Segregation of cells entering into distinct differentiation
pathways is an essential step for morphogenesis. One such
morphogenetic event is neural crest segregation from the neural
tube. Neural crest precursors are born in the ectodermal
epithelium constituting the tip of the neural fold. During and
after the closure of the neural fold, neural crest cells emerge
from the future roof plate region of the neural tube, undergoing
epithelial-mesenchymal transformation. Many factors and
genes, such as Pax3 (Tremblay et al., 1995), slug (Nieto et al.,
1994), AP-2 (Zhang et al., 1996; Schorle et al., 1996), and
Wnt-1/3a (Ikeya et al., 1997) are expressed in the dorsal most
region of the neural tube, and have been shown to be involved
in the generation of neural crest cells. However, little is known
about the mechanism by which these cells can physically
disperse from the neural tube epithelium.
We previously demonstrated that multiple subtypes of
cadherin cell-cell adhesion molecules were dynamically
expressed during neural crest development in the chick
embryo. The early dorsal ectoderm expresses L-CAM (chicken
E-cadherin) (Thiery et al., 1984). During neural plate
invagination, the L-CAM expression is gradually replaced by
that of N-cadherin (Hatta and Takeichi, 1986). At the same
time, cadherin-6B (cad6B) begins to be expressed in the
invaginating neural plate, most strongly at the neural crest-
generating area (Nakagawa and Takeichi, 1995). In the neural
tube that has just closed, N-cadherin and cad6B are co-
expressed in the dorsal portion. When neural crest cells emerge
from the neural tube, these cadherins become scarcely
detectable, instead, cadherin-7 (cad7) appears (Nakagawa and
Takeichi, 1995). The cad7 expression persists during migration
of crest cells until their homing to appropriate sites; thereafter,
its expression becomes restricted to particular neural crest
Cadherins are homophilic adhesion receptors, and essential
for stable intercellular adhesion (Takeichi, 1995). An important
feature of cadherin interactions is that each subtype of the
Development 125, 2963-2971 (1998)
Printed in Great Britain © The Company of Biologists Limited 1998
During the emergence of neural crest cells from the neural
tube, the expression of cadherins dynamically changes. In
the chicken embryo, the early neural tube expresses two
cadherins, N-cadherin and cadherin-6B (cad6B), in the
dorsal-most region where neural crest cells are generated.
The expression of these two cadherins is, however,
downregulated in the neural crest cells migrating from the
neural tube; they instead begin expressing cadherin-7
(cad7). As an attempt to investigate the role of these
changes in cadherin expression, we overexpressed various
cadherin constructs, including N-cadherin, cad7, and a
dominant negative N-cadherin (cN390∆), in neural crest-
generating cells. This was achieved by injecting adenoviral
expression vectors encoding these molecules into the lumen
of the closing neural tube of chicken embryos at stage 14.
In neural tubes injected with the viruses, efficient infection
was observed at the neural crest-forming area, resulting in
the ectopic cadherin expression also in migrating neural
crest cells. Notably, the distribution of neural crest cells
with the ectopic cadherins changed depending on which
constructs were expressed. Many crest cells failed to escape
from the neural tube when N-cadherin or cad7 was
overexpressed. Moreover, none of the cells with these
ectopic cadherins migrated along the dorsolateral
(melanocyte) pathway. When these samples were stained
for Mitf, an early melanocyte marker, positive cells were
found accumulated within the neural tube, suggesting that
the failure of their migration was not due to differentiation
defects. In contrast to these phenomena, cells expressing
non-functional cadherins exhibited a normal migration
pattern. Thus, the overexpression of a neuroepithelial
cadherin (N-cadherin) and a crest cadherin (cad7) resulted
in the same blocking effect on neural crest segregation from
neuroepithelial cells, especially for melanocyte precursors.
