Requirement for anti-dorsalizing morphogenetic protein in organizerpatterning

Roland Dosch, Christof Niehrs*

Division of Molecular Embryology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany

Received 16 September 1999; accepted 20 September 1999

Abstract

The amphibian Spemann organizer is subdivided in trunk and head organizer and it is unclear how this division is regulated. The Xenopus

trunk organizer expresses anti-dorsalizing morphogenetic protein (ADMP), a potent organizer antagonist. We show that ADMP represses

head formation during gastrulation and that its expression is activated by BMP antagonists. A speci®cally acting dominant-negative ADMP

anteriorizes embryos and its coexpression with BMP antagonists induces secondary embryonic axes with heads as well as expression of head

inducers. Unlike other BMPs, ADMP is not inhibited by a dominant-negative BMP type I receptor, Noggin, Cerberus and Chordin but by

Follistatin, suggesting that it utilizes a distinct TGF-b receptor pathway and displays differential sensitivity to BMP antagonists. The results

indicate that ADMP functions in the trunk organizer to antagonize head formation, thereby regulating organizer patterning. q 2000 Elsevier

Science Ireland Ltd. All rights reserved.

Keywords: Anti-dorsalizing morphogenetic protein; Head formation; Organizer; BMP, Follistatin; Xenopus; Wnt inhibitors

1. Introduction

The Spemann organizer or upper dorsal blastopore lip of

the amphibian gastrula is of central importance for the

establishment of the vertebrate body plan. The organizer

can be subdivided into head-, trunk- and tail organizer as

revealed by their inducing potential: The early gastrula

organizer induces a secondary embryonic axis containing

a head, the midgastrula organizer induces predominantly

secondary trunks and transplants at the end of gastrulation

evoke secondary tails (Spemann, 1931; Mangold, 1933).

The subdivision into distinct head and trunk organizers is

strongly supported by studies in the mouse and zebra®sh

where mutations in a number of genes leads to embryos

without head but relatively normal trunks (Bouwmeester

and Leyns, 1997; Beddington and Robertson, 1999;

Mullins, 1999). Tail formation may involve the combined

action of Notch and Wnt signals (Beck and Slack, 1999).

The molecular mechanism of the trunk organizer resides in

its secretion of potent BMP inhibitors such as Chordin,

Noggin and Follistatin which antagonize the ventro-poster-

iorizing action of Bone Morphogenetic Proteins (BMPs)

(Harland and Gerhart, 1997). The distinguishing feature of

the head organizer is that it not only involves BMP inhibi-

tors but also secretion of inhibitors for posteriorizing Wnt

signals, such as Cerberus, Frzb and Dkk1 (Niehrs, 1999).

Cerberus which simultaneously inhibits BMP, Wnt as well

as Nodals is able to induce ectopic heads on its own follow-

ing mRNA microinjection, while the Wnt inhibitors domi-

nant-negative Wnt8, Frzb and Dkk1 require co-expression

of a BMP antagonist for head induction. Consistent with

their role in head induction, cerberus, frzb and dkk1 are

predominantly expressed in prospective anterior endoderm

or prechordal plate. In contrast, BMP inhibitors are

expressed in all organizer derivatives: cerberus in anterior

endoderm and noggin, follistatin and chordin in the prechor-

dal plate and chordamesoderm.

How is this molecular subdivision between head and

trunk organizer regulated? There is evidence that head and

trunk organizer inhibit each other: Nodal and Wnt signalling

are required for trunk formation but inhibit head induction.

Head organizer in turn secretes Nodal- and Wnt inhibiting

factors, thereby establishing a boundary between head and

trunk organizer (Piccolo et al., 1999). A complication for the

view that organizer function requires inhibition of BMP

signalling is the expression in the organizer of anti-dorsa-

lizing morphogenetic protein (admp), a member of the BMP

family, closely related to BMP3. Despite its dorsal expres-

sion, admp overexpression elicits ventralization in Xenopus

(Moos et al., 1995) and chick embryos (Joubin and Stern,

1999). However, these gain-of-function studies left unre-

Mechanisms of Development 90 (2000) 195±203

0925-4773/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved.

PII: S0925-4773(99)00245-2

www.elsevier.com/locate/modo

* Corresponding author.

solved the question of the physiological role played by

ADMP since overexpression of any BMP leads to ventro-

posteriorization. For example, the secondary axes induced

by Wnt8 mRNA injection obscure its role in trunk forma-

tion, which was only revealed by a combination of zygotic

overexpression and of loss-of-function experiments (Chris-

tian and Moon, 1993; Hoppler et al., 1996).

Therefore, we have investigated the requirement for

ADMP signalling during gastrulation by generating a domi-

nant-negative variant which acts speci®cally on ADMP. Our

data indicate that ADMP plays a role in subdividing head

and trunk organizer. admp is predominantly expressed in

trunk organizer and its inhibition promotes head formation,

including ectopic head induction in conjunction with BMP

inhibitors. Furthermore, we show that ADMP displays

differential sensitivity to BMP antagonists, as it is effec-

tively inhibited only by Follistatin. We propose that

ADMP is part of the genetic network which regulates the

mutual antagonism between head and trunk organizer.

2. Results

2.1. Zygotic expression of admp leads to dorso-anterior

defects

In previous experiments ventro-posteriorization of

embryos was observed following ADMP overexpression

before midblastula transition (MBT) by mRNA injection.

To test if ADMP is also effective after MBT, the time of

endogenous expression, we injected pCS1ADMP, whose

transcription becomes activated after MBT. Fig. 1A shows

that injected embryos fail to form a head and do not express

neural marker genes anterior to the otic vesicle (white

arrowheads) like Xanf (pituitary, Zaraisky et al., 1992),

otx2 (forebrain/midbrain, Blitz and Cho, 1995; Pannese et

al., 1995) and en2 (midbrain-hindbrain boundary, Hemmati-

Brivanlou et al., 1991). The hindbrain still expresses krox20

(Bradley et al., 1993) in rhombomere 5 while the expression

in rhombomere 3 is absent in injected embryos. The data

suggest that ADMP is able to antagonize head formation

during gastrulation.

