DEVELOPMENTAL DYNAMICS 213:199–206 (1998)
Inhibitory Action of BMPs on Pax1 Expression and on
Shoulder Girdle Formation During Limb Development
CLEMENTINE HOFMANN,1* GARYFALIA DROSSOPOULOU,2 ANDREW McMAHON,3
RUDI BALLING,1 AND CHERYLL TICKLE2
1GSF-Forschungszentrum für Umwelt und Gesundheit, Institut für Säugetiergenetik, Neuherberg, Germany
2Department of Anatomy and Developmental Biology, University College and Middlesex School of Medicine, London, England
3Department of Molecular and Cellular Biology, The Biolabs, Harvard University, Cambridge, Massachusetts
ABSTRACT Pax1 expression in vertebrate
limb buds is confined to cells in a discrete ante-
rior proximal domain (Timmons et al. [1994] De-
velopment 120:2773–2785; Ebensperger et al. [1995]
Anat. Embryol. 191:297–310). In dorsoventral pat-
terning of Drosophila, expression of pox meso, an
insect gene with high sequence similarity to Pax1,
is repressed by decapentaplegic (dpp) in dorsal
mesoderm and, thus, is restricted to a discrete
ventral domain (Staehling-Hampton et al. [1994]
Nature 372:783–786). In the chick wing, cells ex-
pressing a vertebrate homolog of dpp, bone mor-
phogenetic protein 4 (Bmp4), abut the Pax1 do-
main, suggesting a similar relationship between
homologous genes in both vertebrates and inver-
tebrates. Here, we show that two BMPs (BMP4,
and BMP2, also highly related to dpp) can repress
Pax1 in the developing chick wing. Chick wing
bud cells expressing Pax1 give rise to the shoul-
der girdle. Cells in an equivalent position in the
mouse forelimb also express Pax1, and Pax1 mu-
tant mice display shoulder girdle defects. Simi-
larly in chick embryos, girdle defects are pro-
duced by treatments with signalling molecules
that lead to expression of BMPs, which subse-
quently reduce Pax1 expression in the limb bud.
Recently, BMP4 has been shown to inhibit Pax1
expression in the developing trunk (Monsoro-
Burq et al. [1996] Development 122:3607–3616)
and Pax9 expression in developing teeth (Neu-
büser et al. [1997] Cell 90:247–255). Thus, a prop-
erty of BMPs appears to be to regulate pox meso
homologs negatively and, thus, limit their expres-
sion domains. Dev. Dyn. 1998;213:199–206.
r 1998 Wiley-Liss, Inc.
Key words: chick embryo; vertebrate limb; shoul-
der girdle; bone morphogenetic pro-
teins
INTRODUCTION
Pax genes are known to be involved in organogenesis
and early differentiation programs in embryogenesis
(for review see Dahl et al., 1997; Gruss and Walther,
1992; Noll, 1993). Pax1 is expressed in the very anterior
and proximal part of the early vertebrate limb bud
r 1998 WILEY-LISS, INC.
(Ebensperger et al., 1995; Timmons et al., 1994). Fate
map studies show that this region of wing mesenchyme
contributes to the shoulder girdle (Bowen et al., 1989;
Saunders, 1948; Vargesson et al., 1997). Furthermore,
mice which have defects in the Pax1 locus show shoul-
der girdle malformations (Balling, 1994; Dietrich and
Gruss, 1995; Timmons et al., 1994; Wilm et al., unpub-
lished).
In Drosophila, decapentaplegic (dpp) activity is re-
quired in the dorsal part of the embryo for repression of
genes whose expression is normally confined to ventral
mesoderm (e.g., pox meso), and is required for activa-
tion of dorsally restricted mesodermal genes, like tin-
man and bagpipe (Staehling-Hampton et al., 1994). dpp
is highly related to the vertebrate bone morphogenetic
proteins Bmp2 and -4, and pox meso has high sequence
similarity to vertebrate Pax1 (Bopp et al., 1989;
Deutsch et al., 1988; Noll, 1993). An inhibitory action of
BMP4 on Pax1 and Pax3 expression has been shown
recently when BMP4-expressing cells were implanted
into dorsal axial structures in chick (Monsoro-Burq et
al., 1996).
BMPs were originally identified by their ability to
induce formation of ectopic bone when applied subcuta-
neously in rodents (Sampath and Reddi, 1981; Urist,
1965; Wozney et al., 1988), but many studies implicate
them and their relatives in the regulation of patterning
processes and organogenesis in both vertebrates (for
review see Hogan, 1996) and invertebrates (Ferguson
and Anderson, 1992; Gelbart, 1989; Lecuit et al., 1996;
Nellen et al., 1996).