These findings suggest that the regulation of cadherin
expression or its activity at the neural crest-forming area
plays a critical role in neural crest emigration from the
Key words: Adenoviral vector, Cadherin, Melanocyte precursor,
Neural crest, Cell adhesion, Cell migration, Mouse
Neural crest emigration from the neural tube depends on regulated cadherin
Shinichi Nakagawa and Masatoshi Takeichi*
Department of Biophysics, Faculty of Science, Kyoto University, Kitashirakawa, Sakyo-ku, Kyoto 606-8502, Japan
*Author for correspondence (e-mail: email@example.com)
Accepted 6 May; published on WWW 9 July 1998
cadherins preferentially binds like molecules. However, N-
cadherin binds neither cad6B nor cad7 (Nakagawa and
Takeichi, 1995). Cad6B shows a partial affinity to cad7, but
cells expressing these cadherins segregate from one another
within their chimeric aggregates, suggesting that their
heterophilic interaction is less stable than the homophilic
interaction of each subtype (Nakagawa and Takeichi, 1995).
Based on these observations, we proposed that the changes in
cadherin expression during neural crest development play a
role in the segregation of the neural crest cells from the neural
tube. For example, downregulation of N-cadherin and cad6B
in the crest cells may be a prerequisite for their detachment
from neural tube cells.
To test the above idea, we sought to perturb the cadherin
expression in neural crest cells by overexpressing various
cadherin constructs in the chicken embryonic neural tube by
use of adenoviral expression vectors. This approach was
technically successful; we could efficiently label not only
dorsal neural tube cells but also crest cells with ectopic
cadherin molecules. Our results show that the overexpression
of N-cadherin or cad7 suppressed the emergence of neural crest
cells from the neural tube, but the migration of crest cells that
had already detached from the neural tube was not affected by
the ectopic cadherin expression. These findings support the
idea that the regulated cadherin expression, occurring during
neural crest emigration, is a necessary step for this
MATERIALS AND METHODS
To raise antibodies against the extracellular domain of chicken cad6B
or cad7, we prepared chimeric proteins consisting of this domain and
the human immunoglobulin Fc region: A cDNA fragment encoding
the N-terminal 605 amino acids of cad6B or the N-terminal 597 amino
acids of cad7 was amplified by PCR, and cloned into pEF-Fc (Suda
and Nagata, 1994). These plasmids were introduced into COS-7 cells
by electroporation, and the cells were incubated for 24 hours in
Dulbecco’s MEM (DMEM, Nissui, Japan) supplemented with 10%
fetal calf serum. They were then washed twice with a Hepes-buffered
(pH 7.4) balanced salt solution (HBSS), and further incubated in a
serum-free DMEM for 48 hours. The fusion proteins were secreted
into the conditioned culture medium, and they were purified on
protein-A Sepharose CL4B columns (Pharmacia Biotech). We usually
obtained 1 µg of purified proteins from 1 ml of the conditioned
medium. Mice (BDF1, the hybrid strain of C57/Bl6 and DBA2) were
immunized at 4-week intervals by intraperitoneal injection of the
purified proteins (50 µg/mouse) that had been emulsified with RIBI
adjuvant (Immunochem Research). The splenocytes of the mice were
then fused with P3-X63-Ag-U1 myeloma cells 3 days after the final
boost. Culture supernatants of the hybridomas thus isolated were
screened for reactivity to lysates of L-cells transfected with cad6B or
cad7 cDNA (Nakagawa and Takeichi, 1995); and by this procedure
two monoclonal antibodies (IgGs), termed CCD6B-1 and CCD7-1,
which specifically recognize cad6B and cad7, respectively, were
obtained. Rabbits were also immunized at 4-week intervals by
subcutaneous injection of the same fusion proteins (200 µg/rabbit)
emulsified with RIBI adjuvant (Immunochem Research), and a
polyclonal antiserum specific for cad6B, designated as anti-6B, was
Immunostaining and immunoblotting
Antibodies used were as follows: rat monoclonal antibody NCD-2
(Hatta and Takeichi, 1986); mouse monoclonal antibody HNK1
(Becton Dickinson 347390); mouse monoclonal antibody M2 (Kodak
IB13010), rabbit polyclonal antibody against Mitf (Mochii et al.,
1998); Cy3-conjugated anti-mouse, -rat, and -rabbit IgGs (Chemicon
AP124C, AP136C, AP132C); FITC-conjugated anti-mouse IgM
(ZYMED, 62-6811); HRP-conjugated anti-mouse and -rabbit IgGs
(Amersham NA9310, NA934); biotinylated anti-mouse and -rabbit
IgGs (Amersham RPN1001, RPN1004); streptavidin-labeled FITC
(Amersham RPN1232); Texas Red-conjugated anti-rabbit IgG
(CAPPEL 55675). Embryos were fixed with 4% paraformaldehyde in
HBSS at 4°C for 2 hours.