2.2. Trunk inducers activate admp expression

Despite its ventro-posteriorizing effects in gain-of-func-

tion experiments, admp shows predominant expression in

trunk organizer (Moos et al., 1995 and Fig. 6B). To analyze

if it is also regulated by trunk inducers we tested admp

induction by RT-PCR in ventral marginal zone explants

following mRNA microinjection of various BMP inhibitors.

Noggin (Smith and Harland, 1992), Chordin (Sasai et al.,

1994), dominant-negative BMP receptor type I (tBR) (Graff

et al., 1994; Suzuki et al., 1994) and dominant-negative

BMP7 (CmBMP7) (Hawley et al., 1995) are able to induce

admp (Fig. 1B) as well as chordin, a trunk organizer marker.

The ventral marker vent1 (Gawantka et al., 1995) served as

control and is downregulated by the trunk inducers tested.

These results con®rm that admp is regulated like a typical

Spemann organizer gene and requires BMP inhibition.

2.3. ADMP signals by distinct BMP receptors

Since the ADMP overexpression phenotype is similar to

that elected by other BMPs (Cho and Blitz, 1998) we inves-

tigated if ADMP utilizes the same signalling receptors. Fig.

2A shows that microinjected mRNA of a dominant-negative

BMP type I receptor (tBR), known to interact with BMP2, -

4 and -7 (Graff et al., 1994; Suzuki et al., 1994; Wang et al.,

1997a), expectedly inhibits BMP4-mediated induction of

the ventral markers vent1 and -2 (Onichtchouk et al.,

1996) in animal cap explants. In contrast, ADMP is still

R. Dosch, C. Niehrs / Mechanisms of Development 90 (2000) 195±203196

Fig. 1. (A) ADMP signalling represses head formation during gastrulation.

Four-cell embryos were either injected with 100 pg per blastomere pCS21

(Control) or pCS1ADMP (ADMP) and processed for in situ hybridisation

at stage 35 for the expression of the neural marker genes krox20, otx2, en2

and Xanf as indicated. Embryos are shown in a lateral view with the anterior

side facing right, dorsal side up. The arrowheads mark the position of the

otic vesicle. (B) Induction of admp by BMP-Inhibitors. Four-cell embryos

were radially injected into all blastomeres with 1.5 ng preprolactin (PPL) as

control, 250 pg dominant-negative BMP receptor (tBR), 50 pg noggin, 250

pg chordin or 750 pg dominant-negative bmp7 (CmBMP7) mRNA per

blastomere, as indicated. Dorsal (DMZ) or ventral (VMZ) marginal zones

were explanted from early gastrulae and cultured until analysis for the

expression of the indicated marker genes at stage 12 by RT-PCR. H4:

Histone 4 serves as loading control. Note induction of admp by all BMP-

inhibitors tested. (-RT, minus reverse transcriptase control, WE, whole

embryo control).

able in the presence of tBR to induce vent1 and -2 expres-

sion. Likewise, co-injection of tbr mRNA partially rescued

embryos ventralized by bmp4 (Dale et al., 1992; Jones et al.,

1992) but not by admp mRNA (Fig. 2B). Thus, although

ADMP induces similar phenotypes and target genes as

BMP4, it utilizes a distinct receptor pathway.

2.4. A dominant-negative variant interferes speci®cally with

ADMP signalling

The dorso-anterior defects observed following overex-

pression of ADMP are consistent with a role in antagonizing

head formation but the physiological relevance of this effect

is unclear since this overexpression phenotype is typically

observed with all BMPs. To investigate the requirement for

ADMP in axis formation we mutated its putative tetrabasic

cleavage site, a mutation known to create speci®c dominant-

negative TGF-b growth factors by preventing maturation of

functional growth factor dimers (Wittbrodt and Rosa, 1994;

Hawley et al., 1995; Nishimatsu and Thomsen, 1998; Osada

and Wright, 1999; Sun et al., 1999) (Fig. 3A). In animal cap

explants this cleavage mutant (CmADMP) speci®cally inhi-

bits signalling by ADMP but neither by BMP4 (vent1 induc-

tion, Fig. 3B) nor by Nodal-related1 (Jones et al., 1995)

(otx2, chordin induction, Fig. 3C). Likewise, co-injection

of cmadmp mRNA rescues embryos ventralized by admp

but not by bmp4 mRNA (Fig. 3D, Table 1), suggesting that

it acts as a speci®c dominant-negative variant.

2.5. CmADMP overexpression causes anteriorized embryos

Further evidence for the speci®city of CmADMP is the

observation that its mRNA microinjection into ventral blas-

tomeres does neither induce secondary embryonic axes nor

any other phenotype, unlike all other known BMP inhibi-

tors, including CmBMP7 (Fig. 4A, top panel), an analogous

cleavage mutant of BMP7 (Hawley et al., 1995). However,

when injected radially in all blastomeres or into two dorsal

blastomeres (not shown) CmADMP leads to embryos with

enlarged heads, with big eyes and cement glands (Fig. 4A,

middle panel). Whole-mount in situ hybridization revealed

an expansion of the cement gland marker Xag1 (Sive et al.,

1989) (Fig. 4A, bottom panel) and otx2 (not shown) but little

effect on the midbrain-hindbrain marker en2 (Fig. 4A).

Typically, the trunks of CmADMP-injected embryos are

moderately shortened and are still able to develop a tail.

In contrast, mRNA injection of cmbmp7 doses that result

in strong shortening of the trunk and stalled tail formation

have little effect on head formation, nor do they expand

cement glands (Fig. 4A). Note however, that maximal

doses of CmBMP7 result in radially dorsalized embyros,

(e.g. similar to Fig. 2B, lower left panel), a phenotype

never observed with CmADMP. We conclude that

CmADMP interferes speci®cally with ADMP, resulting in

anteriorization.

