Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
DEVELOPMENTAL BIOLOGY 192, 537–550 (1997)
ARTICLE NO. DB978788
Differential Regulation of chordin Expression
Domains in Mutant Zebrafish
1
Valarie E. Miller-Bertoglio,*
,
† Shannon Fisher,* Alejandro Sa
´nchez,*
Mary C. Mullins,‡ and Marnie E. Halpern*
,
†
*Department of Embryology, Carnegie Institution of Washington, Baltimore, Maryland 21210;
†Department of Biology, Johns Hopkins University, Baltimore, Maryland 21210; and
‡Department of Cell and Developmental Biology, University of Pennsylvania,
Philadelphia, Pennsylvania 19104-6058
Patterning along the dorsal–ventral (D –V) axis of Xenopus and Drosophila embryos is believed to occur through a conserved
molecular mechanism, with homologous proteins Chordin and Short gastrulation (Sog) antagonizing signaling by bone
morphogenetic protein 4 (BMP-4) and Decapentaplegic (Dpp), respectively. We have isolated a zebrafish gene that is highly
homologous to chordin and sog within cysteine-rich domains and exhibits conserved aspects of expression and function.
As in Xenopus embryos, zebrafish chordin is expressed in the organizer region and transiently in axial mesoderm. Injection
of zebrafish chordin mRNA to the ventral side of Xenopus embryos induced secondary axes. Ectopic overexpression in
zebrafish resulted in an expansion of paraxial mesoderm and neurectoderm at the expense of more lateral and ventral
derivatives, producing a range of defects similar to those of dorsalized zebrafish mutants (Mullins et al., 1996). In accordance
with the proposed function of chordin in D–V patterning, dorsalized zebrafish mutants showed expanded domains of
chordin expression by midgastrulation, while some ventralized mutants had reduced expression; however, in all mutants
examined, early organizer expression was unaltered. In contrast to Xenopus, zebrafish chordin is also expressed in paraxial
mesoderm and ectoderm and in localized regions of the developing brain, suggesting that there are additional roles for
chordin in zebrafish embryonic development. Surprisingly, paraxial mesodermal expression of chordin appeared unaltered
in spadetail mutants that later lack trunk muscle (Kimmel et al., 1989), while axial mesodermal expression was affected.
This finding reveals an unexpected function for spadetail in midline mesoderm and in differential regulation of chordin
expression during gastrulation.
q1997 Academic Press
INTRODUCTION
gin, and Follistatin, all of which dorsalize mesoderm and
induce neurectoderm (Smith and Harland, 1992; Hemmati-
Dorsal marginal cells of the early amphibian embryo can Brivanlou et al., 1994; Sasai et al., 1994).
influence mesodermal differentiation along the dorsal–ven- More recently, development of ventral derivatives such as
tral axis and direct presumptive ectoderm to take on a neu- blood and epidermis has been found to occur not by default,
ral rather than epidermal fate (refer to Lemaire and Kodja- but by the activity of ventralizing factors such as bone mor-
bachian, 1996; Thomsen, 1997). Transplantation of this dor- phogenetic proteins (BMPs). Ectopic expression of BMP-4 in
sal region, referred to as the Spemann organizer, from one Xenopus, for example, causes an expansion of ventral meso-
gastrula to the ventral side of another, results in the induc- derm at the expense of more lateral and dorsal mesoderm
tion of a second embryonic axis consisting of both donor (Jones et al., 1992; Schmidt et al., 1995a) and rescues embryos
and host cells (Spemann and Mangold, 1924). It was long- dorsalized by LiCl (Fainsod et al., 1994). Conversely, disrup-
held that the organizer directs patterning of adjacent dorso- tion of BMP signaling leads to an increase in dorsal mesoderm
lateral mesoderm and ectoderm through the action of se- and neurectoderm (Graff et al., 1994; Hawley et al., 1995;
creted molecules, and in the absence of these factors, ven- Schmidtet al., 1995a; Steinbeisser et al., 1995).Furtherstudies
tral cell fates would be assumed. Candidate proteins for have shown that both Chordin and Noggin proteins directly
mediating organizer activity in vivo include Chordin, Nog- bind BMP-4 and prevent it from activating its receptor (Piccolo
et al., 1996; Zimmerman et al., 1996). These findings suggest
that dorsally derived proteins function in vivo by antagonizing
1
This paper is dedicated in memory of Dr. Jane Oppenheimer
(1911–1996) for her pioneering studies on the teleost organizer.
the ventralizing action of BMP-4.
537
0012-1606/97 $25.00
Copyright q1997 by Academic Press
All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
538
Miller-Bertoglio et al.
The role of Chordin in antagonizing BMP signaling is chordin homologue and examined expression in early zebra-
fish development. Zebrafish chordin is initially expressedconserved in evolution (Holley et al., 1995; Sasai et al.,
1995). In Drosophila melanogaster, short gastrulation (sog; shortly after midblastula transition and, as in Xenopus,
comes to be strongly expressed in the organizer region atFranc
¸ois et al., 1994; Biehs et al., 1996), a chordin homo-
logue (Franc
¸ois and Bier, 1995; Schmidt et al., 1995b), antag- early gastrulation stages. The effect of ectopic chordin
mRNAin wild-type (WT) embryos andthe expanded expres-onizes the function of decapentaplegic (dpp), a bmp-4 ho-
mologue (Padgett et al., 1987). In addition, the proteins sion domains that we find in dorsalized zebrafish mutants
areconsistent withthe dorsalizing action of chordin in vivo.themselves can functionally substitute for one another.
Xenopus embryos ventralized by ultraviolet light can be In contrast to Xenopus chordin, expression of the zebra-
fish gene is not restricted to the organizer and its deriva-rescued by injection of sog mRNA, whereas injection of
chordin mRNA into D. melanogaster mimics some of the tives, but is also transiently expressed in the developing
brain and in paraxial mesoderm and ectoderm. To probeventralizing effects of ectopic sog expression ( Holley et al.,
1995; Schmidt et al., 1995b). the functions of these additional chordin domains, we have
examined the effect of other zebrafish mutations on chordinAs in amphibians, gastrulae of other vertebrates possess
organizing centers which presumably mediate their effects expression. Inspadetail mutants, which lack myoD expres-
sion at gastrulation (Weinberg et al., 1996), and later lackthrough similar mechanisms (Kintner and Dodd, 1991; Ho,
1992; Storey et al., 1992; Beddington, 1994; Shih and Fraser, trunk somiticmesoderm (Kimmel et al., 1989;Hoand Kane,
1990), expression of chordin in paraxial mesoderm appears1996). In developing mouse embryos, BMP-4 is also essen-
tial for mesoderm formation; however, a detailed pheno- normal. However, expression of chordin is not maintained
in the mutant midline, revealing a function for spadetailtypic analysis of BMP-4 null mutants has been difficult
since they are most often early embryonic lethals. Mice in axial mesoderm and in the differential regulation of
chordin.deficient for BMP-4 which survive gastrulation, possibly
through rescue by diffusion of maternal BMP-4, exhibit se-
lective mesodermal defects, including a lack of blood is-
lands in the visceral yolk sac (Winnier et al., 1995).
MATERIALS AND METHODS
In zebrafish and other teleost fishes, the dorsal embryonic
shield has properties similar to those of the Xenopus orga-
Embryo Culture and Staging
nizer. Transplantation of cells from the embryonic shield
Techniques for the care and breeding of zebrafish were followed
to the ventral side of another embryo results in the forma-
as described (refer to Westerfield, 1993). Embryos were collected
tion of a second axis (Oppenheimer, 1936; Ho, 1992; Shih
from single pair matings, maintained in embryo medium (15 mM
and Fraser, 1996). Furthermore, the embryonic shield has
NaCl, 0.5 mM KCl, 1mM CaCl
2
, 1mM MgSO
4
, 0.15 mM KH
2
PO
4
,
been found to express many of the same genes expressed in
0.05 mM Na
2
HPO
4
, 0.7 mM NaHCO
3
) at 28.57C, and staged ac-
the Xenopus organizer region (Stachel et al., 1993; Stra
¨hle
cording to hours (h) postfertilization and morphological criteria
et al., 1993; Schulte-Merker et al., 1994; Thisse et al., 1994;
(Kimmel et al., 1995).
Talbot et al., 1995, Toyama et al., 1995).
From recent mutagenesis screens in zebrafish, mutations
Zebrafish Mutants
have been isolated that alter specification of the dorsal–
ventral axis at gastrulation (Hammerschmidt et al., 1996b;
All studies on WT fish were carried out using the AB line from
Mullins et al., 1996; Solnika-Krezel, 1996; Fisher et al.,
Oregon (Fritz et al., 1996; provided by C. Walker). The gamma
1997), thereby providing a genetic means for exploring the
ray-induced mutations cyclops
b16
(Hatta et al., 1991), no tail
b160
mechanisms underlying vertebrate embryonic patterning.