For cryosections, embryos were immersed in a graded series of
sucrose solutions (12%, 15%, 18% sucrose in HBSS), embedded in
Tissue Tek (Miles), and frozen in liquid nitrogen. Cryostat sections of
10 µm thickness were collected, and dried on silane-coated glass
slides. Sections were rehydrated in TBS (50 mM Tris, 150 mM NaCl,
1 mM CaCl2), and permeabilized by incubation in −20°C methanol
for 5 minutes. Non-specific binding of antibodies was blocked by
immersing the sections in TBS containing 5% skim milk (Difco).
When using HNK1 as the primary antibody, TBS containing 2% BSA
(Sigma) was used as a blocking solution instead. For double-label
immunostaining, sections were first incubated with a primary
antibody for 1 hour, followed by incubation with an appropriate
secondary fluorescence-labeled antibody for 30 minutes, and finally
with a second set of primary and secondary antibodies. For staining
with HNK1, the incubation time was extended to 16 hours.
Fluorescence was visualized with a epifluorescence microscope
(Zeiss) or a laser-scanning confocal microscope (Bio-Rad).
For whole-mount immunostaining, embryos were fixed in 4%
paraformaldehyde/HBSS for 2 hours. Fixed embryos were washed
twice with HBSS, and incubated for 6 hours in TBS containing 0.2%
Triton X-100, 5% heat-inactivated FBS, and 0.5% peroxide. Embryos
were then incubated overnight with an appropriately diluted primary
antibody, and washed six times (1 hour each time) with TBS
containing 0.2% Triton X-100. HRP-conjugated antibodies were used
as secondary antibodies, and the antibody incubation and washing
procedures were the same as for the primary antibodies. After having
been visualized with DAB and photographed under a binocular
microscope (Nikon SMZ-U), the samples were sectioned with a
microslicer (D. S. K. DTK-1000, Kyoto) at a thickness of 100 µm,
cleared with 75% glycerol, and mounted on a glass slide. These
sections were examined under a Zeiss microscope using Nomarski
For immunoblotting, embryos at stage 14 were lysed with
Laemmli’s sample buffer and boiled for 5 minutes. The samples were
run in a 7.5% SDS-polyacrylamide gel, and transferred onto a
nitrocellulose membrane. The blots incubated with appropriate
antibodies were processed for chemiluminescence with the ECL
detection system (Amersham).
The adenoviral shuttle vectors encoding various cadherin cDNAs
were constructed as follows: The whole coding regions of N-cadherin,
cN390∆, and cad7 were amplified by PCR using Z10T6, cN390∆/BS,
and pBSSK-nk2, respectively as a template (Hatta et al., 1988;
Fujimori and Takeichi, 1993; Nakagawa and Takeichi, 1995). For
cN/CBR(−), a DNA fragment corresponding to the first 855 amino
acids of chicken N-cadherin was amplified by PCR. All primers were
designed to introduce a SplI site followed by Kozak’s consensus
sequence in the 5′ terminal region, and BglII and HpaI sites just before
the stop codon. The amplified fragments were digested with SplI and
HpaI, and cloned into the SplI-HpaI site of pCMV-pA (provided by
K. Moriyoshi), which were successively digested with BglII, and an
adapter (see below) was inserted into this site to yield pCMV-
pCMV-cN/CBR(−)/FLAG-pA. The adapter was made by annealing
the following two primers
after phosphorylation: 5′-
S. Nakagawa and M. Takeichi
2965 Cadherin-dependent neural crest emigration
ctgttcattcgccggcg-3′. These adapter sequences contained the FLAG
epitope DYKDDDDK (Hopp et al., 1988) and NotI site. L-cells were
transiently transfected with these pCMV constructs, and the cells were
then stained with the anti-FLAG antibody M2 to ensure proper protein
expression. An adenovirus shuttle vector plasmid pAdV-CA-pA was
made by inserting the CA promoter sequence (Niwa et al., 1991),
