Developmental Cell 10, 647–656, May, 2006 ª2006 Elsevier Inc.DOI 10.1016/j.devcel.2006.04.004
The Cell Surface Membrane Proteins Cdo and Boc
Are Components and Targets of the Hedgehog
Signaling Pathway and Feedback Network in Mice
Jong-Sun Kang,2Robert S. Krauss,2
and Andrew P. McMahon1,*
1Department of Molecular and Cellular Biology
16 Divinity Avenue
Cambridge, Massachusetts 02138
2Brookdale Department of Molecular, Cell,
and Developmental Biology
Mount Sinai School of Medicine
One Gustave L. Levy Place
New York, New York 10029
Cdo and Boc encode cell surface Ig/fibronectin super-
family members linked to muscle differentiation. Data
here indicate they are also targets and signaling com-
ponents of the Sonic hedgehog (Shh) pathway. Al-
by Hedgehog (HH) signaling, in the neural tube Cdo is
expressed within the Shh-dependent floor plate while
Boc expression lies within the dorsal limit of Shh sig-
sion of Boc or Cdo results in a Shh-dependent, cell
autonomous promotion of ventral cell fates and a non-
tities consistent with Shh sequestration. Cdo and Boc
bind Shhthrough ahigh-affinity interaction withaspe-
cific fibronectin repeat that is essential for activity. We
propose a model where Cdo and Boc enhance Shh
signaling within its target field.
Hedgehog (HH) signals regulate the specification of
complex patterns within embryonic fields as diverse as
imaginal discs in Drosophila larvae and the neural tube
and limb of vertebrate embryos (McMahon et al., 2003).
In the neural tube, the induction of all ventral cell identi-
ties requires direct Sonic hedgehog (Shh) signaling; the
of Shh ligand (reviewed in Briscoe and Ericson, 2001;
Jessell, 2000). Shh is initially released from the midline
notochord underlying the ventral neural plate/tube and
later from the floor plate. The floor plate, a population
of ventral midline support cells within the neural tube,
isitself atarget ofShhsignaling thatrequiresthehighest
from these sources forms a gradient that extends over
the ventral half of the neural tube (Gritli-Linde et al.,
In these patterning processes, feedback mechanisms
acting at the level of ligand binding play a critical role in
determining both the number and full range of ventral
cell types (reviewed in Ingham and McMahon, 2001).
Patched-1 (Ptch1) encodes the vertebrate HH receptor
while Hedgehog-interacting protein-1 (Hhip1) encodes
an unrelated membrane-associated protein that simi-
larly binds all mammalian HH ligands. Ptch1 and Hhip1
are upregulated in response to HH signaling; their feed-
back functions serve to modify the range of signaling
and regional response of target cells (Chuang and
McMahon, 1999; Jeong and McMahon, 2005). A third
HH binding factor, Growth arrest-specific-1 (Gas1), is
thought to inhibit Shh signaling; Gas1 is itself repressed
in response to HH signaling (Lee et al., 2001). Here we
present evidence that Cdo and Boc, which encode cell
surface bound members of the Ig/fibronectin domain
superfamily, are novel feedback components that act
in a different manner, to enhance Shh signaling within
subregions of Shh’s neural target field.
Results and Discussion
Cdo and Boc Are Targets of Shh Signaling that Cell
Autonomously Enhance Shh Signaling
To attempt to identify novel, general feedback compo-
nents, we compared transcriptional profiles (data not
HH signaling is either normal (wild-type embryos), ab-
sent (Smoothened [Smo] mutant embryos [Zhang et al.,
2001]), or enhanced (Ptch1 mutant embryos [Goodrich
et al., 1997]), with profiles generated from microdis-
sected tissues from later stage embryos where Shh sig-
naling is lost (head and limb fractions from E10.5 Shh
mutantembryos [St-Jacques etal., 1998]). Among those
genes encoding cell surface or secreted proteins down-
regulated in response to Shh (enhanced expression in
Smo and Shh mutants and repressed in Ptch1 mutants),
we identified Gas1, as expected, and two genes that en-
code related members of an Ig/fibronectin repeat-con-
taining superfamily of cell surface, membrane-spanning
proteins, Cdo (sometimes Cdon [Kang et al., 1997]) and
Boc (Kang et al., 2002).
family, consisting of an ectodomain comprised of four
lar domain (Kang et al., 1997, 2002). Interestingly, Cdo
mutant mice exhibit a microform holoprosencephaly,
reminiscent of a partial loss of Shh signaling (Cole and
in Drosphila that lists a Boc/Cdo relative, CG9211, as a
putative effector of HH signaling (Lum et al., 2003).
Cdo and Boc expression were examined in the devel-
oping mouse and chick embryo. In both, Cdo and Boc
expression are excluded from most HH-signaling do-
mains, consistent with negative regulation by HH signal-
ing (Figure 1 and Figure S1 [see the Supplemental Data
available with this article online], data not shown, and
both genes are expressed dorsally, whereas in the limb,
mesenchymal expression is restricted to the anterior
two-thirds. On removal of HH signaling in Smo and Shh
mutants,Cdo andBocexpressionisenhanced, expand-
ing ventrally in the somites and neural tube and to the
posterior margin of the limb (Figure 1). In contrast, nor-
mal expression is lost, or markedly downregulated,
both when HH signaling is derepressed in Ptch1 mu-
tants (Figure 1A) or ectopically activated following
expression of a constitutively active allele of Smo
(SmoM2, Figure 1B) (Jeong et al., 2004). Thus, Cdo and
Boc appear to be negative targets of HH regulation in
multiple HH-responsive tissues.
ship between Cdo and Boc expression domains and HH
signaling is more complex. Cdo is transiently expressed
at low levels within the notochord, a midline structure
that produces, responds to, and requires Shh signaling
(Figure S1) (Chiang et al., 1996; Echelard et al., 1993).
