May 2012

Accepted 18 May 2012Available online 25 May 2012

Keywords:

Gastrulation

Pre-gastrulation morphogenesis

Notochord

Fate maps

Dorsal marginal zone

Meso-endoderm induction

Neural induction

Urodele

Anura

06/$ - see front matter & 2012 Elsevier Inc. A

x.doi.org/10.1016/j.ydbio.2012.05.019

esponding author. Fax: þ81 965 53 1389.

ail addresses: kaneda@kumamoto-nct.ac.jp (T

umamoto-nct.ac.jp (J.D. Motoki).

a b s t r a c t

Studies of meso-endoderm and neural induction and subsequent body plan formation have been

analyzed using mainly amphibians as the experimental model. Xenopus is currently the predominant

model, because it best enables molecular analysis of these induction processes. However, much of the

embryological information on these inductions (e.g., those of the Spemann–Mangold organizer), and on

the morphogenetic movements of inductively interacting tissues, derives from research on non-model

amphibians, especially urodeles. Although the final body pattern is strongly conserved in vertebrates,

and although many of the same developmental genes are expressed, it has become evident that there

are individually diverse modes of morphogenesis and timing of developmental events. Whether or not

this diversity represents essential differences in the early induction processes remains unclear. The aim

of this review is to compare the gastrulation process, induction processes, and gene expressions

between a urodele, mainly Cynops pyrrhogaster, and an anura, Xenopus laevis, thereby to clarify

conserved and diversified aspects. Cynops gastrulation differs significantly from that of Xenopus in

that specification of the regions of the Xenopus dorsal marginal zone (DMZ) are specified before the

onset of gastrulation, as marked by blastopore formation, whereas the equivalent state of specification

does not occur in Cynops until the middle of gastrulation. Detailed comparison of the germ layer

structure and morphogenetic movements during the pre-gastrula and gastrula stages shows that the

entire gastrulation process should be divided into two phases of notochord induction and neural

induction. Cynops undergoes these processes sequentially after the onset of gastrulation, whereas

Xenopus undergoes notochord induction during a series of pre-gastrulation movements, and its

traditionally defined period of gastrulation only includes the neural induction phase. Comparing the

structure, fate, function and state of commitment of each domain of the DMZ of Xenopus and Cynops has

revealed that the true form of the Spemann–Mangold organizer is suprablastoporal gsc-expressing

endoderm that has notochord-inducing activity. Gsc-expressing deep endoderm and/or superficial

endoderm in Xenopus is involved in inducing notochord during pre-gastrulation morphogenesis, rather

than both gsc- and bra-expressing tissues being induced at the same time.

& 2012 Elsevier Inc. All rights reserved.

Introduction

Gastrulation is a set of evolutionarily conserved morphoge-netic movements in the early development of a wide variety ofvertebrates. During gastrulation in amphibians, the cells of thedorsal marginal zone (DMZ) involute through the blastopore toform an archenteron roof (ARF) that underlies the future central

ll rights reserved.

. Kaneda),

nervous system (CNS). The three-dimensional germ layer struc-ture of the embryo is established during gastrulation, and twomajor inductive interactions are required. One is meso-endoderminduction and establishment of the anteroposterior and dorso-ventral regional characteristics of the induced mesoderm. Theother is neural induction, in which induced and regionally specificmesoderm within the ARF or associated with it, spatially andtemporally interacts with the presumptive neuroectoderm toform the CNS. Thus, the early patterning of the embryo involvesa complex set of spatial and temporal inductions, morphogeneticmovements and regional interactions.

Spemann and Mangold (1924) first discovered in urodeles theorganizational activity of the early gastrula dorsal blastoporal lip

Fig. 1. Structure of the dorsal marginal zone (DMZ) of Xenopus and Cynops early gastrulae. The Xenopus early gastrula DMZ has superficial and deep layers. At the late

blastula/early gastrula stage (A), the superficial layer is subdivided into prospective archenteron roof (ARF) endoderm, part of the prospective dorsal mesoderm and

prospective bottle cells. The deep layer has sub-classes of leading edge meso-endoderm (LEM) of the future pharyngeal endoderm/prechordal plate and notochord. At stage

10þ when bottle cells are formed (B), the deep layer has already started to involute, independent of bottle cell formation, by pre-gastrulation morphogenesis and the Cleft

of Brachet is formed. At this stage, the involuting LEM starts to vertically contact the surface neuroectoderm (Modified from Bauer et al., 1994; Harland and Gerhart, 1997;

Arendt and Nubler-Jung, 1999; Shook et al., 2004 and Keller and Shook, 2004). (C) At the early gastrula stage when bottle cells are formed, the Cynops DMZ is defined as the

area between the blastopore and the limits of involution. The DMZ is a single-cell-layered structure except at its most vegetal (future anterior) part, and is divided into a

vegetal (future anterior) prechordal region and an animal (future posterior) presumptive notochord, arranged in a planar juxtaposition. (Modified from Suzuki et al., 1997).

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–182

(DLP1), which can induce an ectopic secondary axis when trans-planted into another part of the embryo. These basic concepts ofthe organizer were then widely investigated by comprehensiveembryological experiments using mainly urodele embryos as themodel (reviewed by Nieuwkoop et al., 1985; Hamburger, 1988).The heritage of embryological knowledge derived from urodeleanexperimental embryology is translated into the current amphi-bian model of the anuran, Xenopus laevis, which best enablesanalysis of the molecular nature of the meso-endoderm andneural induction processes. However, there may be some mis-interpretation of basic embryological processes, because Xenopus

and urodeles differ significantly in the structure of their respec-tive DMZ, the state of commitment of the early gastrula DMZ, themode of morphogenetic movement and the spatiotemporal inter-actions between the ARF and the overlying presumptive neuroec-toderm (for reviews, see Nieuwkoop, 1996, 1997).

Comparative studies of gastrulation mechanisms (Shook andKeller, 2008a, b), developmental profiles (Ninomiya et al., 2001;Collazo and Keller, 2010) and gene expression patterns (Beckhamet al., 2003; del Pino et al., 2007; Moya et al., 2007; Nath andElinson, 2007; Venegas-Ferrın et al., 2010) of several members ofthe Urodele and Anura families have been performed and theresults compared with those for Xenopus. From these it hasbecome evident that the gastrulation mechanism and/or thetiming of gene expression in the embryo differs according to theembryo’s size, developmental speed and/or amount and omni-presence of yolk (for reviews, see Arendt and Nubler-Jung, 1999;Elinson and Beckham, 2002; Solnica-Krezel, 2005; Callery, 2006).For example, in frogs with slow development, the mesoderm maybecome specified later than in Xenopus (del Pino et al., 2007;Moya et al., 2007). Recently, the gene regulatory network (GRN)

1 DLP is often used as a synonym for ‘‘organizer’’ or ‘‘DMZ’’. However, in many

earlier experiments, the stage criterion, size and location of the DLP were not

always defined. Throughout the text we use DLP to mean the upper blastopore

generally, at any given stage, and DMZ as the stage- and site-defined upper

blastopore region of the beginning gastrula.

for endoderm and mesoderm specification and differentiation,and its essential role in body plan formation, have been eluci-dated (e.g., Swiers et al., 2010; Rankin et al., 2011; for review, seeDavidson and Erwin, 2006). These results suggest that there aregenerally conserved but individually diverse modes of morpho-genesis even among the anuran species. However, whether or notthis diversity represents essential differences in the early pattern-ing of the embryo remains obscure. Although the final antero-posterior and dorsoventral body patterns and maps of geneexpression domains are strongly conserved in all vertebrates(Elinson and Kezmoh, 2010), it is necessary to analyze thesimilarities and dissimilarities of these basic processes to deepenour understanding of the fundamental principles of the earlypatterning of the embryos. In this review, we aim to verify thesimilarities and dissimilarities of the gastrulation process, and thespatially and temporally controlled mesoderm and neural induc-tion processes, mainly from an embryological point of view,between a urodele, primarily the Japanese newt Cynops pyrrho-

gaster (formerly Triturus pyrrhogaster) and an anura, Xenopus

laevis, both of which have extensive data and histories ofinvestigation.

Structure, prospective fate, morphogenesis and functionalgene expression domains of the urodelean and anuran earlygastrula DMZ

The spatial location or topological architecture of the DMZvaries among vertebrates but the essential feature is well con-served (Arendt and Nubler-Jung, 1999; Solnica-Krezel, 2005).In amphibians, the early gastrula DMZ occupies an arc-shapedregion restricted to the dorso-vegetal part of the embryo.The DMZ is a mixture of multiple heterologous domains withdiffering prospective fates, self-differentiation and functions.However, the early gastrula DMZ of both Xenopus and Cynops showsconsiderable differences in its germ layer structure (Figs. 1 and 2and Table 1).

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–18 3

Structure, prospective fate and morphogenesis of the early

gastrula DMZ

Xenopus laevis and other anurans

At the late blastula and early gastrula stages (Fig. 1A, B), theXenopus DMZ (the dorsal sector of the involuting marginal zone,Keller, 1975, 1976) is a multilayered structure in which at leasttwo layers are recognized: a superficial epithelial layer (super-ficial layer) and a deep mesenchymal layer (deep layer). Althoughthe degree of multilayering varies among anurans (Vogt, 1929;Keller, 1975, 1976; Dettlaff, 1983; Smith and Malacinski, 1983;Campos Casal and Manes, 1999), the multilayered DMZ structureis a characteristic feature in the Anura family. The deep layer isdivided into sub-classes of the leading edge meso-endoderm(LEM) for the future pharyngeal endoderm/prechordal plate andaxial mesoderm for future notochord and somites (Keller, 1975,1976; Smith and Slack, 1983; Shook et al., 2004; for reviews, seeGerhart, 2001; Keller and Shook, 2004; Shook and Keller, 2008a, b).The suprablastoporal superficial layer consists of suprablastoporalendoderm of both the future ARF and walls, presumptive bottlecells in the most vegetal part and a small amount of presumptivemesoderm (notochord and somites). The sub-blastoporal super-ficial layer comprises the archenteron floor. Anurans have rela-tively little superficial mesoderm compared with urodeles, andXenopus has the least amount of superficial mesoderm (Minsukand Keller, 1997; Keller and Shook, 2004). Fate mapping revealsthat the dorsal mesoderm is derived mainly from the deep layerand partially from the superficial layer in both Xenopus (Shooket al., 2004) and Rana pipiens (Delarue et al. 1994). About 20% ofpresumptive notochord and about 5% of presumptive mid toposterior somites are distributed in the superficial layer inXenopus (Shook et al., 2004).

Xenopus shows characteristic morphogenetic movements ofthe DMZ (Fig. 2A). At the late blastula to early gastrula stages,prior to the onset of gastrulation as defined by bottle cellformation (Fig. 1A), cells of the deep layer start to involuteindependently of bottle cell formation (Nieuwkoop andFlorschutz, 1950; Keller, 1976) by ‘‘pre-gastrulation movements’’such as the pre-gastrulation epiboly movement (Keller, 1980;Bauer et al., 1994, Papan et al., 2007) and vegetal rotation(Winklbauer and Schurfeld, 1999). At the early gastrula stagewhen bottle cells have been formed (Fig. 1B), the LEM has alreadystarted to involute and its anterior part has reached beneath theforebrain/midbrain part of the presumptive neuroectoderm(Poznanski and Keller, 1997), and the Cleft of Brachet is formed.Slightly later, the presumptive notochord of the deep layer alsostarts to involute and then bottle cells form in the most vegetalregion of the superficial layer (Fig. 1B), and then the superficiallayer starts to involute (Fig. 2A). During gastrulation, the arch-enteron is formed and the involuted superficial layer lines the ARF(see reviews, Keller and Shook, 2004). As will be discussed indetail later, these processes indicate that the onset of gastrulationin Xenopus has two steps: involution of the deep layer LEM bypre-gastrulation movements before bottle cell formation, andinvolution of the deep and superficial layers after bottle cellformation.

During gastrulation, bottle cells are located at the anterior endof the archenteron, and then spread bilaterally until finally theyare located beneath the anterior region of the deep presumptivenotochord at the late gastrula stage (Fig. 2A, b–d). Duringgastrulation, the ARF is composed mainly of endoderm andpartially of presumptive mesoderm. Presumptive notochordaland pre-somitic cells originate from the superficial layer ingressesand move into the underlying deep presumptive notochord andsomites that originated from the deep layer during neurulation(Shook et al., 2004), demonstrating that the dorsal mesodermal

tissues of Xenopus have a dual origin with the same fate. Inaddition to these, LEM and presumptive notochord show differentmovements during gastrulation. Involuted LEM migrates ante-riorly and spreads bilaterally. The presumptive notochord invo-lutes and then extends anteroposteriorly by convergence andextension (CE) movements. After that, the presumptive notochordelongates posteriorly and bilaterally shears the pre-somitic meso-derm during the neurula stages (for reviews, see Keller and Shook,2004; Shook and Keller, 2008a).

Cynops pyrrhogaster and other urodeles

Fig. 1C shows the DMZ structure of the Cynops stage 11beginning gastrula, in which bottle cells has just formed at theblastopore site (Cynops stages are according to Okada andIchikawa, 1947) and Fig. 2B shows the gastrulation process ofCynops.

The structure of the urodelean early gastrula shows significantdifferences when compared with that of Xenopus. Animal cappresumptive ectoderm consists of a layer of cells about 3–4 deepin the late blastula and early gastrula of Ambystoma mexicanum

(Slack, 1984). In Cynops, the animal cap of the blastula to gastrulaembryo is a single-cell-layered structure (Komazaki, 1992; Imoh,1988). The Cynops early gastrula DMZ is defined as the areabetween the pigment line (bottle cells) and the limits of involu-tion (Fig. 3A, angle of about 501 from the pigment line; Kanedaand Hama, 1979), and the Cynops early gastrula DMZ has a rathermore simple structure than that of Xenopus with two majordifferences (Table 1). One is that the Cynops early gastrula DMZis a single-cell-layered structure except for the most vegetalregion near the blastopore (Figs. 1C, 3A, Imoh, 1988; Suzukiet al., 1997). During gastrulation, the DMZ involutes throughblastopore as a monolayered sheet (Fig. 3F), and forms the ARF(Hama et al., 1985; Imoh, 1988).

The other difference is the distribution of presumptive phar-yngeal endoderm, prechordal plate and dorsal mesoderm. Urode-lean embryos have a larger presumptive mesoderm andpresumptive pharyngeal endoderm/prechordal plate than Xeno-

pus, and all these components are entirely located on theembryo’s surface (Fig. 4A). Presumptive pharyngeal endoderm/prechordal plate and notochord are located in the vegetal (futureanterior) and animal (future posterior) halves of the DMZ,respectively (Figs. 1C, 3, 4). Fate mapping has been unable todefine the exact location of the presumptive prechordal plate.Because it is hard to assign an exact location of the prechordalplate, or to distinguish precisely the limits between the prechor-dal plate and notochord on the involuting and extending ARF, thepresumptive prechordal plate has been expediently placed in anintermediate region between the future pharyngeal endodermand notochord on many urodelean fate maps (Vogt, 1929;Pasteels, 1942; Nakamura, 1942; Hama, 1978; see Cynops earlygastrula fate map, Fig. 4A). Bottle cells form at the most vegetalpart of the surface DMZ prior to the onset of involution (Fig. 1C).This tissue arrangement of the Cynops DMZ resembles that of theavian Koller’s sickle region, in which anterior definitive endo-derm, presumptive notochord and presumptive neuroectodermare planarly arranged (Arendt and Nubler-Jung, 1999).