followed by insertion of a poly(A) signal into the EcoRV site of pAdV
(Moriyoshi et al., 1996). The whole coding sequences including the
FLAG epitope were cut from the pCMV constructs with SplI and NotI,
blunted using the Klenow fragment, and cloned into the HpaI site of
pAdV-CA-pA to yield pAdV-CA-cN/FLAG-pA,
cN390∆/FLAG-pA, pAdV-CA-c7/FLAG-pA, pAdV-CA-cN/CBR(−)/
Construction of recombinant adenoviruses
AdV-CA-lacZ, adenovirus expressing β-galactosidase driven by a
CAG promoter, was a kind gift from K. Moriyoshi at Kyoto
University. Construction of recombinant viruses was performed
according to the methods described previously (Moriyoshi et al.,
1996). Recombinant adenoviruses were obtained by homologous
recombination in HEK 293 cells (ATCC CRL 1573), which were
maintained in DMEM supplemented with 10% FBS. HEK 293 cells
cultured in 6-well plates (Nunc) were co-transfected with viral
genome fragments (0.2 µg) and linearized adenoviral shuttle vector
plasmids (1 µg) using LipofectAMINE (GIBCO). On the next day,
the cells were divided and placed into collagen-coated 24-well plates
(IWAKI, Japan). Ten days later, some wells were full of dead cells,
caused by viral propagation, of which debris was screened for proper
protein expression by immunostaining with the anti-FLAG antibody
M2. We thus obtained AdV-Ncad, AdV-cN390∆, AdV-cad7, and
AdV-cN/CBR(−), expressing these cadherins under the control of the
CAG promoter (Niwa et al., 1991). The recombinant adenoviruses
were amplified, and purified by CsCl2 step-gradient centrifugation
(Kanegae et al., 1994). FBS was added to the purified adenovirus
solutions at a final concentration of 10%. Aliquots of the virus
solutions were stored at −80˚C until used.
In ovo injection of adenoviruses
Fertilized eggs of chickens from Yamagishi’s Farm (Kyoto, Japan)
were incubated at 38°C for 48 hours to allow embryos to develop to
stage 14 (Hamburger and Hamilton, 1951). The posterior neuropore
of the chicken embryos from this farm closed earlier than that of
White Leghorn embryos, which was an essential requirement to obtain
higher infection efficiency. The tip of pulled glass micropipettes was
broken by forceps and front-filled with adenovirus solution. Fast
Green (Sigma) was added to the virus solution at a final concentration
of 0.2% to help visualization of the injected solution. The injection
pipette was inserted obliquely into the neural tube at the caudal-most
level by means of a micromanipulator (Narishige). The virus solution
up to 100 nl was very slowly injected by air pressure until the solution
filled the entire lumen of the neural tube. In most cases, diffused virus
solution was seen to reach the level of the midbrain. Occasionally the
virus solution leaked from the neural tube. In such cases, the embryos
were discarded. The injected embryos were further incubated at 38°C
until they were fixed for immunostaining.
Changes in cadherin expression during neural crest
We determined the protein expression patterns for three
cadherins, N-cadherin, cad6B, and cad7, during neural crest
development in the chicken embryos, since previous
information was limited to mRNA expression for the latter two.
Two mouse monoclonal antibodies, CCD6B-1 and CCD7-1,
which specifically recognize chicken cad6B and cad7,
respectively, were prepared; and a rabbit antiserum specific for
cad6B, designated as anti-6B, was also generated. In western
blotting, CCD6B-1/anti-6B and CCD7-1 specifically detected
a major 120 kDa protein and 105 kDa protein, respectively
(Fig. 1). These antibodies as well as NCD-2, specific for
chicken N-cadherin (Hatta and Takeichi, 1986), were used for
subsequent immunostaining studies. HNK1 antibody was also
used to identify migrating neural crest cells (Tucker et al.,
1984; Bronner-Fraser, 1986).