BMPs are expressed in developing limbs at various
stages throughout their development. In early limb
buds, Bmp2 is expressed in posterior mesenchyme
while Bmp4 is expressed, in addition, extensively in
anterior mesenchyme. Bmp7 is expressed weakly in
anterior and also, more strongly, in posterior mesen-
chyme. All three BMPs are expressed throughout the
apical ridge (Francis et al., 1994; Francis-West et al.,
Grant sponsor: National Institutes of Health; Grant number:
NS33642.
*Correspondence to: Clementine Hofmann, GSF-Forschungszen-
trum für Umwelt und Gesundheit, Institut für Säugetiergenetik,
Ingolstädter Landstr. 1 85764 Neuherberg, Germany.
E-mail: chofmann@gsf.de
Received 14 April 1998; Accepted 19 June 1998
200 HOFMANN ET AL.
Fig. 1. Comparison of relative expression domains of bone morphoge-
netic proteins (Bmps) and Pax1 in stage 22 forelimb buds as detected by
whole mount in situ hybridization. A: Bmp2 is expressed in posterior
mesenchyme and along the apical ectodermal ridge. B: Bmp4 is ex-
pressed in the anterior mesenchyme adjacent to Pax1 domain (compare
with D), and is also present in posterior mesenchyme and apical
1995; and own results). We tested the possibility that
BMP signalling restricts Pax1 gene expression in devel-
oping limb buds. We found that application of BMPs
reduces expression of Pax1 in anterior mesenchyme
and leads to shoulder girdle defects. Sonic hedgehog
(SHH) is capable of inducing Bmp expression in the
limb bud (Laufer et al., 1994) when applied to the
anterior limb mesenchyme, and this then leads to
reduction of Pax1 expression and concomitant shoulder
girdle defects.
RESULTS
Normal Expression of Bmps and Pax1
in the Limb
In order to delineate relative expression domains of
Pax1 and Bmp genes, and to determine whether Bmp2
and Bmp4 are expressed in a pattern consistent with a
potential function of restricting Pax1 to a specific region
within the bud, we analyzed expression patterns of
Bmp2, -4, and -7 and Pax1 in early chick limb buds by in
situ hybridization. At stage 22/23 (Hamburger and
Hamilton, 1951), all three Bmps are expressed through-
out the apical ectodermal ridge (AER). In the mesen-
chyme, Bmp2 is expressed posteriorly (Fig. 1A); Bmp7
is present posteriorly in the bud, and to a much lesser
extent, anteriorly at the tip of the bud (Fig. 1C). Bmp4
is expressed in posterior mesenchyme, and its mesenchy-
ectodermal ridge (AER). In (C), Bmp7 expression domains are mainly in
posterior mesenchyme, to a lower extent in anterior mesenchyme, and in
the AER. D shows the expression of Pax1 located deep in the anterior and
proximal limb mesenchyme and adjacent to the Bmp4 expression do-
main. Specimens are photographed from dorsal, posterior is to the left,
anterior to the right. Scale bar ⫽ 200 µm.
mal domain also extends to the anterior bud margin
(Fig. 1B). Pax1, whose expression in the limb is estab-
lished at stage 22/23, is expressed in the mesenchyme
at the very anterior and proximal region of the limb bud
and appears to abut Bmp4 expression in the mesen-
chyme (Fig. 1D). Although it should be kept in mind
that we are looking at Bmp transcripts and not BMP
proteins, the expression patterns suggest that BMP4
signalling by anterior mesenchyme cells and/or BMP2,
-4, and-7 signalling by the apical ectodermal ridge could
restrict Pax1 expression to a proximal part of the bud.
BMPs Inhibit Expression of Pax1 in Anterior
Limb Mesenchyme
To find out whether BMPs can regulate Pax1 expres-
sion in the wing bud, we implanted beads soaked in
BMP4 and/or BMP2 at anterior and proximal positions
in the mesenchyme of wing buds at stages 19–24.
Implantation of beads soaked in BMP4 (0.33 mg/ml
soaking concentration, n ⫽ 9) at the anterior margin of
chick wing buds at these different developmental stages
leads to a marked reduction in or even complete loss of
Pax1 expression (see Fig. 2A, compare with Fig. 1D).