(Halpern et al., 1993), and short tail
b180
(C. Kimmel, personal com-
munication); the spontaneous mutations floating head
n1
(Halpern
A number of mutations result in a dorsalized phenotype, in
et al., 1995; Talbot et al., 1995), spadetail
b104
(Kimmel et al., 1989),
which dorsal and dorsolateral derivatives are expanded with
and ogon
m60
(Solnika-Krezel et al., 1996); and the ethylnitrosourea
a corresponding reduction in more ventral derivatives. In
(ENU)-induced mutations swirl
ta72
,somitabun
dtc24
,snailhouse
ty68a
contrast, ventralizing mutations are characterized by an in-
(Mullins et al., 1996), and mercedes
tm305
(Hammerschmidt et al.,
crease in ventral derivatives (blood and epidermis) and a
1996b) were used for chordin expression analyses. Intercrosses of
decrease in dorsolateral derivatives (somites and neurecto-
heterozygous fish produced approximately 25% homozygous mu-
derm). Analysis of the ventralizing mutation dino (Ham-
tant progeny, with the exception of swr
ta72
and sbn
dtc24
which are
merschmidt et al., 1996a,b) strongly corroborates the model
zygotic semidominant alleles (refer to Mullins et al., 1996). All
for dorsal–ventral pattern formation derived from the am-
mutations were maintained in their original genetic background
phibianand Drosophila studies(Holley and Ferguson, 1997).
through intercrosses of heterozygous parents, except for sbn which
is also a maternal dominant mutation and was maintained by out-
In fact, recent molecular characterization confirms that the
crossing heterozygous males.
defect of dino (renamed chordino) is due to a lesion in the
zebrafish chordin homologue (Schulte-Merker et al., 1997).
In the absence of chordin, ventralizing signals are inappro-
Isolation of a Zebrafish chordin Gene
priatelyactive in more dorsal regionsof the embryo, thereby
producing the ventralized chordino mutant phenotype.
Degenerate oligonucleotides, 5*-TT(C/T)GG(A/G/C/T)GT(A/G/
C/T)ATG(C/T)A (C/T)TG-3*and 5*-GG(A/G)CA(A/G/C/T)GT(C/
We have tested the dorsalizing activity of the zebrafish
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
539
Chordin Expression in Mutant Zebrafish
T)TT(A/G)CA(A/G)CA-3*, were prepared from DNA sequences in tein (GFP) mRNA(20 ng/
m
l each) was injected into the ventral side
of one- to two-cell embryos. Following injection, embryos werethe first cysteine-rich (CR) repeat of Xenopus chordin and Drosoph-
ila sog and used for polymerase chain reaction (PCR) amplification incubated in 3% Ficoll for approximately 2 h and transferred to
dechlorinated water. Embryos were examined for the appearancefroma zebrafish gastrula stagecDNA library (provided byM. Rebag-
liati and I. Dawid). Reaction conditions were as follows: initial of the dorsal blastopore lip and, at various times, were scored for
abnormal blastopores and the development of secondary axes.denaturation at 967C for 2 min, 727C for 10 min, followed by 36
cycles of 947C for 1 min, 507C for 1.5 min, 727C for 15 s, and a Zebrafish embryos (one- to four-cell stage) were pressure injected
(8 nl) with a mixture of chordin mRNA and GFP mRNA (25–100final incubation of 727C for 10 min. The resulting 161-bp fragment
was gel purified using the Qiaex II gel extraction kit (Qiagen), ream- and 100 ng/
m
l, respectively). At gastrulation, injected embryos were
examined using a stereomicroscope (MZ12, Leica) with fluores-plified, and subcloned using the Original TA cloning kit (In-
vitrogen). The zebrafish fragment was used as a probe to screen a cence and GFP filter attachments (Kramer Scientific). At the four-
to six-somite stage, the degree of dorsalization was assessed usinggastrula stage cDNA library prepared in
l
ZAPII (provided by S.
Ekker), and two clones, 1.12S and 11.13, were obtained. the morphological criteria described by Mullins et al. (1996). The
most severely dorsalized embryos (classes 4 and 5) were fixed in
4% paraformaldehyde and processed for whole-mount in situ hy-
bridization using either pax2 (Krauss et al., 1991) and myoD (Wein-
Northern Analysis
berg et al., 1996) or pax2 and ntl (Schulte-Mercker et al., 1992)
Total RNA was isolated from staged zebrafish embryos using antisense RNA probes. Less severely affected embryos (classes 1–
Trizol (Gibco BRL), and aliquots (10
m
g) from each stage were sepa- 3) were allowed to develop and were reexamined at 24 h.
rated by formaldehyde–agarose gel electrophoresis and transferred
to Magnagraph nylon membrane (Micron Separations, Inc.). High
specific activity probes (ú10
9
dpm/
m
g) were prepared using the Pri-
RESULTS
meIt II system (Stratagene) from a HinfII restriction fragment (601
bp) encompassing the chordin CR3 and CR4 repeats. After hybrid-
Isolation of a Zebrafish chordin Homologue
ization, the blot was washed to a stringency of 0.11SSC, 0.5% SDS
at 657C and visualized with a Storm Fluorimager system (Molecular
Degenerate DNA primers were designed within four
Dynamics).
highly conserved CR regions (CR1–CR4) of the Xenopus
chordin and Drosophila sog coding sequences and used in
PCR amplifications with a zebrafish cDNA library as tem-
RNA in Situ Hybridization
plate. A resulting amplification product (161 bp) was found
Zebrafish chordin (clone 11.13 in pBluescript) was linearized
to be 45% homologous to Xenopus CR1 at the nucleotide
with SpeI for synthesis of antisense mRNA by T7 RNA polymerase
level (Fig. 1A). A full-length coding sequence of the zebra-
(Promega), in the presence of digoxigenin-UTP (Boehringer-Mann-
fishgene (Fig.1A) was derived from two cDNA clones(1.12S
heim). Whole-mount in situ hybridization was performed according
and 11.13) that were obtained by screening a gastrula-stage
to Thisse et al. (1993). Double-labeling in situ hybridization tech-
cDNA library. Translation of the single open reading frame
niquesusing fluorescein-UTP (Boehringer-Mannheim) fora chordin
generates a predicted protein of 940 amino acids which is 54
probe and digoxigenin-UTP for a krox-20 probe (Oxtoby and Jowett,
and 29% identical to the Xenopus and Drosophila proteins,
1993) were performed as previously described (Hauptmann and
Gerster, 1994).
respectively (Fig. 1B). Proteins of all three species contain
the four highly conserved and similarly spaced CR repeat
regions.
Embedding and Sectioning of Zebrafish Embryos
A single chordin transcript of approximately 4 kb was
detected by Northern blot analysis prior to gastrulation and
Following in situ hybridization, embryos were postfixed in 4%
a maximal level of expression was reached at 70% epiboly
paraformaldehyde, dehydrated through a graded ethanol series, and
(Fig. 1C). Expression persisted throughout early somitogen-
infiltrated overnight with London Resin Gold (Ted Pella, Inc.), then
transferred to fresh resin. The resin was polymerized by adding
esis but was greatly decreased by the 20-somite stage (19 h)
benzoin methyl ether (0.5%) and UV irradiating overnight. Sections
and not detected at 24 h.
(1.5
m
m) were cut using a Reichert ultramicrotome equipped with
a diamond knife.
Dynamic Pattern of chordin Expression during
Gastrulation
RNA Synthesis and Microinjection
In Xenopus embryos, chordin is first expressed in the
To produce an expression vector for in vitro synthesis of full-
dorsal lip and subsequently in tissues derived from the orga-
length chordin RNA, the 2.8-kb coding region of clone 1.12S was
nizer, including the prechordal plate, notochord, and chord-
amplified by long-range PCR using Pfu DNA polymerase (Stra-
oneural hinge (Sasai et al., 1994). As in Xenopus, the expres-
tagene). Oligonucleotide primers included BamHI restriction sites
sion pattern of zebrafish chordin is highly dynamic and is
for ligation of the partially digested PCR product into linearized
initially confined to the organizer region, the dorsal embry-
pCS2/vector (Turner et al., 1994). Capped chordin mRNA was
onic shield.
synthesized using the mMessage mMachine kit (Ambion).
Expression of zebrafish chordin was first detected by in
Xenopus laevis oocytes were fertilized in vitro, and following
situ hybridization approximately 1 h following the midblas-
cortical rotation, viable embryos were dejellied in 2% L-cysteine.
A mixture (5–10 nl) of chordin mRNA and green fluorescent pro-
tulatransition (MBT) and 1 2/3h prior to the onset of gastru-
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
540
Miller-Bertoglio et al.