In Cynops, and other urodeles (Vogt, 1929), the first sign of theonset of gastrulation is formation of bottle cells at the blastoporesite. In Xenopus, the deep layer starts involuting by pre-gastrula-tion movements prior to bottle cell formation (Fig. 2A). However,in urodeles, bottle cells involute first and they are always locatedat the anterior end of the involuting archenteron during gastrula-tion. Following bottle cell involution, the presumptive pharyngealendoderm/prechordal plate and presumptive notochord sequen-tially involute and form the single-cell-layered ARF in Cynops

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–184

(Figs. 2B, 3F). In Xenopus, the surface of the ARF is lined with ARFendoderm originating from the superficial layer, but in Cynops

and other urodeles, such as A. mexicanum (Shook et al., 2002), thesurface of the ARF is not covered with endoderm during gastrula-tion. The Cynops ARF is progressively covered with lateral endo-dermal crest derived from the lateral wall of the archenteron thatrolls up during neurulation (Fig. 3G). Ignoring the multilayeredstructure of the DMZ, the spatial architecture of the late blastula/early gastrula DMZ of Xenopus, in which bottle cells have not yetformed, is roughly similar to that of the Cynops early gastrulaDMZ (Fig. 1A, C). However, the structure of the Xenopus earlygastrula DMZ, in which bottle cells have formed (Fig. 1B), isfundamentally different from that of Cynops, because of the pre-gastrulation morphogenesis of Xenopus.

It has been demonstrated that the vegetal (lower) half of theCynops early gastrula DMZ (LDMZ) forms the anterior half of the ARF

that self-differentiates only into endodermal tissues, such as phar-yngeal endoderm (Kaneda et al., 2009). The animal (upper) half ofthe DMZ (UDMZ) forms the entire notochord after involution. Theanterior half domain of the involuting ARF is, therefore, designatedas the fore-notochordal endodermal roof (FNE). As shown later, theCynops ortholog of goosecoid (Cygsc) is expressed in the LDMZ at theearly gastrula stage (Figs. 4 and 5) but the expression is progres-sively restricted to the intermediate region between the FNE andnotochord at the late gastrula to neurula stage (Figs. 5 and 6). Thus,the presumptive prechordal plate does not have a definite pre-sumptive location in the early Cynops gastrula, and the LDMZ shouldbe identified as a prechordal region by its gene expression patternand prospective fate (Table 1, Fig. 1C), as proposed in avian embryosby Foley et al. (1997).

The LDMZ and UDMZ of Cynops show different morphogeneticmovements during gastrulation (Fig. 2C). Like Xenopus, bottle cells

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–18 5

and presumptive pharyngeal endoderm/prechordal plate elongateanteriorly and spread bilaterally after involution. Dorsal CEmovements in the axial mesoderm are relatively weak and occurlater in A. mexicanum than in Xenopus (Keller and Jansa, 1992;Shook et al., 2002). Cynops presumptive notochord shows thesame morphogenetic behavior; it elongates anteriorly after invo-lution, and at the neurula to tail-bud stages it in turn starts toextend posteriorly and bilaterally shears the still involutingpresumptive tail somite (Hama, 1978, see Fig. 2C). These regionaldifferences in the morphogenetic movements of the Cynops ARFmay be controlled by the same morphogenetic mechanics as inXenopus.

Gene expression domains of the early gastrula DMZ

The gene expression profiles in Xenopus have been extensivelyanalyzed, but only limited observations are available for urodeles.In Xenopus, although some gene expressions overlap in the earlygastrula DMZ, goosecoid (gsc, Cho et al., 1991), chordin (Sasai et al.,1994), Xlim-1 (Taira et al., 1997), noggin (Smith and Harland,1992) and Cerberus (Bouwmeester et al., 1996) are expressedmostly in the lower half of the late blastula/early gastrula DMZ(Fig. 5A). The upper half of the DMZ expresses Xnot and Xbra

(Smith et al., 1991). Based on these results, the Xenopus earlygastrula DMZ has been divided into two domains according to thegene expressions: the lower gsc-expressing domain of LEM andupper Xbra/Xnot-expressing domain of presumptive notochord(Vodicka and Gerhart, 1995; Zoltewicz and Gerhart, 1997;Winklbauer and Schurfeld, 1999). The gsc-expressing domain inXenopus, and also in chick and fish embryos, remains only in cellsthat involute early, and later reach the prechordal plate region(e.g., Cho et al., 1991; Sasai et al., 1994; Izpisua-Belmonte et al.,1993; De Robertis, 2009). Xlim-1 expression follows similarly(Taira et al., 1997). At the late gastrula to neurula stage, stronggsc expression is observed in the prechordal plate. Chordin andXlim-1 expressions overlap with that of gsc and their expressionsextend to the anterior notochord (Sasai et al., 1994; Taira et al.,1997). On the other hand, at the late blastula/early gastrula stage,the Xbra-expressing domain is located entirely on the deepmarginal zone and its expression is progressively restricted tothe entire notochord after involution (Smith et al., 1991,Zoltewicz and Gerhart, 1997; Conlon and Smith, 1999). Thesegene expression patterns in Xenopus are in good agreement with

Fig. 2. Gastrulation process of Xenopus and Cynops, and morphogenetic movement of th

laevis. Prior to the onset of traditionally defined gastrulation, as marked by blastopo

gastrulation epiboly movement (red arrow in a; Keller, 1980; Bauer et al., 1994; Papan e

During pre-gastrulation morphogenesis, the germ layer structure of the DMZ is drastica

notochord (Nt, light magenta) start to involute independently of bottle cells (dark green)

and vegetal (future anterior) part of the notochord have already started to involute, an

neuroectoderm has thus begun at this stage. As involution proceeds (c, d), the LEM and

are formed at the blastopore site (Blp, b). Following bottle cell involution, the surface la

cells are located beneath the anterior end region of the involuted notochord throughou

(d). Presumptive mesoderm originating from superficial layer ingress and move into the

2004, Keller and Shook, 2004.); (B) Gastrulation process of Cynops pyrrhogaster. At 20 1C

stage 11 (b), and gastrulation is almost completely finished about 30 h from the onset o

Hama, 1979). It takes approximately 12 h from stage 11 (b) to the mid-gastrula (stage 12

Cynops DMZ is subdivided into UDMZ (presumptive notochord, Pre-Nt, light magenta) a

green). Prior to the onset of gastrulation, bottle cells (dark green) are formed at the most

UDMZ sequentially involute and form a single-cell-layered ARF (b–d). Thus, bottle ce

gastrulation. At the mid-gastrula stage (c), involuted LDMZ underlies the surface UDMZ

Involuted notochord starts to vertically interact with surface presumptive neuroectode

onward, the LDMZ is segregated into fore-notochordal endodermal roof (FNE) and Cygs

(DLP) material of Cynops during gastrulation. The DLP of each gastrula stage was vital

traced during gastrulation (Redrawn from Hama, 1978). After involution, the LDMZ-e

anteriorly and spreads bilaterally, occupying the anterior half of the ARF at the mid-gas

and e). The UDMZ-equivalent DLP (light magenta and gray at stage 12b and 12c) involut

movement (c, d). Finally, it occupies the posterior half of the ARF (d-e) and forms n

posteriorly and bilaterally shears the still involuting tail somites (e, gray and brown).

the self-differentiation and organizing activity of each domain ofthe early gastrula DMZ (for reviews, see Gerhart, 2001;Bouwmeester, 2001; De Robertis, 2009).

In Cynops, goosecoid (Cygsc), noggin (Cynog), chordin (Cychd),Lim-1 (CyLim-1) and VegT (CyVegT) expressions are preferentiallyrestricted to the LDMZ; their expression does not extend to theUDMZ or the sub-blastoporal region (Sone et al., 1997; Yokotaet al., 1998; Doi et al., 2000; Kaneda et al., 2009; Motoki et al.,unpublished; see Figs. 4B, 5B and Table 1). All these geneexpressions appear on the surface of the LDMZ. As shown later,the Cygsc-expressing region coincides with the area that hassecondary axis-inducing activity. Cygsc expression starts in thelate blastula, and reaches its maximum in the early gastrula (Soneet al., 1997). Cygsc-expressing LDMZ completely involutes by themid-gastrula (stage 12b/c) and then the Cygsc expression isprogressively restricted and confined to the intermediate regionbetween the FNE and notochord beneath the anterior tip of theneural groove in the late gastrula to mid-neurula (Figs. 5B, 6A).Although the germ layer structures are considerably different, theCynops LDMZ is roughly homologous to the Xenopus prechordalplate/pharyngeal endoderm region of the deep DMZ and in part tothe endodermal portion of the superficial layer of the DMZ(Table 1, Figs. 1 and 2).

In Xenopus, chordin is expressed in the anterior part of the ARFduring gastrulation, and at the neurula stages there is broad butstrong expression in the whole prechordal plate region and theanterior part of the notochord (Sasai et al., 1994). Our preliminaryobservations show that, in the early gastrula stages, Cychd isexpressed in the LDMZ. Unlike Xenopus, however, Cychd expres-sion progressively expands to the notochord, and at the neurulastage the expression is restricted to the entire notochord. Theexpression is negligible or absent in the FNE that underlies theforebrain. At this stage, the Cychd expression pattern is nearly thesame as that of Cybra (Fig. 6B, Motoki et al. unpublished). InXenopus, Wnt-8 and BMP-2/4 are expressed in the ventro-vegetalpart of the embryo from the mid-blastula transition (MBT) andthey cooperatively pattern the mesoderm (e.g., Hoppler andMoon, 1998). In the early gastrula of A. mexicanum, Wnt-8 isexpressed in the ventro-vegetal region of the early gastrula(Bachvarova et al., 2001). Our observations (Motoki et al. unpub-lished) show that Cynops Wnt-8 (Cywnt-8) is expressed by vegetalendoderm around the vegetal pole and the expression extends tothe sub-blastoporal endoderm (Fig. 4B) at the early gastrula stage.

e dorsal lip of each stage of the Cynops gastrula. (A) Gastrulation process of Xenopus

re formation, Xenopus undergoes pre-gastrulation movements such as the pre-

t al., 2007), and vegetal rotation (blue arrow in a; Winklbauer and Schurfeld,1999).

lly changed, and leading edge meso-endoderm (LEM, light green) and presumptive

formation. At the early gastrula stage (b) when the bottle cells are formed, the LEM

d the Cleft of Brachet (CB) is formed. Vertical contact between the LEM and future

the deep notochord elongate anteriorly. Slightly after LEM involution, bottle cells

yer (dark yellow) starts to involute to form the roof of the archenteron (c, d). Bottle

t gastrulation (b, c), and they spread bilaterally and forms respreading bottle cells

underlying deep mesoderm layer during neurulation. (Modified from Shook et al.,

, involution of the DMZ starts about 3 h after the first appearance of bottle cells at

f involution (detailed staging criteria for Cynops gastrulae are shown in Kaneda and

b/c) and 24 h to stage 13b/c (d). At the late blastula/early gastrula stages (a, b), the

nd LDMZ (presumptive pharyngeal endoderm/prechordal plate, Pre-PhEþPrC, light

vegetal region of the LDMZ (b). Following bottle cell involution, the LDMZ and then

lls are always located at the anterior end of the involuting archenteron during

. The UDMZ is induced to notochord (Nt) during the early to mid-gastrula stages.

rm (Pre-Neu; dark blue) and induces neuroectoderm (Neu). From the late gastrula

c-expressing prechordal plate (PrC, d). (C) Morphogenetic movement of dorsal lip

stained at the same size (0.4�0.4 mm) and the movement of the dye mark was

quivalent DLP (DLP of stages 11 and 12a, light green and light yellow) elongates

trula to early neurula stages (a–d). It then spreads bilaterally to form the FNE (c, d

es and elongates anteriorly and then anteroposteriorly by a convergent–extension

otochord. During the late gastrula to neurula stages (d-e), notochord elongates

Table 1Prospective fate, self-differentiation and organizing specificity of each domain of the dorsal marginal zone (DMZ) of Xenopus and Cynops early gastrulae.

Subdomain Xenopus Cynops

Pharyngeal

endoderm

� Location: � Location:

Deep layer of the DMZ Vegetal half of the surface DMZ (LDMZ)

Leading edge endoderm that starts involution prior to

bottle cell formation

� Fate: � Fate:

Anterior part of the involuting deep layer Anterior half of the ARF (FNE)

Pharyngeal endoderm and foregut endoderm Pharyngeal endoderm and foregut endoderm

� Gene expression: � Gene expression:

gsc, Lim-1, cerberus, chordin gsc, Lim-1, chordin, noggin, VegT

� Organizing activity: � Organizing activity:

Induces head structure (forebrain, midbrain) Induces trunk–tail structure depending on its

notochord-inducing activity in the early gastrula, but the

activity disappears soon after involution

Prechordal

plate

� Location: � Location:

Vegetal to the presumptive notochord in the deep DMZ Not defined exactly

(intermediate region between presumptiveLeading edge meso-endoderm that starts involution prior to

bottle cell formation pharyngeal endoderm and notochord of the surface DMZ)

� Fate: � Fate:

Prechordal endomesoderm Prechordal endo-mesoderm

Intermediate region between pharyngeal endoderm and notochord At late gastrula to neurula stages, confined to intermediate region between

notochord and pharyngeal endoderm (FNE)

� Gene expression: � Gene expression:

gsc, Lim-1, cerberus, chordin, noggin

(same as pharyngeal endoderm)

gsc, Lim-1, chordin, noggin, VegT

(same as pharyngeal endoderm)

� Organizing activity: � Organizing activity:

Induces head structure (forebrain, midbrain) Same as pharyngeal endoderm

Notochord � Location: � Location:

Animal half in the deep DMZ Animal half of the surface DMZ (UDMZ)

� Fate: � Fate:

Notochord Posterior half of the ARF

Notochord

� Gene expression: � Gene expression:

Chordin, noggin, Xbra, Xnot etc. Not detected at early to mid-gastrula stage

Cybra expression starts at mid-gastrula

� Organizing activity: � Organizing activity:

Trunk–tail organizer activity Not detected at early gastrula stage

Induces trunk–tail structure Trunk–tail organizing activity is evoked after

mid-gastrula stage

Superficial

epithelial

layer

� Location: � Nonexistent

Surface, suprablastoporal region of the DMZ.