With beads soaked in bovine serum albumin/phosphate-
buffered saline (BSA/PBS) as controls (n ⫽ 6), Pax1
expression is not affected (Fig. 2B). Implantation of
beads soaked in BMP2 (1 mg/ml [n⫽ 24 total], and 0.33
 BMPs INHIBIT Pax1 EXPRESSION AND SHOULDER FORMATION
mg/ml [n⫽ 6] soaking concentration) also markedly
reduced Pax1 expression, regardless of whether the
soaking concentration of BMP2 was high or low. Figure
2C shows an example in which BMP2 application was
201
carried out at stage 20 and Pax1 expression has com-
pletely disappeared (compare with Fig. 2D showing the
contralateral untreated bud). Figure 2E is another
example in which treatment was carried out at stage 20
and Pax1 is strongly reduced, with only a small and
weak expression domain still detectable. When bead
implants were performed at stage 21/22, Pax1 is still
strongly reduced but never completely abolished (see
below).
The precise position in which BMPs are applied does
not appear to be important. When BMP2 beads were
placed more distally under the ridge in stage 19–22
(n ⫽ 7) in order to mimick BMP signalling by the AER,
Pax1 expression was still reduced or absent (Fig. 2I).
Upon application of both BMP2 and BMP4 (2 beads
added, each soaked in concentration of 0.33 mg/ml;
n ⫽ 2), Pax1 expression was again very reduced (Fig.
2F). In the older case (when treatment was performed
at stage 21/22) a small domain of faint staining is still
left, suggesting that there is no synergistic effect when
two beads soaked with different BMPs are applied.
Synergism may occur when BMPs are coexpressed
allowing heterodimer formation.
We observed a whitish area close to the bead in most
BMP4-treated specimens, but this was not so promi-
nent with BMP2. This area was not an even circular
patch centered around the bead, but instead extended
distally along the anterior margin of the bud under the
ridge with a straight distal edge slanting at an angle
towards the proximal part of the bud. Neither BSA/PBS
buffer- soaked beads (this work) nor beads soaked in
arginine-histidine dilution buffer, induced a similar
white area (Francis-West and Tickle, unpublished obser-
vations). This area could represent cell death, which
has been shown to be triggered by BMP-soaked beads
(Macias et al., 1997).
In order to exclude the possibility that gene activity is
therefore prevented or reduced in general by BMP
application, whole-mount in situ hybridization with
Msx-1 was performed on treated embryos (n ⫽ 5) and on
two embryos implanted with BSA/PBS-bead (data not
shown). Compared to contralateral sides and BSA/PBS
Fig. 2. Bone morphogenetic protein (BMP)4 and -2 inhibit Pax1
expression when applied to anterior-proximal limb mesenchyme. Distribu-
tion of Pax1 transcripts in wing buds implanted with beads soaked in
BMP4 (A), BMP2 (C, E, G, I) or BMP4 and -2 together (F). A heparin bead
soaked in 0.33 mg/ml BMP4 (A), or in 1 mg/ml BMP2 (C, E, G) was placed
at the anterior mesenchymal region of the right wing bud 24 hours before
fixation. B represents a control in which a bead soaked in phosphate-
buffered saline (PBS)/0.1% bovine serum albumin (BSA) was applied.
D shows the contralateral, untreated left bud of (C); in H the contralateral
side of G is shown. In I a bead soaked in 1 mg/ml BMP2 is stapled
underneath the apical ectodermal ridge (AER). In F a case is shown
where a BMP2 (0.33 mg/ml) and a BMP4 (0.33 mg/ml) bead were applied.
Arrows mark bead implantation sites and retained or decreased Pax1
expression domains. Specimens are photographed from dorsolateral,
posterior is to the left, anterior to the right (in D anterior is to the left,
posterior to the right), except in G and H where specimens are photo-
graphed from ventral, anterior is up, posterior to the bottom. Scale bars ⫽
500 µm.
202 HOFMANN ET AL.
TABLE 1. Shoulder Girdle Malformations With Bone
Morphogenic Protein (BMP) Beads
Implantation at stage 20
1 mg/ml BMP-2 8/8 Proximal end of coracoid
and scapula affected
0.333 mg/ml BMP-2 1/1 Break in scapula
0.337 mg/ml BMP-4 1/1 Both coracoid and scapula
affected
0.333 mg/ml BMP-2 1/1 Break in scapula
0.337 mg/ml BMP-4 1/1
BSA/PBSa-control 0/2 No shoulder defects
Implantation at later
stages
1 mg/ml BMP-2
Stage 23 1/1 Proximal parts of scapula
and coracoid missing
Stage 24/25 0/1 No shoulder defects
aBSA/PBS, bovine serum albumin/phosphate-buffered saline.
control embryo, no difference in Msx-1 expression was
observed.