2.6. Simultaneous inhibition of BMP and ADMP signalling

induces head formation

To corroborate the head-promoting effects of CmADMP

we co-expressed it with BMP inhibitors. Microinjection of

BMP inhibitors in ventral blastomeres leads to the induction

of trunk organizer (Harland and Gerhart, 1997) (including

admp expression, Fig. 1B) with embryos forming incom-

plete secondary embryonic axes, lacking a head. While

mRNA injection of tbr or cmbmp7 result in incomplete

secondary embryonic axes, co-injection with cmadmp

induces heads with one eye and cement gland as well as a

short trunk (Fig. 4B; Table 2). Neither higher doses of any

individual BMP inhibitor nor co-injection of BMP inhibitors

resulted in secondary head formation (Table 2) as shown

previously (Glinka et al., 1997). CmADMP can even

enhance partial head formation induced by co-injection of

tBR with dominant-negative Wnt8 (Hoppler et al., 1996),

which typically yields secondary heads with only one eye

(Glinka et al., 1997) (Fig. 4B), since triple injection with

CmADMP resulted in complete heads containing two well

separated eyes (Fig. 4B, bottom right panel; Table 2). We

conclude that ADMP inhibition is required for head forma-

tion.

2.7. Simultaneous inhibition of BMP and ADMP signalling

induces the expression of Wnt inhibitors

We have previously shown that inhibition of Wnt signal-

ling is necessary to induce secondary axes containing head

R. Dosch, C. Niehrs / Mechanisms of Development 90 (2000) 195±203 197

Fig. 2. (A) ADMP signalling is not inhibited by a dominant-negative BMP

receptor. Eight-cell embryos were injected into the four animal blastomeres

with 500 pg preprolactin (PPL) as control, 150 pg bmp4, 150 pg admp or 1

ng constitutively-active BMP receptor (CABR) mRNA together with 250

pg preprolactin (2) or truncated BMP receptor (tBR) (1) mRNA per

blastomere. Animal caps were cut from blastulae and analyzed at stage

10.5 for the expression of the indicated marker genes by RT-PCR. Note

that ADMP unlike BMP4 induces vent expression in the presence of tBR.

(B) Phenotype of embryos injected radially with the same mRNA concen-

trations as in (A).

structures (Glinka et al., 1997; Glinka et al., 1998). To

investigate whether head induction by CmADMP involves

regulation of Wnt inhibitors or occurs by an independent

route we analyzed the expression of the head organizer

genes frzb (Leyns et al., 1997; Wang et al., 1997b) and

dkk1 (Glinka et al., 1998) in gastrulae co-injected ventrally

with CmADMP and tBR. While tBR injection alone induces

no ectopic frzb and little dkk1, co-injection with CmADMP

upregulates both genes ectopically on the ventral side, in the

anterior endomesoderm (Fig. 5). The notochord marker not2

(Gont et al., 1993) is also superinduced by CmADMP/tBR

co-injection while ectopic expression of the midline marker

pintallavis (Ruiz i Altaba and Jessell, 1992) was not signif-

icantly enhanced (Fig. 5). Co-injection of tBR with dnWnt8

upregulated these markers similarly to CmADMP/tBR. The

superinduction of not2 both by dnWnt8/tBR and CmADMP/

tBR is consistent with the notochord-promoting effects of

Wnt-antagonists (Hoppler et al., 1996; Glinka et al., 1997).

R. Dosch, C. Niehrs / Mechanisms of Development 90 (2000) 195±203198

Table 1

CmADMP speci®cally rescues ventralization by ADMPa

mRNA injected (ng per blastomere) Ventralized (%) Normal (%) Enlarged head (%) Gastrulation defectsb (%) No. of embryos

LacZ (0.5) 0 94.8 0 5.2 77

ADMP (0.25) 1 LacZ (0.25) 100c 0 0 0 82

ADMP (0.25) 1 CmADMP (0.125) 0 91.4 0 8.6 70

ADMP (0.25) 1 CmADMP (0.25) 0 19.5 64.9 15.6 77

BMP4 (0.25) 1 LacZ (0.25) 100d 0 0 0 47

BMP4 (0.25) 1 CmADMP (0.125) 100e 0 0 0 86

BMP4 (0.25) 1 CmADMP (0.25) 98.7f 1.3 0 0 77

LacZ (0.25) 1 CmADMP (0.125) 0 8.6 69 22.4 58

LacZ (0.25) 1 CmADMP (0.25) 0 7.7 58.5 33.9 65

a Four-cell embryos were microinjected into all blastomeres with the indicated mRNAs. After 3 days at room temperature embryos were scored for the

indicated phenotypes (as shown in Fig. 3D). A minimum of two independent experiments was carried out for every injection.b Gastrulation defects give rise to spina bi®da.c Average DAI 1.25 (dorso-anterior index (Kao and Elinson, 1988)).d Average DAI 0.8.e Average DAI 0.9.f Average DAI 2.3.

Fig. 3. Speci®c inhibition of ADMP signalling by a dominant-negative form of ADMP (CmADMP) promotes head formation. (A) Amino acid sequence of

ADMP showing the introduced mutations in the putative tetrabasic cleavage side to generate dominant-negative ADMP. (B, C) Eight-cell embryos were

microinjected into the animal blastomeres with (B) 250 pg lacZ as a control, 250 pg admp or bmp4 mRNA per blastomere and either with 250 pg lacZ ( 2 ) or

increasing doses of cmadmp mRNA (125 pg and 250 pg per blastomere) or (C) 250 pg nodal-related1 or ppl with 250 pg cmadmp or ppl mRNA as indicated.

Animal caps were cut from blastulae and analyzed at stage 11 for the expression of the indicated marker genes by RT-PCR (-RT, minus reverse transcriptase

control, WE, whole embryo control). (D) Phenotypes of embryos injected radially at eight-cell stage with 250 pg admp or bmp4 mRNA per blastomere. Co-

injection of 125 pg cmadmp mRNA rescues ADMP but not BMP4 induced phenotypes (see Table 1). Embryos are shown in lateral view with anterior facing

left.