In the trunk of embryos at stage 16, N-cadherin was
expressed throughout the neural tube, although its expression
was reduced at the dorsal-most region where neural crest cells
were being generated (Fig. 2A) (Hatta and Takeichi, 1986;
Akitaya and Bronner-Fraser, 1992). In this dorsal-most portion,
cad6B was strongly expressed (Fig. 2B), overlapping the weak
N-cadherin expression. When neural crest cells had left the
neural tube, both of these cadherins were downregulated. They
instead expressed cad7 (Fig. 2C). Double-staining for cad7 and
the HNK1 antigen showed that most of HNK1-positive cells
were also cad7 positive (Fig. 2C,D), suggesting that the
majority of migrating crest cells expressed this cadherin. At
subsequent stages, e.g., at stage 22, cad7 and the HNK1 antigen
became co-expressed by cells migrating towards the space
between the dermamyotome and overlying ectoderm (Fig. 2E,F
arrows), which were likely melanocyte precursors entering the
dorsolateral pathway (Erickson et al., 1992). The cad7 was also
expressed in the neural tube at the boundary zone between the
alar and basal plates. These expression patterns of cadherin
proteins were basically identical to those of their mRNAs
previously reported (Nakagawa and Takeichi, 1995), although
the immunostaining turned out to be more sensitive than the
mRNA in situ hybridization, resulting in the detection of more
Fig. 1. Immunoblot detection of cad6B and cad7 from a lysate of
stage 14 embryos. CCD7-1 detected the 105 kDa cad7 band.
CCD6B-1 and anti-6B detected the 120 kDa cad6B band. Lower
bands are probably degradation products.
For closer observations of the changing process of the
cadherin expression pattern, we employed confocal microscopy
(Fig. 2G-I). Staining for N-cadherin and cad6B indicated that
both of them were distributed throughout the lateral contact
sites between dorsal neural tube cells, although more cadherin
proteins were accumulated at the luminal side of the neural tube,
where the adherens junctions are localized between
neuroepithelial cells. Major cad7 signals were detected only in
cells that had detached from the neural tube (Fig. 2H,I, arrows),
although some faint signals were also detectable in the adherens
junction area (Fig. 2H,I, arrowheads). To confirm this feature
of cad7 expression, we examined thick sections at the dorsal-
most region of the neural tube stained for cad7 by conventional
immunofluorescence microscopy at higher magnification (Fig.
2J,K). Cad7 signals were indeed already present
in the neural tube, but these were confined to the
luminal junctional sites, never extending to the
main body of neuroepithelial cells. Strong cad7
signals covering the entire cell body appeared
only after crest cells had become isolated from the
Ectopic β-galactosidase or cadherin
expression induced by adenoviral
To investigate the role of the above changes in
development, we designed experiments to induce
over or ectopic expression of various cadherin
constructs in neural crest precursors as well as in
surrounding neural tube cells by use of adenoviral
expression vectors. To test adenoviral vectors for
usefulness in our system, we injected those
carrying the β-galactosidase gene (provided by K.
Moriyoshi, Kyoto University), whose expression
was under a control of the CAG promoter (Niwa
et al., 1991), into the lumen of the neural tube of
stage 14 embryos. The injected embryos were
fixed at various times, and subjected to β-gal
staining as whole-mount samples. In embryos
fixed at 4 hours, no ectopic expression of β-gal
was observed (Fig. 3A). At 10 hours, however,
enzymatic activity of β-gal was detected in cells
scattered around the dorsal portion of the
embryos as well as in those aligned along their
dorsal midline (Fig. 3B). A similar pattern of
positive cells was observed at 16 hours also,
although they were more intensely labeled (Fig.
3C). At 24 hours, dorsal root ganglia (DRG)
became positive (Fig. 3D). These results indicate
that the virus-derived protein expression was
initiated at a time between 4 and 10 hours after
viral injection and that its level increased up to 24
hours of incubation. Thereafter, the expression
level gradually decreased presumably due to a
dilution of this replication-incompetent virus by
cell proliferation (data not shown).
For more precise localization of β-gal-positive
cells, embryos at 24 hours after viral injection,
which had reached stage 20, were sectioned and
immunostained for β-gal. In the neural tube,
during neural crest
infection was observed in various regions of the tube; however,
the most reproducible, intense infection occurred at the dorsal-
most portion where premigratory neural crest cells reside (Fig.
3E). This infection resulted in β-gal labeling of migrating
neural crest cells also, as confirmed by double-staining for β-
gal and the HNK1 epitope: for example, at the hindlimb level,
a cluster of β-gal-positive cells was found migrating down the
typical ventral pathway for neural crest migration (Fig. 3E);
and these cells were also HNK1 positive (Fig. 3F). Some of
the myotome cells occasionally expressed β-gal, possibly due
to some leakage of viruses from the injected sites. In these
experiments, more than half of the HNK1-positive neural crest
cells expressed the lacZ gene, although the infection efficiency
varied with the embryos and experiments.