In all the cases (described above) when bead implan-
tations were performed at stages 19–22 (stage 19/20,
n ⫽ 4; stage 20/21, n ⫽ 16; stage 21/22, n ⫽ 13), Pax1
expression is strongly reduced or disappears com-
pletely. However, when implantations were carried out
in animals of stages 23–25 in development (n ⫽ 8), we
observed that BMPs exhibit a much weaker effect on
Pax1 expression (Fig. 2G, compare with the contralat-
eral side in 2H). At this stage, we implanted BMP beads
as in the other cases quite proximal, i.e., close to cells
which were already expressing Pax1 in the proximal-
anterior part of the limb. This decreased influence of
BMPs on Pax1 expression suggests that the effect of
BMPs on Pax1 has to be before or as early as the onset
of Pax1 expression (at approximately stage 21/22) and
is less effective if Pax1 expressing domain is already
established.
Shoulder Girdle Defects
Since Pax1 deficient mice display shoulder girdle
defects, we investigated whether reduction of Pax1
expression, which follows application of BMP2 and -4
protein, is accompanied by shoulder girdle pattern
defects. Table 1 lists the results and shows that, in all
cases, when BMP beads were implanted at the anterior
margin of early wing buds, shoulder defects were seen
in the wings (11/11 cases). The malformations were in
the joint region where scapula and coracoid connect to
the humerus, and usually both coracoid and scapula
showed defects. In the most severe cases, the coracoid
was absent, and the proximal part of the scapula was
also affected (Fig. 3A), while, in the least affected cases,
only the proximal part of the scapula was misshapen
(Fig. 3B). Shoulder girdle defects were sometimes asso-
ciated with loss of anterior limb skeletal elements,
radius and/or digit 2, e.g., in Figure 3C. A scapula defect
was also induced with a lower concentration of BMP2
(0.33 mg/ml), illustrated in Fig. 3D; and with beads
soaked in 0.1 mg/ml BMP2, 2 out of 10 wings had
defective shoulder joints and in 7 others the shape of
the coracoid was altered (data not shown). Control
beads never lead to shoulder girdle defects.
Signals That Activate BMPs Lead to Shoulder
Girdle Abnormalities
In previous experiments, we showed that retinoic
acid application to the anterior margin of wing buds at
stage 18 leads to changes in shoulder girdle pattern
(Oliver et al., 1990). Since Bmp2 can be activated by
retinoic acid (RA; Francis et al., 1994), it seems likely
that the shoulder defects are caused by RA-induced
BMP production which in turn reduces Pax1 expres-
sion. Sonic hedgehog can also induce BMPs in anterior
limb mesenchyme (Laufer et al., 1994; Yang et al., 1997)
and therefore, we examined shoulder girdle structures
in wings which developed after application of SHH
protein (Table 2).
After implantation of beads soaked in high concentra-
tions of SHH protein (16 mg/ml) at the anterior margin
of stage 20 wing buds, 30% of wings had girdle defects;
the coracoid tended to be thin proximally. With beads
soaked in lower concentrations of SHH (1–2 mg/ml), the
coracoid can be thicker than normal or even branched.
It should be noted that these coracoid duplications are
found at concentrations of SHH that usually induce an
additional digit 3 (see Yang et al., 1997), but can occur
in wings without any extra digits.
Furthermore, when we applied high doses of SHH (16
mg/ml) at stage 17/18, as shown in Figure 4, the Pax1
domain of expression is reduced in anterior limb mesen-
chyme.
DISCUSSION
Pax1 is expressed in a distinct domain in the anterior
proximal part of the limb bud. Distally the Pax1 domain
abuts mesenchyme cells expressing Bmp4; Bmp2, Bmp4,
and Bmp7 are all expressed throughout the apical
ridge. When BMPs are applied to the anterior margin of
the wing bud, Pax1 expression is reduced and shoulder
girdle defects occur. Shoulder girdle defects and reduc-
tion in Pax1 expression are also seen following applica-
tion of SHH, which also induces Bmp expression. The
observation that Pax1 is not only a key player in the
formation of sclerotome and vertebral column but also
shoulder girdle elements, has been made through the
study of undulated mice and has been reported by
several authors (Dietrich and Gruss, 1995; Grüneberg,
1963; Timmons et al., 1994). Pax1 is expressed at the
very anterior part of the vertebrate limb bud. Cells in
this position have been shown to give rise to the
shoulder girdle by fate map studies in chick (Bowen et
al., 1989; Saunders, 1948; Vargesson et al., 1997) and
by ␤-galactosidase staining in Pax1-deficient mice, car-
rying the bacterial lacZ reporter gene under the control
of the endogenous Pax1 regulatory elements (Dahl et
al., unpublished). Here we have shown that Pax1, a
vertebrate relative of pox meso (Noll, 1993), is sup-
pressed by ectopically applied BMPs. Similarly, in
 BMPs INHIBIT Pax1 EXPRESSION AND SHOULDER FORMATION 203

Fig. 3. Shoulder girdle defects in bone morphogenetic protein (BMP)-
treated chick limbs. A heparin bead soaked in 1 mg/ml BMP2 (A, B), 0.33
mg/ml BMP4 (C), and two beads (D), one soaked in BMP2 and one in
BMP4 (both in 0.33 mg/ml soaking concentration) were placed at the
anterior mesenchymal region of the right wing buds 7 days before fixation
and skeletal preparation. Shoulder girdles have defects in all cases
examined in the region of the joint which connects humerus to scapula
and coracoid as compared to the control left-hand wing buds. Frequently,
parts of scapula and coracoid or the whole coracoid are missing and the
humerus is thickened. Note that in some cases more distal elements are
also affected, such as radius (in A and C) or digit 2 (C). c, coracoid; h,
humerus; r, radius; s, scapula; u, ulna, 2,3,4, digits; arrowhead indicates
missing digit 2, arrows indicate missing radius.