� Fate:

Bottle cells (most vegetal region)

ARF endoderm

Part of notochord and somite mesoderm that ingress to

deep mesoderm during neurulation.

� Gene expression:

Gsc, Lim-1, chordin, noggin

� Organizing activity:

Notochord-inducing activity

Trunk–tail organizer activity

Data regarding the prospective fate, self-differentiation, gene expression and organizing specificity of Xenopus are summarized from Harland and Gerhart (1997), Gerhart

(2001), De Robertis, (2009) and Keller and Shook (2004). ARF: archenteron roof; FNE: fore-notochordal endodermal roof.

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–186

Cywnt-8 expression does not extend beyond the pigment line,thus Cygsc-expressing LDMZ and Cywnt-8-expressiing endodermare clearly separated by the blastopore. Cywnt-8-expressingendoderm involutes through the ventral lip and forms the floorendoderm of the archenteron and the yolk plug during gastrula-tion (Fig. 5B).

The most critical difference is the brachyury expression pat-tern. In Xenopus, ring-shaped Xbra expression is clearly detectedbefore the onset of gastrulation. However, Cybra expression isnever detected in any part of the Cynops beginning gastrula, but isfirst detectable from the mid-gastrula stage onward in the surfacepresumptive notochord region of the upper blastopore (Fig. 9C).

At the yolk-plug stage, Cybra expression is restricted to involutedand elongated notochord and surface presumptive tail mesodermaround the blastopore (Sone et al., 1997; Doi et al., 2000; Kanedaet al., 2009). At the neurula stage, Cybra expression is detected inthe entire notochord (Figs. 5B, 6C). Brachyury expression is clearlyshown to couple with dorsal mesoderm (notochord) induction invertebrates (Smith et al., 1991). These results indicate criticaldifferences in the state of commitment of the early DMZ of Cynops

and Xenopus. The Xenopus early gastrula DMZ is already welldetermined according to the prospective fate at the early gastrulastage; however, dorsal mesoderm induction and regional specifi-cation of the mesoderm occurs during gastrulation in Cynops

Fig. 3. Fine structure of Cynops early gastrula dorsal marginal zone (DMZ) and the structure of the involuted archenteron roof. (A) At the late blastula/early gastrula stage,

the animal cap and DMZ are single-cell-layered structures, except for the most vegetal (future anterior) part of the LDMZ region. The Cynops embryo is approximately

2.2–2.4 mm in diameter. (Reproduced with permission from Suzuki et al., 1994, Dev. Growth Differ. 39, 135–141); (B) External form of a stage 11 beginning gastrula. The

crescent-shaped pigment line (PL) is formed by the accumulation of bottle cells at the dorso-vegetally restricted blastopore (Blp) site. (C) Fine structure of the blastopore

(Blp) with bottle cells (red dotted circle) at stage 11. (D) Involution of the DMZ about 6 h after the onset of gastrulation (approximately stage 12a). Bottle cells are seen in

the red dotted circle. (Figs. C and D reproduced with permission from Sakaguchi et al., 2002, Int. J. Dev. Biol., 46, 793–800); (E) Spemann and Mangold (1924) inserted the

DLP into the blastocoel of the host embryo and analyzed the organizing activity of the DLP. Imoh et al. (1998) transplanted the LDMZ into the surface layer of the host

embryo. The LDMZ is isolated with bottle cells and part of the sub-blastoporal endoderm from a FDA-labeled embryo, transplanted into the surface layer of the ventral part

of a stage 11 gastrula and then the FDA–LDMZ transplanted embryo develops to the late gastrula stage. Transplanted LDMZ induces a complete secondary archenteron roof

(ARF) in which the notochord of the secondary axis is entirely derived from host cells, while the transplanted LDMZ forms the fore-notochordal endodermal roof (FNE) of

the secondary ARF. (Reproduced with permission from Imoh et al., 1998, Dev. Growth Differ. 40, 439–448). (F) Single-cell-layered structure of the late gastrula ARF. It is

divided into notochord (Nt) and FNE. Dotted line is the approximate position of the boundary between the notochord and FNE. Neu, neuroectoderm. (G) Cross-section of

the early neurula ARF. Notochordal plate (Nt) occupies the median line of the ARF. Lateral mesoderm (LM) is covered with lateral endodermal crest (LEC). During

neurulation, LEC progressively covers the Nt, and the ARF is finally covered with endoderm at the mid- to late neurula stage. Blue dotted line represents the boundary

between neuroectoderm and LM. Red dotted line represents the boundary between LM and LEC.

Fig. 4. Fate map and gene expression domains of Cynops early gastrula. (A) Fate map of Cynops late blastula/early gastrula embryo (right-lateral view, redrawn from

Nakamura, 1942). Like other urodeles (Vogt, 1929; Pasteels, 1942), Cynops has a large surface prospective mesoderm. Presumptive pharyngeal endoderm (PhE), prechordal

plate (PrC), notochord (NT) and tail somites (Pre-Sm) are planarly located on the embryo’s surface. Presumptive anterior and trunk somites (Pre-Sm) are located lateral to

the presumptive notochord. (B) Gene expression map of early Cynops gastrula (dorso-vegetal view). At the early gastrula stage, expressions of Cygsc, Cychd, Cylim-1, Cynog

and CyVegT are preferentially restricted to the LDMZ (light green), and their expressions do not extend below the blastopore. CyNodal (Ito et al., 2006) is expressed around

the marginal zone, except for the LDMZ region. Cynops Wnt-8 is expressed in the ventro-vegetal part of the embryo (gray) around the vegetal pole (Vp), but Wnt-8

expression does not extend beyond the blastopore (Blp). Cybra is not expressed in any part of the embryo at this stage. The presumptive notochord (UDMZ) has epidermal

self-differentiation capacity at this stage and no inducing activity. (Gene expression patterns are from Motoki et al., unpublished).

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–18 7

(Kaneda et al., 2002), as indicated by the self-differentiation andinducing activities of the DMZ (Figs. 7 and 8). The situation shouldbe the same as in A. mexicanum in which Axbra expression startsat the mid-gastrula stage onward (Swiers et al., 2010).

Summarizing these results, at the early gastrula stage, theXenopus DMZ is divided into two gene expression domains of aXbra/Xnot-expressing notochord and a gsc-expressing LEM,whereas the Cynops early gastrula has only Cygsc-expressing

Fig. 5. Morphogenesis of the gene expression domains during gastrulation in Xenopus and Cynops. (A) Xenopus laevis: Pre-gastrulation and gastrulation morphogenesis of

gsc-expressing leading edge meso-endoderm (LEM, light green) and Xbra/Xnot-expressing notochord (light magenta) are shown. Gene expression domains of the DMZ are

established before bottle cell (dark green) formation, and at the late blastula/early gastrula stage (stage 9/10), the DMZ is divided into gsc-expressing LEM in the lower half

and Xbra/Xnot-expressing notochord in the upper half. BMP-4 and Wnt-8 are expressed in the ventral marginal zone (gray). As gastrulation proceeds, gsc-expressing LEM

and Xbra/Xnot-expressing notochord involute and extend. At the late gastrula stage (St. 10.5/11), when gastrulation is almost finished, gsc-expressing LEM located in the

anterior region and Xbra/Xnot-expressing notochord in the posterior region. Red arrow at stages 8 and 9/10 indicates pre-gastrulation epiboly movement and blue arrow in

stage 9/10 indicates vegetal rotation. (B) Cynops pyrrhogaster: At the early gastrula stage (St. 11), the Cybra-expressing domain is not yet established and only Cygsc-

expressing LDMZ (light green) exists. During the early to mid-gastrula stages (stages 11–12b/c), the Cygsc-expressing LDMZ planarly induces Cybra-expressing notochord

(light magenta) in the UDMZ, and at the mid-gastrula stage (St. 12b/c) the Cygsc-expressing and Cybra-expressing domains are established. At the mid- to late gastrula

stages (St. 12c–13b/c), Cygsc expression gradually disappears from the anterior to posterior region of the ARF, and finally Cygsc expression is confined to the intermediate

region between Cybra-expressing notochord and the FNE (St. 13b/c). Cychd expression also gradually disappears and its expression moves to the notochord. Cywnt-8 is

expressed in the vegetal and dorso-vegetal region of the embryo (gray) at the late blastula/early gastrula stage. During gastrulation, Cywnt-8 expressing endoderm

involutes and forms the floor endoderm of the archenteron and yolk plug (St. 11–13b/c).

Fig. 6. Gene expression patterns at the neurula stage of Cynops. (A) Cygsc, (B)

Cychd and (C) Cybra expression at the neurula stage (stage 18) of Cynops. Cygsc

expression is restricted to the fore-notochordal endodermal roof (FNE), just

anterior to the notochord. Cychd and Cybra expressions are restricted to the entire

notochord. Arrowhead indicates the anterior end of the notochord. (A and C from

Kaneda et al., 2009; B is from Motoki et al., unpublished.).

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–188

LDMZ and a Cybra-expressing domain is not detected (Table 1,Figs. 4 and 5).

The expressions of Bra and Lim-1 proteins were analyzed inseveral anuran families and compared with Xenopus (del Pino, 1996;del Pino et al., 2007; Moya et al., 2007; Venegas-Ferrın et al., 2010).Lim-1 was simultaneously detected in the prechordal plate andanterior notochord at the mid-gastrula stage of the rapidly

developing embryos of Xenopus and Engystomops randi. In contrast,only the prechordal plate expressed Lim-1 in the slowly developingembryos of Colostethus machalilla. The notochord became Lim-1-positive after blastopore closure in C. machalilla and Gastrotheca

riobambae embryos. Like Xenopus, a wide ring of Bra expression isdetected in the deep and surface layers of the marginal zone at theblastula stages of C. machalilla or in the early gastrula of G. riobambae

(del Pino, 1996). Bra distribution shows a considerably diversepattern in relation to notochord elongation (del Pino et al., 2007).These observations indicate that the DMZ has a different schedule ofgene expression even among frog families. However, the structure,fate, morphogenesis and organizing activities of each DMZ domainof these frogs are not yet clear (Venegas-Ferrın et al., 2010). It istherefore difficult, at present, to compare these results with those ofurodeles. However, Bra expression in the early gastrula of thesefrogs indicates that there are essential differences in the mesoderminduction process and the mode of DMZ formation between anuransand urodeles.

Self-differentiation and organizing activity of the early gastrula DMZ

Although the embryonic stage, isolation size, spatial localiza-tion and methods of identifying the self-differentiation and

Fig. 7. Self-differentiation and inducing activity of each part of the dorsal

marginal zone (DMZ) of the early gastrula of Cynops. Early gastrula (stage 11)

DMZ is divided into three parts along the midline according to the presumptive

fate of pharyngeal endoderm (A), trunk–tail notochord (C) and a mixture of

pharyngeal endoderm, putative prechordal plate and most anterior part of the

trunk notochord (B). Each part (0.3�0.4 mm) is isolated and cultured alone

(isolation culture) to examine self-differentiation, or its inducing activity is

assayed by the sandwich assay method in which part A, B or C is wrapped

between two sheets of animal cap ectoderm from stage 11 and cultured for

2 weeks. To achieve immediate adhesion of the test tissue to the presumptive

ectoderm, the sandwich explant was weighted by another embryo. Average

number of cells was 88, 133 and 145 for parts A, B and C, respectively. Notochord

and muscle differentiation are mainly observed in part B in the isolation culture.

Presumptive trunk–tail notochord (part C) preferentially self-differentiates into

epidermis. In the sandwich culture, well-organized trunk–tail structures asso-

ciated with notochord, somites and spinal cord are induced mainly by part B. Head

(mainly hind brain) or headþtrunk–tail structures are obtained from parts A and

B, but the frequency is very low. Future trunk–tail notochord (part C) has only

slight inducing activity. (Redrawn from data of Kaneda and Hama,1979 and

Hama et al.,1985.).

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–18 9

organizing capacities vary among studies according to the definingcriteria used, the self-differentiation and organizing activities ofthe early gastrula DMZ have been extensively analyzed in manyurodeles (e.g., Spemann and Mangold, 1924; Holtfreter, 1938a;Holtfreter-Ban, 1965; Kaneda and Hama, 1979; Kaneda, 1980,1981; Slack, 1984; Cleine and Slack, 1985; Hama et al., 1985;Delarue et al., 1992; Yamamoto and Suzuki, 1994; Imoh et al.,1998; Kaneda et al., 2002, 2009) and in anura species (e.g.,

Holtfreter, 1938b; Smith and Slack, 1983; Stewart and Gerhart,1990, 1991; Shih and Keller, 1992; Domingo and Keller, 1995;Lane and Keller, 1997; Manes and Campos Casal, 1997; Fujii et al.,2002). By inserting the DLP into the blastocoel, earlier experi-ments revealed that the DLP of the early gastrula has secondaryaxis organizing activity in which the inserted DLP differentiatesinto notochord, somites and endoderm, while the secondaryneural axis originates from the host embryo (e.g., Spemann andMangold, 1924). In some cases, notochord and/or somites are alsoderived from host cells. Nevertheless, the early gastrula DLP isdefined as an organizer (the Spemann–Mangold organizer), which is‘‘a cell population capable of releasing inducers to adjacent cells and

self-differentiating into dorsal mesoderms such as notochord and

somites’’. This definition has led to the concept that the organizerhas been determined to self-differentiate into notochord at theearly gastrula stage.

The self-differentiation of the Xenopus organizer, defined as anarc of about 60o in the early gastrula DMZ (Stewart and Gerhart,

1990, 1991), has been analyzed by grafting experiments(e.g., Smith and Slack, 1983). In these experiments, the DMZ itselfdifferentiates into a variety of tissues such as notochord, headmesenchyme and endoderm. However, these studies used ratherlarge isolates of DMZ, which included almost the entire DMZ.Fujii et al. (2002) isolated the early gastrula DMZ of a stage 10Xenopus, which is much smaller (about 0.3�0.3 mm) than theDMZ studied in previous works, and found that the central region(core region) of the DMZ only induced anterior structures, while itdifferentiated into notochord and endoderm. Furthermore, theydemonstrated that the sub-blastoporal region, which has endo-dermal differentiation, has no inducing activity. The prospectivefate of the ‘‘core region’’ is not exactly defined; however, theseresults demonstrate that the inducing activity of the Xenopus

stage 10 early gastrula is restricted to the DMZ, and that the DMZhas already been specified to the regions that induce head andtrunk–tail, as indicated by the gene expression patterns.

Holtfreter (1938a, b) comprehensively investigated the self-differentiation of the urodelean (Triton alpestris, T. taeniatus andA. mexicanum) and anuran (Rana esculenta and Bombinator pachy-

pus) early gastrula DMZ. Holtfreter-Ban (1965) further analyzedthe self-differentiation of the Cynops and Ambystoma tigrinum

early gastrula DMZ. She demonstrated that the upper animal partof the early gastrula DMZ of both Cynops and A. tigrinum

preferentially differentiates into epidermis. The lower vegetalpart tends to differentiate into endodermal derivatives. Mesoder-mal tissues are obtained from the middle and lower portions, butthe frequency of notochord and somite differentiation is unex-pectedly low. In addition, neural tissues often differentiate fromall parts of the DMZ.