Further, Cdo is weakly expressed at the ventral-mid line
of the neural tube coincident with Shh-mediated induc-
tion of the rostral brain and caudal floor plate (arrow in
Figure 1C and Figure S1; and see the accompanying pa-
per by Zhang et al.  in this issue of Developmental
Cell). Finally, regions of active Shh signaling (as judged
by upregulation of the general transcriptional targets
Ptch1 and Gli1) overlap the ventral boundary of Boc ex-
pression in the neural tube, and posterior boundaries
of Boc and Cdo expression in the limb mesenchyme
(Figure S2) (Gritli-Linde et al., 2002, 2001). Thus, Cdo
and Boc may play active roles within specific HH-signal-
To address this possibility, we determined whether
Cdo and Shh genetically interact. Cdo2/2mutants have
a mild holoprosencephalic phenotype; midline struc-
tures are lost, and left and right nasal processes, while
separate structures, are positioned closer to the midline
(Figure 2A) (Cole and Krauss, 2003). Although Shh2/2
embryos exhibit an extreme holoprosencephalic pheno-
type, Shh+/2embryos are comparable to wild-type
(Chiang et al., 1996). When Shh gene dosage is lowered
in a Cdo mutant background (Shh+/2; Cdo2/2), the Cdo
phenotype is dramatically enhanced; the nasal pro-
mark of Shh deficiency (Figure 2A) (Chiang et al., 1996;
Mulieri et al., 2000). The observed genetic interaction
suggests that Cdo may normally promote Shh signaling.
Given Cdo expression in the floor plate, a structure in-
duced by high levels of Shh signaling (McMahon et al.,
2003), we characterized ventral patterning in the neural
tube of these mutants. During normal floor plate devel-
opment there is a transitory period wherein ventro-me-
dial progenitors are Nkx2.2+and Foxa2+. At later stages,
Foxa2 is restricted to the definitive floor plate and
Nkx2.2 to ventro-lateral vp3 progenitors of the V3 class
ofspinal interneurons (for reviews,see Briscoeand Eric-
the floor plate; activation requires the activity of Foxa2,
control regions (Jeong and Epstein, 2003). At E10.5,
dent cellpopulationsin the neural tube ofboth wild-type
and Shh+/2embryos (Figure 2B and data not shown).
However, in Cdo mutants, few midline cells are Foxa2+
Figure 1. Boc and Cdo Are Negatively Regu-
lated by Hedgehog Signaling
In situ hybridization analysis of Boc and Cdo
expression in the developing mouse embryo.
(A) E8.5 (8–10 somite stage) mouse embryos
show enhanced or ectopic expression of
Cdo and Boc in Shh and Smo mutants, and
repression in Ptch1 mutants. Anterior to left,
(B) Boc and Cdo expression expand into the
posterior mesenchyme of Shh mutant fore-
limb buds at E10.5, but are broadly repressed
on activation of HH signaling following ec-
topic expression of SmoM2 throughout the
limb mesenchyme. Anterior at top, dorsal
(C) Cdo and Boc expression in the E10.5 neu-
ral tube at the forelimb level. Dorsal Cdo and
Boc expression is upregulated and their ex-
pression domains expand ventrally in the
Shh mutant neural tube. Cdo also shows ex-
pression in the floor plate (arrow) and noto-
chord (arrowhead); the former is lost in Shh
mutants (see also Figure S1). Dorsal at top.
only; most remain both Foxa2+and Nkx2.2+(Figure 2B).
The total number of Foxa2+cells is also reduced (Fig-
is a corresponding decrease in the Shh-producing floor
type is enhanced; in some embryos a few remaining
Shh+, Foxa2+cells are present at the midline and all
also reduced (Figure 2C and data not shown). However,
vpMN, Olig2+motor neuron progenitors that are posi-
tioned more dorsally are unaffected (Figure 2B). The re-
duction in vp3, Nkx2.2+progenitors most likely reflects
reduced levels of normal floor plate-derived Shh, be-
cause in Gli2 mutants, FP specification is lost and a re-
duction in vp3 progenitors is also observed (Ding et al.,
Figure 2. Genetic Interactions between Cdo and Shh
(A) Frontal views of facial development in E10.5 mouse embryos. The facial phenotype of Cdo mutants is greatly enhanced on lowering of Shh
gene dosage. Paired olfactory pits (arrowed), normally separate structures, fuse at the midline in Shh+/2; Cdo2/2embryos.
(B) Immunostaining of the floor plate (Shh+, Foxa2+), vp3 (Nkx2.2+), and pMN (Olig2+) progenitors in the E10.5 mouse neural tube at the forelimb
level. Dorsal is at top.
(C) Quantitation of numbers of Foxa2+and Nkx2.2+cells above (n = 4). Bars represent standard deviation. ** indicates a significant difference in
the number of Foxa2+cells between Shh+/2and Cdo2/2embryos (p = 0.0007), while * indicates differences in the number of Foxa2+and Nkx2.2+
cells between Cdo2/2and Shh+/2; Cdo2/2embryos (p = 0.0097 and p = 0.0006, respectively).
Cdo and Boc Binding of Shh Enhances Shh Signaling
specification and interacts with the Shh signaling path-
way in this process.
To further address interactions between Boc, Cdo,
and the Shh pathway, we ectopically expressed Boc
and Cdo in the neural tube of the developing chick em-
bryo. Shh patterns the presumptive spinal cord by mod-
ulating the expression of transcriptional regulators that
determine specific neural cell fates (Briscoe and Eric-
son, 2001; Jessell, 2000). For example, class I genes,
such as Pax6 and Pax7, are repressed by Shh signaling,
while class II genes, which include Nkx2.2 and Olig2, are
activated (Briscoe et al., 2000). Ectopic expression of
cDNA constructs encoding either full-length (fl) Cdo or
Boc, or truncated forms (t) of both factors that lack
cell autonomous repression of Pax6 and dorsal expan-
sion of Nkx2.2+vp3 progenitors, Olig2+motor neuron
progenitors (vpMN), and Foxa2+floor plate (Figures
3A–3C, Figure S3, and data not shown). Importantly,
where different ectopic cell identities are observed in
the same section, progenitors show a normal, relative
distribution. For example, ectopic Olig2+vpMN progen-
itors always lie dorsal to ectopic Nkx2.2+vp3 progeni-
tors (Figure 3B and data not shown).