TABLE 2. Shoulder Girdle Defects With Sonic dorsalized. In wildtype Drosophila embryos, the inhibi-
Hedgehog (SHH) Beads Placed Anteriorly Into Wing tory effect of dpp restricts pox meso to its normal region
Buds at Stage 20
of ventral expression (Staehling-Hampton et al., 1994).
Affected Our studies suggest that Bmps might restrict expres-
Soaking embryos/ sion of Pax1 to the anteriormost region of the vertebrate
concentration total limb, but it is not clear whether this is mediated by the
(mg/ml) number Defects
mesenchymal domain of Bmp4 or by epithelial signal-
16 4/13 Mainly thin or reduced coracoid at ling of several Bmps.
most proximal joint
5 1/8 Clear abnormality in coracoid thin When Pax1 expression is reduced by ectopic applica-
proximally tion of signalling molecules in embryonic chick wings,
2 4/10 Defects still appear in coracoid, but morphological defects in the shoulder girdle were ob-
it tends to be thick or even served which parallel the defects seen in Pax1 mutant
branched proximally
1 7/32 Defects in proximal coracoid,
mice. In the mouse mutant, the clavicle and the acro-
branched or fused with scapula mion (part of the scapula), which together with the
0.75 0/9 No shoulder abnormalities noticed head of the humerus form the shoulder joint, are
0.5 2/7 Branched and thick coracoid reduced or abnormally shaped. In the chick, similar
0.1 1/20 Slight thickening on coracoid proxi- malformations are seen at the shoulder joint, which in
mally
0.01 0/3 No shoulder abnormalities this species comprises the scapula and coracoid ele-
ments at the head of the humerus. In Pax1 mutant
mice, the acromion is most severely affected and the
Drosophila, ectopically expressed dpp in ventral meso- clavicle only slightly (Dietrich and Gruss, 1995; Grüne-
derm or adjacent ectoderm, is sufficient to abolish pox berg, 1963; Timmons et al., 1994). In the chick, both
meso expression in ventral mesoderm, and as a conse- coracoid and scapula are affected in most cases, while
quence of such dpp overexpression, the epidermis is sometimes the scapula shows defects but the coracoid is
204 HOFMANN ET AL.
Fig. 4. SHH represses Pax1 expression in the limb. An Affigel bead
soaked in sonic hedgehog (SHH)-protein (16 mg/ml) was implanted to the
anterior margin of the forelimb anlage of a stage 17 chick embryo 48 hours
before fixation. A: In situ hybridization with Pax1 probe shows a decrease
in expression in the anterior forelimb. Posterior is to the left, anterior to the
normal. Thus, these malformations in chick embryos
may be slightly more severe than in Pax1 mouse
mutants.
However, in both cases, chick and mouse, the region
which directly connects to the humerus is affected, and
it is striking that both functional inactivation of Pax1 in
transgenic mice and signalling molecule-mediated re-
ductions of Pax1 expression in chick embryos lead to
similar shoulder girdle abnormalities.
Chick limbs treated with BMPs also sometimes lack
distal elements, such as digit 2 or radius, suggesting
that BMPs may have other effects, and other genes
could be affected. In a recent report, BMPs were shown
to cause apoptosis in limb buds (Yokouchi et al., 1996),
but we found that expression of Msx-1 is still apparent.
It is not clear in any system, including Drosophila,
whether BMPs restrict Pax expression by inducing cell
death. In addition, it is not known what happens to the
prospective shoulder girdle cells in the Pax1 mutant.