12-15-97 04:49:58 dbal
541
Chordin Expression in Mutant Zebrafish
lation in a region presumed to form the future dorsal one- dorsal blastopore lips or secondary axes seen in uninjected
or GFP mRNA-injected control embryos.third of the embryo (oblong stage; data not shown). By
sphere stage (4 h postfertilization), the chordin-expressing In most chordin-injected embryos, an incomplete second-
ary axis emerged from the primary axis (compare Figs. 3Adomain encompassedapproximately one-quarter of the cir-
cumference of the embryo (Figs. 2A and 2B) and by the onset and 3B); however, partial secondary axes were occasionally
formed on the ventral side (Fig. 3C). Somites and the dorsalof gastrulation coincided with the dorsal embryonic shield
(Figs. 2C and 2D). As gastrulation proceeded chordin was tail fin were easily distinguished in embryos where the sec-
ond axis formed opposite the primary axis. Sectioning (nÅstrongly expressed in the midline of the newly forming axis
and in new bilateral domains flanking the dorsal shield 6) revealed the presence of neural tissue (data not shown);
however, notochord cells were not detected. Formation of(Figs. 2E–2H).
Sectioning of gastrulas (75% epiboly) revealed that the anterior structures, such as the eyes or cement glands, was
never observed.strong midline expression was confined to deep cells (Figs.
2D and 2H) which likely corresponded to mesoderm of the The dorsalizing activity of zebrafish Chordin was also
tested by ectopic overexpression in zebrafish embryos. Full-prechordal plate and presumptive notochord, based on partial
overlap between chordin and goosecoid expression (data not length chordin mRNA (200–800 pg) was coinjected with
GFP mRNA into the yolk of one- to four-cell stage embryos.shown). By late epiboly stages, chordin was no longer ex-
pressed in anterior axial mesoderm but was transiently ex- At early somite stages, injected embryos were examined for
morphological phenotypes. The most severely dorsalizedpressed in adaxial cells (Figs. 2I and 2J). In contrast to the
midline, chordin expression in lateral domains encompassed embryos (equivalent to classes 4–5 of Mullins et al., 1996)
were elongated in shape and had laterally expanded somitesall ectodermal and mesodermal cell layers, with the exception
of the cell layer closest to the yolk which did not express (Figs. 3E and 3F). Less severely affected embryos were char-
acterized at later stages of development (refer to Table 2)chordin (Fig. 2H). These domains spread in a lateral and vege-
tal direction during epiboly and were largely confined to the by the lack of the ventral tail fin (classes 1–3), slightly
enlarged tailbud (classes 2 and 3), and fused posterior so-tailbud by somitogenesis stages (Figs. 2J and 2K).
In the central nervous system (CNS), chordin expression mites (class 3).
In strongly dorsalized embryos (nÅ20), expression patternswas first detected at 100% epiboly in discrete regions of the
forebrain, midbrain, and hindbrain (Fig. 2L). Double in situ of pax2 (Krauss et al., 1991), myoD (Weinberg et al., 1996),
and ntl (Schulte-Merker et al., 1992) were considerably alteredhybridization for chordin and krox-20, which is expressed
in rhombomeres 3 (r3) and 5 (r5), demonstrated that expres- compared to mock-injected embryos (Figs. 3D–3I). Somite ex-
pression of myoD typically extended around the entire embry-sion of chordin was dynamic in the hindbrain. Although
there was considerable variability between embryos of the onic circumference (Figs. 3E and 3F), as did expression of pax
2in the midbrain. Expression of pax2 in the pronephric ductssame stage, chordin was generally expressed in r2 through
r5 at 100% epiboly and only in r2 and r4 at the 3- to 5- (Fig. 3G) was also ventrally displaced or entirely lacking (nÅ
10; Figs. 3H and 3I). Analysis of ntl expression in the presump-somite stage. By the 8-somite stage, chordin expression was
greatly diminished in the brain and was barely detectable tive notochord revealed varying, but limited, degrees of lateral
expansion (Fig. 3H). Together, the results indicate thatchordinin 12- to 15-somite stage embryos (data not shown). overexpression promotes differentiation of dorsolateral cell
fates at the expense of ventral or ventrolateral fates, but is not
sufficient to induce ectopic development of the most dorsal
Dorsalizing Activity of Zebrafish chordin in Frog
mesodermal derivatives.
and Fish Embryos
To test whether zebrafish Chordin possessed activities
Expanded and Reduced chordin Expression in
similar to the Xenopus and Drosophila proteins, full-length
Dorsalized and Ventralized Zebrafish Mutants
zebrafish chordin mRNA was injected into the ventral side
of one- to two-cell stage Xenopus embryos. Morphologically To explore the action of chordin further, we examined
the relationship between chordin expression and zebrafishrecognizable secondary axes were formed in 45% of injected
embryos (nÅ89; Table 1). In no case were supernumerary mutations previously implicated in dorsal–ventral pattern
FIG. 1. Sequence conservation and expression of zebrafish chordin. (A) The alignment compares the amino acid sequence of zebrafish
Chordin with Xenopus Chordin and Drosophila Short gastrulation. Stars and dots indicate identical amino acids and conservative amino acid
substitutions, respectively. Cysteine-rich repeats are outlined in boxes and the region corresponding to the CR1 PCR amplification product is
shown in gray. (B) The numbers indicate the percentages of amino acid identity within and outside the CR repeats (shaded) in the three proteins
(proteins are not drawn to scale). (C) Northern blot analysis demonstrates that a single chordin transcript is present at low levels in the late
blastula, increases during gastrulation, and persists through early somitogenesis. RNA was collected from embryos prior to midblastula transition
(mat) and at sphere stage (sph), 30% epiboly (30%), 50% epiboly (shld), 70% epiboly (70%), 6 somites (6s), 20 somites (20s), and 24 h postfertiliza-
tion (24h). For this blot, the probe was a HinfII restriction fragment (601 bp) that included CR3 and CR4.
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
542
Miller-Bertoglio et al.
FIG. 2. Dynamic expression of chordin in the developing zebrafish. (A, B) Zebrafish chordin was detected at the future dorsal side of
the late blastula (sphere stage: A, animal pole view; B, dorsal view). (C, D) By early gastrulation, expression was confined to deep cells in
the embryonic shield (arrowhead). D is a sagittal section through the shield at 50% epiboly. (E–H) As gastrulation proceeded, expression
was found transiently in the axis (arrowhead) and new bilateral domains of expression appeared (open arrowhead) in all superficial
(ectoderm) and deep (mesoderm) cell layers, except the layer closest to the yolk (open arrow, H). E is shield stage at 50% epiboly, F is 30
min after E, and G and H are at 70% epiboly. H is a transverse section through the axis and flanking regions. (I–K) In late gastrulae,
chordin was expressed in cells adjacent to the notochord and downregulated in the axis (open arrowheads). Bilateral expression domains
persisted in the most posterior region only. I is at 90% epiboly and J and K are at 100% epiboly. (L, M) At late gastrulation, chordin was
also expressed in discrete domains in the forebrain, midbrain, and hindbrain (in blue). Double labeling with a probe for krox-20, which is
expressed in rhombomeres 3 and 5 (in pink), indicated that chordin expression in the hindbrain is very dynamic, present in different
rhombomeres over time. Scale bars Å150
m
m (A–C, E–G, and I–M), 25
m
m (D), and 50
m
m (H).
formation at gastrulation. The mutations we focused on fell all of the dorsalized mutants that we examined, chordin
expression appeared normal in the dorsal organizer (throughinto two classes: those that produce dorsalized phenotypes
(swirl, snailhouse, somitabun, Mullins et al., 1996) and shield stage; Fig. 4G and data not shown) and in the axial
midline at later stages (Fig. 4B). However, as epiboly pro-thosethat produceventralized phenotypes (mercedes, Ham-
merschmidt et al., 1996b; ogon, Solnika-Krezel et al., 1996; ceeded, the lateral domains of chordin expression were
more ventrally expanded than in WT, which was most no-and short tail, V. Miller-Bertoglio and M. E. Halpern unpub-
lished observations). ticeable in animal or vegetal pole views (Figs. 4C–4F and
4H–4L). By 70% epiboly, chordin expression extendedEmbryosmutant forswirl (swr), somitabun (sbn),or snail-
house (snh) are characterized by the expansion of dorsally around the entire embryonic margin of swr mutants (23%,
9/39) and of more severely affected sbn mutants (33%, 16/derived tissues such as neurectoderm and somitic meso-
derm, as well as a decrease in more ventral derivatives such 45; Figs. 4D and 4J). In snh mutants (24%, 31/127), and in
less severely affected sbn embryos (67%, 29/45), chordinas the pronephric ducts and blood (Mullins et al., 1996). In
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
543
Chordin Expression in Mutant Zebrafish
TABLE 1
Induction of Secondary Axis in Xenopus Embryos
chordin RNA
a
Number Number Number with
(pg) Total number
b
abnormal
c
normal
d
second axis
e
(%)
0ú150 0 ú150 0 0
100 34 2 12 20 59
200 55 7 28 20 36
a
All embryos were coinjected with GFP mRNA.