Pax7 broadly marks dorsal cell identities, the ventral-
most of which lie at the normal limit of Shh signaling and
Linde et al., 2001; Wijgerde et al., 2002; C. Chamberlain
and A.P.M., unpublished data). Ectopic expression of
expressed normally; see Figure S2), but not in more dor-
sal positions, leads to a cell autonomous repression of
when Boc or Cdo are extensively expressed within the
Shh target field just ventral to the normal Pax7 domain,
a cell nonautonomous expansion of Pax7+cells is ob-
served into the normal Shh target field (Figure 3A, far
left panel, and Figure 3E).
These results lead to several conclusions. First, ex-
pression of Boc and Cdo promotes the adoption of more
ventral neural identities than is appropriate for cells at
a given D-V position. However, the relative position of
ectopic cell identities to one another is normal, suggest-
ing that polarity cues are still observed, as expected if
Cdo- and Boc-mediated inductions are Shh dependent.
Further, while elevated levels of Cdo and Boc can re-
press Pax7+fates, repression is only observed close to
ent and active based on direct analysis of Shh protein
distribution and the expression of Ptch1 and Gli1, tran-
scriptional targets of the pathway. In all these instances
we only observe a cell autonomous action of Cdo and
Boc, consistent with their directly modulating Shh sig-
naling input to the ectopically expressing cells. Second,
the cell nonautonomous appearance of more dorsal cell
fates above a strong area of Boc or Cdo expression is
suggestive of phenotypes observed when Shh is se-
questered by ectopic expression of Hhip1 (Stamataki
et al., 2005). This may indicate that either Boc or Cdo
Figure 3. Ectopic Expression of Boc and Cdo
Promotes Shh-Dependent Cell Fate Specifi-
cation in the Shh Target Field
Human Boc- or mouse Cdo-encoding cDNAs
were electroporated unilaterally into the
chick neural tube at HH stage 11/12. Neural
tube patterning was analyzed at forelimb
levels at HH stage 21/22 with cell-type-spe-
cific antibody markers. Dorsal is at the top
in all panels. Electroporated cells coexpress
GFP and are green in all panels.
(A and B) Both full-length (fl) and cytoplasmi-
cally truncated (t) forms of Boc and Cdo acti-
vate ectopic Nkx2.2 and Olig2 in more dorsal
positions in the Shh target field, repress Pax6
within the Shh target field, and repress both
Pax6 and Pax7 (arrowhead in [A]) just dorsal
to the D-V intersect. Cell-non-autonomous
ventral expansion of Pax7 is also detected
(arrow in [A]). When Olig2 and Nkx2.2 are ex-
amined in the same section, ectopic Olig2+
(arrowhead in [B]) cells lie dorsal to ectopic
Nkx2.2+(arrow in [B]) progenitors. The lower
concentration of Cdo plasmid injected (ne-
cessitated by a pronounced growth arrest
at higher concentrations) leads to a weaker
GFP signal that is somewhat masked in over-
(C) Boc expression in the Nkx2.2 domain
leads to silencing of Nkx2.2 and ectopic
(D) A cell autonomous repression of Pax7
dorsal cell identities (upper open arrowhead).
(E) A cell-non-autonomous ventral expansion
following ectopic expression of Boc (green
themselves bind Shh, or their activity promotes Shh
retention indirectly. Third, both Cdo and Boc appear to
have similar properties, as eachgenerates asimilar phe-
notype. Although Cdo and Boc can associate with each
other (Kang et al., 2002), this association does not ap-
pear to be necessary to promote ventralization. Fourth,
in their molecular action, the intracellular domain is dis-
pensable for ventralizing activity, whereas the trans-
membrane domain is not (data not shown). Together,
domains of Cdo and Boc bind to and sequester Shh
ligand, thereby enhancing Shh signaling cell autono-
mously where ligand is available but also potentially
limiting Shh movement to more dorsal positions in the
normal target field.
To address whether the action of Cdo and Boc are in-
deed specific for HH signaling, we performed coelectro-
poration studies with HH pathway-specific components
that are known to act at the level of ligand binding and
ligand-dependent feedback regulation of membrane
signaling. Ptch1Dloop2encodes a modified form of the
HH-receptor Ptch1 in which removal ofone of two extra-
cellular loops prevents ligand binding (Briscoe et al.,
2001). As ligand binding to Ptch1 is required to block
Ptch1-mediated inhibition of Smo activity, and Smo ac-
tivity is required for the specification of all HH-depen-
dent cell fates, expression of Ptch1Dloop2specifically
inhibits ligand-dependentsignaling atthelevelofPtch1-
Smo. As expected if the ectopic induction of ventral cell
identities by Cdo and Boc is dependent on Shh ligand-
based signaling, coexpression of Ptch1Dloop2with Boc
results in a cell autonomous inhibition of Boc-mediated
ventralization (Figure 4A, lower panel). Importantly,
where cells express only Boc, ectopic ventral cell iden-
tities are observed (Figure 4A, upper panel). Expression
of Ptch1Dloop2is associated with ventral cells ectopically
activating the dorsal marker Pax7; inhibition of Pax7
is the lowest Shh threshold response reported to date:
less than 500 pM of Shh is sufficient in in vitro assays
to abolish Pax7 expression, whereas greater than 4 nM
is required for floor plate induction (Ericson et al.,
1997). As expected, ectopic ventral expression of
Ptch1Dloop2results in ectopic Pax7+cells (Figure 4B, up-
per panel). However, coexpression of Boc suppresses
this phenotype (Figure 4B, lower panel). These results
suggest that, where ligand is available, Boc enhance-
itory effects of Ptch1Dloop2, providing sufficient, minimal
level signaling to enable Pax7 repression, but insuffi-
cient for ectopic induction of Nkx2.2+vp3 progenitors.