Recently it was shown that, in PDGF␣ receptor-
deficient mice, the Pax1 expression domain in the limbs
is missing and, concomitantly, shoulder girdle defects
occur comparable to those in Pax1 mutant mice (Sori-
ano, 1997). Thus, the question arises whether platelet
derived growth factors (PDGFs) or other growth fac-
tors, like fibroblast growth factors (FGFs), are potential
inducers of Pax1. It has recently been reported that
FGF8 induces Pax9, another Pax gene which is related
to pox meso, in mandibular mesenchyme, and that this
induction is inhibited by BMPs (Neubüser et al., 1997).
Our data show that BMP and SHH application at the
anterior region of the wing bud lead to shoulder girdle
defects. Previously, we found that retinoic acid applica-
tion produces girdle defects (Oliver et al., 1990). A likely
mechanism explaining shoulder girdle defects observed
in SHH-treated animals is that SHH induces ectopic
expression of Bmp-2 (Laufer et al., 1994) and, in turn,
this inhibits Pax1 expression. A similar explanation
could also be advanced for retinoic acid which has been
shown to activate Bmp expression (Francis et al., 1994).
right. B: In the contralateral control of the same embryo Pax1 transcripts
are localized in a round shaped, distinct domain. Anterior is to the left,
posterior to the right. Arrows mark Pax1 expression domains in forelimbs.
Specimens are photographed from lateral. Scale bar ⫽ 500 µm.
Retinoic acid also induces Shh (Riddle et al., 1993),
which in turn could activate Bmp expression. Thus, the
common basis for production of the shoulder girdle
defects appears to be a BMP-mediated reduction in
Pax1 expression. In normal limb development, SHH
induces Bmp-2 expression posteriorly in the limb and
Bmp-4 , which is expressed in anterior mesenchyme
independently of SHH, could act to negatively regulate
Pax1. We therefore propose that ectopic SHH represses
Pax1 expression by inducing BMP2 (its normal target)
which would enhance the activity of the highly related
BMP4. It should be noted that in the trunk, SHH
activates Pax1 in early paraxial mesoderm in vivo
(Johnson et al., 1994) and in culture (Fan et al., 1995;
Fan and Tessier-Lavigne, 1994). In SHH-deficient mice,
Pax1 expression is decreased in the sclerotomes, but
expression in limbs is fully retained (Chiang et al.,
1996) showing that SHH does not act as a negative
regulator of Pax1 in the normal limb.
EXPERIMENTAL PROCEDURES
Chick Embryos
Fertilized chicken eggs were obtained from the Geflue-
gelklinik Oberschleissheim, Germany, or from Poyndon
Farm, Waltham Cross, Herts, UK, and were incubated
in a humidified atmosphere at 38°C. Embryos were
staged according to Hamburger and Hamilton (1951),
then used for grafting experiments as described below
or dissected into sterile phosphate-buffered saline (PBS:
136.8 mM NaCl, 2.7 mM KCl, 1.5 mM KH2PO4, 10.2
mM Na2HPO4, pH 7.4) and processed for in situ hybrid-
ization.
Generation of Probes and In Situ Hybridization
Digoxigenin-labelled RNA probes for in situ hybridiza-
tion to whole-mount preparations were synthesized as
previously described (Hofmann et al., 1996; Wilkinson,
1993). Antisense RNA probes specific for Bmp2 and 4
transcripts were synthesized as described in Francis et
al. (1994). Chick Bmp7 was provided by B. Houston (see
 BMPs INHIBIT Pax1 EXPRESSION AND SHOULDER FORMATION
Houston et al., 1994). Pax1 specific transcripts were
synthesized from a quail cDNA template as described
by Ebensperger et al. (1995). Chick Msx-1 was provided
by A. Lumsden and riboprobes made as described
(Graham et al., 1993).
Application of BMP Protein to Chick Wing Buds
Recombinant human BMP2 and BMP4 protein was
provided by Dr. Vicky Rosen, Genetics Institute, Cam-
bridge, Massachusetts, at a concentration of 2.27 mg/ml
and 0.337 mg/ml, respectively, containing 0.5 M argi-
nine, 10 mM histidine, pH 6.5. BMP2 was diluted in
PBS/0.1% BSA (according to Genetics Institute) to final
concentrations of 1 mg/ml, 0.333 mg/ml, and 0.1 mg/ml.
BMP4 was taken undiluted. Heparin-acrylic beads
(200–250 µm, H-5263 Sigma) were soaked for 1 hour in
2 µl of BMP2 or BMP4 solution. Control beads were
soaked in PBS/0.1% BSA. Beads were placed into slits
cut vertically or horizontally in the anterior margins of
stage 19–25 wing buds. Embryos were allowed to
develop for 24 hours, then fixed in 4% paraformalde-
hyde (PFA) in PBS, dehydrated in ascending methanol
series and processed for in situ hybridization analysis;
some cases were allowed to develop for a further 6–7
days, and skeletal pattern was then determined by
alcian green staining.