‘‘Self-differentiation’’ can mean ‘‘differentiation of a given

embryonic part according to its normal fate’’ or ‘‘differentiation

according to the state of determination of the embryonic part at a

given stage and place’’ (Hamburger, 1988). The results ofHoltfreter-Ban (1965) are incompatible with the definition ofthe ‘‘organizer’’. All endodermal, dorsal mesodermal and neuraltissues are differentiated when the DMZ is isolated in a largerpiece that includes all parts of the DMZ. A relatively large DLP wasisolated and used in earlier experiments, in which dorsal meso-dermal tissues should have been induced by induction within thegrafted DLP. Holtfreter-Ban (1965) thus proposed that the ‘‘self-

differentiation capacity of the Cynops and Ambystoma organizer is

not yet determined, even at early gastrula stage, and still stays at

the labile state’’. Her proposal was later confirmed in Cynops.The Cynops DMZ was divided into three parts and the respectiveself-differentiation and organizing activities analyzed (Fig. 7). Themost vegetal part (presumptive pharyngeal endoderm) exclu-sively self-differentiates into an endodermal cell mass, butinduces the trunk–tail structure. The most animal part (futuretrunk–tail notochord) preferentially differentiates into atypicalepidermis and has almost no inducing activity. The intermediatepart that is fated to the posterior pharyngeal endoderm, putativeprechordal plate and anterior part of the notochord differentiatesinto an endodermal cell mass and/or notochord, and induces thetrunk–tail axis (Fig. 7, see Hama et al., 1985; Kaneda and Hama,1979). Much later, dividing the Cynops early gastrula DMZ intothe UDMZ and LDMZ further confirmed that the LDMZ self-differentiates into an endodermal cell mass, but has potentnotochord-inducing activity (Fig. 8), and that the UDMZ hasepidermal self-differentiation capacity (Kaneda, 1981; Kanedaand Suzuki, 1983; Suzuki et al., 1984; Kaneda et al., 2002, 2009;Sakaguchi et al., 2002).

Grafting experiments (Yamamoto and Suzuki, 1994) confirmthe secondary axis organizing activity of the LDMZ, in whichgrafted LDMZ differentiates into endoderm but induces a com-plete secondary axis, and that the secondary axis-inducing

Fig. 8. Gene expression, self-differentiation and notochord-inducing activity of the dorsal marginal zone (DMZ) of the early gastrula of Cynops. (A) The early gastrula (stage

11) DMZ of Cynops is divided into two domains of UDMZ (future trunk–tail notochord) and LDMZ (future pharyngeal endoderm/prechordal plate). (B–E) The UDMZ and

LDMZ are separately isolated and Cygsc and Cybra expressions analyzed. The UDMZ express neither Cygsc nor Cybra at stage 11 (B, C). The expression is unchanged when

isolated UDMZ is cultured in vitro for 24 h. LDMZ expresses Cygsc but not Cybra (D, E). When the LDMZ is isolated and cultured in vitro for 24 h, Cygsc expression is

maintained. (F) Keller sandwich of early gastrula DMZ. The early gastrula DMZ was isolated and cultured in the Keller sandwiches for 24 h at 20 1C until the control embryo

developed to the late gastrula (stage 13b/c). Cybra expression was planarly induced in the UDMZ part of the explants. Approximate position of the UDMZ is marked by

asterisks (B–F reproduced from Kaneda et al., 2009.); (G, I, J) Self-differentiation of the LDMZ in the isolation culture. LDMZ differentiates to an endodermal cell mass. Any

mesodermal or ectodermal tissues are differentiated. (I and J reproduced with permission from Sakaguchi et al., 2002, Int. J. Dev. Biol., 46, 793–800). (H) UDMZ self-

differentiates to epidermis. (K) When the LDMZ is combined with a small piece of animal cap ectoderm, the animal cap ectoderm is induced to notochord (Nt) and muscle

cells (Mus), while LDMZ itself differentiates into a mass of endodermal cells.

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–1810

activity of Cynops DMZ is restricted to the suprablastoporal regionof an area 301 from the blastopore along the animal–vegetal axisand 601 laterally from the dorsal-midline, an area identical to thegsc-expressing LDMZ. Morepver, grafting the Cynops LDMZ andXenopus early gastrula DMZ into the ventral marginal zone, Imohet al. (1998) compared the secondary axis organizing activity. Asshown in Fig. 2E, they found that the grafted Cynops LDMZ itselfforms the FNE and differentiates into pharyngeal endoderm andanterior gut endoderm, but could also organize a completesecondary axis, whereas the DMZ of Xenopus organizes a ratherpoor or site-restricted secondary axis. Thus, Imoh et al. (1998)proposed that the early gastrula DMZ of both Xenopus and Cynops

has critical differences in its specifications; the Xenopus DMZ iscommitted much earlier than that of Cynops.

As summarized above, these results and the absence of Cybra

expression show that, in Cynops, the mesoderm is not yet inducedeven at the early gastrula stage. At that stage, the early gastrulaDMZ is simply divided into two domains of UDMZ, which is fatedto notochord and a Cygsc-expressing LDMZ, which has notochord-inducing activity.

Mesoderm induction before and after the onset of gastrulation

Is bottle cell formation coupled with mesoderm induction?

In general, parts of the embryo that are initially located at aremote distance can interact with each other by means ofgastrulation movements. During gastrulation, internalization ofthe marginal zone begins with blastopore formation in allchordates (Arendt and Nubler-Jung, 1999; Solnica-Krezel, 2005;

Shook and Keller, 2008a). In Xenopus, bottle cells are progressivelyformed by cells of the lower half of the suprablastoporal endo-derm layer around the entire circumference of the marginal zone,but neither bottle cells nor the blastopore forms on the ventralside in Cynops (Doi et al., 2000). Cell lineage analyses havesuggested that blastopore (bottle cell)-forming activity is notalways defined in a particular blastomere at the cleavage stagein either urodeles or Xenopus (Holtfreter, 1944; Nakamura andKishiyama, 1971; Doucet-de Bruıne, 1973; Smith and Slack, 1983;Gimlich, 1985; Takasaki, 1987; Hardin and Keller, 1988; Baueret al., 1994; Suzuki et al., 2002), indicating that the bottle cells areinduced before the onset of gastrulation. It can therefore beassumed that bottle cell induction and mesoderm induction occurin a temporally synchronized manner.

Many studies have evaluated the synchronicity between bottlecell formation and mesoderm induction in the DMZ (Hardin andKeller, 1988; Black, 1989; Kurth and Hausen, 2000). In Xenopus, itappears that dorsal extension movement and axis formation areinhibited in an antimorphic gsc-injected embryo, but bottle cellformation is not (Ferreiro et al., 1998). On the other hand, Kurthand Hausen (2000) found that ectopic bottle cells are induced inActivin- or Xnr1-injected ectoderm, and they proposed that theprocesses of both bottle cell and mesoderm formation are closelyinterlinked in Xenopus. These findings indicate that simultaneousmesoderm induction and bottle cell induction in the DMZ isnecessary to initiate gastrulation and normal axis formation of theembryo. These induction processes may depend on a separate setof maternal molecules, as suggested for Xenopus (Sakai, 1996,2008; Nagano et al., 2000) and for Cynops (Doi et al. 2000).

In Cynops, ultraviolet-irradiated eggs form an abnormal blas-topore and thus dorsal axis formation is arrested (Doi et al., 2000;

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–18 11

Suzuki et al., 2002), while the notochord-inducing activity of theLDMZ remains active (Suzuki et al., 2002). Suzuki et al. (2002)treated Cynops embryos with anti-morphogenesis reagents suchas suramin (aspecific FGF-inhibitor, Gerhart et al., 1991; Grunz,1992; Oschwald et al., 1993; Cardellini et al., 1994; Wallingfordet al., 1997; Kaneda et al., 2002; Sakaguchi et al., 2002) ornocodazole (inhibitor for microtubule polymerization, Lane andKeller, 1997). Suramin injection and nocodazole treatment inhib-ited involution of the DMZ, but the blastopore formed normally.Cybra expression was activated in the nocodazole-treatedembryos but not in the suramin-injected embryos. The noto-chord-inducing activity of the LDMZ in the nocodazole-treatedgastrulae remained active, but the LDMZ of the suramin-injectedembryos completely lost its notochord-inducing activity. Theseresults strongly indicate that blastopore formation and dorsalmesoderm induction in the DMZ occur independently. Cynops

gastrulation indispensably requires bottle cells (Suzuki et al.,2002), but in Xenopus, involution of the deep layer occursindependently to bottle cell formation (Fig. 2A). Bottle cellsappear to make a significant but not the major contribution togastrulation in Xenopus (Keller, 1976; Hardin and Keller, 1988).These observations indicate that Cynops and Xenopus bottle cellshave different functions during gastrulation movements.

Pre-gastrulation movements and the onset of gastrulation

The movements of embryonic cells that occur before the onsetof gastrulation, as defined by bottle cell formation, have beeninvestigated in many amphibians. In Xenopus, these compriseepiboly (Keller, 1980; Bauer et al., 1994; Papan et al., 2007),unipolar ingression (Zust and Dixon, 1975), vegetal rotation(Winklbauer and Schurfeld, 1999) and the movements of pre-chordal plate migration and notochord CE morphogenesis(Nieuwkoop and Florschutz, 1950; Keller, 1975, 1976). Pre-gas-trulation movements mostly start from the mid-blastula stageonward. By these movements, the internal tissue arrangementsare drastically changed and the deep layer involutes indepen-dently of blastopore formation in Xenopus.

In contrast, urodeles undergo these pre-gastrulation movementsonly moderately or not at all (Schechtman, 1934; Nieuwkoop andFlorschutz, 1950; Harris, 1964; for reviews, see Nieuwkoop, 1996,1997). Komazaki (1992) demonstrated that the animal cap pre-sumptive ectoderm of the Cynops blastula and gastrula undergoesepiboly movement. By epibolic extension, the single-cell-layeredanimal cap is formed, but there are no structural changes of theembryo. These movements may not be so critical for DMZ config-uration in the Cynops embryo, and the simple planar (animal–vegetal) juxtaposition of the DMZ is retained before and after theonset of gastrulation (Fig. 1C). Consequently, retaining its initialstructure the surface DMZ starts to internalize through the blas-topore as a monolayer sheet (Figs. 2B, 3F).

The significance of the ‘‘pre-gastrula epiboly movement’’ onthe formation of the multilayered Xenopus DMZ has beendescribed. Using cell lineage tracing (Bauer et al., 1994) or time-lapse microscopic magnetic resonance imaging (Papan et al.,2007), it has been demonstrated that the tissue movementsleading to the formation of the three-dimensional multilayeredDMZ begin by stage 8 or at least by stage 9 in Xenopus.Pre-gastrula epiboly drastically moves animal cap tissue intothe DMZ, but not into the ventral marginal zone (Papan et al.,2007). Following this, the deep layer involutes independently ofblastopore formation by a series of pre-gastrulation movements(Fig. 5). Vertical interaction between the presumptive neuroecto-derm and involuting deep layer is thus established prior toformation of the blastopore. Traditionally defined gastrulationmovement starts after blastopore/bottle cell formation at stage

10/10þ in Xenopus. Archenteron formation and the verticalinteraction between neuroectoderm and the involuted ARF isfinished by stage 10.25 to 10.5.

In Xenopus, the pre-gastrulation movements are much moreextensive than commonly thought (Bauer et al., 1994; De Robertiset al., 1994; Papan et al., 2007), and substantially constitute theonset of gastrulation. Embryological evidence indicates thatXenopus gastrulation should be revised to beginning at stage8 as a two-step process: ‘‘pre-blastoporal gastrulation’’, whichstarts at stage 8 when the pre-gastrulation movements start, andthe traditionally defined gastrulation (‘‘traditional gastrulation’’)in which marginal zone cells involute through the blastopore.During pre-blastoporal gastrulation, dorsal mesoderm inductionin the DMZ should occur, because functional domains character-ized by their specific gene expression profile, self-differentiation,inducing activities and capacity for dorsal CE movements areestablished before the onset of traditional gastrulation.

On the other hand, urodeles synchronously undergo bottle cell(blastopore) formation and traditional gastrulation. In addition tothe pre-gastrulation movements, significant differences betweenanuran and urodelean gastrulation mechanics are also identified:no autonomous dorsal CE movements of the axial mesodermoccur in Pleurodeles (Shi et al., 1989), and the dorsal CE mechan-isms in the axial mesoderm are relatively weak but occur later inurodeles (Keller and Jansa, 1992; Shook et al., 2002). The planarcell polarity (PCP) pathway occurs in the trunk mesoderm(notochord), but not in the head mesoderm. As the dorsal CEmovements induced by the Wnt/PCP pathway are a specificcharacter of notochord (mesoderm) cells (Keller and Danilchik,1988; Keller and Shook, 2004), it is evident that urodelessequentially undergo, first, bottle cell/blastopore formation andthe onset of traditional gastrulation, and second, dorsal meso-derm induction, much later than in anura.

Meso-endoderm induction in the marginal zone

Nieuwkoop (1969a, b) and Nakamura et al. (1970) found inA. mexicanum and Xenopus that the vegetal endodermal hemisphereinduces dorsal mesoderm in the presumptive ectoderm. Thus itwas proposed that the Spemann–Mangold organizer is induced inthe DMZ by this ‘‘meso-endoderm induction’’. Further investigationrevealed that meso-endoderm inducing activity of the vegetalhemisphere is activated after MBT at stage 8 in Xenopus (Wylieet al., 1996) and thus meso-endoderm induction starts after MBTas proposed (Nagano et al., 2000; Vonica and Gumbiner, 2007;reviewed by Sakai, 2008). Furthermore, Nakamura et al. (1971)demonstrated that the inducing activity of the Xenopus DMZ firstappears at stage 9 and that the organizing activities of theXenopus DMZ are fixed until stage 10. As discussed earlier, pre-blastoporal gastrulation in Xenopus starts at stage 8 and theresults indicate that meso-endoderm induction of the Xenopus

DMZ occurs during pre-blastoporal gastrulation, although thenotochord-inducing activity of the superficial epithelium is main-tained during traditional gastrulation (Shih and Keller, 1992).

Several maternal molecules are found to be key molecules formeso-endoderm induction in Xenopus. Two models are currentlyproposed to explain this process. One is the three-signals model(Smith and Slack, 1983; Dale and Slack, 1987), in which localizeddistribution of maternal VegT (Zhang et al., 1998) and Wnt/ß-cateninin the oocyte cortex (Wylie et al., 1996; Moon and Kimelman, 1998)activate the GRN for meso-endoderm induction, and finally, signalsemanating from dorso-vegetal cells induce and specify the DMZ. Theother proposed model in Xenopus and Bufo is that the organizer iscell-autonomously specified by synergic action of maternal cyto-plasmic determinants without inducing signals from dorso-vegetalcells (Sakai, 1996; Manes and Campos Casal, 2002; Fujii et al., 2002;

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–1812

Katsumoto et al., 2003; for review, see Sakai, 2008). Although thesedeterminants are distributed more widely in the egg, a similarsituation is also indicated in Cynops (Doi et al., 2000).