b
Total number of injected embryos surviving gastrulation and neurulation.
c
Number of embryos which did not neurulate properly.
d
Number of injected embryos which developed a single, morphologically normal axis.
e
Number of injected embryos which developed a secondary axis.
expression encompassed approximately 75% of the circum- viously observed in ntl mutant gastrulas (Melby et al.,
1997).ference of the margin (Figs. 4I and 4L). By 95% epiboly,
however, this expression expanded around the entire mar- In floating head (flh) and spadetail (spt) mutants, chordin
expression was unchanged in early gastrulation, but wasgin of snh embryos, equivalent to the pattern observed in
swr mutants (Fig. 4F and data not shown). noticeably altered by late gastrula stages (Fig. 5). At 90%
epiboly, chordin was strongly expressed in the posteriorVentralized mutants develop a phenotypic syndrome op-
posite to that of dorsalized embryos, with a narrower axis region of the flh mutant midline (25%, 11/44), a region
where expression was downregulated in WT embryos of theand reduced rather than expanded expression of neurecto-
dermal and dorsal mesodermal markers. A variable increase same age (Figs. 5A and 5B). The high level of expression
that persisted in the midline of flh mutants likely resultedin ventral derivatives such as blood and ectoderm, including
extra fin folds in the tail, is also observed (Hammerschmidt from the respecification of axial mesodermal cells to parax-
ial mesoderm that has been proposed to be the basis of theet al., 1996b, Solnica-Krezel et al., 1996; V. Miller-Bertoglio
and M. E. Halpern, unpublished observations). flh mutant phenotype (Halpern et al., 1995; Melby et al.,
1996). Thus, as with the fused somites that later form in flhIn mercedes (mes) mutants that exhibit a weakly ven-
tralized phenotype, chordin expression was indistinguish- mutants, the lateral chordin expression domains appeared
continuous across the midline.able from WT (Fig. 4M). In ogon (ogo) mutants, expression
was normal in the embryonic shield; however, an overall The spadetail (spt) mutation is thought to cause the mis-
specification of trunk mesoderm, resulting in defective cel-decrease inthe axialand lateralchordin expressiondomains
was observed by 75% epiboly (25%, 7/28) and was more lular convergence toward the dorsal side of the embryo, and
an increase in cells in the tailbud at the expense of trunkpronouncedby 90% epiboly (25%, 22/88; Fig. 4O) compared
to WT (Fig. 4N). In short tail (stl) mutants, alterations in somites (Kimmel et al., 1989; Ho and Kane, 1990). We ex-
pected that the paraxial domains of chordin expression (Fig.the chordin expression pattern were not observed until 80%
epiboly, when the lateral chordin expression domains 5C) would be altered in spt mutants, possibly contributing
to the defect in trunk somite formation; however, this didflanking the axis were found to be less broad and thus more
dorsally restricted (Figs. 4P and 4Q). not prove to be the case (Fig. 5D). Rather, axial mesodermal
expression of chordin, which was normal at early gastrula-
tion, was lost by 80% epiboly in spt mutants (26%, 24/92;
Selective Regulation of chordin Expression
compare Figs. 5E and 5F), revealing an unexpected role for
spt
/
in the gastrula midline.The involvement of chordin in tissue patterning was ex-
amined further using zebrafish mutations that perturb the
development of specific dorsal or dorsolateral derivatives
such as the ventral neural tube, notochord, or somites.
DISCUSSION
In cyclops (cyc) mutants that display reduced axial meso-
derm and neurectoderm and notably lack the CNS floor We have cloned and characterized a zebrafish gene that
encodes a protein highly homologous to Chordin and Sogplate later in development (Hatta et al., 1991; Thisse et al.,
1993; Yan et al., 1995), chordin expression appeared normal in four similarly spaced cysteine-rich repeat regions. The
cysteine-rich domains characterize a superfamily of extra-throughout gastrulation (data not shown). Although no tail
(ntl) mutant embryos fail to form notochords, expression of cellular matrix or cell-surface proteins, members of which
have been shown to bind TGF-
b
(Paralkar et al., 1991; Mur-chordin was also relatively normal in progeny from ntl//
intercrosses. Variably diffuse and wider expression domains phy-Ullrich et al., 1992), and more recently BMP homo- and
heterodimers (Piccolo et al., 1996). The amino acid identitywere found at the midline of some embryos (data not
shown), consistent with defects in cellular convergence pre- between the fish, fly, and frog proteins is relatively low
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
544
Miller-Bertoglio et al.
FIG. 3. Dorsalizing activity of zebrafish Chordin in Xenopus and zebrafish embryos. (A) Uninjected Xenopus embryo (stage 37/38); (B)
Xenopus embryo coinjected with 10 nl of zebrafish chordin and GFP RNA (20 ng/
m
l) on the ventral side at the two-cell stage; (C) Xenopus
embryo coinjected ventrally at the one-cell stage with 5 nl of zebrafish chordin and GFP RNA (20 ng/
m
l). (D–I) In situ hybridization using
pax2 and myoD probes (D–F) or pax2 and ntl probes (G–I) of uninjected zebrafish embryos (D and G) or embryos injected with chordin
RNA (E, F, H, and I) at the one- to four-cell stage during segmentation. All injected embryos shown correspond to class d of Table 2. (E,
F) Severely dorsalized embryos were characterized by their elongated shape, laterally expanded domain of pax2 expression at the midbrain–
hindbrain junction (arrow), and ventrally extending somites (open arrows). (H) In some injected embryos, the ntl expression domain
(arrowheads) in the notochord was widened relative to controls (G). D–H are dorsal views. (I) A ventral view of the same embryo as H
shows the expansion of pax2 expression in the brain around the entire circumference of the embryo (arrow). Pronephric duct expression
of pax2 (arrow, G) was often missing in injected embryos.
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$8788 12-15-97 04:49:58 dbal
545
Chordin Expression in Mutant Zebrafish
TABLE 2
Ectopic chordin Expression Dorsalizes Wild-Type Zebrafish Embryos
Number
chordin RNA
a
Total number morphologically Number mildly Number strongly Dorsalized
(pg) of embryos
b
wild-type
c
dorsalized
d
dorsalized
e
(%)
098 — — — 0
200 62 25 20 17 60
400 64 11 24 29 83
800 99 20 42 37 80
a
All embryos were coinjected with GFP mRNA.
b
Total number of injected embryos which survived gastrulation.
c
Embryos with a ventral tail fin at 24 h.
d
Embryos classified as C1 and C2 after Mullins et al. (1996).
e
Embryos classified as C3–C5.
outside of the cysteine-rich regions, yet all three proteins of the amphibian organizer (Oppenheimer, 1936; Ho, 1992;
Shih and Fraser, 1996).are involved in establishing cell identity along the dorsal–
ventral axis of developing embryos. As with injections of sog RNA (Schmidt et al., 1995b),
injection of zebrafish chordin RNA to the ventral side ofSeveral lines of evidence support a conserved function for
zebrafish Chordin as an antagonist of ventralizing signals Xenopus embryos produced a partial secondary axis. While
somites and dorsal fins were visible in some induced axes,at gastrulation. These include the timing and location of
chordin expression, the dorsalizing activity resulting from notochords and anterior structures, such aseyesand cement
glands, were not observed. Ectopic expression of Xenopusectopic expression of the zebrafish gene in frog and fish
embryos, the alteration of expression in ventralized and dor- chordin, however, often resulted in the formation of noto-
chord in secondary axes (Sasai et al., 1994), suggesting thatsalized zebrafish mutants, and the embryonic phenotypes
produced by deletion or point mutations of chordin (Fisher the frog protein is capable of inducing notochord. The zebra-
fish and fly proteins may have a lower overall activity inet al., 1997, and unpublished observations; Schulte-Merker
et al., 1997). Xenopus, which could account for the absence of notochord
in the ectopic secondary axis, or have different properties
than the frog protein.