S. Nakagawa and M. Takeichi
Fig. 2. Immunostaining detection of cadherins in neural crest cells.
(A-D,G-K) Stage-16 embryos; (E,F) stage 22 embryos. (A) N-cadherin. (B) cad6B.
(C-F) Double-staining for cad7 (C,E) and HNK1 (D,F). Arrows in E and F point to
cells entering the dorsolateral pathway, which express both cad7 and HNK1 antigen.
dm, dermamyotome. (G-I) Double-staining for N-cadherin (red) and cad6B (green)
(G), N-cadherin (red) and cad7 (green) (H), and cad6B (red) and cad7 (green) (I).
Photographed under confocal microscopy. In H and I, arrows indicates cells that
have just detached from the neural tube, which express cad7 only, and arrowheads,
sites in which both cadherins are detected as yellow signals. (J,K) Nomarsky optics
micrographs of the dorsal-most region of the neural tube (J), and cad7 staining of
the same section (K). Note that cad7 signals are localized in cells at the luminal side
of the roof plate in addition to a cell located outside the neural tube.
2967 Cadherin-dependent neural crest emigration
Next, we generated recombinant viruses expressing N-
cadherin (AdV-Ncad) and cad7 (AdV-cad7), whose carboxy
terminus was attached to the FLAG tag (Fig. 4), and injected
them into stage 14 embryos, as described above. At 24 hours
after the injection, the embryos were stained for the tag
epitope. As expected from the above experiments, migrating
neural crest cells as well as cells localized in the roof plate of
the neural tube were intensely labeled with the anti-FLAG
antibody (Fig. 3G,H). Strong expression of the ectopic
cadherins was observed not only in the roof plate but also in
migrating crest cells (Fig. 3G,H arrows).
Alteration of neural crest behavior by ectopic
In the above experiments, we did not notice significant effects
of the ectopic cadherin expression on neural crest behavior.
However, when the injected embryos were incubated longer, a
dramatic difference in neural crest patterning became evident
between the embryos expressing β-gal and the full-length
cadherin constructs. In embryos incubated for 30 hours after
the injection of viruses carrying the β-gal gene, cells
expressing this gene were detected not only in DRG but also
along the dorsolateral pathway, as revealed by whole-mount
staining of the embryos (Fig. 5A) as well as by sectioning of
them (Fig. 5B, arrows). However, in embryos expressing
ectopic N-cadherin or cad7, no infected cells were found
migrating along the dorsolateral pathway (Fig. 5C-F), as
judged by staining with anti-FLAG antibody. Notably, a
number of infected cells had accumulated in the midline of the
neural tube, some of which had invaded the lumen of the neural
tube (Fig. 5D,F arrowheads). In these embryos, the number of
FLAG-positive cells that had been incorporated into the DRG
also tended to be slightly reduced.
As controls for the above experiments, we constructed
recombinant viruses expressing cN390∆ (AdV-cN390∆), and
cN/CBR(−) (AdV-cN/CBR(−)). cN390∆ is a mutant N-
cadherin with a deletion at the extracellular domain, but
having the normal intracellular domain; it exhibits a dominant
negative effect on the activity of endogenous cadherins
(Kintner 1992; Fujimori and Takeichi, 1993). cN/CBR(−) is
another mutant form of N-cadherin lacking only the catenin-
binding region. When these constructs were injected, cN390∆
expression did not induce any abnormal clustering of infected
cells in the dorsal neural tube, allowing many of them to
migrate along the dorsolateral pathway (Fig. 5G,H). Cells
expressing cN/CBR(−) showed an intermediate phenotype;
some of them migrated along the dorsolateral pathway, but a
considerable number of them remained localized in the
midline of the neural tube (Fig. 5I,J).
To confirm the ectopic protein expression for N-cadherin
and cad7 in the above experiments, we immunostained
embryos incubated for 30 hours after viral injection for these
proteins. As expected, both proteins were overexpressed in the
roof plate region of the neural tube (Fig. 6). As mentioned
above, in these neural tubes, some roof plate cells invaded their
lumen (Fig. 6A,B arrows). These findings demonstrate that
overexpressed N-cadherin or cad7 inhibited the escape of a
particular population of neural crest precursors from the neural
tube. This population likely includes melanocyte precursors, as
only these crest derivatives are capable of migrating in the
dorsolateral path (Erickson and Goins, 1995).