VegT shows diverse distribution and timing of its expression inembryos. Maternal VegT is localized in the vegetal cortex of theoocyte in both Xenopus (Zhang et al., 1998) and Rana pipiens (Nathet al., 2005), but at the animal pole region in Eleutherodactylus

coqui (Beckham et al., 2003). The timing of VegT expression alsovaries in vertebrates. Xenopus expressed both maternal andzygotic VegT (e.g., Fukuda et al., 2010), but only zygotic VegT isexpressed in urodeles, chicks, mice and humans (Elinson andBeckham, 2002). In A. mexicanum, AmVegT mRNA is zygoticallyexpressed (Nath and Elinson, 2007). Our preliminary experimentsalso indicate that CyVegT is zygotically expressed after the lateblastula stage (Figs. 4 and 5; Motoki et al., unpublished). Despitethis diversity, it appears that localized distribution of VegT in theegg or embryo triggers the GRN for meso-endoderm induction viaseveral members of the FGF and TGF-ß families of growth factorssuch as Nodal-related protein and Mix (e.g., Loose and Patient,2004; Luxardi et al., 2010). The early gastrula DMZ includesmultiple heterologous domains with different fates and functions(Figs. 1 and 2). These domains are obviously specified by themeso-endoderm induction process, but which subdomain theVegT/Nodal signaling network is targeted to establish remainsunclear. However, at least in Cynops, the suprablastoporalgsc-expressing LDMZ, which has potent notochord-inducing activity,should be the target tissue of this induction, because notochord(dorsal mesoderms) is not yet induced in the early gastrula.

Notochord induction before and after the onset of gastrulation

In Xenopus, although the size and spatial location variesaccording to the criteria used, the stage 10 early gastrula DMZ(Smith and Slack, 1983) or the core region of the DMZ (Fujii et al.,2002) shows notochord and endoderm differentiation. Gsc isexpressed after MBT and reaches its maximum level just beforethe onset of traditional gastrulation. Xbra expression starts afterMBT and is clearly detected at stage 9, but not at stage 8/8.5(e.g., Smith et al., 1991; Eimon and Harland, 2002; Fukuda et al., 2010,see Fig. 5A). Thus, it is possible that in Xenopus the gsc-expressingtissues such as the deep layer LEM and superficial endoderm areinduced earlier than the Xbra-expressing tissues, rather than bothgsc- and Xbra-expressing tissues being induced at the same time.

In Cynops, the entire prospective notochord is not yet inducedto become notochord until the mid-gastrula (Figs. 5, 7, 8).To clarify when and how notochord is induced, the spatiallocation of the presumptive notochord can be traced at eachstage of the gastrulae and the temporal changes in self-differ-entiation and neural-inducing activity analyzed. It is clear thatnotochord/somite self-differentiation capacity and trunk–tailneural-inducing activity of the presumptive notochord is evokedfrom the mid-gastrula stage onward (Kaneda and Hama, 1979).The notochord- and Cybra-inducing activity of the LDMZ disap-pears soon after involution or culture in vitro (Kaneda, 1980,1981; Suzuki et al., 1984; Kaneda et al., 2002; see Fig. 10). At themid-gastrula when the involuted LDMZ underlies the surfacepresumptive notochord ((Fig. 9B, C), it has already been inducedto become notochord. Using a Keller sandwich of the earlygastrula DMZ confirmed that the Cygsc-expressing LDMZ planarlyinduces Cybra expression in the UDMZ (Fig. 8F). Thus, it isproposed that notochord is induced by planar induction signalsfrom the LDMZ during the early to mid-gastrula stages (Kanedaet al., 2009), which indicates that formation of the LDMZ is thedirect target of meso-endoderm induction, and that the LDMZ inturn acts as the source of notochord-inducing signals.

Taking all these findings into consideration, the differences inthe state of determination of the early Cynops and Xenopus

gastrula DMZ and the process for notochord induction can besummarized as follows.

(1)

At the onset of traditional gastrulation, dorsal mesoderm is notyet induced in any part of the Cynops embryo. Only a domainthat has notochord-inducing activity exists at the suprablas-toporal endodermal prechordal region (i.e., LDMZ). The activityis restricted to this region and does not extend to the sub-blastoporal endoderm region (Fig. 2B). Similar activity isdistributed over a rather broad region of the suprablastoporalregion in Xenopus, including superficial endoderm and deeplayer LEM, all of them having notochord-inducing activity(Stewart and Gerhart, 1990, 1991; Shih and Keller, 1992).

(2)

In Xenopus, the gsc-expressing deep layer LEM and the super-ficial endoderm will be induced and established earlier thanXbra/Xnot-expressing notochord (Fig. 5A). These gsc-expres-sing tissues play an essential role in inducing notochordduring pre-blastoporal gastrulation. Thus, notochord inductionin Xenopus is completed prior to the onset of traditionalgastrulation. Cynops undergoes these processes after the onsetof traditional gastrulation (Figs. 2, 5B).

(3)

LDMZ planarly induces notochord in the neighboring cellsduring the early phase of traditional gastrulation in Cynops

(Figs. 8 and 9). On grafting the LDMZ, a complete secondary axisis organized (Yamamoto and Suzuki, 1994; Imoh et al. 1998),indicating that grafted LDMZ can induce an anteroposteriorregional difference in inducing activity of the secondary inducednotochord (Fig. 3E). Secondary axis-forming gastrulation is thusinduced by both the grafted LDMZ and induced notochord. Thisdemonstrates that the initial action of the LDMZ is necessaryand sufficient to trigger the sequential morphological networkfor the body plan formation (Sakaguchi et al., 2002).

(4)

Although the spatial and temporal locations vary, notochordor dorsal mesoderm-inducing activity of endoderm aroundthe blastopore may be conserved in vertebrates. Cynops, andperhaps other urodeles, evolved this as the LDMZ.

As summarized above, Xenopus undergoes notochord (dorsalmesoderm) induction prior to the onset of involution by a seriesof pre-gastrulation movements. During this process, planar andvertical intra-DMZ interactions facilitate establishment of theregional specification of the multilayered DMZ.

Although differences of the early gastrula DMZs are classifiedas above, there should be common mechanisms among them.To unify these differences, at least two problems need to be con-sidered. One is the rearrangement of the organizing activity of thelate blastula to early gastrula DMZ. In addition to the traditionalconcept of head and trunk–tail organizers, Xenopus embryologydefines several organizers, centers and/or regions that havespecific organizing activities on the blastula and gastrula embryo.For example, the blastula organizer, gastrula organizer, Nieuw-koop center, surface endodermal epithelium and deep endodermare classified as leading tissues in Xenopus embryogenesis (forreview, see Gerhart, 2001). The equivalent tissue has beeninvestigated in other vertebrates (e.g., for reviews, see Arendtand Nubler-Jung, 1999; Stern, 2001; Solnica-Krezel, 2005); how-ever, because these tissues are located in a restricted area of arather small embryo and are spatially continuous, and becausethey change their spatial location and function with time, it isdifficult to introduce unifying criteria even in the Anura family.The Cynops LDMZ equivalent tissue in Xenopus may be superficialendodermal epithelium and deep endoderm, however, re-evalua-tion and rearrangement of the multiple organizers and centers in

Fig. 9. Planar notochord induction in the upper dorsal marginal zone during the early phase of Cynops gastrulation. (A, B) Spatial and temporal process of anteroposterior

regional patterning of the involuting and extending ARF. At the early gastrula stage, the Cygsc-expressing LDMZ is identified as the ‘‘prechordal region’’ and the notochord

is not yet induced in the UDMZ. During early to mid-gastrulation, the LDMZ planarly induces Cybra-expressing notochord in the UDMZ. As involution proceeds, the LDMZ

loses its notochord-inducing activity and forms the fore-notochordal endodermal roof (FNE). Cygsc expression of the LDMZ is progressively restricted to the FNE just

anterior to the induced notochord, and finally confined to the prechordal plate. (C) Spatial and temporal expression patterns of Cybra and Cygsc during gastrulation.

At stage 11, Cygsc expression is restricted to the surface region of the suprablastoporal endodermal prechordal region (LDMZ); however, Cybra expression is not identified

in any part of the embryo. Cybra expression is first induced in the surface presumptive notochord at the mid-gastrula stage (stage 12b). At this stage, Cygsc-expressing cells

involute and underlie the surface Cybra-expressing region. Despite the Cygsc expression, the anterior two-thirds of the ARF at this stage show no inducing activity. At the

late gastrula stage (stage 13c), Cygsc expression is progressively restricted to the FNE anterior to the induced notochord. At this stage, Cybra is expressed in the posterior

one-third of the ARF and the surface region around the blastopore. In neurulae, Cybra expression is restricted to notochord, and the Cygsc-expressing region is confined to

the prechordal plate beneath the anterior neural plate. Arrows indicate the anterior margin of the involuting ARF. In B and C, left and right represent the anterior and

posterior directions, respectively. blp: blastopore; yp: yolk plug. (B and C reproduced from Kaneda et al., 2009.).

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–18 13

Xenopus is required, especially around stage 8 at the onset of pre-blastoporal gastrulation.

The other problem is the criterion of staging the embryo.Traditionally, the stage at which the visible blastopore or pigmentline has just formed is defined as the first stage of the gastrula.Xenopus stage 10/10þ and Cynops stage 11 are thus defined as thefirst stage of the gastrula (Fig. 1). Although the final body pattern isstrongly conserved, and although many of the same developmentalgenes are expressed, there are essential differences between Xenopus

and Cynops in self-differentiation, gene expression patterns andorganizing activity of each part of the embryo. It is worthwhileidentifying the equivalent stages between different species using acombination of the gene expression profiles in the area of definiteprospective fate and the morphological characters. Using gsc andBra expressions as a marker, De Robertis et al. (1994) haveproposed that the stages of maximal gsc expression are equivalentin many organisms. Furthermore, the tissue with maximal gsc

expression can be identified as homologous, such as the DMZ inXenopus and the early Hensen’s node in chicks. In addition tothese definitions, it may be most important to define the criteriafor the substantive onset of gastrulation.

According to these definitions, it is reasonable to conclude thatXenopus stage 10/10þ is equivalent to the Cynops mid-gastrula(stage 12b/c), the stage at which notochord has been induced andneural induction between the induced notochord and futureneuroectoderm begins (Kaneda et al., 2009). As discussed againlater, Xenopus traditional gastrulation should be specified as aprocess to realize neural patterning (‘‘neural induction phase’’)by inductive interactions between the already patterned ARF andthe overlaying ectoderm, as suggested by Koide et al. (2002).Of course, other important events occur in this period, such asinternalization of all the endoderm leading to archenteron for-mation, rearrangement of much of the mesoderm, and in Xenopus

perhaps further induction of the somatic mesoderm by BMPantagonists such as Noggin and Chordin occurs. On the otherhand, the Cynops stage 11 beginning gastrula is equivalent tostage 8/9 of Xenopus, because pre-blastoporal gastrulation startsat this stage. Sequential events for notochord (dorsal mesoderms)induction and patterning of the induced notochord that occurduring the early to mid-gastrula stages (‘‘notochord inductionphase’’) in Cynops should occur at stage 8 to 10 in Xenopus. It isthus necessary to rearrange the criteria for the real onset of

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–1814

gastrulation to elucidate and verify the similarities and dissim-ilarities of the basic morphogenetic process.

Fig. 10. Change in the inducing activity of the lower dorsal marginal zone (LDMZ)

from trunk–tail to head. (A) Change in the organizing activity of the LDMZ during

cultivation in vitro. The LDMZ was isolated and cultured for up to 30 h at 20 1C.

After cultivation, the organizing activity of the LDMZ was assayed by sandwich

assay method. At 0–6 h of culture, the LDMZ induced trunk–tail structures (green

bar) associated with notochord (blue line). The trunk–tail-inducing activity of the

LDMZ disappeared during cultivation, and the activity changed to induction of

head (forebrain) by cultivation for more than 12 h. Head-inducing activity was

transiently maintained and then disappeared with more prolonged cultivation.

(Redrawn from the data of Suzuki et al., 1984); (B) Schematic representation of the

change in the inducing activity of the LDMZ from trunk–tail to head. At the early

gastrula stage, the LDMZ expresses Cygsc, CyLim, Cychd and Cynog, but not Cybra.

In the sandwich assay, LDMZ induces notochord and the induced notochord in

turn organizes the trunk–tail structures. The notochord- or Cybra-inducing activity

of the LDMZ is lost by in vitro cultivation or suramin treatment. Expression of

Cychd and Cynog is maintained in the LDMZ during cultivation in vitro, so these

gene products may function in the neural-inducing activity, and can induce head

structures. But the activity is transiently maintained. In the normal embryo, the

loss of notochord-inducing activity of the LDMZ occurs soon after involution until

the mid-gastrula stage. (Redrawn from Kaneda et al., 2002). (For interpretation of

the references to color in this figure legend, the reader is referred to the web

version of this article.)

What is the true nature of the Spemann–Mangold organizer?

Dividing the urodelean neurula ARF into several parts, thesecondary axis-inducing activity (Mangold, 1933) or neural-indu-cing activity (Sala, 1955; Johnen, 1956a, 1956b) of each part of theneurula ARF was analyzed. The results show that most anteriorquarter (anterior endoderm and prechordal plate) has littleinducing activity, but the head and trunk–tail-inducing activityis broadly distributed from the second quarter (anterior noto-chord) to the posterior end of the ARF. These results support theexistence of separate head and trunk–tail organizers, as firstproposed by Spemann (1931). However, these experiments weredone at the neurula stage, in which almost all of the basicmorphogenesis has been completed. More recently, the role ofthe Spemann–Mangold organizer in the anteroposterior pattern-ing of the trunk–tail in Xenopus has been analyzed (Jansenet al., 2007; Durston et al., 2010) and it has been proposedthat interaction between the Spemann–Mangold organizer andnon-organizer mesoderm during gastrulation establishes ananteroposterior axis in Xenopus. However, it remains unclearwhen and how the final regional difference in the inducingactivities of the ARF is established, when neural induction isinitiated, and the cellular source of signals for anteroposteriorCNS patterning.