Organizer Activity of Zebrafish Chordin
The amphibian organizer is believed to be composed of
A Role for chordin in the Dorsolateral Territory of
three domains which have distinct inductive capacities in
the Zebrafish Gastrula
the head, trunk, and tail (Gont et al., 1993; refer to Lemaire
and Kodjabachian, 1996). In Xenopus, chordin is initially Ectopic expression of chordin in the zebrafish embryo,
through yolk injections of RNA, resulted in an expansionexpressed in the organizer and later in derivatives of each
of thesedomains, includingthe prechordalplate,notochord, of dorsolateral mesodermal and neurectodermal derivatives
at the expense of more ventral derivatives. In the most se-and chordoneural hinge (Sasai et al., 1994). The zebrafish
homologue is first expressed in the late blastula in a re- verely dorsalized embryos, expression of neural plate and
somite markers was fully expanded to the ventral side ofstricted domain of cells that give rise to the dorsal embry-
onicshield, the region presumed tobe the teleost equivalent the embryo and pax2-expressing pronephric ducts derived
FIG. 4. Altered dorsolateral chordin expression domains in dorsalized and ventralized zebrafish mutants. (A–F) Expression of chordin
in dorsalized swr mutants (B, D, F) was expanded ventrally (arrowheads) compared to WT sibs (A, C, E). A and B are animal pole views
and C and D are dorsal views at 75% epiboly. E and F are vegetal pole views at 90% epiboly. (G) At shield stage, dorsal organizer expression
of chordin was indistinguishable in sbn mutants and WT siblings (dorsal view) and in embryos from intercrosses of all other dorsalizing
mutations examined (not shown). (H–L) Animal pole views. Expression was ventrally expanded (arrowheads) compared to WT (H, K) in
both sbn (I and J) and snh (L) mutants at 75% epiboly. As described in the text, all embryos from sbn intercrosses had altered expression
patterns which fell into two classes (I was the pattern of two-thirds and J was the pattern of one-third of embryos). (M–Q) Dorsal views.
(M) chordin expression was indistinguishable in weakly ventralized mes mutants and WT siblings at 75% epiboly. (N, O) However,
expression was decreased overall in ventralized ogo mutant embryos (O) compared to WT sibs (N) at 90% epiboly. (P, Q) Dorsolateral
domains of chordin expression (arrowheads) were also more dorsally restricted in stl mutants (Q) compared to WT siblings (P) at 80%
epiboly. Asterisks indicate the dorsal midline. Scale bar Å20
m
m.
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
546
Miller-Bertoglio et al.
nation of BMP-4 signaling by a dominant negative receptor
expands the lateral mesoderm and neural plate without an
increase of dorsal midline derivatives (Schmidt et al.,
1995a). One interpretation is that there is a lateral limit
beyond which axial mesoderm cannot form, either because
genes necessary for notochord specification are not ex-
pressed outside of the dorsal-most region or because genes
expressed lateral to the midline limit the extent to which
axial mesoderm can be induced.
Analysis of chordin expression in dorsalized and ven-
tralized zebrafish mutants further supports the hypothesis
that the dorsolateral region of ectoderm and mesoderm is
the principal territory where chordin exerts its effects in
the zebrafish gastrula. In the dorsalized mutants swr, sbn,
and snh, chordin expression in the embryonic shield was
unchanged at the onset of gastrulation, and only later was
an expansion of the lateral expression domains observed.
Conversely, at late gastrulation, the ventralized mutants
ogo and stl showed an overall decrease in intensity and size
of the lateral domains of expression, respectively. Thus,
none of the mutations we examined affected the early ex-
pression of chordin in the organizer region. Rather, changes
in the later dorsolateral expression domains correlated well
with the phenotypic classes: expanded in dorsalized and
decreased in ventralized mutants.
The most compelling evidence for the dorsolateral terri-
tory of the gastrula being the focus of chordin action comes
from the characterization of embryonic phenotypes that re-
sult from deletions and point mutations of chordin (Ham-
merschmidt et al., 1996a,b; Fisher et al., 1997, and unpub-
lished observations; Schulte-Merker et al., 1997). Embryos
mutant for chordin show a pronounced reduction in tissues
derived from the dorsolateral region of the gastrula and a
corresponding increase in ventral derivatives, while the for-
FIG. 5. Domains of chordin expression are selectively affected in
flh and spt mutants. (A, B) At 90% epiboly, chordin expression
mation of the organizer and its subsequent differentiation
decreases in the posterior midline of WT embryos (A), but is in-
are less affected. We conclude, that at least for primary
creased in this region in flh mutants (B). (C–F) At 80% epiboly,
gastrulation, zebrafish chordin does not play a major role
chordin expression is not maintained in the axis of spt mutant
in defining the organizer region itself or in promoting the
embryos (D, F), compared to WT sibs (C, E). Scale bar Å20
m
m for
development of organizer derivatives, but is required for
A–D and 40
m
m for E and F.
correct patterning of adjacent dorsolateral derivatives, by
specifically antagonizing ventralizing signals in this region
of the embryo.
from lateral mesoderm were absent. More mildly dorsalized
embryos had laterally expanded somites and small, ven-
The Molecular Basis for chordin Action
trally displaced domains of pax2 pronephric duct expres-
sion. Injected embryos often had wider notochords and an In Xenopus embryos, Chordin has been shown to bind
BMPsand preventthem fromactivating theirreceptors(Pic-elongated shape, indicative of abnormal cell movements at
gastrulation. colo et al., 1996), and most likely exerts its effect in the
zebrafish gastrula by a similar mechanism. To date, thereThe observed classes of effects from chordin overex-
pression closely paralleled the phenotypes of zebrafish dor- has been no direct measurement of Chordin diffusion in the
embryo, although it has been shown to be a secreted proteinsalized mutants identified in recent mutagenesis screens
(Mullins et al., 1996). Embryos mutant for swirl, somita- (Piccolo et al., 1996). Recent work in Xenopus and zebrafish
supportsa modelwhereby BMP activity is present as agradi-bun, and snailhouse have variably expanded somites and
neurectoderm and wider notochords. However, as in the ent in mesoderm and ectoderm and that the level of activity
conveyspositional information and sets thelimits for devel-most severely dorsalized mutants, chordin overexpression
alone was insufficient to promote a complete expansion of opment of specific cell types (Dosch et al., 1997; Neave et
al., 1997; Wilson et al., 1997). The gradient ofBMP activitythe most dorsal derivatives such as notochord. A similar
phenomenon is found in Xenopus embryos, in which elimi- in the zebrafish has also been proposed to arise through
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
547
Chordin Expression in Mutant Zebrafish
interactions with Chordin (Hammerschmidt et al., 1996a); ment of maternally derived sbn
/
in dorsal–ventral pat-
terning.thus,in chordin mutants, BMP activity wouldexpand unop-
posed and the gradient would be altered to favor develop-
ment of ventral derivatives. Conversely, in the presence
of increased Chordin, as with overexpression, the gradient
Additional Sites for chordin Activity in the
would be altered in the opposite direction, favoring develop-
Zebrafish
ment of dorsal derivatives.
Homologues of bmp-4 and bmp-2 have been cloned and BMP family members are involved in tissue patterning
their expression has been examined in the zebrafish embryo beyond gastrulation, influencing differentiation of the neu-
(Chin et al., 1997; Nikaido et al., 1997). Both genes are ral tube and somites, organogenesis of numerous organ sys-
expressedin ventral and ventrolateral regionsof the gastrula tems, and skeletal morphogenesis (Dudley et al., 1995; Luo
and have ventralizing activity when ectopically expressed. et al., 1995; Pourquie
´et al., 1996; Reissmann et al., 1996;
However, in contrast to Xenopus, zebrafish bmp-4 is also Storm and Kingsley, 1996; Zhang and Bradley, 1996; Zou
expressed dorsally, in the embryonic shield and its deriva- and Niswander, 1996; Macias et al., 1997; Schultheiss et al.,
tive, the prechordal plate. bmp-2 is expressed earlier and is 1997; and refer to Hogan, 1996). In addition to expression in
not expressed dorsally, leading to the suggestion that it is the dorsal shield and midline mesoderm of the zebrafish
more of a functional equivalent to bmp-4 in Xenopus (Ni- gastrula, similar to that described for Xenopus (Sasai et al.,
kaido et al., 1997). Regions of bmp-2/4 coexpression could 1994), chordin transcripts were detected in restricted re-
also be sites of heterodimer formation. While the identity gions of the presumptive forebrain and midbrain –hindbrain
of the active BMPs in the zebrafish gastrula remains un- junction and in a highly dynamic pattern in developing
known, the recent discovery that swr mutations are lesions hindbrain rhombomeres. Although neither bmp-4 nor bmp-
in the bmp-2 gene (M. Mullins, unpublished observations) 2expression has been detected in the CNS of tailbud and
confirmsthe importanceof BMP-2in dorsal–ventral pattern early somite stage zebrafish embryos (Chin et al., 1997; our
formation. unpublished observations), Chordin could function in pat-
Ectopic overexpression of either bmp-2 or bmp-4 in zebra- terning the brain by antagonizing other BMPs. For example,
fish embryos eliminates notochord (Nikaido et al., 1997), BMP-7 has been shown to influence both dorsal–ventral
suggestingthat in chordin mutants that develop trunk noto- polarity and growth in the mouse hindbrain at a comparable
chord, another BMP antagonist must be present to prevent stage in embryonic development (Arkell and Beddington,
total ventralization of the embryo. Support for the presence 1997). As in the early gastrula, binding to Chordin would
of additional factors that antagonize ventral signals comes be a rapid mechanism for limiting the range of BMP action
from the isolation of other zebrafish mutations that produce in the developing zebrafish brain.