Note that the ectopic expression of these genes occurs

always at the same anterior-posterior (a-p) level as the endo-

genous expression, suggestive of a-p prepatterning in the

gastrula. We conclude that simultaneous inhibition of

ADMP and BMP signalling induces the expression of

head-promoting Wnt inhibitors.

2.8. ADMP signalling is not inhibited by Chordin and

Noggin but by Follistatin

Expression of admp in the organizer raises the question of

how this BMP is able to escape the inhibitory action of BMP

antagonists. Whole-mount in situ hybridisation shows that

admp is expressed in chordamesoderm overlapping with

chordin unlike frzb and dkk1, which are predominantly

expressed in leading edge endomesoderm and prechordal

plate (Leyns et al., 1997; Wang et al., 1997b; Glinka et

al., 1998) (Fig. 6B). We therefore tested relevant BMP

antagonists in animal caps for their ability to interfere

with ADMP signalling (Fig. 6A). ADMP and BMP4 both

induce vent gene expression and co-injection of BMP4 with

all tested BMP inhibitors, including Chordin, Noggin,

Follistatin (Hemmati-Brivanlou et al., 1994) and Cerberus

(Bouwmeester et al., 1996), prevents vent gene induction.

However, ADMP signalling is unaffected by Noggin and

Cerberus, only weakly inhibited by high doses of Chordin

but strongly inhibited by Follistatin. Thus, ADMP can

largely escape the inhibitory effect of Chordin and Noggin

but it is selectively antagonized by Follistatin. Fig. 6B

shows that the expression domains of follistatin and admp

partially overlap, except for the most posterior chordame-

soderm, and anterior chordamesoderm/prechordal plate

where only admp is expressed.

3. Discussion

The control of organizer regionalization is of fundamental

importance for vertebrate axis formation and is thought to

involve a mutual antagonism between head- and trunk orga-

nizer (Piccolo et al., 1999): Nodal and Wnt signals present

in trunk organizer inhibit head and promote trunk formation,

while the head organizer secretes inhibitors of both growth

factors, including Cerberus, Frzb and Dkk1. Our results

argue for a model (Fig. 6C) where the trunk organizer antag-

onizes the head organizer also by ADMP: It is expressed in

the trunk organizer at gastrula stage and is both necessary

and suf®cient to inhibit head formation.

3.1. ADMP has properties distinct from other BMPs

ADMP was originally isolated as a gene speci®cally

expressed in the Xenopus organizer. It shows ventro-poster-

iorizing effects following mRNA overexpression (Moos et

al., 1995) similar to those observed with other ventralizing

BMPs, such as BMP2,-4,-7 and -7 related (Harland and

Gerhart, 1997). However there are clear differences between

ADMP and these BMPs. (i) By protein sequence compar-

ison ADMP belongs to the BMP3 subfamily, which is

distinct from either the BMP2/4 and the BMP7 subfamilies

(Hogan, 1996). (ii) Unlike other BMPs, ADMP signalling is

not blocked by a dominant-negative type IA BMP (Alk3)

receptor. In addition to Alk3, BMP2,-4 and -7 can also

signal via BMPR-IB (Alk6) and BMP7 may signal by

type I Activin receptors (Alk1 and -2) (MassagueÂ, 1998)

and may be candidates for ADMP type I receptors. A domi-

nant-negative type II BMP receptor has been reported that

induces a secondary embryonic axis containing an eye

(Frisch and Wright, 1998), suggesting that it may be able

to inhibit both BMP and ADMP signalling. However, this

R. Dosch, C. Niehrs / Mechanisms of Development 90 (2000) 195±203 199

Fig. 4. (A) CmADMP overexpression causes enlarged head formation. Top

row, phenotypes of embryos injected into two ventral blastomeres at the

four-cell stage with 1.5 ng ppl (preprolactin) as control (left), 375 pg

cmadmp (middle) or 375 pg cmbmp7 (right) mRNA per blastomere.

Embryos are shown in dorsal view, anterior facing left. Note that

CmADMP does not induce a secondary embryonic axis in contrast to

CmBMP7. Middle and bottom row, four-cell embryos injected radially

with 1.5 ng ppl (left), 750 pg cmadmp (middle) or 40 pg cmbmp7 (right)

mRNA per blastomere. Middle row, embryos are shown in lateral view,

anterior facing left. Note the enlarged head in CmADMP injected embryos.

Bottom row, albino embryos were analyzed for the expression of Xag1

(cement gland marker) and en2 (midbrain-hindbrain boundary marker).

Embryos are shown in anterior view, dorsal facing up. (B) Phenotypes of

embryos injected into two ventral blastomeres at the four-cell stage with 1.5

ng ppl, 125 pg tbr, 150 pg cmbmp7 or 125 pg tbr plus 150 pg dominant-

negative wnt8 (dnWnt8) mRNA either with 375 pg ppl (top row) or

cmadmp (bottom row) mRNA. Embryos are shown in dorsal or lateral

(second column) view with anterior facing left. Note that inhibition of

ADMP and BMP signalling causes head formation with one eye while

inhibition of ADMP, BMP and Wnt signalling (lower right panel) causes

head formation with two eyes.

receptor is not co-expressed with ADMP, making a physio-

logical role in ADMP signalling unlikely. (iii) Unlike other

BMPs, ADMP is not signi®cantly inhibited by Noggin,

Chordin or Cerberus (see below).

Members of the BMP3 family are expressed during orga-

nogenesis (Takahashi and Ikeda, 1996; Takao et al., 1996)

and studies in the rat and the mouse have shown that BMP3

promotes bone differentiation (Reddi, 1998). In light of the

differences between ADMP and other BMPs it should be

interesting to compare the roles of the BMP3 subfamily with

that of other BMPs in this context.