In support ofthis model, electroporation with a constitu-
tive repressor form of Gli3 (Gli3R) that is insensitive
to Shh signaling results in a cell autonomous ventral
expansion of Pax7+cells that cannot be inhibited by
coelectroporation with Boc (Figure S4). Hhip1 encodes
a second, membrane-associated feedback antagonist
that binds ligand directly (Chuang and McMahon,
regulate HH signaling cell autonomously in HH respond-
ing cells. We also coexpressed Boc with Hhip1; how-
ever, ectopic expression of Hhip1 alone leads to a
severe growth defect and an apparent loss of viability
in ventral progenitors, precluding further study (data
Cdo and Boc Bind Shh Directly through a Specific
The data above are most readily explained if Cdo and
Boc play a direct role in Shh signaling, rather than indi-
bers of the Tgf-b superfamily that play opposite roles to
Shh in patterning the dorsal neural tube. Together, the
Shh ligand-dependent action of Cdo and Boc and their
cell surface localization suggest that they could act
through binding of Shh ligand. To address this possibil-
ity, we examined binding of a secreted form of an N-
Shh::AP fusion protein (the N-terminal signaling moiety
of Shh fused at its C terminus to alkaline phosphatase
[AP]) to Cos7 cells expressing Hhip1 (positive control),
Boc, Cdo, and as a negative control, Frizzled-3 (Fz3),
a putative Wnt receptor (Figures 5A and 5B). Only cells
Figure 4. Coexpression of Boc with Ptch1Dloop2Abrogates the
Effects of Boc-Mediated Enhancement of Ventral Cell Fate Specifi-
Plasmid constructs encoding cytoplasmically truncated Boc(t) (GFP
coexpression) or Ptch1Dloop2(DsRed coexpression) were electropo-
rated into the chick neural tube, and the expression of (A) vp3
(Nkx2.2+) and (B) dorsal (Pax7+) progenitors was assayed by immu-
nohistochemistry as indicated. Control DNAs consist of base vec-
tors producing only GFP or DsRed.
Cdo and Boc Binding of Shh Enhances Shh Signaling
expressing Hhip1, Boc, and Cdo bind N-Shh::AP, bind-
ing is Shh-dependent (i.e., AP alone does not bind),
and both Boc and Cdo bind N-Shh::AP as effectively as
Hhip1 (Chuang and McMahon, 1999). We next examined
whether binding represents a direct association of Shh
with either Boc or Cdo. When N-Shh::AP and epitope-
tagged, secreted forms of Boc or Cdo (BocDTMCD and
natants are assayed, we detect N-Shh::AP/BocDTMCD
and N-Shh::AP/CdoDTMCD complexes, indicating that
Shh binds to both Boc and Cdo ectodomains (Figures
5C and 5D). The use of secreted forms reduces the pos-
binding or contribute directly to the complex.
Figure 5. Cdo and Boc Bind Shh
(A) Binding of N-Shh::AP to Cdo- and Boc-expressing cells. Scale bar, 50 mm.
(B) Quantitation of N-Shh::AP binding to Cdo and Boc. Error bars represent the mean 6 SD of four identical treatment groups. Significant differ-
ences between binding of N-Shh::AP and AP conditioned medium (CM) are indicated by * (two-tailed student t test, p < 0.01).
(C and D) (C) Immunoprecipitation of Boc and (D) immunoprecipitation of Cdo extracellular domains (BocDTMCD and CdoDTMCD, respectively)
with Shh. Cos7 cells were transfected with BocDTMCD or CdoDTMCD alone (lane 1), or cotransfected with N-Shh::AP (lane 2) or AP (lane 3).
Complexes were immunoprecipitated from supernatants with anti-AP beads. Epitope-labeled BocDTMCD was detected following Western
blot analysis of immunoprecipates with anti-Myc antibody, and epitope-tagged CdoDTMCD with anti-HA antibody. N-Shh::AP was detected
with anti-Shh antibody or anti-AP; the latter was also used to detect AP.
(E) Dissociation constant (Kd) measurements. Saturation binding curves and Scatchard analysis (insets) of NShh::AP binding to Hhip1 (left), Boc
(middle), and Cdo (right). Each point on the graphs represents the average of three identical treatment groups.
Figure 6. Mapping of Shh Binding Domains in Cdo and Boc
(A) Cos7 cells transfected with N-Shh::AP alone (lane 1) or cotransfected with various Cdo ectodomain Fc fusion constructs (lanes 2–6). Com-
plexes were immunoprecipitated from conditioned medium with Protein A agarose. Top panel: Detection of N-Shh::AP with anti-Shh. Bottom
panel: Identification of Fc fusion proteins with anti-human IgG. Asterisks highlight the various Fc fusion proteins as confirmed by comparison
with the migration of molecular weight markers (left).
(B) Quantitation of relative Shh binding to each Cdo-Fc fusion proteins. Binding is expressed as a ratio of Shh band intensity/Cdo-Fc band
(C) Schematic of Cdo mutant constructs. Full-length Cdo (top) is contrasted with constructs expressing only the FNIII(3) domain of the Cdo
tical constructs were also generated for Boc.
(D) Binding of NShh::AP to Cdo and Boc constructs containing the third FNIII repeat [CdoFNIII(3)TMCD and BocFNIII(3)TMCD, respectively] or
lacking the third FNIII repeat [CdoDFNIII(3) and BocDFNIII(3), respectively]. Scale bar, 50 mm.
(E) Quantitation of N-Shh::AP binding to full-length and truncated Cdo and Boc constructs.
Error bars represent the mean 6 SD of four identical treatment groups.
Cdo and Boc Binding of Shh Enhances Shh Signaling
To determine the affinity of the Shh-Boc and Shh-Cdo
interactions, saturation binding experiments were per-
formed using Cos7 cells transfected with Boc, Cdo, or
Hhip1 for comparison (Figure 5E). Calculation of the dis-
sociation constants (Kd) for NShh::AP binding to Boc
and Cdo yielded similarly high affinities (approximately
3 and 4 nM, respectively), while a Kdof 1 nM for Hhip1
is in close agreement with previously published data
(Chuang and McMahon, 1999).