ventralized phenotypes (Hammerschmidt et al., 1996b; Sol- Another distinctive feature of chordin expression in the
nica-Krezel et al., 1996; V. Miller-Bertoglio and M. E. Hal- zebrafish were the lateral domains that arise secondarily to
pern, unpublished observations). Noggin and Follistatin the dorsal embryonic shield or organizer expression. These
(Smith and Harland, 1992; Hemmati-Brivanlou et al., 1994) domains persisted throughout gastrulation and tailbud
are candidates for this function in the zebrafish gastrula. stages to the late somite stage, corresponding to a period
In addition to BMPs, other signaling pathways have been when elongation of the body axis and secondary neurulation
implicated in dorsal–ventral patterning of the zebrafish gas- occur (refer to Kimmel et al., 1995). Expression in seg-
trula, including those involving Wnt (Kelly et al., 1995a,b) menting embryos was highest in the caudal region of the
and FGF (Fu
¨rthauer et al., 1997) family members. At early embryo, where dorsally and ventrally derived cells were
gastrulation, fgf-8 is expressed in a gradient emanating from shown to exhibit different migratory behaviors and give rise
the dorsal side of the embryo and upon overexpression, fgf- to a different complement of tissue fates (Kanki and Ho,
8promotes development of dorsolateral mesoderm and neu- 1997). We hypothesize that the role of the lateral chordin
rectoderm in a manner similar to chordin, most likely expression domains is equivalent to that of the initial orga-
through the inhibition of bmp-2 and bmp-4 expression nizer expression, to function within a tail organizing center
(Fu
¨rthauer et al., 1997). This suggests that fgf-8 functions to maintain correct dorsal–ventral polarity of the extending
in a common pathway; however, it remains to be shown caudal trunk and tail. The lateral expression domains may
whether it mediates its effects on BMPs directly or indi- therefore reflect one of the key differences between zebra-
rectly through regulation of chordin or by other mecha- fish andXenopus gastrulation, namelythe migrationof cells
nisms. over a single yolk mass that occurs in teleost epiboly which
Identification of a number of loci that when mutated dor- results in the eventual fusion of the dorsal and ventral sides
salize zebrafish embryos also indicates that there are other of the blastoderm margin (Kimmel et al., 1995; Kanki and
genes involved in the pathway and reveals a role for mater- Ho, 1997). Indeed, chordin may fulfill an expanded role in
nal components (Mullins et al., 1996). In particular, the the zebrafish tailbud, since in contrast to the trunk,
observation of two types of abnormal chordin expression chordino (Hammerschmidt et al., 1996b) as well as other
patterns in embryos from sbn//intercrosses correlates well mutant alleles (S. Fisher, unpublished observations) some-
with the dominant maternal effect described for this muta-
tion (Mullins et al., 1996) and further supports the involve- times cause a loss of tail notochord.
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
548
Miller-Bertoglio et al.
Beddington, R. S. (1994). Induction of a second neural axis by the
Selective Regulation of chordin Expression
mouse node. Development 120, 613–620.
Through examination of chordin expression with respect
Biehs, B., Franc
¸ois, V., and Bier, E. (1996). The Drosophila short
to zebrafish mutations that perturb formation of specific
gastrulation gene prevents Dpp from autoactivating and sup-
dorsal or dorsolateral derivatives, we have begun to explore
pressing neurogenesis in the neurectoderm. Genes Dev. 10,
whether chordin functions in other aspects of tissue pat-
2922–2934.
Chin, A. J., Chen, J., and Weinberg, E. S. (1997). Bone morphogene-
terning. In flh mutants, dorsal axial mesoderm is believed
tic protein-4 characterizes inductive boundaries in organs of the
to assume paraxial mesodermal identity and muscle forms
developing zebrafish. Dev. Genes Evol. 207, 107–114.
in the midline in place of notochord (Halpern et al., 1995;
Dosch, R., Gawantka, V., Delius, H., Blumenstock, C., and Niehrs,
Melby et al., 1996). Consistent with having an altered iden-
C. (1997). Bmp-4 acts as a morphogen in dorsoventral mesoderm
tity, the most posterior axial mesoderm in flh mutants ex-
patterning in Xenopus. Development 124, 2325–2334.
pressed chordin at a high level more comparable to adjacent
Dudley, A. T., Lyons, K. M., and Robertson, E.J. (1995). A require-
paraxial mesoderm, at a stage when expression is decreased
ment for bone morphogenetic protein-7 during development of
in the WT midline. In flh mutants, the prolonged and in-
the mammalian eye and kidney. Genes Dev. 9, 2795–2807.
tense chordin expression in the posterior midline may ac-
Fainsod, A., Steinbeisser, H., and De Robertis, E. M. (1994). On the
count for the patterning defects or loss of tissues that do
function of BMP-4 in patterning the marginal zone of the Xeno-
pus embryo. EMBO J. 13, 5015–5025.
not themselves express flh, such as the caudal blood vessels
Fisher, S., Amacher, S. L., and Halpern, M. E. (1997). Loss of
(Talbot et al., 1995).
cerebum function ventralizes the zebrafish embryo. Develop-
In spt mutants, cells that would normally form trunk
ment 124, 1301–1311.
somitic mesoderm are misspecified and accumulate in the
Franc
¸ois, V., and Bier, E. (1995). Xenopus chordin and Drosophila
tail (Kimmel et al., 1989; Ho and Kane, 1990). At gastrula-
short gastrulation genes encode homologous proteins function-
tion, spt mutants also fail to express myoD, a marker of
ing in dorsal–ventral axis formation. Cell 80, 19–20.
paraxial mesoderm (Weinberg et al., 1996). We examined
Franc
¸ois, V., Solloway, M., O’Neill, J. W., Emery, J., and Bier, E.
whether defects in the earlier expression of chordin in para-
(1994). Dorsal–ventral patterning of the Drosophila embryo de-
xial mesodermal domains that flank the gastrula midline
pends on a putative negative growth factor encoded by the short
contributed to the spt mutant phenotype. Expression of
gastrulation gene. Genes Dev. 8, 2602–2616.
Fritz, A., Rozowski, M., Walker, C., and Westerfield, M. (1996).
chordin was normal in the early spt embryo, but as gastrula-
Identification of selected gamma-ray induced deficiencies in ze-
tion proceeded, paraxial expression domains remained nor-
brafish using multiplex polymerase chain reaction. Genetics 144,
mal while expression in axial mesoderm was not main-
1735–1745.
tained as in WT. This result reveals an unexpected function
Fu
¨rthauer, M., Thisse, C., and Thisse, B. (1997). A role for FGF-8
in the gastrula midline for spt, a gene previously implicated
in the dorsoventral patterning of the zebrafish gastrula. Develop-
in development of trunk paraxial mesoderm. Maintenance
ment 124, 4253–4264.
of chordin expression in axial mesoderm may be a direct
Gont, L. K.., Steinbeisser, H., Blumberg, B., and De Robertis, E.M.
function of spt or, alternatively, may require signals pro-
(1993). Tail formation as a continuation of gastrulation: The mul-
vided by paraxial mesoderm which is lacking in spt mu-
tiple cell populations of the Xenopus tailbud derive from the late
tants. Analysis of chordin expression in mutants such as
blastopore lip. Development 119, 991–1004.
Graff, J. M., Thies, R. S., Song, J. J., Celeste, A. J., and Melton, D. A.
flh and spt illustrates the ways in which the pathway for
(1994). Studies with a Xenopus BMP receptor suggest that ventral
dorsal–ventral patterning at gastrulation could converge
mesoderm-inducing signals override dorsal signals in vivo. Cell
with genetic pathways for tissue specification.
79, 169–179.
Halpern, M. E., Ho, R. K., Walker, C., and Kimmel, C. B. (1993).
Induction of muscle pioneers and floor plate is distinguished by
ACKNOWLEDGMENTS
the zebrafish no tail mutation. Cell 75, 99–111.