Application of SHH-Protein
Affigel CM beads (BioRad, Richmond, CA) were
soaked in different concentrations (0.01 to 16 mg/ml) of
SHH protein and applied to the anterior wing margin at
stage 17 and also at stage 20–21 in some cases, and
grafted as described for BMP-protein application (see
also Yang and Drossopoulou et al., 1997).
ACKNOWLEDGMENTS
We thank Drs. B. Houston, A. Lumsden for providing
cDNAs, Dr. V. Rosen for BMP-proteins, Heike Welzel for
technical assistance, and Drs. Edgar Dahl, Kenji Imai
and Beat Lutz for critical reading of and helpful com-
ments on the manuscript. Work in A.M.’s laboratory is
supported by a grant (NS33642) from the NIH.
REFERENCES
Balling R. The undulated mouse and the development of the vertebral
column. Is there a human PAX-1 homologue? Clin. Dysmorphol.
1994;3:185–191.
Bopp D, Jamet E, Baumgartner S, Burri M, Noll M. Isolation of two
tissue-specific Drosophila paired box genes, Pox meso and Pox neuro.
EMBO J. 1989;8:3447–3457.
Bowen J, Hinchliffe JR, Horder TJ, Reeve AMF. The fate map of the
chick forelimb-bud and its bearing on hypothesized developmental
control mechanisms. Anat. Embryol. 1989;179:269–283.
Chiang C, Litingtung Y, Lee E, Young KE, Corden JL, Westphal H,
Beachy PA. Cyclopia and defective axial patterning in mice lacking
Sonic hedgehog gene function. Nature 1996;383:407–413.
Dahl E, Koseki H, Balling R. Pax genes and organogenesis. BioEssays
1997;19:755–765.
Deutsch U, Dressler GR, Gruss P. Pax 1, a member of a paired box
homologous murine gene family, is expressed in segmented struc-
tures during development. Cell 1988;53:617–625.
Dietrich S, Gruss P. undulated phenotypes suggest a role of Pax-1 for
205
the development of vertebral and extravertebral structures. Dev.
Biol. 1995;167:529–548.
Ebensperger C, Wilting J, Brand-Sabieri B, Mizutani Y, Christ B,
Balling R, Koseki H. Pax-1, a regulator of sclerotome development is
induced by notochord and floor plate signals in avian embryos. Anat.
Embryol. 1995;191:297–310.
Fan C-M, Tessier-Lavigne M. Patterning of mammalian somites by
surface ectoderm and notochord: Evidence for sclerotome induction
by a hedgehog homolog. Cell 1994;79:1175–1186.
Fan C-M, Porter JA, Chiang C, Chang DT, Beachy PA, Tessier-Lavigne
M. Long-range sclerotome induction by sonic hedgehog: Direct role
of the amino-terminal cleavage product and modulation by the cyclic
AMP signaling pathway. Cell 1995;81:457–465.
Ferguson EL, Anderson KV. Decapentaplegic acts as a morphogen to
organize dorsal-ventral pattern in the Drosophila embryo. Cell
1992;71:451–461.
Francis PH, Richardson MK, Brickell PM, Tickle C. Bone morphoge-
netic proteins and a signalling pathway that controls patterning in
the developing chick limb. Development 1994;120:209–218.
Francis-West PH, Robertson KE, Ede DA, Rodriguez C, Izpisua-
Belmonte JC, Houston B, Burt DW, Gribbin C, Brickell PM, Tickle
C. Expression of genes encoding bone morphogenetic proteins and
sonic hedgehog in talpid (ta3) limb buds: Their relationships in the
signalling cascade involved in limb patterning. Dev. Dyn. 1995;203:
187–197.
Gelbart WM. The decapentaplegic gene: A TGF-␤ homologue control-
ling pattern formation in Drosophila. Development 1989;122(suppl):
65–74.
Graham A, Heyman I, Lumsden A. Even-numbered rhombomeres
control the apoptotic elimination of neural crest cells from odd-
numbered rhombomeres in the chick hindbrain. Development 1993;
119:233–245.
Grüneberg H. The Pathology of Development. Oxford: Blackwell,
1963.
Gruss P, Walther C. Pax in development. Cell 1992;69:719–722.
Hamburger V, Hamilton HL. A series of normal stages in the develop-
ment of the chick embryo. J. Morphol. 1951,88:49–92.
Hofmann C, Luo G, Balling R, Karsenty G. Analysis of limb patterning
in BMP-7 deficient mice. Dev. Genet. 1996;19:43–50.