Having established that Boc and Cdo interact with
Shh with high affinity, we performed domain-mapping
analysis to define the region of Boc and Cdo that binds
to Shh (Figure 6). Immunoprecipitation experiments us-
ing Fc-fusion constructs that contain either the Ig-like
domains or FNIII domains of Cdo indicate that the fibro-
nectin repeat-containing region plays the major role in
Shh binding (Figure 6A). Furthermore, analysis of con-
structs expressing each FNIII domain singly suggests
that most binding can be ascribed to the FNIII(3) do-
main, the most highly conserved of these repeats (Fig-
ures 6A and 6B). To further confirm the importance of
this domain in Cdo and Boc, NShh::AP binding assays
were performed using constructs whose extracellular
domains consist of only the third FNIII repeat (CdoF-
NIII(3)TMCD and BocFNIII(3)TMCD) or the entire extra-
cellular domain except for the third FNIII repeat (CdoDF-
the FNIII(3) domain of Cdo and Boc is both necessary
and sufficient to specifically mediate NShh::AP binding
to Cos7 cells, indicating that this region plays a critical
role in these interactions.
Cdo and Boc Binding to Shh Is Necessary, but Not
Sufficient, to Ectopically Activate Shh Signaling
To test whether the FNIII(3) domains of Cdo and Boc are
necessary to augment Shh signaling, chick electropora-
tion experiments were performed with a series of Cdo
and Boc constructs (Figure 7and Figure S5). Expression
of full-length Cdo or Boc results in ectopic activation
of Nkx2.2 (Figure 7A, top left panels, and Figure 7B, left
panels, respectively), as does expression of Cdo con-
or FNIII(2) (Figure 7A, bottom left panels). In contrast,
Cdo or Boc lacking FNIII(3) fails to ectopically activate
Nkx2.2, despite strong ventral expression of these vari-
ants (Figure 7A, bottom right panel, and Figure 7B, right
panels, respectively). Despite the clear requirement for
FNIII(3) in Shh binding and activity, the FNIII(3) domain
alone of either Boc or Cdo is not sufficient to reproduce
Cdo- or Boc-dependent phenotypes within the neural
tube(FigureS5).Thus, CdoandBocmostlikely promote
requiring other regions of their extracellular domains.
Understanding these interactions may provide some in-
sight into why some Shh binding proteins, such as Boc
while others, such as Hip1, function as negative regula-
tors.These studies,together withthoseintheaccompa-
nying paper (Zhang et al., 2006), identify Boc and Cdo
ing pathway. A report that a related gene is required for
normal HH signaling in Drosophila tissue culture cells
suggests this function is conserved (Lum et al., 2003).
analyses, we propose a model wherein Cdo and Boc en-
hanceHHsignalingattwocritical positions withinapos-
ing levels are required in FP specification and (2) at the
fringes of a HH target field, close to the D-V intersect
in the neural tube and possibly also at the anterior limit
of signaling in the limb bud. At these latter positions,
where ligand levels are expected to be low, this mecha-
ally, the negative regulation of Boc and Cdo expression
by HH signaling would restrict Cdo/Boc-mediated en-
hancement ofHH signaling tothe relevant region, estab-
lishing by this feedback system an appropriate domain
for theiraction. While the current model is clearly specu-
lative, the future analysis of Boc mutants and Cdo; Boc
compound mutants, along with further biochemical and
cellular analyses, should provide important tests of
Our data provide two critical mechanistic insights;
first, Cdo and Boc both bind Shh via their FNIII(3) do-
mains, and second, Cdo and Boc binding to Shh is nec-
essary for enhancement of Shh signaling. We suggest
Figure 7. The FNIII(3) Domains in Both Cdo and Boc Are Necessary
for Enhancement of Shh Signaling in the Developing Chick Neural
in chick neural tubes electroporated with full-length Cdo (top left
panels), CdoDFNIII(1) (top right panels), CdoDFNIII(2) (bottom left
panels), or CdoDFNIII(3) (bottom right panels). Green cells indicate
GFP expression in electroporated cells.
following electroporation with full-length Boc (left panels) or BocDF-
NIII(3) (right panels).
sentation of active ligand to Ptch1, or that binding coun-
teracts feedback-mediated sequestration of ligand and
ligand turnover, increasing effective levels of signaling
sent a new class of factors in the increasingly complex
Shh feedback network; expression of each is broadly
negatively regulated by HH signaling, but their activity
stimulates HH signaling. Importantly, the accompanying
work of Zhang et al. (2006) identifies Cdo as a modulator
genes as potential interacting factors in Shh-related
The transcriptional profiling will be described in detail elsewhere (TT
and APM, in preparation). Briefly, RNA was prepared from 6–8 and
10–13 somite stage wild-type, Ptch12/2(Goodrich et al., 1997) and
Smo2/2(Zhang et al., 2001) embryos, and from head and limb
buds isolated from E10.5 wild-type and Shh2/2(St-Jacques et al.,
1998)embryos.RNAs wereusedinstandardprocedures togenerate
probes for analysis of transcript expression on Affymetrix U74Av2
and M430 A and B microarrays. Data were statistically analyzed us-
ing Resolver software (Rosetta).
Mouse experimentswerecarriedout largely on a129background as
in the original Cdo report; hence the ‘‘weak’’ midline defects in the
Cdo2/2embryos in this study. The Shh mutant allele on a 129/Sv;
C57BL6/J; CBA/J hybrid background was crossed with Cdo+/2
stock (129/Sv; C57BL6) and the phenotypes of littermates examined
in this mixed background (Cdo2/2, n = 14; Cdo2/2; Shh+/2, n = 13).
The strongest midline defects in Cdo2/2littermates were always ob-
served in those carrying the Shh null allele.
Generation of Cdo and Boc Constructs
All constructs were generated using standard molecular biology
procedures (Maniatis et al., 1982). Briefly, cytoplasmic truncations
of Boc and Cdo [Boc(t) and Cdo(t)] deleted aa 79–1115 of human
Boc and aa 986–1251 of mouse Cdo. Soluble versions of Boc and
Cdo that lack both the transmembrane and cytoplasmic domains
(BocDTMCD and CdoDTMCD) were truncated at aa 855 and aa
958 of Boc and Cdo, respectively. Constructs encoding the third
FNIII repeat, transmembrane, and cytoplasmic domains of Boc
[BocFNIII(3)TMCD] and Cdo [CdoFNIII(3)TMCD] were fused to their
respective signal peptides at aa 712 of Boc and aa 830 of Cdo. Con-
structs encoding Boc and Cdo that lack the third FNIII repeat
[BocDFNIII(3) and CdoDFNIII(3), respectively] deleted aa 710–809
of Boc and aa 832–919 of Cdo. All constructs were cloned into pCIG.