A crucial embryological observation was that the inducingactivity of the Cynops DMZ changes on invagination (Okada andTakaya, 1942a). The Cynops early gastrula DMZ originally inducestrunk–tail structures, whereas the same tissue, if isolated afterinvolution, induces the head. This change in the organizingactivity of the early gastrula DMZ has been confirmed and it hasbeen revealed that the change is mimicked when the DMZ isisolated and cultured in vitro for a period of time (Okada andTakaya, 1942b; Okada and Hama, 1943, 1944; Suzuki et al., 1984;for review, see Takaya, 1978; Slack and Tannahill, 1992; Okada,1994). Importantly, a similar experiment was performed usingRana nigromaculata and its early gastrula DMZ induced the head(Okada and Takaya, 1942a), which demonstrated that the indu-cing activity of the early gastrula DMZ is determined considerablyearlier in Rana than in Cynops. These phenomena were fullyconfirmed later in A. mexicanum (Hoessels, 1957; for review, seeNieuwkoop, 1997). The change in the inducing activity of theearly gastrula LDMZ has been further confirmed in Cynops (Suzukiet al., 1984; Yokota et al., 1998; Kaneda et al., 2002). Furthermore,Ariizumi and Asashima (1994, 1995) and Ninomiya et al. (1998,1999) treated Cynops and Xenopus animal cap ectoderm with ahigh dose of activin A and then cultured it in vitro for a period oftime. There was a change in its neural-inducing activity fromtrunk–tail to head. Although the change in the inducing activity ofthe Xenopus DMZ has not been investigated, these results indicatethat Xenopus DMZ shows the same change in its organizingactivity in vivo.

The LDMZ of the early Cynops gastrula is confirmed to havenotochord- and Cybra-inducing activities (Fig. 8). The notochord-or Cybra-inducing activity of the LDMZ is lost during normalinvolution, cultivation in vitro or with suramin treatment, buttransiently retains weak but definite head-organizing activity(Suzuki et al., 1984; Fig. 10A). It is confirmed that Cygsc, Cychd,

Cynog and CyLim-1 expressions of the LDMZ are retained duringin vitro cultivation (Yokota et al., 1998; Kaneda et al., 2002). Theseresults show that the LDMZ induces notochord first, and theinduced notochord in turn organizes the trunk–tail structure(Fig. 10B). Thus, the uninvoluted LDMZ apparently shows

trunk–tail organizing activity. On the other hand, after involutionor culture in vitro, the notochord- and Cybra-inducing activities ofthe LDMZ disappear, and it changes its organizing activity. Whenthe late gastrula DLP of Cynops, which has been committed tonotochord, is isolated and cultured in vitro, a change in inducingactivity is scarcely observed (unpublished). These observationssuggest that the change in organizing activity of the LDMZ is

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–18 15

explained by the loss of notochord- and Cybra-inducing activity(Fig. 10B). Thus, the change in the organizing activity of theorganizer is a specific feature of the LDMZ. Furthermore, the trueform of the Spemann–Mangold organizer is indicated to besuprablastoporal endoderm that has notochord-inducing activity,at least in Cynops, and perhaps also in other urodeles.

The other questions are when neural induction is initiated, andthe cellular source of signals for anteroposterior CNS patterning.Although the stage criteria and positioning of the ARF differ amongstudies, the anteroposterior regional inducing activities of theinvoluting and extending ARF have been investigated in Cynops

(Okada and Hama, 1945; Suzuki et al., 1975; Kaneda et al., 2009),and in A. mexicanum and P. waltlii (Eyal-Giladi, 1954). Dividing theCynops gastrula ARF into several parts, the anterior one-third (Suzukiet al., 1975) or anterior half (Okada and Hama, 1945; Kaneda et al.,2009) of the mid- to late gastrula ARF is shown to have no neural-inducing activity. These results demonstrate that the neural-indu-cing activity of the involuting and extending ARF of the mid- to lategastrula of Cynops is restricted to the involuting presumptivenotochord (posterior half of the ARF), which has been induced tobecome notochord by planar signals from the LDMZ. Vertical contactbetween the induced notochord and the overlying presumptiveneuroectoderm occurs from the mid-gastrula stage onward (Fig. 9B),indicating that vertical neural induction starts at the mid-gastrulastage in Cynops. On the other hand, Poznanski and Keller (1997)demonstrated that vertical neural induction could begin within thefirst half hour of traditional gastrulation (between stages 10 and10þ) in Xenopus.

In general, the prechordal plate is thought to be the source ofhead-organizing signals in amphibians (e.g., Niehrs, 2004). How-ever, it has been shown that the prechordal region does not haveneural-inducing activity in the avian embryo (Foley et al., 1997).In Cynops, the Cygsc-expressing prechordal region forms the FNEafter involution (Figs. 2, 5, 8) but the FNE has no inducingactivities. These results suggest that not the prechordal plate(region) but the notochord plays an essential role in neuralinduction. These observations in Cynops, and perhaps also inother urodeles, show that traditional gastrulation has two phases:the early phase of notochord induction in the early to mid-gastrula stage and the later phase of neural induction at themid- to late gastrula stage (Kaneda et al. 2009). This suggests thatneural induction in Cynops is initiated at the mid-gastrula stage,and the final CNS pattern is determined from the neurula stageonward as indicated in Xenopus (Saha and Grainger, 1992). On theother hand, Xenopus undergoes the notochord induction phaseduring pre-blastoporal gastrulation and the neural inductionphase during traditional gastrulation.

In addition, the developing notochord induces neural tissue of asomewhat broader regional character than would be expected fromits final position in the embryo in Cynops. This indistinct regionaldifference in the neural-inducing activity of the presumptive noto-chord indicates that forebrain-inducing activity is not restricted to aspecific region, but is distributed over a broad region of the entirepresumptive notochord during the gastrula stages (Kaneda et al.,2009). Therefore, the traditional concept of fixed tissues of the headand trunk–tail organizers needs to be re-evaluated, especially for thehead organizer. Changing their position and function with time, andinteracting with each other, the inducing activities of the ARF areprogressively established during gastrulation.

Conclusion

Amphibian embryology has developed at least three generalconcepts of body plan formation. One is the significance ofspatially and temporally controlled sequential inductive

interactions. Uneven distribution of maternal molecules in thefertilized egg activates the GRN for dorsoventral and animal–vegetal regional specification in the embryo. The early gastrulaDMZ is thus specified and in turn triggers the subsequentprocess that forms the anteroposterior CNS pattern. The secondis the significance of morphogenetic movements. Cells that arecommitted to differentiate into their own fate acquire thecapacity for cell-autonomous movements, such as CE move-ments. By these acquired capacities, cells initially locatedremotely can interact with each other by topological changesin the embryonic germ layers. The third is the synergy ofsequential inductions and cell movements. Synchronicity ofthe morphogenetic movements and the spatiotemporally con-trolled induction process are necessary to establish the correctbody pattern. Experiments analyzing these processes have beenextensively performed using many amphibian species as mod-els, and earlier works paid special attention to the thirdconcept. Compiling the results from individual divergent spe-cies, the unifying features of the organizer or the processes forearly body pattern formation have been theorized. Undoubt-edly, Xenopus has advantages for analyzing the cellular andmolecular nature of morphogenetic movements, and meso-endoderm and neural induction processes. However, Xenopus

shows a rather extreme mode of internal and external gastrula-tion, germ layer structure, origin of internal mesoderm andthe timing of dorsal mesoderm induction. It is thereforedifficult but useful to compare the experimental results fromXenopus with those from other animals such as urodeles, and itis also important to note that comparisons among amphibiansprovide insights for extending these results to the othervertebrates.

The current understanding of Xenopus development providesa comparative background for the analysis of amphibian devel-opment. Recent works on the comparative gene expressionpatterns provide evidence that the molecular mechanismsunderlining early embryogenesis are conserved to a large extent.However, it has also become evident that timing and/or spatialgene expression patterns and the onset of morphogenetic pro-cesses are unexpectedly diverse among individual species.Despite the growing evidence, it is not yet clear whether ornot this diversity represents essential embryological differences.This situation may arise largely because of the difficulties inidentifying and comparing the equivalent stage or homologoustissues in each organism that has distinct germ layer construc-tion, timing of mesoderm and neural induction, and initialsigns of the onset of gastrulation. Combining the morphologicalcharacters and the gene expression patterns, we proposethat amphibian gastrulation is divided into two phases: anotochord induction phase and a neural induction phase. Uro-deles sequentially undergo these phases after the onset oftraditional gastrulation, whereas Xenopus undergoes the noto-chord induction phase during pre-blastoporal gastrulation. InXenopus, traditional gastrulation is specifically a process torealize neural induction as well as to internalize endoderm,rearrange mesoderm and reorganize the germ layer. Thus, themolecular and cellular characteristics of the Xenopus earlygastrula would be equivalent to the mid-gastrula of urodeles.On the other hand, the event that occurs during the early tomid-gastrula stage of urodeles represents the events that occurduring the pre-gastrula stage of Xenopus (Fig. 5). Furthermore,comparison of the structure, self-differentiation, fate andfunction of each domain of the early gastrula DMZ indicatesthat the true form of the Spemann–Mangold organizer issuprablastoporal gsc-expressing endoderm that has notochord-inducing activity. There is clearly still much to learn from theembryo.

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–1816

Acknowledgments

We thank Dr. Masao Sakai (Kagoshima Univ.) for his readingand comments on the manuscript, and Drs. Takashi Ariizumi(Tamagawa Univ. Tokyo) and Chikara Hashimoto (JT BiohistoryResearch Hall, Takatsuki, Osaka) for their reading of the manu-script. Thanks are also due to all members of our laboratory fortheir help. T. K. sincerely thanks his late professor, Tadao Hama(Nagoya Univ.), and professor Pieter D. Nieuwkoop (Univ. Utrecht)for their instruction in urodele embryology. Special thanks arealso due to Professor Emeritus, Akio S. Suzuki (Kumamoto Univ.)and Hiroshi Imoh (Miyazaki Univ.) for their continuous encour-agement and discussions. We also acknowledge the manyresearchers whose work we could not cite because of spaceconstraints. We sincerely thank the unknown reviewers for theirunusually detailed criticisms and helpful suggestions.

References

Arendt, D., Nubler-Jung, K., 1999. Rearranging gastrulation in the name of yolk:evolution of gastrulation in yolk-rich amniote eggs. Mech. Dev. 81, 3–22.

Ariizumi, T., Asashima, M., 1994. In vitro control of the embryonic form of XenopusIaevis by activin A: time and dose-dependent inducing properties of activin-treated ectoderm. Dev. Growth Differ 36, 499–507.

Ariizumi, T., Asashima, M., 1995. Head and trunk–tail organizing effects of thegastrula ectoderm of Cynops pyrrhogaster after treatment with activin A.Roux’s Arch. Dev. Biol. 204, 427–435.

Bachvarova, R.F., Masi, T., Hall, L., Johnson, A.D., 2001. Expression of Axwnt-8 andAxszl in the urodele, axolotl: comparison with Xenopus. Dev. Genes Evol. 211,501–505.

Bauer, E.V., Huang, S., Moody, S.A., 1994. The cleavage stage origin of Spemann’sOrganizer: analysis of the movements of blastomere clones before and duringgastrulation in Xenopus. Development 120, 1179–1189.

Beckham, Y.M., Nath, K., Elinson, R.P., 2003. Localization of RNAs in oocytes ofEleutherodactylus coqui, a direct developing frog, differs from Xenopus laevis.Evol. Dev. 5, 562–571.

Black, S.D., 1989. Experimental reversal of the normal dorsal-ventral timing ofblastopore formation does not reverse axis polarity in Xenopus laevis embryos.Dev. Biol. 134, 376–381.

Bouwmeester, T., Kim, S.-H., Sasai, Y., Lu, B., De Robertis, E.M., 1996. Cerberus is ahead-inducing secreted factor expressed in the anterior endoderm of Spe-mann’s organizer. Nature 382, 595–601.

Bouwmeester, T., 2001. The Spemann–Mangold organizer: the control of fatespecification and morphogenetic rearrangements during gastrulation in Xenopus.Int. J. Dev. Biol. 45, 251–258.

Cardellini, P., Polo, C., Coral, S., 1994. Suramin and heparin: aspecific inhibitors ofmesoderm induction in the Xenopus laevis embryo. Mech. Dev. 45, 73–87.

Callery, E.M., 2006. There’s more than one frog in the pond: a survey of theAmphibia and their contributions to developmental biology. Semin. Cell Dev.Biol. 17, 80–92.

Campos Casal, F.C., Manes, M.E., 1999. Ectopic mesodermal expression promotedby the eight-cell stage Bufo arenarum subequatorial cytoplasm. Dev. GenesEvol. 209, 421–426.

Cho, K.W., Blumberg, B., Steinbeisser, H., De Robertis, E.M., 1991. Molecular natureof Spemann’s organizer: the role of the Xenopus homeobox gene goosecoid. Cell 67,1111–1120.

Cleine, J.H., Slack, J.M.W., 1985. Normal fates and states of specification of differentregions in the axolotl gastrula. J. Embryol. Exp. Morph. 86, 247–269.

Collazo, A., Keller, R., 2010. Early development of Ensatina eschscholtzii: anamphibian with a large, yolky egg. Evo. Devo. 1, 6.

Conlon, F.L., Smith, J.C., 1999. Interference with Brachyury function Inhibitsconvergent extension, causes apoptosis, and reveals separate requirementsin the FGF and Activin signalling pathways. Dev. Biol. 213, 85–100.

Dale, L., Slack, J.M.W., 1987. Regional specification within the mesoderm of earlyembryos of Xenopus laevis. Development 100, 279–295.

Davidson, E.H., Erwin, D.H., 2006. Gene regulatory networks and the evolution ofanimal body plans. Science 311, 796–800.

De Robertis, E.M., 2009. Spemann’s organizer and the self-regulation of embryonicfields. Mech. Dev. 126, 925–941.

De Robertis, E.M., Fainsod, A., Gont, L.K., Steinbeisser, H., 1994. The evolution ofvertebrate gastrulation. Dev. Suppl., 117–124.

Dettlaff, T.A., 1983. A study of the properties, morphogenetic potencies andprospective fate of outer and inner layers of ectodermal and chordamesoder-mal regions during gastrulation, in various anuran amphibians. J. Embryol.Exp. Morph. 75, 67–86.

del Pino, E.M., 1996. The expression of Brachyury (T) during gastrulation in themarsupial frog Gastrotheca riobambae. Dev. Biol. 177, 64–72.

del Pino, E.M., Venegas-Ferrın, M., Romero-Carvajal, A., Montenegro-Larrea, P.,Saenz-Ponce, N., Moya, I.M., Alarcon, I., Sudou, N., Yamamoto, S., Taira, M.,

2007. A comparative analysis of frog early development. Proc. Nat. Acad. Sci.U.S.A. 104, 11882–11888.

Delarue, M., Sanchez, S., Johnson, K.E., Darrib�ere, T., Boucaut, J.-C., 1992. A fate mapof superficial and deep circumblastoporal cells in the early gastrula ofPleurodeles waltl. Development 114, 135–146.

Delarue, M., Johnson, K.E., Boucaut, J.-C., 1994. Superficial cells in the earlygastrula of Rana pipiens contribute to mesodermal derivatives. Dev. Biol.165, 702–715.