TGF-bs can form heterodimers (MassagueÂ, 1998) but the

state of heterodimerization of ADMP is unknown. Since

heterodimerisation occurs intracellularly, the heterodimer

partner has to be co-expressed with admp. Candidate

TGF-bs co-expressed with ADMP are Nodals, BMP7R

(Hawley et al., 1995; Wang et al., 1997a) and possibly

BMP4 in the prechodal plate of late gastrulae. However, it

appears unlikely that ADMP forms heterodimers with these

TGF-bs since dominant-negative CmADMP does not inter-

fere with their signalling after mRNA injection in ventral

mesoderm (no secondary embryonic axis) or ectoderm (Fig.

3). Another heterodimerisation partner for ADMP may be

the recently isolated DerrieÁre, which is expressed in the

marginal zone of Xenopus, whose overexpression like that

of ADMP leads to microcephaly (Sun et al., 1999).

However, a dominant-negative DerrieÁre induces gastrula-

tion defects, a phenotype typically not observed with

CmADMP, and futhermore does not enlarge head structures

unlike CmADMP. In conclusion, none of these TGF-bs

appears as a likely partner interacting with ADMP.

3.2. The role of ADMP in organizer patterning

The gastrula organizer can be subdivided into head and

trunk organizer and our data indicate that this subdivision is

R. Dosch, C. Niehrs / Mechanisms of Development 90 (2000) 195±203200

Fig. 5. ADMP inhibits head organizer genes. Four-cell embryos were

injected into two opposite blastomeres with 500 pg ppl or 125 pg tbr

with either 375 pg ppl or cmadmp, or 125 pg ppl with 375 pg cmadmp,

or 125 pg tbr with 150 pg dominant-negative wnt8 mRNA. Co-injection of

250 pg nlslacZ mRNA coding for b -galactosidase carrying a nuclear loca-

lization sequence identi®ed after b-galactosidase staining at stage 11 those

embryos that were injected ventrally. Only these embryos were sectioned

sagitally and processed for whole-mount in situ hybridisation for not2,

pintallavis, frzb or dkk1 expression as indicated. The dorsal side is to the

right with animal facing up. Note that injection of tBR with CmADMP

superinduces the expression of dkk1, frzb and not2 while pintallavis expres-

sion is not enhanced (arrowheads).

Table 2

Head induction by simultaneous repression of BMP and ADMP signallinga

Ventrally injected mRNA (ng per blastomere) Incomplete

secondary axis (%)

Secondary axis

with one eye (%)

Complete secondary

axis (%)

Normal

(%)

No. of

embryos

PPL (1.5) 0.4 0 0 99.6 234

tBr (0.125) 1 PPL (0.75) 57.8 0 0 42.2 128

tBr (0.25 - 0.5) 59.2 0 0 40.8 265

tBr (0.125) 1 CmADMP (0.375) 25 53 0 22 100

PPL (0.25) 1 CmADMP (0.375) 0 0 0 100 226

tBr (0.125) 1 dnWnt8 (0.15) 6.3 71.8 1.4 20.4 142

tBr (0.125) 1 CmADMP (0.375) 1 dnWnt8 (0.15) 7.3 40.2 37.8 14.6 82

CmBMP7 (0.375) 1 PPL (0.75) 71.6 0 0 28.4 134

CmBMP7 (0.5 - 1.5) 53.6 0 0 46.4 235

CmBMP7 (0.75) 1 tBR (0.25) 50 0 0 50 38

CmBMP7 (0.375) 1 CmADMP (0.75) 23.1 61.5 0 15.4 65

CmBMP7 (0.25) 1 dnWnt8 (0.15) 15.5 63.1 0 21.4 103

Chordin (0.05) 1 PPL (2.5) 59.3 0 0 40.7 199

Chordin (0.1±0.5) 64.1 0 0 35.9 167

Chordin (0.05) 1 tBR (0.25) 57.9 0 0 42.1 38

Chordin (0.05) 1 CmADMP (0.75±2.5) 20.3 34 0 45.8 212

Chordin (0.05) 1 dnWnt8 (0.15) 13.9 49.4 0 36.7 180

Noggin (0.025±0.1) 46 0.6 0 53.4 161

Noggin (0.025) 1 CmADMP (0.375±1.5) 27.3 48.5 0 24.2 33

a Four-cell embryos were microinjected into two ventral blastomeres with the indicated mRNAs. After three days at room temperature embryos were scored

for the formation of secondary embryonic axis without eyes (incomplete), with one eye or two separated eyes (complete; as shown in Fig. 4B). A minimum of

two independent experiments was carried out for every injection.

regulated by ADMP. We show that ADMP is required to

antagonize head formation: CmADMP injection leads to

dorsoanteriorized embryos, whose hallmark are enlarged

heads and shortened trunks and CmADMP co-injection

with BMP inhibitors induces ectopic heads on the ventral

side, which is paralled by the induction of head-inducing

Wnt-antagonists. Thus, one reason for the failure of anti-

BMPs (trunk inducers) to induce complete secondary

embryonic axes is due to concomittant induction of admp,

which in turn represses head organizer genes. Consistent

with a role in head repression, admp is predominantly

expressed in chordamesoderm (trunk organizer). These

results argue for a model (Fig. 6C) where ADMP represses

head inducing anti-Wnts in the trunk organizer. We propose

that ADMP functions together with Wnt and Nodal signals

(Piccolo et al., 1999) to repress expression of head inducers.

However, ADMP may not solely act through inhibition of

Wnt inhibitors, since CmADMP can even synergize in co-

injections with dominant-negative Wnt8 to form more

complete heads (Fig. 4B).

We show that ADMP escapes the inhibitory function of

Cerberus, Chordin and Noggin but not of Follistatin. admp

and follistatin show largely overlapping expression during

mid-and late gastrula in chordamesoderm, but unlike follis-

tatin, admp is also expressed in posterior chordamesoderm

and reaches further anteriorly, into prechordal plate. The

role of Follistatin may be to attenuate ADMP signalling to

levels suf®cient for repression of head- but not of trunk

organizer. Targeted disruption of mouse follistatin reveals

no early embryonic head or trunk defects (Matzuk et al.,

1995). Possibly, follistatin-related-protein, which is

expressed like follistatin in Xenopus gastrulae (Okabayashi

et al., 1999) acts redundantly in this context.