Cdo-Fc fusion proteins were generated by PCR amplification of
each of the indicated regions as follows: CdoIg(1–5)-Fc (aa 1–575),
CdoFNIII(1–3)-Fc (aa 534–959), CdoFNIII(1)-Fc (aa 534–711), CdoF-
NIII(2)-Fc (aa 662–814), and CdoFNIII(3)-Fc (aa 802–959). The FNIII
constructs were then fused in-frame to native mouse Cdo start co-
don and signal sequence (aa 1–41), followed by the cloning of all
constructs into the Igtag vector (Bergemann et al., 1995).
AP Binding Assays
These experiments were performed essentially as described previ-
ously (Flanagan et al., 2000). Briefly, Cos-7 cells were transfected
with either AP or N-Shh::AP alone, or cotransfected with full-length
mouse Hhip1 (Chuang and McMahon, 1999), mouse Cdo (Kang
et al., 1997), human Boc (Kang et al., 2002), or mouse Fz3. Bound
AP protein was visualized with BM purple AP substrate (Roche) for
cell surface staining, or with AP yellow liquid substrate (Sigma) to
quantify AP binding in cell extracts. Saturation binding curves and
Scatchard analysis were performed using GraphPad Prism version
4.03 forWindows (GraphPadSoftware,San Diego, CA). Kdmeasure-
ments were determined by nonlinear regression analysis of the sat-
uration binding data.
Cos-7 cells were transfected using Lipofectamine 2000 (Invitrogen),
and conditioned medium was collected 48 hr after transfection. Im-
munoprecipitation of AP and N-Shh::AP from conditioned medium
was performed by incubation with anti-AP agarose beads (Sigma)
overnight at 4ºC on a rotator. Beads were washed three times with
buffer (50 mM Tris [pH 7.6], 500 mM NaCl, 1% TritonX-100), resus-
pended in Laemmli sample buffer, heated at 95ºC for 5 min, and an-
alyzed by SDS-PAGE and Western blot analysis. AP and N-Shh::AP
weredetected withrabbit anti-APantibody(Biomeda).Mycepitope-
tagged BocDTMCD was detected with mouse anti-myc antibody
(clone 9E10; Developmental Studies Hybridoma Bank). HA-tagged
CdoDTMCD was identified with mouse anti-HA antibody (Covance).
Rabbit anti-Shh antibody has been described previously (Bumcrot
et al., 1995).
In Situ Hybridization and Immunofluorescence
scribed on wild-type and mutant embryos (Wilkinson,1992). Section
in situ hybridization was carried out on 30 mm sections with digoxi-
genin probes at forelimb-levels. Immunofluorescence analysis was
performed on 10 mm frozen sections; image collection was carried
out on a Zeiss LSM510 confocal microscope. The following anti-
bodies were used: rabbit anti-Olig2 (1:5000), mouse anti-Foxa2
(1:5), rabbit anti-Nkx2.2 (1:4000), mouse anti-Pax7 (1:20), mouse
anti-Pax6 (1:20), rabbit anti-Olig2 (1:5000), mouse anti-Shh (1:25,
Developmental Studies Hybridoma Bank), Alexa 568 or 633 goat
anti-rabbit or goat anti-mouse (1:300, Molecular Probes) and rabbit
anti-DsRed (1:400, BD Bioscience).
Boc or Cdo and their derivatives and Gli3R (a gift of S. Vokes) were
cloned into pCIG vector (Megason and McMahon, 2002) to enable
coexpression of Boc and Cdo with GFP to visualize electroporated
cells. Ptch1Dloop2(Briscoe et al., 2001) was cloned into pCIR. In
this vector, the GFP-encoding cDNA of pCIG was replaced by one
encoding Red fluorescent protein (DsRed-Express, Invitrogen). For
electroporation, Qiagen-purified, supercoiled plasmid DNA was
injected into the neural tubes of Hamburger-Hamilton (HH) stage
11–12 chicken embryos (Hamburger and Hamilton, 1992). Boc and
Cdo were injected at concentrations of 1.0 or 0.7 mg/ml in PBS, re-
spectively, with 50 ng/ml Fast Green. In coelectroporation studies,
both DNAs were at a concentration of 1.5 mg/ml. Electrodes were
made from 0.5 mm diameter platinum wire (Aldrich) and were 5
mm long and spaced 5 mm apart. Electrodes were placed flanking
the neural tube, covered with a drop of PBS, and pulsed five times
at 25 V for 50 ms with a BTX Electroporator (Genetronics). Embryos
were recovered after 48 hr at HH stage 21–22 and fixed for immuno-
histochemistry. Each analysis was repeated a minimum of 20 times.
We thank Drs. T. Jessell, H. Takebayashi, and D. Rowitch for anti-
bodies for neural tube analyses, Jeremy Nathans for the Fz3 clone,
James Briscoe for Ptch1Dloop2, and Steve Vokes for Gli3R. We are
gratefultoBiogen, inparticulartoBlake Pepinsky forthegiftofmod-
ified Shh protein used herein. FoxA2, Nkx2.2, Pax6, Pax7, and Shh
antibodies were obtained from the Developmental Studies Hybrid-
oma Bank developed under the auspices of the NICHD and main-
tainedbyTheUniversity ofIowa, Department ofBiological Sciences,
Iowa City, IA. Work in A.P.M.’s laboratory was supported by a grant
from the NIH (R37 NS033642). T.T. was partially supported by
a grant-in-aid for research abroad from the Ministry of Education,
Culture, Sports, Science and Technology in Japan. B.L.A. was sup-
ported by postdoctoral fellowship #PF0512501DDC from the Amer-
ican Cancer Society. Work in R.S.K.’s lab was supported by grants
from the NIDCR, March of Dimes, and T.J. Martell Foundation.