Doi, J., Niigaki, H., Sone, K., Takabatake, T., Takeshima, K., Yasui, K., Tosuji, H.,Tsukahara, J., Sakai, M., 2000. Distribution of dorsal-forming activity in pre-cleavage embryos of the Japanese newt, Cynops pyrrhogaster: effects ofdeletion of vegetal cytoplasm, UV irradiation, and lithium treatment. Dev.Biol. 223, 154–168.

Domingo, C., Keller, R., 1995. Induction of notochord cell intercalation behaviorand differentiation by progressive signals in the gastrula of Xenopus laevis.Development 121, 3311–3321.

Doucet-de Bruıne, M.H.M., 1973. Blastopore formation in Ambystoma mexicanum.Roux’s Arch. Dev. Biol. 173, 136–163.

Durston, A.J., Jansen, H.J., Wacker, S.A., 2010. Time-space translation regulatestrunk axial patterning in the early vertebrate embryo. Genomics 95, 250–255.

Eimon, P.M., Harland, R.M., 2002. Effects of heterodimerization and proteolyticprocessing on Derri�ere and Nodal activity: implications for mesoderm induc-tion in Xenopus. Development 129, 3089–3103.

Elinson, R.P., Beckham, Y., 2002. Development in frogs with large eggs and theorigin of amniotes. Zoology 105, 105–117.

Elinson, R.P., Kezmoh, L., 2010. Molecular Haeckel. Dev. Dyn. 239, 1905–1918.Eyal-Giladi, H., 1954. Dynamic aspects of neural induction in amphibia. Arch. Biol.

65, 179–259.Ferreiro, B., Artinger, M., Cho, K.W.Y., Niehrs, C., 1998. Antimorphic goosecoids.

Development 125, 1347–1359.Foley, A.C., Storey, K.G., Stern, C.D., 1997. The prechordal region lacks neural

inducing ability, but can confer anterior character to more posterior neuroe-pithelium. Development 124, 2983–2996.

Fujii, H., Nagai, T., Shirasawa, H., Doi, J., Yasui, K., Nishimatsu, S., Takeda, H., Sakai,M., 2002. Anteroposterior patterning in Xenopus embryos: egg fragment assaysystem reveals a synergy of dorsalizing and posteriorizing embryonicdomains. Dev. Biol. 252, 15–30.

Fukuda, M., Takahashi, S., Haramoto, Y., Onuma, Y., Kim, Y.-G., Yeo, C.-Y., Ishiura, S.,Asashima, M., 2010. Zygotic VegT is required for Xenopus paraxial mesodermformation and is regulated by Nodal signaling and Eomesodermin. Int. J. Dev.Biol. 54, 81–92.

Gerhart, J.C., Doniach, T., Stewart, R., 1991. Organizing the Xenopus organizer. In:Keller, R., Clark Jr., W.H., Griffin, F. (Eds.), Gastrulation, Movements, Patterns,and Molecules. Plenum Press, New Yolk and London, pp. 57–77.

Gerhart, J.C., 2001. Evolution of the organizer and the chordate body plan. Int. J.Dev. Biol. 45, 133–153.

Gimlich, R.L., 1985. Cytoplasmic localization and chordamesoderm induction inthe frog embryo. J. Embryol. Exp. Morph. 89, 89–111.

Grunz, H., 1992. Suramin changes the fate of Spemann’s organizer and preventsneural induction in Xenopus laevis. Mech. Dev. 38, 133–142.

Hama, T., 1978. Dynamics of the organizer. New findings on the regionality andmorphogenetic movement of the organizer. In: Nakamura, O., Toivonen, S.(Eds.), Organizer-A Milestone of a Half-Century From Spemann. Elsevier/NorthHolland Biomedical Press, pp. 71–90.

Hama, T., Tsujimura, H., Kaneda, T., Takata, K., Ohara, A., 1985. Inductive capacitiesfor the dorsal mesoderm of the dorsal marginal zone and pharyngealendoderm in the very early gastrula of the newt, and presumptive pharyngealendoderm as an initiator of the organization center. Dev. Growth Differ. 27,419–433.

Hamburger, V., 1988. The Heritage of Experimental Embryology: Hans Spemannand the Organizer. Oxford Univ. Press, Oxford.

Hardin, J., Keller, R.E., 1988. The behavior and function of bottle cells duringgastrulation of Xenopus laevis. Development 103, 211–230.

Harland, H., Gerhart, J., 1997. Formation and function of Spemann’s organizer.Annu. Rev. Cell Dev. Biol. 13, 611–667.

Harris, T.M., 1964. Pregastrular mechanisms in the morphogenesis of the sala-mander Ambystoma maculatum. Dev. Biol. 10, 247–268.

Hoessels, E.L.M.J., 1957. Evolution de la plaque prechordale d’ambystoma mex-icanum; sa differenciation propre et sa puissance inductricependant lagastrulation. PhD Thesis, Univ. Utrecht. (In French).

Holtfreter-Ban, H., 1965. Differentiation Capacities of Spemann’s Organizer Inves-tigated in Explants of Diminishing Size, PhD Thesis, Univ. Rochester.

Holtfreter, J., 1938a. Differenzierungspotenzen isolierter Teile der UrodelenGastrula. W. Roux’ Arch. Entw. Mech. Org. 138, 522–656.

Holtfreter, J., 1938b. Differenzierungspotenzen isolierter Teile der Anuren Gas-trula. W. Roux’ Arch. Entw. Mech. Org 138, 657–738, in German.

Holtfreter, J., 1944. A study of the mechanics of gastrulation. J. Exp. Zool. 95,171–212.

Hoppler, S., Moon, R.T., 1998. BMP-2/-4 and Wnt-8 cooperatively pattern theXenopus mesoderm. Mech. Dev. 71, 119–129.

Imoh, H., 1988. Formation of germ layers and roles of the dorsal lip of theblastopore in normally developing embryos of the newt Cynops pyrrhogaster. J.Exp. Zool. 246, 258–270.

Imoh, H., Yamamoto, Y., Terahara, T., Moody, S.A., Suzuki, A.S., 1998. Timingand mechanisms of mesodermal and neural determination revealed by

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–18 17

secondary embryo formation in Cynops and Xenopus. Dev. Growth Differ. 40,439–448.

Ito, Y., Oinuma, T., Takano, K., Komazaki, S., Obata, S., Asashima, M., 2006. CyNodal,the Japanese newt nodal-related gene, is expressed in the left side of thelateral plate mesoderm and diencephalon. Gene Exp. Pattern 6, 294–298.

Jansen, H.J., Wacker, S.A., Nabila Bardine, N., Durston, A.J., 2007. The role of theSpemann organizer in anterior–posterior patterning of the trunk. Mech. Dev.124, 668–681.

Johnen, A.G. 1956a. Experimental studies about the temporal relationships in theinduction process. I. Experiments on Amblystoma mexicanum. Proc. Kon. Ned.Acad. V. Wet. A’dam. Ser. C, 59, 554–561.

Johnen, A.G. 1956b. Experimental studies about the temporal relationships in theinduction process. II. Experiments on Triturus vulgaris. Proc. Kon. Ned. Acad. V.Wet. A’dam. Ser. C, 59, 652-660.

Kaneda, T., 1980. Studies on the formation and state of determination of the trunkorganizer in the newt, Cynops pyrrhogaster. II. Inductive effect from theunderlying cranial archenteron roof. Dev. Growth Differ. 22, 841–849.

Kaneda, T., 1981. Studies on the formation and state of determination of the trunkorganizer in the newt Cynops pyrrhogaster. III. Tangential induction in thedorsal marginal zone. Dev. Growth Differ. 23, 553–564.

Kaneda, T., Hama, T., 1979. Studies on the formation and state of determination ofthe trunk organizer in the newt Cynops pyrrhogaster. Roux’s Arch. Dev. Biol.187, 25–34.

Kaneda, T., Suzuki, A.S., 1983. Studies on the formation and state of determinationof the trunk organizer in the newt, Cynops pyrrhogaster. IV. The association ofneural inducing activity with the mesodermization of the trunk organizer.Roux’s Arch. Dev. Biol. 192, 8–12.

Kaneda, T., Miyazaki, K., Kudo, R., Goto, K., Sakaguchi, K., Matsumoto, M., Todaka, S.,Yoshinaga, K., Suzuki, A.S., 2002. Regional specification of the head and trunk–tailorganizers of a urodele (Cynops pyrrhogaster) embryo is patterned duringgastrulation. Dev. Biol. 244, 66–74.

Kaneda, T., Iwamoto, Y., Motoki, J.D., 2009. Origin of the prechordal plate and patterningof the anteroposterior regional specificity of the involuting and extending arch-enteron roof of a urodele, Cynops pyrrhogaster. Dev. Biol. 334, 84–96.

Katsumoto, K., Arikawa, T., Doi, J., Fujii, H., Nishimatsu, S., Sakai, M., 2003.Cytoplasmic and molecular reconstruction of Xenopus embryos: synergy ofdorsalizing and endo-mesodermalizing determinants drives early axial pat-terning. Development 131, 1135–1144.

Keller, R.E., 1975. Vital dye mapping of the gastrula and neurula of Xenopus laevis. I.Prospective areas and morphogenetic movements of the superficial layer. Dev.Biol. 42, 222–241.

Keller, R.E., 1976. Vital dye mapping of the gastrula and neurula of Xenopus laevis.II. Prospective areas and morphogenetic movements of the deep layer. Dev.Biol. 51, 118–137.

Keller, R.E., 1980. The cellular basis of epiboly: a SEM study of deep-cell rearrangementduring gastrulation in Xenopus laevis. J. Embryol. Exp. Morph. 60, 201–234.

Keller, R.E., Danilchik, M., 1988. Regional expression, pattern and timing ofconvergence and extension during gastrulation of Xenopus laevis. Development103, 193–209.

Keller, R.E., Jansa, S., 1992. Xenopus gastrulation without a blastocoel roof. Dev.Dyn. 195, 162–176.

Keller, R., Shook, D., 2004. Gastrulation in amphibians. In: Stern, C.D. (Ed.),Gastrulation, From Cells to Embryo. Cold Spring Harbor Lab. Press, Cold SpringHarbor, New Yolk, pp. 171–203.

Koide, T., Umesono, K., Hashimoto, C., 2002. When does the anterior endomeso-derm meet the anteriormost neuroectoderm during Xenopus gastrulation? Int.J. Dev. Biol. 46, 777–783.

Komazaki, S., 1992. Epibolic extension of the presumptive ectodermal layer ofembryos of the newt Cynops pyrrhogaster before and during gastrulation. J.Exp. Zool. 263, 414–422.

Kurth, T., Hausen, P., 2000. Bottle cell formation in relation to mesodermalpatterning in the Xenopus embryo. Mech. Dev. 97, 117–131.

Lane, M.C., Keller, R., 1997. Microtubule disruption reveals that Spemann’sorganizer is subdivided into two domains by the vegetal alignment zone.Development 124, 895–906.

Loose, M., Patient, R., 2004. A genetic regulatory network for Xenopus mesendo-derm formation. Dev. Biol. 271, 467–478.

Luxardi, G., Marchal, L., Thome, V., Kodjabachian, L., 2010. Distinct Xenopus Nodalligands sequentially induce mesendoderm and control gastrulation move-ments in parallel to the Wnt/PCP pathway. Development 137, 417–426.

Izpisua-Belmonte, J.C., De Robertis, E.M., Storey, K.G., Stern, C.D., 1993. Thehomeobox gene goosecoid and the origin of organizer cells in the early chickblastoderm. Cell 74, 645–659.

Manes, M.E., Campos Casal, F., 1997. Inductive interactions required for mesoder-mal differentiation in Bufo arenarum gastrulae. Dev. Genes Evol. 207, 90–96.

Manes, M.E., Campos Casal, F., 2002. Mesodermal and axial determinants con-tribute to mesoderm regionalization in Bufo arenarum embryos. Dev. GenesEvol. 212, 374–379.

Mangold, O., 1933. Uber die Induktionsfahigkeit der verschiedenen Bezirke derNeurula von Urodelen. Naturwissenschaften 21, 761–766, in German.

Minsuk, A.B., Keller, R.E., 1997. Surface mesoderm in Xenopus: a revision of thestage 10 fate map. Dev. Genes Evol. 207, 389–401.

Moon, R.T., Kimelman, D., 1998. From cortical rotation to organizer gene expres-sion: toward a molecular explanation of axis specification in Xenopus.BioEssays 20, 536–545.

Moya, I.M., Alarcon, I., del Pino, E.M., 2007. Gastrulation of Gastrotheca riobambaein comparison with other frogs. Dev. Biol. 304, 467–478.

Nagano, T., Ito, Y., Tashiro, K., Kobayakawa, Y., Sakai, M., 2000. Dorsal inductionfrom dorsal vegetal cells in Xenopus occurs after mid-blastula transition. Mech.Dev. 93, 3–14.

Nakamura, O., 1942. Die Entwicklung der Hinteren Korperhalfte bei Urodelen.Annot. Zool. Japonenses. 21, 169–236. (in German). (available online at CiNii,/http://ci.nii.ac.jp/naid/110003368550/en/S).

Nakamura, O., Kishiyama, K., 1971. Prospective fates of blastomeres at the 32 cellstage of Xenopus laevis embryos. Proc. Japan Acad. 47, 407–412. (available onlineat J–Stage. /https://www.jstage.jst.go.jp/article/pjab1945/47/4/47_4_407/_pdfS.

Nakamura, O., Takasaki, H., Ishihara, M., 1970. Formation of the Organizer fromcombinations of Presumptive Ectoderm and Endoderm. I. Proc. Japan Acad.47, 313–318. (available online at J–Stage, /https://www.jstage.jst.go.jp/article/pjab1945/47/3/47_3_313/_pdfS).

Nakamura, O., Takasaki, H., Okumoto, T., Iida, H., 1971. Differentiation duringcleavage in Xenopus laevis. II Development of inductive activity of theorganizer. Proc. Japan Acad. 47, 203-208. (available online at J–Stage,/https://www.jstage.jst.go.jp/article/pjab1945/47/2/47_2_203/_pdfS).

Nath, K., Boorech, J.L., Beckham, Y.M., Burns, M.M., Elinson, R.P., 2005. Status ofRNAs, localized in Xenopus laevis oocytes, in the frogs Rana pipiens andEleutherodactylus coqui. J. Exp. Zool. (Mol. Dev. Evol.) 304B, 28–39.

Nath, K., Elinson, R.P., 2007. RNA of AmVegT, the axolotl orthologue of the Xenopusmeso-endodermal determinant, is not localized in the oocyte. Gene Exp.Pattern 7, 197–201.

Niehrs, C., 2004. Regionally specific induction by the Spemann–Mangold organi-zer. Nat. Rev. Genet. 5, 425–434.

Nieuwkoop, P.D., 1969a. The formation of the mesoderm in urodelean amphibiansI. Induction by the endoderm. W. Roux’s Arch. 162, 341–373.