These results highlight how differential target speci®city

of BMP antagonists may contribute to organizer patterning

and offers one explanation for their apparent redundancy.

4. Experimental procedures

4.1. Embryos and explants

In vitro fertilization, embryo and explant culture were

carried out as described (Dosch et al., 1997). Embryos

R. Dosch, C. Niehrs / Mechanisms of Development 90 (2000) 195±203 201

Fig. 6. (A) Differential sensitivity of ADMP and BMP4 signalling to BMP antagonists. Eight-cell embryos were co-injected into the four animal blastomeres

with 150 pg bmp4 or admp and either 1 ng preprolactin (2), 250 or 500 pg noggin, 500 pg or 1ng admp, 500 pg chordin or follistatin or 250 pg cerberus mRNA

per blastomere, as indicated. PPL, control injection of 1.5 ng preprolactin mRNA. Animal caps were cut from blastulae and analyzed at stage 11 for the

expression of vent1 by RT-PCR. H4, histone 4 as loading control. -RT, minus reverse transcriptase control, WE, whole embryo control. (B) admp is expressed

in the trunk organizer. Single- (top) or double (bottom) in situ hybridizations of late gastrula or early neurula embryos (stages 12 or 13 as indicated in the lower

right) are shown. Top row, expression of chordin, follistatin and admp. Sagittal sections are shown, dorsal side up, anterior to the left. Bottom row, admp (red),

dkk1 and frzb (blue). Left, sagittal section, dorsal side up, anterior to the left. Right, two whole-mounts in anterior view, dorsal side up. (C) Model of ADMP

function in head organizer repression. A stage 12 gastrula is shown schematically as in the whole-mounts (B).

were staged according to (Nieuwkoop and Faber, 1967).

Ventral or dorsal marginal zones (VMZs and DMZs) were

explanted at stage 10 to 10.5 in 0.5£ modi®ed Barth's solu-

tion and then cultivated for development until the stages

indicated. Animal caps were explanted at stage 8 and culti-

vated in 0.5£ modi®ed Barth until the required stage for

RNA preparation.

4.2. Whole-mount in situ hybridization and b -galactosidase

staining

b-Galactosidase staining (Lemaire et al., 1995) and

whole-mount in situ hybridizations were carried out as

described (Harland, 1991). Double stainings of whole-

mount in situ hybridizations were modi®ed according to

(Reifers et al., 1998). After staining with Fast Red

(Roche) and two 5 min washes in PBSw (PBS plus 0.1%

Tween), embryos were re-®xed with 4% paraformaldehyde

in PBS. Following another two 5 min washes in PBSw,

alkaline phosphatase of the anti-¯uorescein antibody was

inactivated by heating the embryos 2 h at 688C. Embryos

were washed in TN-buffer (100 mM Tris pH 7.5, 150 mM

NaCl) for 5 min, blocked with TN 1 1% blocking reagent

(Roche) for 2 h and then incubated with anti-digoxigenin

antibody over night. Staining with BM-purple was as

described (Dosch et al., 1997).

4.3. Constructs and microinjection experiments

Full-length admp was ampli®ed by PCR from neurula

cDNA and cloned into pCS21 (Rupp et al., 1994) and

pBluescript after EcoRI-XhoI digestion with the following

primers: fw:GGGGAATTCCTTGATGAGATGGACCTT-

AGG (EcoRI site underlined) rev:GGGCTCGAGT-

TAGTGGCACCCGCAGCTGCC (XhoI site underlined).

Both plasmids were veri®ed by DNA sequencing. The puta-

tive ADMP cleavage site, RLGR (5 0-TCCGAACCATCT-

3 0) was altered to GVDG (5 0-GGCGTCGACGGA) by the

strategy described in (Hawley et al., 1995) with the follow-

ing gene speci®c primers: fw:GGGGTCGACGGATCAG-

TAGAAGAAGATGGACAA; rev:GGGGTCGACGCCTG-

TTCTGTTTGAAGTTGGTGC-3 0 (SalI site underlined) to

generate pCS1CmADMP. The mutation in the cleavage

site was con®rmed by sequencing. Plasmids were linearized

and capped mRNA was transcribed using the Megascript

Kit (Ambion) as follows: ppl, admp, cmadmp, dnwnt8,

nodal-related1 (Asp718, SP6), tbr (EcoRI, SP6), noggin

(NcoI, SP6), cmbmp7 (XhoI, SP6), nlslacZ (XbaI, SP6),

cabr (HincII, SP6), chordin (S®I, T3), bmp4 (XhoI, T3).

For animal cap assays eight-cell embryos were injected

into the four animal blastomeres. For phenotypes or VMZ/

DMZ assays the embryos were injected radially into all

blastomeres at the four-cell stage with the indicated doses.

4.4. RT-PCR

PCR assays with reverse transcription (RT-PCR) were

carried out in the exponential phase of ampli®cation as

described (Dosch et al., 1997). PCR primers used were

histone4, otx2, chordin (Glinka et al., 1997), admp (Moos

et al., 1995), vent2 (Onichtchouk et al., 1996), vent1, gsc

(Dosch et al., 1997). All RT-PCR experiments were done at

least twice.

Acknowledgements

We thank T. Bouwmeester, A. BraÈndli, K. Cho, A. Fain-

sod, R. Harland, R. Moon, H. Steinbeisser and N. Ueno for

reagents and M. Brand for protocolls. This work was

supported by the Deutsche Forschungsgemeinschaft.

References

Beck, C.W., Slack, J.M., 1999. A developmental pathway controlling

outgrowth of the Xenopus tail bud. Development 126, 1611±1620.

Beddington, R.S.P., Robertson, E.J., 1999. Axis development and early

asymmetry in mammals. Cell 96, 195±209.

Blitz, I.L., Cho, K.W.Y., 1995. Anterior neuroectoderm is progressively

induced during gastrulation: the role of the Xenopus homeobox gene

orthodenticle. Development 121, 993±1004.