F.C. was supported by a predoctoral grant from the HHMI and
a training grant from the NCI. A.P.M. has intellectual property based
around Hedgehog technology and is an advisor to Curis.
Cdo and Boc Binding of Shh Enhances Shh Signaling
Received: December 12, 2005
Revised: March 24, 2006
Accepted: April 5, 2006
Published online: April 27, 2006
Bergemann, A.D., Cheng, H.J., Brambilla, R., Klein, R., and Flana-
gan, J.G. (1995). ELF-2, a new member of the Eph ligand family, is
segmentally expressed in mouse embryos in the region of the hind-
brain and newly forming somites. Mol. Cell. Biol. 15, 4921–4929.
Briscoe, J., and Ericson, J. (2001). Specification of neuronal fates in
the ventral neural tube. Curr. Opin. Neurobiol. 11, 43–49.
Briscoe, J., Pierani, A., Jessell, T.M., and Ericson, J. (2000). A homeo-
domain protein code specifies progenitor cell identity and neuronal
fate in the ventral neural tube. Cell 101, 435–445.
Briscoe, J., Chen, Y., Jessell, T.M., and Struhl, G. (2001). A hedge-
hog-insensitive form of patched provides evidence for direct long-
range morphogen activity of sonic hedgehog in the neural tube.
Mol. Cell 7, 1279–1291.
Bumcrot, D.A., Takada, R., and McMahon, A.P. (1995). Proteolytic
processing yields two secreted forms of sonic hedgehog. Mol.
Cell. Biol. 15, 2294–2303.
Chiang, C., Litingtung, Y., Lee, E., Young, K.E., Corden, J.L., West-
phal, H., and Beachy, P.A. (1996). Cyclopia and defective axial pat-
terning in mice lacking Sonic hedgehog gene function. Nature 383,
Chuang, P.T., and McMahon, A.P. (1999). Vertebrate Hedgehog
signalling modulated by induction of a Hedgehog-binding protein.
Nature 397, 617–621.
Cole, F., and Krauss, R.S. (2003). Microform holoprosencephaly in
mice that lack the Ig superfamily member Cdon. Curr. Biol. 13,
Cooper, M.K., Porter, J.A., Young, K.E., and Beachy, P.A. (1998).
Teratogen-mediated inhibition of target tissue response to Shh
signaling. Science 280, 1603–1607.
Ding, Q., Motoyama, J., Gasca, S., Mo, R., Sasaki, H., Rossant, J.,
and Hui, C.C. (1998). Diminished Sonic hedgehog signaling and
lack of floor plate differentiation in Gli2 mutant mice. Development
Echelard, Y., Epstein, D.J., St-Jacques, B., Shen, L., Mohler, J.,
ber of a family of putative signaling molecules, is implicated in the
regulation of CNS polarity. Cell 75, 1417–1430.
Ericson, J., Rashbass, P., Schedl, A., Brenner-Morton, S., Kawa-
kami, A., van Heyningen, V., Jessell, T.M., and Briscoe, J. (1997).
Pax6 controls progenitor cell identity and neuronal fate in response
to graded Shh signaling. Cell 90, 169–180.
Flanagan, J.G., Cheng, H.J., Feldheim, D.A., Hattori, M., Lu, Q., and
Vanderhaeghen, P. (2000). Alkaline phosphatase fusions of ligands
or receptors as in situ probes for staining of cells, tissues, and em-
bryos. Methods Enzymol. 327, 19–35.
Goodrich, L.V., Milenkovic, L., Higgins, K.M., and Scott, M.P. (1997).
Altered neural cell fates and medulloblastoma in mouse patched
mutants. Science 277, 1109–1113.
Gritli-Linde, A., Lewis, P., McMahon, A.P., and Linde, A. (2001). The
whereabouts of a morphogen: direct evidence for short- and graded
long-range activity of hedgehog signaling peptides. Dev. Biol. 236,
Gritli-Linde, A., Bei, M., Maas, R., Zhang, X.M., Linde, A., and McMa-
hon, A.P. (2002). Shh signaling within the dental epithelium is neces-
Hamburger, V., and Hamilton, H.L. (1992). A series of normal stages
in the development of the chick embryo. 1951. Dev. Dyn. 195, 231–
Ingham, P.W., and McMahon,A.P. (2001). Hedgehog signaling in an-
imal development: paradigms and principles. Genes Dev. 15, 3059–
Jeong, Y., and Epstein, D.J. (2003). Distinct regulators of Shh tran-
scription in the floor plate and notochord indicate separate origins
for these tissues in the mouse node. Development 130, 3891–3902.
Jeong, J., and McMahon, A.P. (2005). Growth and pattern of the
mammalian neural tube are governed by partially overlapping feed-
back activities of the hedgehog antagonists patched 1 and Hhip1.
Development 132, 143–154.
Jeong, J., Mao, J., Tenzen, T., Kottmann, A.H., and McMahon, A.P.
(2004). Hedgehog signaling in the neural crest cells regulates the
patterning and growth of facial primordia. Genes Dev. 18, 937–951.
Jessell, T.M. (2000). Neuronal specification in the spinal cord: induc-
tive signals and transcriptional codes. Nat. Rev. Genet. 1, 20–29.
Kang, J.S., Gao, M., Feinleib, J.L., Cotter, P.D., Guadagno, S.N., and
Krauss, R.S. (1997). CDO: an oncogene-, serum-, and anchorage-
regulated member of the Ig/fibronectin type III repeat family.
J. Cell Biol. 138, 203–213.
Kang, J.S., Mulieri, P.J., Hu, Y., Taliana, L., and Krauss, R.S. (2002).
BOC, an Ig superfamily member, associates with CDO to positively
regulate myogenic differentiation. EMBO J. 21, 114–124.