Nieuwkoop, P.D., 1969b. The formation of the mesoderm in urodelean amphibiansII. The origin of the dorso-ventral polarity of the mesoderm. W, Roux’s Arch.163, 298–315.

Nieuwkoop, P.D., 1996. What are the key advantages and disadvantages of urodelespecies compared to anurans as a model system for experimental analysis ofearly development? Int. J. Dev. Biol. 40, 617–619.

Nieuwkoop, P.D., 1997. Short historical survey of pattern formation in the endo-mesoderm and the neural anlage in the vertebrates: the role of vertical andplanar inductive actions CMLS,. Cell. Mol. Life Sci. 53, 305–318.

Nieuwkoop, P.D., Florschutz, P.A., 1950. Quelques caracteres speciaux de lagas-trulation et de la neurulation de l’oeuf de Xenopus laevis Daud. et de quelquesautres anoures. Arch. Biol. 61, 113–150, In French.

Nieuwkoop, P.D., Johnen, A.G., Albers, B., 1985. The Epigenetic Nature of EarlyChordate Development. Inductive Interaction and Competence. CambridgeUniv. Press, Cambridge.

Ninomiya, H., Ariizumi, T., Asashima, M., 1998. Activin-treated ectoderm hascomplete organizing center activity in Cynops embryos. Dev. Growth Differ. 40,199–208.

Ninomiya, H., Takahashim, S., Tanegashima, K., Yokota, C., Asashima, M., 1999.Endoderm differentiation and inductive effect of activin-treated ectoderm inXenopus. Dev. Growth Differ. 41, 391–400.

Ninomiya, H., Zhang, Q., Elinson, R.P., 2001. Mesoderm formation in Eleutherodac-tylus coqui: body patterning in a frog with a large egg. Dev. Biol. 236, 109–123.

Okada, T.S., 1994. Experimental embryology in Japan, 1930–1960. A historicalbackground of developmental biology in Japan. Int. J. Dev. Biol. 38, 135–154.

Okada, Y.K., Hama, T., 1943. Examination of regional differences in the inductiveactivity of the organizer by means of transplantation into ectodermal vesicles.Proc. Imp. Acad. Japan 19, 48-53. (available online at J–Stage, /https://www.jstage.jst.go.jp/article/pjab1912/19/1/19_1_48/_pdfS).

Okada, Y.K. Hama, T., 1944. On the Different Effects of the Amphibian Organizerfollowing Culture, Transplantation, and Heat Treatment. Proc. Imp. Acad. Japan20, 36-40. (available online at J–Stage, /https://www.jstage.jst.go.jp/article/pjab1912/20/1/20_1_36/_pdfS).

Okada, Y.K., Hama, T., 1945. Regional differences in the inductive capacity of thedorsal roof of the archenteron of the urodele, Triturus pyrrhogaster. Proc. JapanAcad. 21, 240–247. (available online at J–Stage, /https://www.jstage.jst.go.jp/article/pjab1945/21/3-10/21_3-10_240/_pdfS).

Okada, Y.K., Ichikawa, M., 1947. A new normal table of the development of Triturus(Cynops) pyrrhogaster. Jpn. J. Exp. Morphol. 3, 1–6, in Japanese.

Okada, Y.K., Takaya, H., 1942a. Experimental investigation of regional differencesin the inductive capacity of the organizer. Proc. Imp. Acad. Japan 18, 505–513.(available online at J–Stage, /https://www.jstage.jst.go.jp/article/pjab1912/18/8/18_8_505/_pdfS).

Okada, Y.K., Takaya, H., 1942b. Further studies upon the regional differentiation ofthe inductive capacity of the organizer. Proc. Imp. Acad. Jpn. 18, 514–519.(available online at J–Stage, /https://www.jstage.jst.go.jp/article/pjab1912/18/8/18_8_514/_pdfS).

Oschwald, R., Clement, J.H., Knohel, W., Grunz, H., 1993. Suramin preventstranscription of dorsal marker genes in Xenopus laevis embryos, isolateddorsal blastopore lips and activin-A induced animal caps. Mech. Dev. 43,121–133.

Papan, C., Boulat, B., Velan, S.S., Fraser, S.E., Jacobs, R.E., 2007. Formation of thedorsal marginal zone in Xenopus laevis analyzed by time-lapse microscopicmagnetic resonance imaging. Dev. Biol. 305, 161–171.

Pasteels, J., 1942. New observations concerning the maps of presumptive areas ofthe young amphibian gastrula. (Amblystoma and Discoglossus). J. Exp. Zool. 89,255–281.

T. Kaneda, J.D. Motoki / Developmental Biology 369 (2012) 1–1818

Poznanski, A., Keller, R., 1997. The role of planar and early vertical signaling inpatterning the expression of Hoxb-1 in Xenopus. Dev. Biol. 184, 351–366.

Rankin, S.A., Kormish, J., Kofron, M., Jegga, A., Zorn, A.M., 2011. A gene regulatorynetwork controlling hhex transcription in the anterior endoderm of theorganizer. Dev. Biol. 351, 297–310.

Sala, M.J., 1955. Distribution of activating and transforming influences in thearchenteron roof during the induction of the nervous system in Amphibians.Proc. Kon. Ned. Acad. Sci. A’dam., Ser. C, 58, 635–647.

Saha, M.S., Grainger, R.M., 1992. A labile period in the determination of theanterior-posterior axis during early neural development in Xenopus. Neuron 8,1003–1014.

Sakaguchi, K., Kaneda, T., Matsumoto, M., Imoh, H., Suzuki, A.S., 2002. Establish-ment of the organizing activity of the lower endodermal half of the dorsalmarginal zone is a primary and necessary event for dorsal axis formation inCynops pyrrhogaster. Int. J. Dev. Biol 46, 793–800.

Sakai, M., 1996. The vegetal determinants required for the Spemann organizermove equatorially during the first cell cycle. Development 122, 2207–2214.

Sakai, M., 2008. Cell-autonomous and inductive processes among three embryonicdomains control dorsal-ventral and anterior-posterior development of Xeno-pus laevis. Dev. Growth Differ. 50, 49–62.

Sasai, Y., Lu, B., Steinbeisser, H., Geissert, D., Gont, L.K., De Robertis, E.M., 1994.Xenopus chordin: A novel dorsalizing factor activated by organizer-specifichomeobox genes. Cell 79, 779–790.

Schechtman, A.M., 1934. Unipolar ingression in Triturus torosus: a hithertoundescribed movements in the pregastrular stages of a urodele. Univ. Calif.Publ. Zool. 39, 303–310.

Shi, D.-L., Darrib�ere, T., Johnson, K.E., Boucaut, J.-C., 1989. Initiation of mesodermalcell migration and spreading relative to gastrulation in the urodele amphibianPleurodeles waltl. Development 105, 351–363.

Shih, J., Keller, R., 1992. The epithelium of the dorsal marginal zone of Xenopus hasorganizer properties. Development 116, 887–899.

Shook, D.R., Majer, C., Keller, R., 2002. Urodeles remove mesoderm from thesuperficial layer by subduction through a bilateral primitive streak. Dev. Biol.248, 220–239.

Shook, D.R., Majer, C., Keller, R., 2004. Pattern and morphogenesis of presumptivesuperficial mesoderm in two closely related species, Xenopus laevis andXenopus tropicalis. Dev. Biol. 270, 163–185.

Shook, D.R., Keller, R., 2008a. Epithelial type, ingression, blastopore architectureand the evolution of chordate mesoderm morphogenesis. J. Exp. Zool. (Mol.Dev. Evol.) 310B, 85–110.

Shook, D.R., Keller, R., 2008b. Morphogenic machines evolve more rapidly than thesignals that pattern them: Lessons from amphibians. J. Exp. Zool. (Mol. Dev.Evol.) 310B, 111–135.

Slack, J.M.W., 1984. In vitro development of isolated ectoderm from axolotlgastrulae. J. Embryol. Exp. Morph. 80, 321–330.

Slack, J.M.W., Tannahill, D., 1992. Mechanism of anteroposterior axis specificationin vertebrates. Lessons from the amphibians. Development 114, 285–302.

Smith, J.C., Malacinski, G.M., 1983. The origin of the mesoderm in an anuran,Xenopus laevis, and a urodele, Ambystoma mexicanum. Dev. Biol. 98, 250–254.

Smith, J.C., Slack, J.M.W., 1983. Dorsalization and neural induction: properties ofthe organizer in Xenopus laevis. J. Embryol. Exp. Morph. 78, 299–317.

Smith, J.C., Price, B.M.J., Green, J.B.A., Weigel, D., Herrmann, B.G., 1991. Expressionof a Xenopus homolog of Brachyury (T) is an immediate-early response tomesoderm induction. Cell 67, 79–87.

Smith, W.C., Harland, R.M., 1992. Expression cloning of noggin, a new dorsalizingfactor localized to the Spemann organizer in Xenopus embryos. Cell 70, 829–840.

Solnica-Krezel, L., 2005. Conserved patterns of cell movements during vertebrategastrulation. Curr. Biol. 15, R213–R228.

Sone, K., Takeshima, K., Takahashi, T.C., Takabatake, T., 1997. Temporal alterationof dose-dependent response to activin in newt animal-cap explants. Dev.Genes Evol. 207, 147–155.

Spemann, H., 1931. Uber den Abteil vom Implantat und Wirtskeime an derOrientierung und Beschaffenheit der induyierten Embryonalanlage. W. Roux’Arch. EntwMech. Org. 123, 389–517, in German.

Spemann, H., Mangold, H., 1924. Uber Induktion von Embryonalanlagen durchImplantation artfremder Organisatoren. W. Roux’ Arch. EntwMech. Org. 100:599–638. (Translated into english; Induction of embryonic primordia byimplantation of organizers from a different species. Int. J. Dev. Biol. 1-26, 2001).

Stern, C.D., 2001. Initial patterning of the central nervous system: how manyorganizers? Nat. Rev. Neurosci. 2, 92–98.

Stewart, R.M., Gerhart, J.C., 1990. The anterior extent of dorsal development of theXenopus embryonic axis depends on the quantity of organizer in the lateblastula. Development 109, 363–372.

Stewart, R.M., Gerhart, J.C., 1991. Induction of notochord by the organizer inXenopus. Roux’s Arch. Dev. Biol. 199, 341–348.

Suzuki, A.S., Kuwabara, K., Kuwabara, Y., 1975. Temporal relations betweenextension of archenteron roof and realization of neural induction duringgastrulation of newt embryo. Dev. Growth Differ. 17, 343–353.

Suzuki, A.S., Mifune, Y., Kaneda, T., 1984. Germ layer interactions in patternformation of amphibian mesoderm during primary embryonic induction. Dev.Growth Differ. 26, 81–94.

Suzuki, A.S., Yamamoto, Y., Imoh, H., 1997. Direct evidence of an essential role forextended involution in the specification of a dorsal marginal mesoderm duringCynops gastrulation. Dev. Growth Differ. 39, 135–141.

Suzuki, A.S., Sakaguchi, K., Katagiri, M., Ueki, M., Kaneda, T., 2002. Blastoporeformation and dorsal mesoderm induction are independent events in earlyCynops embryogenesis. Dev. Growth Differ. 44, 303–313.

Swiers, G., Chen, Y.-H., Johnson, A.D., Loose, M., 2010. A conserved mechanism forvertebrate mesoderm specification in urodele amphibians and mammals. Dev.Biol., 138–152.

Taira, M., Saint-Jeannet, J.-P., Dawid, I.B., 1997. Role of the Xlim-1 and Xbra genes inanteroposterior patterning of neural tissue by the head and trunk organizer.Proc. Nat. Acad. Sci. U.S.A. 94, 895–900.

Takasaki, H., 1987. Fates and roles of the presumptive organizer region in the 32-cell embryo in normal development of Xenopus laevis. Dev. Growth Differ. 29,151–152.

Takaya, H., 1978. Dynamics of the organizer. Morphogenetic movements andspecificities in induction and differentiation of the organizer. In: Nakamura, O.,Toivonen, S. (Eds.), Organizer—a Milestone of a Half-Century from Spemann.Elsevier/North Holland Biomedical Press, pp. 49–70.

Venegas-Ferrın, M., Sudou, N., Taira, M., del Pino, E., 2010. Comparison of Lim1expression in embryos of frogs with different modes of reproduction. Int. J.Dev. Biol. 54, 195–202.

Vodicka, M.A., Gerhart, J.C., 1995. Blastomere derivation and domains of geneexpression in the Spemann organizer of Xenopus laevis. Development 121,3505–3518.

Vogt, W., 1929. Gestaltungsanalyse am Amphibienkeim mit Ortlicher Vitalfabung. II.Gastrulation und Mesodermbildung bei Urodelen und Anuren. W. Roux’ Arch.EntwMech. Org. 120, 384-706. (Translated into Japanese by T. Hama, 1992.‘‘Morphogenetic movement and organogenesis in the amphibian embryos’’.Gakkai Shuppan Center, Tokyo).

Vonica, A., Gumbiner, B.M., 2007. The Xenopus Nieuwkoop center and Spemann–Mangold organizer share molecular components and a requirement formaternal Wnt activity. Dev. Biol. 312, 90–102.

Wallingford, J.B., Sater., A.K., Uzman, J.A., Danilchik, M.V., 1997. Inhibition ofmorphogenetic movement during Xenopus gastrulation by injected sulfatase:implications for anteroposterior and dorsoventral axis formation. Dev. Biol.187, 224–235.

Wylie, C., Kofron, M., Payne, C., Anderson, R., Hosobuchi, M., Joseph, E., Heasman, J.,1996. Maternal beta-catenin establishes a ‘dorsal signal’ in early Xenopusembryos. Development 122, 2987–2996.

Winklbauer, R., Schurfeld, M., 1999. Vegetal rotation, a new gastrulation move-ment involved in the internalization of the mesoderm and endoderm inXenopus. Development 126, 3703–3713.

Yamamoto, Y., Suzuki, A.S., 1994. Two essential processes in the formation of adorsal axis during gastrulation of Cynops embryo. Roux’s Arch. Dev. Biol. 204,11–19.

Yokota, C., Ariizumi, T., Asashima, M., 1998. Patterns of gene expression in the coreof Spemann’s organizer and activin-treated ectoderm in Cynops pyrrhogaster.Dev. Growth Differ. 40, 335–341.

Zoltewicz, J.S., Gerhart, J.C., 1997. The Spemann organizer of Xenopus is patternedalong its anteroposterior axis at the earliest gastrula stage. Dev. Biol. 192,482–491.

Zhang, J., Houston, D.W., King, M.L., Payne, C., Wylie, C., Heasman, J., 1998. The roleof maternal VegT in establishing the primary germ layers in Xenopus embryos.Cell 94, 515–524.

Zust, B., Dixon, K.E., 1975. The effect of u.v. irradiation of the vegetal pole ofXenopus laevis eggs on the presumptive primordial germ cells. J. Embryol. Exp.Morph. 34, 209–220.