Bouwmeester, T., Leyns, L., 1997. Vertebrate head induction by anterior

primitive endoderm. BioEssays 19, 855±863.

Bouwmeester, T., Kim, S.-H., Sasai, Y., Lu, B., De Robertis, E.M., 1996.

Cerberus is a head-inducing secreted factor expressed in the anterior

endoderm of Spemann's organizer. Nature 382, 595±601.

Bradley, L.C., Snape, A., Bhatt, S., Wilkinson, D.G., 1993. The structure

and expression of the Xenopus Krox-20 gene: conserved and divergent

patterns of expression in rhombomeres and neural crest. Mech Dev 40,

73±84.

Cho, K.W., Blitz, I.L., 1998. BMPs. Smads and metalloproteases: extra-

cellular and intracellular modes of negative regulation. Curr Opin Genet

Dev 8, 443±449.

Christian, J.L., Moon, R.T., 1993. Interactions between Xwnt-8 and

Spemann organizer signaling pathways generate dorsoventral pattern

in the embryonic mesoderm of Xenopus. Genes Dev 7, 13±28.

Dale, L., Howes, G., Price, B.M., Smith, J.C., 1992. Bone morphogenetic

protein 4: a ventralizing factor in early Xenopus development. Devel-

opment 115, 573±585.

Dosch, R., Gawantka, V., Delius, H., Blumenstock, C., Niehrs, C., 1997.

Bmp-4 acts as a morphogen in dorsoventral mesoderm patterning in

Xenopus. Development 124, 2325±2334.

Frisch, A., Wright, C.V.E., 1998. XBMPRII, a novel Xenopus type II

receptor mediating BMP signalling in embryonic tissues. Development

125, 431±442.

Gawantka, V., Delius, H., Hirschfeld, K., Blumenstock, C., Niehrs, C.,

1995. Antagonizing the Spemann organizer: role of the homeobox

gene Xvent-1. Embo J 14, 6268±6279.

Glinka, A., Wu, W., Onichtchouk, D., Blumenstock, C., Niehrs, C., 1997.

Head induction by simultaneous repression of Bmp and wnt signalling

in Xenopus. Nature 389, 517±519.

Glinka, A., Wu, W., Delius, H., Monaghan, P.A., Blumenstock, C., Niehrs,

C., 1998. Dickkopf-1 is a member of a new family of secreted proteins

and functions in head induction. Nature 391, 357±362.

Gont, L.K., Steinbeisser, H., Blumberg, B., De Robertis, E.M., 1993. Tail

formation as a continuation of gastrulation: the multiple cell popula-

tions of the Xenopus tailbud derive from the late blastopore lip. Devel-

opment 119, 991±1004.

Graff, J.M., Thies, R.S., Song, J.J., Celeste, A.J., Melton, D.A., 1994.

R. Dosch, C. Niehrs / Mechanisms of Development 90 (2000) 195±203202

Studies with a Xenopus BMP receptor suggest that ventral mesoderm-

inducing signals override dorsal signals in vivo. Cell 79, 169±179.

Harland, R.M., 1991. In situ hybridization: an improved whole-mount

method for Xenopus embryos. Methods Cell Biol 36, 685±695.

Harland, R.M., Gerhart, J., 1997. Formation and function of Spemann's

organizer. Annu Rev Dev Biol 13, 611±667.

Hawley, S.H., Wunnenberg, S.K., Hashimoto, C., Laurent, M.N., Watabe,

T., Blumberg, B.W., Cho, K.W., 1995. Disruption of BMP signals in

embryonic Xenopus ectoderm leads to direct neural induction. Genes

Dev 9, 2923±2935.

Hemmati-Brivanlou, A., de, L.T.J., Holt, C., Harland, R.M., 1991. Cephalic

expression and molecular characterization of Xenopus En-2. Develop-

ment 111, 715±724.

Hemmati-Brivanlou, A., Kelly, O.G., Melton, D.A., 1994. Follistatin, an

antagonist of activin, is expressed in the Spemann organizer and

displays direct neuralizing activity. Cell 77, 283±295.

Hogan, B.M., 1996. Bone morphogenetic proteins: multifunctional regula-

tors of vertebrate development. Genes Dev. 10, 1580±1594.

Hoppler, S., Brown, J.D., Moon, R.T., 1996. Expression of a dominant-

negative wnt blocks induction of MyoD in Xenopus embryos. Genes

Dev 10, 2805±2817.

Jones, C.M., Lyons, K.M., Lapan, P.M., Wright, C.V., Hogan, B.L., 1992.

DVR-4 (bone morphogenetic protein-4) as a posterior-ventralizing

factor in Xenopus mesoderm induction. Development 115, 639±647.

Jones, C.M., Kuehn, M.R., Hogan, B.L., Smith, J.C., Wright, C.V., 1995.

Nodal-related signals induce axial mesoderm and dorsalize mesoderm

during gastrulation. Development 121, 3651±3662.

Joubin, K., Stern, C., 1999. Molecular interactions continuously de®ne the

organizer during the cell movements of gastrulation. Cell 98, 559±571.

Kao, K.R., Elinson, R.P., 1988. The entire mesodermal mantle behaves as

Spemann's organizer in dorsoanterior enhanced Xenopus laevis

embryos. Dev Biol 127, 64±77.

Lemaire, P., Garrett, N., Gurdon, J.B., 1995. Expression cloning of

Siamois, a Xenopus homeobox gene expressed in dorsal-vegetal cells

of blastulae and able to induce a complete secondary axis. Cell 81, 85±

94.

Leyns, L., Bouwmeester, T., Kim, S.-H., Piccolo, S., De Robertis, E.M.,

1997. Frzb-1 is a secreted antagonist of wnt-signals expressed in the

Spemann organizer. Cell 88, 747±756.

Mangold, O., 1933. UÈ ber die InduktionsfaÈhigkeit der verschiedenen

Bezirke der Neurula von Urodelen. Naturwissenschaften 21, 761±766.

MassagueÂ, J., 1998. TGF beta signal transduction. Annu. Rev. Biochem. 67,

753±791.