Lee, C.S., Buttitta, L., and Fan, C.M. (2001). Evidence that the WNT-
inducible growth arrest-specific gene 1 encodes an antagonist of
sonic hedgehog signaling in the somite. Proc. Natl. Acad. Sci. USA
Lum, L., Yao, S., Mozer, B., Rovescalli, A., Von Kessler, D., Niren-
berg, M., and Beachy, P.A. (2003). Identification of Hedgehog path-
way components by RNAi in Drosophila cultured cells. Science 299,
Maniatis, T., Fritsch, E.F., and Sambrook, J. (1982). Molecular Clon-
ing: A Laboratory Manual (Cold Spring Harbor, NY: Cold Spring Har-
McMahon, A.P., Ingham, P.W., and Tabin, C.J. (2003). Developmen-
tal roles and clinical significance of hedgehog signaling. Curr. Top.
Dev. Biol. 53, 1–114.
Megason, S.G., and McMahon, A.P. (2002). A mitogen gradient of
dorsal midline Wnts organizes growth in the CNS. Development
Mulieri,P.J., Okada, A.,Sassoon, D.A.,McConnell,S.K.,andKrauss,
R.S. (2000). Developmental expression pattern of the cdo gene.Dev.
Dyn. 219, 40–49.
Mulieri, P.J., Kang, J.S., Sassoon, D.A., and Krauss, R.S. (2002). Ex-
pression of the boc gene during murine embryogenesis. Dev. Dyn.
St-Jacques, B., Dassule, H.R., Karavanova, I., Botchkarev, V.A., Li,
hon, A.P. (1998). Sonic hedgehog signaling is essential for hair de-
velopment. Curr. Biol. 8, 1058–1068.
Stamataki, D., Ulloa, F., Tsoni, S.V., Mynett, A., and Briscoe, J.
(2005). A gradient of Gli activity mediates graded Sonic Hedgehog
signaling in the neural tube. Genes Dev. 19, 626–641.
Tian, H., Jeong, J., Harfe, B.D., Tabin, C.J., and McMahon, A.P.
(2005). Mouse Disp1 is required in sonic hedgehog-expressing cells
for paracrine activity of the cholesterol-modified ligand. Develop-
ment 132, 133–142.
Wijgerde, M.,McMahon, J.A., Rule,M., andMcMahon, A.P. (2002).A
direct requirement for Hedgehog signaling for normal specification
of all ventral progenitor domains in the presumptive mammalian spi-
nal cord. Genes Dev. 16, 2849–2864.
Wilkinson, D.G. (1992). In Situ Hybridization: A Practical Approach
(Oxford, New York: IRL Press at Oxford University Press).
Zhang, X.M., Ramalho-Santos, M., and McMahon, A.P. (2001).
Smoothened mutants reveal redundant roles for Shh and Ihh signal-
ing including regulation of L/R symmetry by the mouse node. Cell
Zhang, W., Kang, J.-S., Cole, F., Yi, M.-J., and Krauss, R.S. (2006).
Cdo functions at multiple points in the Sonic Hedgehog pathway
and Cdo-deficient mice accurately model human holoprosence-
phaly. Dev. Cell 10, this issue, 657–665.
The Cell-Surface Membrane Proteins Cdo and Boc
Are Components and Targets of the Hedgehog
Signaling Pathway and Feedback Network in Mice
Toyoaki Tenzen, Benjamin L. Allen, Francesca Cole, Jong-Sun Kang, Robert S. Krauss, and Andrew
Figure S1. In Situ Hybridization Analysis of Cdo and Boc Expression in the Developing
Mammalian Neural Tube
Cdo is expressed weakly, and transiently, in both the floor plate (arrow) and notochord
(arrowhead). Expression in the floor plate region at the forelimb level is first observed at
E9.0. Both floor plate and notochordal expression are lost by E11.5.
Figure S2. Comparison of Boc Expression in the E10.5 Neural Tube with Pax7
A, In situ hybridization analysis of Boc, Cdo, Ptch1, and Gli1 expression in the E10.5
mouse neural tube at the forelimb level. Black arrowheads indicate the dorsal limit of
Ptch1 expression, while brackets indicate the regions of overlap between Ptch1
expression and Boc and Cdo expression, respectively. B, In situ hybridization (Boc) and
immunofluorescence analysis (Pax7 and Olig2) indicates that Boc expression extends
ventral to the Pax7 domain, overlapping the distal limits of Shh signaling. White
arrowheads indicate the ventral limit of the Pax7 expression domain. Brackets mark the
region ventral to the Pax7 expression domain in which Boc expression is detected. C,
Wholemount in situ hybridization analysis of Boc, Cdo, Ptch1, and Gli1 expression in the
E10.5 mouse forelimb bud. Dotted lines indicate the anterior limit of Ptch1 expression;
brackets denote overlap between Ptch1 expression, and Boc and Cdo expression,
respectively. Note that the Boc panel is duplicated from Figure 1.
Figure S3. Boc and Cdo Induce Ectopic Expression of Nkx2.2+ and Olig2+ Cells in a
A, Detailed inspection of the far left panels from Figure 3A, including a separate panel
for Boc expression (GFP, green) demonstrates that all ectopic Nkx2.2+ cell progenitors
are also GFP+. B, Detailed inspection of the second panel from left in Figure 3B,
demonstrates that all ectopic Olig2+ cell progenitors also express Cdo.
Figure S4. Gli3R Induces Cell-Autonomous Expression of Pax7 Independent of Boc
Electroporation of chick neural tubes with a Gli3 repressor construct (Gli3R, left panels),
co-electroporation of Gli3R with full-length Boc (middle panels) or electroporation with
full-length Boc alone (right panels). Electroporated cells are indicated by GFP
expression (green). Immunofluorescent analysis of antibody staining denotes Pax7+ cell
progenitors (red). White arrowheads demarcate areas of Gli3R-mediated cell-
autonomous Pax7 expansion. White arrows indicate cell-non-autonomous expansion of
Pax7 expression in Boc-electroporated neural tubes.
Figure S5. Ectopic Expression of the FNIII(3) Domain of Either Cdo or Boc Is Not Download full-text
Sufficient to Promote Shh-Dependent Cell Fate Specification
Specification of vp3 (Nkx2.2+) and dorsal (Pax7+) neural progenitors is unaltered by
ectopic expression of either CdoFNIII(3)TMCD or BocFNIII(3)TMCD as visualized by
the production of GFP following electroporation of plasmid constructs in the chick neural