Halpern, M. E., Thisse, C., Ho. R. K., Thisse, B., Riggleman, B.,
The authors thank Chen-Ming Fan for technical advice and help- Trevarrow, B., Weinberg, E. S., Postlethwait, J.H., and Kimmel,
ful comments on the manuscript and Lilianna Solnica-Krezel, C. B. (1995). Cell-autonomous shift from axial to paraxial meso-
Charline Walker, and Chuck Kimmel for their generosity in provid- dermal development in zebrafish floating head mutants. Devel-
ing mutant fish lines. We also thank Allison Pinder, Mike Sepanski, opment 121, 4257–4264.
and Christine Norman for expert technical assistance and Michelle Hammerschmidt, M., Serbedzija, G. N., and McMahon, A. P.
Macurak for maintenance of the fish facility. S.F. was supported (1996a). Genetic analysis of dorsoventral pattern formation in
by a NINDS Clinical Investigator Development Award (NS01850- zebrafish: Requirement of a BMP-like ventralizing activity and
02), M.C.M. by NIH RO1-GM56326, and M.E.H. by the Pew Schol- its dorsal repressor. Genes Dev. 10, 2452–2461.
ars Program in Biomedical Research. The GenBank Accession No. Hammerschmidt, M. Pelegri, F., Mullins, M. C., Kane, D. A., van
for zebrafish chordin is AF034606. Eeden, F.J. M., Granato, M., Brand, M., Furutani-Seiki, M.,
Hafter, P., Heisenberg, C., Jiand, Y., Kelsh, R. N., Odenthal, J.,
Warga, R. M., and Nu
¨sslein-Volhard, C. (1996b). dino and mer-
REFERENCES
cedes, two genes regulating dorsal development in the zebrafish
embryo. Development 123, 95–102.
Arkell, R., and Beddington, R. S. P. (1997). BMP-7 influences pat- Hatta, K., Kimmel, C. B., Ho, R. K., and Walker, C. (1991). The
tern and growth of the developing hindbrain of mouse embryos. cyclops mutation blocks specification of floor plate of the zebra-
fish central nervous system. Nature 350, 339–341.Development 124, 1–12.
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
549
Chordin Expression in Mutant Zebrafish
Hauptmann, G., and Gerster, T. (1994). Two-color whole mount in Melby, A. E., Warga, R. M., and Kimmel, C. B. (1996). Specification
of cell fate at the dorsal margin of the zebrafish gastrula. Develop-situ hybridizations on zebrafish and Drosophila embryos. TIG
10, 266. ment 122, 2225–2237.
Mullins, M. C., Hammerschmidt, M., Kane, D. A., Odenthal, J.,Hawley, S.H. B., Wu
¨nnenberg-Stapleton, K., Hashimoto, C.,
Laurent, M., Watabe, T., Blumberg, B. W., and Cho, K. W. Y. Brand, M., van Eeden, J. M., Furutani-Seiki, M., Granato, M.,
Haffter, P., Heisenberg, C., Jiang, Y., Kelsh, R. N.,and Nu
¨sslein-(1995). Disruption of BMP signals in embryonic Xenopus ecto-
derm leads to direct neural induction. Genes Dev. 9, 2923–2935. Volhard, C. (1996). Genes establishing dorsoventral pattern for-
mation in the zebrafish embryo: The ventral specifying genes.
Hemmati-Brivanlou, A., Kelly, O. G., and Melton, D. A. (1994). Fol- Development 123, 81–93.
listatin, an antagonist of activin, is expressed in the Spemann
organizer and displays direct neutralizing activity. Cell 77, 283– Murphy-Ullrich, J. E., Schultz-Cherry, S., and Ho
¨o
¨k, M. (1992).
Transforming growth factor-
b
complexes with thrombospondin.295. Mol. Biol. Cell 3, 181–188.
Ho, R. K. (1992). Axis formation in the embryo of the zebrafish
Brachydanio rerio. Semin. Dev. Biol. 3, 53–64. Neave, B., Holder, N., and Patient, R. (1997). A graded response
to BMP-4 spatially coordinates patterning of the mesoderm and
Ho, R. K., and Kane, D. A. (1990). Cell-autonomous action of zebra- ectoderm in the zebrafish. Mech. Dev. 62, 183–195.
fish spt-1 mutation in specific mesodermal precursors. Nature
348, 728–730. Nikaido, M., Tada, M., Saji, T., and Ueno, N. (1997). Conservation
of BMP signaling in zebrafish mesoderm patterning. Mech. Dev.
Hogan, B. L. M. (1996). Bone morphogenetic proteins in develop- 61, 75–88.
ment. Curr. Opin. Dev. Genet. 6, 432–438.
Holley, S. A., and Ferguson, E.L. (1997). Fish are like flies are like Onichtchouk, D., Gawantka, V., Dosch, R., Delius,H., Hirschfeld,
frogs: Conservation of dorsal–ventral patterning mechanisms. K., Blumenstock, C., and Niehrs, C. (1996). The Xvent-2 homeo-
BioEssays 19, 1–4. box gene is part of the BMP-4 signalling pathway controlling
dorsoventral patterning of Xenopus mesoderm. DevelopmentHolley, S. A., Jackson, P. D., Sasai, Y., Lu, B., De Robertis, E. M.,
Hoffman, F. M., and Ferguson, E.L. (1995). A conserved system 122, 3045–3052.
for dorsal–ventral patterning in insects and vertebrates involving Oppenheimer, J. M. (1936). Transplantation experiments on devel-
sog and chordin. Nature 376, 249–253. oping teleosts (Fundulus and Perca). J. Exp. Zool. 72, 409–437.
Jones, C. M., Lyons, K.M., Lapan, P. M., Wright, V. E., and Hogan, Oxtoby, E., and Jowett, T. (1993). Cloning of the zebrafish krox-20
B. L. M. (1992). DVR-4 (bone morphogenetic protein-4) as a poste- gene (krx-20) and its expression during hindbrain development.
rior-ventralizing factor in Xenopus mesoderm induction. Devel- Nucleic Acids Res. 21, 1087–1095.
opment 115, 639–647. Padgett, R., St. Johnston, R. D., and Gelbart, W. M. (1987). A tran-
Kanki, J. P., and Ho, R. K. (1997). Thedevelopment of the posterior script from a Drosophila pattern gene predicts a protein homolo-
body in zebrafish. Development 124, 881–893. gous to the transforming growth factor-beta family. Nature. 325,
81–84.Kelly, G. M., Greenstein, P., Erezyilmaz, D. F., and Moon, R. T.
(1995a). Zebrafish wnt8 and wnt8b share a common activity but Paralkar, V. M., Vukicevic, S., and Reddi, A. H. (1991). Trans-
are involved in distinct developmental pathways. Development forming growth factor
b
type 1 binds to collagen IV of basement
121, 1787–1799. membrane matrix: Implications for development. Dev. Biol. 143,
303–308.Kelly, G. M., Erezyilmaz, D. F., and Moon, R.T. (1995b). Induction
of a secondary embryonic axis in zebrafish occurs following the Piccolo, S., Sasai, Y., Lu, B., and De Robertis, E. M. (1996). Dorso-
overexpression of beta-catenin. Mech. Dev. 53, 261–273. ventral patterning in Xenopus: Inhibition of ventral signals by
direct binding of Chordin to BMP-4. Cell 86, 589–598.Kimmel, C. B., Kane, D. A., Walker, C., Warga, R.M., and Roth-
man, M. B. (1989). A mutation that changes cell movement and Pourquie
´, O., Fan, C.-M., Coltey, M., Hirsinger, E., Watanabe, Y.,
cell fate in the zebrafish embryo. Nature 337, 358–362. Bre
´ant, C., Francis-West, P., Brickell, P., Tessier-Lavigne, M., and
Le Douarin, N. M. (1996). Lateral and axial signals involved inKimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullman, B., and Schil-
ling, T. F. (1995). Stages of embryonic development of the zebra- avian somite patterning: A role for BMP4. Cell 84, 461–471.
fish. Dev. Dyn. 203, 253–310. Reissmann, E., Ernsberger, U., Francis-West, P. H., Rueger, D.,
Brickell, P. M., and Rohrer, H. (1996). Involvement of bone mor-Kintner, C. R., and Dodd, J. (1991). Henson’s node induces neural
tissue in Xenopus ectoderm. Implications for the action of the phogenetic protein-4 and bone morphogenetic protein-7 in the
differentiation of the adrenergic phenotype in developing sympa-organizer in neural induction. Development 113, 1495–1505. thetic neurons. Development 122, 2079–2088.
Krauss, S., Johansen, T., Korzh, V., and Fjose, A. (1991). Expression
patternof zebrafish pax genes suggests a rolein early brain region- Sasai, Y., Lu, B., Steinbeisser, H., and De Robertis, E.M. (1995).
Regulation of neural induction by the Chd and BMP-4 antagonis-alization. Nature 353, 267–270. tic patterning signals in Xenopus. Nature 376, 333–336.