Matzuk, M.M., Lu, N., Vogel, H., Sellheyer, K., Roop, D.R., Bradley, A.,

1995. Multiple defects and perinatal death in mice de®cient in follista-

tin. Nature 374, 360±363.

Moos, M.J., Wang, S., Krinks, M., 1995. Anti-dorsalizing morphogenetic

protein is a novel TGF-beta homolog expressed in the Spemann orga-

nizer. Development 121, 4293±4301.

Mullins, M.C., 1999. Embryonic axis formation in the zebra®sh. Methods

Cell Biol 59, 159±178.

Niehrs, C., 1999. Head in the Wnt- the molecular nature of Spemann's head

organizer. Trends Genet.

Nieuwkoop, P.D., Faber, J., 1967. Normal Table of Xenopus laevis,

Amsterdam, North Holland.

Nishimatsu, S.-I., Thomsen, G.H., 1998. Ventral mesoderm induction and

patterning by bone morphogenetic proteins. Mech. Dev. 74, 75±88.

Okabayashi, K., Shoji, H., Onuma, Y., Nakamura, T., Nose, K., Sugino, H.,

Asashima, M., 1999. cDNA cloning and distribution of the Xenopus

follistatin-related protein. Biochem Biophys Res Commun 254, 42±

48.

Onichtchouk, D., Gawantka, V., Dosch, R., Delius, H., Hirschfeld, K.,

Blumenstock, C., Niehrs, C., 1996. The Xvent-2 homeobox gene is

part of the BMP-4 signaling pathway controlling dorsoventral pattern-

ing of Xenopus mesoderm. Development 122, 3045±3053.

Osada, S.I., Wright, C.V., 1999. Xenopus nodal-related signaling is essen-

tial for mesendodermal patterning during early embryogenesis. Devel-

opment 126, 3229±3240.

Pannese, M., Polo, C., Andreazzoli, M., Vignali, R., Kablar, B., Barsacchi,

G., Boncinelli, E., 1995. The Xenopus homologue of Otx2 is a maternal

homeobox gene that demarcates and speci®es anterior body regions.

Development 121, 707±720.

Piccolo, S., Agius, E., Leyns, L., Bhattacharyya, S., Grunz, H., Bouwmee-

ster, T., De Robertis, E.M., 1999. The head inducer Cerberus is a multi-

functional antagonist of Nodal. BMP and Wnt signals. Nature 397, 707±

710.

Reddi, A.H., 1998. Role of morphogenetic proteins in skeletal tissue engi-

neering and regeneration. Nat Biotechnol 16, 247±252.

Reifers, F., Bohli, H., Walsh, E.C., Crossley, P.H., Stainier, D.Y., Brand,

M., 1998. Fgf8 is mutated in zebra®sh acerebellar (ace) mutants and is

required for maintenance of midbrain-hindbrain boundary development

and somitogenesis. Development 125, 2381±2395.

Ruiz i Altaba, A., Jessell, T.M., 1992. Pintallavis, a gene expressed in the

organizer and midline cells of frog embryos: involvement in the devel-

opment of the neural axis. Development 116, 81±93.

Rupp, R.A., Snider, L., Weintraub, H., 1994. Xenopus embryos regulate the

nuclear localization of XMyoD. Genes Dev 8, 1311±1323.

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-speci®c homeobox genes. Cell 79, 779±790.

Sive, H., Hattori, K., Weintraub, H., 1989. Progressive determination

during formation of the anteroposterior axis in Xenopus laevis. Cell

58, 171±180.

Smith, W.C., Harland, R.M., 1992. Expression cloning of noggin, a new

dorsalizing factor localized to the Spemann organizer in Xenopus

embryos. Cell 70, 829±840.

Spemann, H., 1931. UÈ ber den Anteil von Implantat und Wirtskeim an der

Orientierung und Beschaffenheit der induzierten Embryonalanlage. W.

Roux Arch Entwicklungsmech 123, 390±516.

Sun, B.I., Bush, S.M., CollinsRacie, L.A., et al., 1999. Derriere: a TGF-beta

family member required for posterior development in Xenopus. Devel-

opment 126, 1467±1482.

Suzuki, A., Thies, R.S., Yamaji, N., Song, J.J., Wozney, J.M., Murakami,

K., Ueno, N., 1994. A truncated bone morphogenetic protein receptor

affects dorsal-ventral patterning in the early Xenopus embryo. Proc Natl

Acad Sci USA 91, 10255±10259.

Takahashi, H., Ikeda, T., 1996. Transcripts for two members of the trans-

forming growth factor-beta superfamily BMP-3 and BMP-7 are

expressed in developing rat embryos. Dev Dyn 207, 439±449.

Takao, M., Hino, J., Takeshita, N., Konno, Y., Nishizawa, T., Matsuo, H.,

Kangawa, K., 1996. Identi®cation of rat bone morphogenetic protein-3b

(BMP-3b), a new member of BMP-3. Biochem Biophys Res Commun

219, 656±662.

Wang, S., Krinks, M., Kleinwaks, L., Moos, M.J., 1997a. A novel Xenopus

homologue of bone morphogenetic protein-7 (BMP-7). Genes Funct 1,

259±271.

Wang, S., Krinks, M., Lin, K., Luyten, F.P., Moos, M., 1997b. Frzb, a

secreted protein expressed in the Spemann organizer, binds and inhibits

Wnt-8. Cell 88, 757±766.

Wittbrodt, J., Rosa, F.M., 1994. Disruption of mesoderm and axis formation

in ®sh by ectopic expression of activin variants: the role of maternal

activin. Genes-Dev. 8, 1448±1462.

Zaraisky, A.G., Lukyanov, S.A., Vasiliev, O.L., Smirnov, Y.V., Belyavsky,

A.V., Kazanskaya, O.V., 1992. A novel homeobox gene expressed in

the anterior neural plate of the Xenopus embryo. Dev Biol 152, 373±

382.

R. Dosch, C. Niehrs / Mechanisms of Development 90 (2000) 195±203 203