Hogan BLM. Bone morphogenetic proteins: Multifunctional regula-
tors of vertebrate development. Genes Dev. 1996;10:1580–1594.
Houston B, Thorp BH, Burt DW. Molecular cloning and expression of
bone morphogenetic protein-7 in the chick epiphyseal growth plate.
J. Molec.Endocrinol. 1994;13:289–301.
Johnson RL, Laufer E, Riddle RD, Tabin C. Ectopic expression of Sonic
hedgehog alters dorsal-ventral patterning of somites. Cell 1994;79:
1165–1173.
Laufer E, Nelson CE, Johnson RL, Morgan BA, Tabin C. Sonic
hedgehog and fgf-4 act through a signaling cascade and feedback
loop to integrate growth and patterning of the developing limb bud.
Cell 1994;79:993–1003.
Lecuit T, Brook WJ, Ng M, Calleja M, Sun H, Cohen S.M. Two distinct
mechanisms for long-range patterning by Decapentaplegic in the
Drosophila wing. Nature 1996;381:387–393.
Macias D, Gañan Y, Sampath TK, Piedra ME, Ros MA, Hurle JM. Role
of BMP-2 and OP-1 (BMP-7) in programmed cell death and skeleto-
genesis during chick limb development. Development 1997;124:1109–
1117.
Monsoro-Burq A-H, Duprez D, Watanabe Y, Bontoux M, Vincent C,
Brickell P, Le Douarin N. The role of bone morphogenetic proteins in
vertebral development. Development 1996;122:3607–3616.
Nellen D, Burke R, Struhl G, Basler K. Direct and long-range action of
a DPP morphogen gradient. Cell 1996;85:357–368.
Neubüser A, Peters H, Balling R, Martin G.R. Antagonistic interac-
tions between FGF and BMP signaling pathways: A mechanism for
positioning the sites of tooth formation. Cell 1997;90:247–255.
Noll M. Evolution and role of Pax genes. Curr. Opin. Gen. Dev.
1993;4:595–605.
Oliver G, DeRobertis EM, Wolpert L, Tickle C. Expression of a
homeobox gene in the chick wing bud following application of
retinoic acid and grafts of polarizing region tissue. EMBO J.
1990;9:3093–3099.
206 HOFMANN ET AL.
Riddle RD, Johnson, RL, Laufer E, Tabin C. Sonic hedgehog mediates
the polarizing activity of the ZPA. Cell 1993;75:1401–1416.
Sampath TK, Reddi AH. Dissociative extraction and reconstitution of
extracellular matrix components involved in local bone differentia-
tion. Proc. Natl. Acad. Sci. USA 1981;78:7599–7603.
Saunders JWJ. The proximo-distal sequence of origin of the parts of
the chick wing and the role of the ectoderm. J. Exp. Zool. 1948;108:
363–403.
Soriano P. The PDGFa receptor is required for neural crest cell
development and for normal patterning of the somites. Development
1997;124:2691–2700.
Staehling-Hampton K, Hoffmann FM, Baylies MK, Rushton E, Bate
M. dpp induces mesodermal gene expression in Drosophila. Nature
1994;372:783–786.
Timmons PM, Wallin J, Rigby PW, Balling R. Expression and function
of Pax 1 during development of the pectoral girdle. Development
1994;120:2773–2785.
Urist MR. Bone: Formation by autoinduction. Science 1965;150:893–
899.
Vargesson N, Clarke JDW, Vincent K, Coles C, Wolpert L, Tickle C.
Cell fate in the chick limb bud and relationship to gene expression.
Development 1997;124:1909–1918.
Wilkinson DG. In situ hybridization. In: Stern CD, Holland PWH, eds.
Essential Developmental Biology. Oxford: Oxford University Press,
1993:257–274.
Wozney JM, Rosen V, Celeste AJ, Mitsock LM, Whitters MJ, Kriz R,
Hewick RM, Wang EA. Novel regulators of bone formation: Molecu-
lar clones and activities. Science 1988;242:1528–1534.
Yang Y, Drossopoulou G, Chuang P-T, Duprez D, Marti E, Bumcrot D,
Vargesson N, Clarke J, Niswander L, McMahon A, Tickle C.
Relationship between dose, distance and time in Sonic Hedgehog-
mediated regulation of anteroposterior polarity in the chick limb.
Development 1997;124:4393–4404.
Yokouchi Y, Sakiyama J-I, Kameda T, Iba H, Suzuki A, Ueno N,
Kuroiwa A. BMP-2/-4 mediate programmed cell death in chicken
limb buds. Development 1996;122:3725–3734.