Lemaire, P., and Kodjabachian, L. (1996). The vertebrate organizer:
Structure and molecules. TIG 12, 525–531. Sasai, Y., Lu, B., Steinbeisser, H., Geissert, D., Gont, L. K., and
De Robertis, E. M. (1994). Xenopus chordin: A novel dorsalizing
Luo, G., Hofmann, C., Bronckers, A. L. J. J., Sohocki, M., Bradley, factor activated by organizer-specific homeobox genes. Cell 79,
A., and Karsenty, G. (1995). BMP-7 is an inducer of nephrogen- 779–790.
esis, and is also required for eye development and skeletal pat-
terning. Genes Dev. 9, 2808–2820. Schmidt, J. E., Suzuki, A., Ueno, N., and Kimelman, D. (1995a).
Localized BMP-4 mediates dorsal/ventral patterning in the early
Macias, D., Gana, Y., Sampath, T. K., Piedra, M. E., Ros, M. A., and Xenopus embryo. Dev. Biol. 169, 37–50.
Hurie, J. M. (1997). Role of BMP-2 and OP-1 (BMP-7) in pro-
grammed cell death and skeletogenesis during chick limb devel- Schmidt, J. E., Franc
¸ois, V., Bier, E., and Kimelman, D. (1995b).
Drosophila short gastrulation induces an ectopic axis in Xeno-opment. Development 124, 1109–1117. pus: Evidence for conserved mechanisms of dorsal–ventral pat-
Melby, A. E., Kimelman, D., and Kimmel, C. B. (1997). Spatial regu- terning. Development 121, 4319–4328.
lation of floating head in the developing notochord. Dev. Dyn.
209, 156–165. Schmidt, J. E., von Dassow, G., and Kimelman, D. (1996). Regula-
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
550
Miller-Bertoglio et al.
tion of dorsal–ventral patterning: The ventralizing effects of the D. (1995). A homeobox gene essential for zebrafish notochord
development. Nature 378, 150–157.
novel Xenopus homeobox gene Vox. Development 122, 1711– Thisse, C., Thisse, B., Halpern, M. E., and Postlethwait, J. H. (1994).
1721. Goosecoid expression in neurectoderm and mesendoderm is dis-
Schulte-Merker, S., Ho, R. K., Herrmann, B.G., and Nu
¨sslein-Vol- rupted in zebrafish cyclops gastrulas. Dev. Biol. 164, 420–429.
hard, C. (1992). The protein product of the zebrafish homologue Thisse, C., Thisse, B., Schilling, T. F., and Postlethwait, J. H. (1993).
of the mouse Tgene is expressed in nuclei of the germ ring Structure of the zebrafish snail1 gene and its expression in wild-
and the notochord of the early embryo. Development 116, 1021– type, spadetail and no tail mutant embryos. Development 119,
1032. 1203–1215.
Schulte-Merker, S., Hammerschmidt, M., Beuchle, D., Cho, K.W., Thomsen, G. H. (1997). Antagonism within and around the orga-
De Robertis. E. M., and Nu
¨sslein-Volhard, C. (1994). Expression nizer: BMP inhibitors in vertebrate body patterning. Trends
of zebrafish goosecoid and no tail gene products in wild-type and Genet. 13, 209–211.
mutant no tail embryos. Development 120, 843–852. Tidman Ault, K., Dirksen, M.-L., and Jamrich, M. (1996). A novel
Schulte-Merker, S., Lee, K. J., McMahon, A. P., and Ham- homeobox gene PV.1 mediates induction of ventral mesoderm
merschmidt, M. (1997). The zebrafish organizer requires in Xenopus embryos. Proc. Natl. Acad. Sci. USA 93, 6415–6420.
chordino. Nature 387, 862–863. Toyama, R., O’Connell, M. L., Wright, C. V., Kuehn, M. R., and
Schultheiss, T. M., Burch, J. B., and Sassar, A.B. (1997). A role for Dawid, I. B. (1995). Nodal induces ectopic goosecoid and lim1
bone morphogenetic proteins in the induction of cardiac myogen- expression and axis duplication in zebrafish. Development 121,
esis. Genes Dev. 11, 451–462. 383–391.
Shih, J., and Fraser, S. E. (1996). Characterizing the zebrafish orga- Turner, D. L., and Weintraub, H. (1994). Expression of acaete –scute
nizer: Microsurgical analysis at the early-shield stage. Develop- homolog 3 in Xenopus embryos converts ectodermal cells to neu-
ment 122, 1313–1322. ral fate. Genes Dev. 8, 1434–1447.
Smith, W. C., and Harland, R. M. (1992). Expression cloning of nog- Weinberg, E. S., Allende, M. L., Kelly, C. S., Abdelhamid, A., Mura-
gin, a new dorsalizing factor localized to the Spemann organizer kami, T., Andermann, P., Doerre, O. G., Grunwald, D. J., and
in Xenopus embryos. Cell 70, 829–840. Riggleman, B. (1996). Developmental regulation of zebrafish
Solnica-Krezel, L., Stemple, D., Mountcastle-Shah, E., Rangini, Z., MyoD in wild-type, no tail and spadetail embryos. Development
122, 217–280.
Neuhauss, S. C. F., Malicki, J., Shier, A. F., Stainier, D. Y. R., Westerfield, M. (Ed.) (1993). ‘‘The Zebrafish Book.’’ Univ. Oregon
Zwartkruis, F., Abdelilah, S., and Driever, W. (1996). Mutations Press, Eugene, OR.
affecting cell fates and cellular rearrangements during gastrula-
tion in zebrafish. Development 123, 67–80. Wilson, P. A., Lagna, G., Suzuki, A., and Hemmati-Brivanlou, A.
Spemann, H., and Mangold, H. (1924). Uber induktion von embryo- (1997). Concentration-dependent patterning of the Xenopus ecto-
nalanlagen du
¨rch implantation artfremder organisatoren. Roux’s derm by BMP4 and its signal transducer Smad1. Development
Arch. Entwmech. Org. 100, 599–683. 124, 3177–3184.
Stachel, S. E., Grunwald, D. J., and Meyers, P. Z. (1993). Lithium Winnier, G., Blessing, M., Labosky, P. A., and Hogan, B. L. M.
perturbation and goosecoid expression identify a dorsal specifi- (1995). Bone morphogenetic protein-4 is required for mesoderm
cation pathway in the pregastrula zebrafish. Development 117, formation and patterning in the mouse. Genes Dev. 9, 2105–
1261–1274. 2116.
Steinbeisser, H., Fainsod, A., Niehrs, C., Sasai, Y., and De Robertis, Yan, Y. L., Hatta, K., Riggleman, B., and Postlethwait, J.H. (1995).
E. M. (1995). The role of gsc and BMP-4 in dorsal–ventral pat- Expression of a type II collagen gene in the zebrafish embryonic
terning of the marginal zone in Xenopus: A loss of function study axis. Dev. Dyn. 203, 363–376.
using antisense RNA. EMBO J. 14, 5230–5243. Yokouchi, Y., Sakiyama, J., Kameda, T., Iba, H., Suzuki, A., Ueno,
Storey, K., Crossley, J. M., De Robertis, E. M., Norris, W.E., and N., and Kuroiwa, A. (1996). BMP-2/4 mediate programmed cell
Stern, C. D. (1992). Neural induction and regionalisation in the death in chicken limb buds. Development 122, 3725–3734.
chick embryo. Development 114, 729–741. Zhang, H., and Bradley, A. (1996). Mice deficient for BMP2 are
Storm, E. E., and Kingsley, D. M. (1996). Joint patterning defects nonviable and have defects in amnion/chorion and cardiac devel-
caused by single and double mutations in members of the bone opment. Development 122, 2977–2986.
Zimmerman, L. B., De Jesu
´s-Escobar, J. M., and Harland, R. M.morphogenetic protein (BMP) family. Development 122, 3969– (1996). The Spemann organizer signal noggin binds and inacti-
3979. vates bone morphogenetic protein 4. Cell 86, 599–606.
Stra
¨hle, U., Blader, P., Henrique, D., and Ingham, P. W. (1993). Zou, H., and Niswander, L. (1996). Requirement for BMP signaling
Axial, a zebrafish gene expressed along the developing body axis, in interdigital apoptosis and scale formation. Science 272, 738–
shows altered expression in cyclops mutant embryos. Genes Dev. 741.
7, 1436–1446.
Talbot, W. S., Trevarrow, B., Halpern, M. E., Melby, A. E., Farr, G., Received for publication June 3, 1997
Postlethwait, J. H., Jowett, T., Kimmel, C. B., and Kimelman, Accepted October 17, 1997
Copyright q1997 by Academic Press. All rights of reproduction in any form reserved.
AID DB 8788 / 6x35h$$$61 12-15-97 04:49:58 dbal