Differentiation is not inhibited by ectopic cadherin
The above inhibition of neural crest emigration along the
dorsolateral pathway could have been brought about by a
suppression of their differentiation specific for this particular
pathway, that is, melanocyte differentiation (LeDouarin, 1982).
To test this possibility, we examined the expression of Mitf, an
early melanocyte marker (Opdecamp et al., 1997; Y.
Wakamatsu, personal communication), in embryos expressing
ectopic N-cadherin or cad7. Embryos were incubated for 48
hours after viral injection to allow a sufficient expression of
Mitf. In control embryos injected with the β-gal vector, Mitf-
positive future melanocytes were observed along the normal
dorsolateral pathway (Fig. 7A). Many of them co-expressed
Mitf and β-gal, as revealed by double-staining for these markers
(Fig. 7A,B), indicating that the ectopic β-gal expression and
also virus infection itself did not affect the Mitf-positive cell
migration. In contrast, in embryos injected with N-cadherin- or
cad7-expressing viruses, cells co-expressing Mitf and FLAG
Fig. 3. Adenovirus-mediated exogenous protein expression in neural
crest cells. (A-D) Dorsal view at the trunk region of embryos injected
with β-gal-expressing adenoviruses, and fixed after 4 hours (A), 10
hours (B), 16 hours (C), and 24 hours (D). The samples were
processed for X-gal staining. (E,F) Double-staining for β-gal (E) and
HNK1 (F) in a section at the hindlimb level of an embryo incubated
for 24 hours after viral injection. Virus infection is observed in
premigratory and migrating neural crest cells that are
immunoreactive for HNK1 as well as in the neural tube.
(G,H) Embryos were injected with AdV-Ncad (G) or AdV-cad7 (H),
and stained for FLAG epitope 24 hours after the injection. Note that
cells strongly expressing exogenous cadherins are migrating along
the ventral pathway (arrowheads).
had accumulated in the dorsal lumen of the neural tube (Fig.
7C,E arrows), although some of the Mitf-positive cells were
FLAG negative (Fig. 7C-F, arrowheads). Some Mitf-positive
cells were also scattered dorsal to the neural tube as in control
embryos (Fig. 7C,E), but none of them expressed exogenous
cadherins (Fig. 7D,F). In embryos injected with cN390∆-
expressing viruses, Mitf-positive cells behaved like those in the
control embryos (Fig. 7G,H). These results suggest that
melanocyte differentiation took place in cells with ectopic N-
cadherin or cad7 but that these cells were unable to escape from
the neural tube and instead invaded the lumen of the tube.
Neural crest emigration from the neural tube involves multiple
processes (Tosney, 1978; Erickson and Weston, 1983; Erickson
and Perris, 1993). Disruption of the attachment of premigratory
crest cells to neuroepithelial cells is one of these processes. The
neural tube cells are connected to one another along their
lateral surfaces. At their luminal side which corresponds to the
apical side of simple epithelia, the adherens junction (AJ)
develops, which probably represents the major sites for the
interconnection of neuroepithelial cells (Duband et al., 1988).
Cadherin adhesion molecules are localized throughout the
lateral contact sites between these cells, being highly enriched
in the AJs. For neural crest cells to escape from the tube, these
cell-cell contacts must be disrupted. We can imagine at least
two possible mechanisms for the disruption of neuroepithelial
junctions. One is a physiological mechanism to down-regulate
the AJ function. Cadherin-based junctions are believed to be a
machinery whose activity can be regulated (Takeichi et al.,
1993). Cadherins are associated with catenins, and this
association is essential for the normal activity of the former
(Aberle et al., 1996; Barth et al., 1997). Physiological
modification of catenins could de-stabilize cadherin-mediated
cell-cell adhesion (Takeichi, 1993). It is of note that cells
expressing chimeric proteins of E-cadherin and α-catenin, to
which β-catenin is unable to bind, are more stably connected
with one another than those expressing the normal
cadherin/catenin complex, suggesting that β-catenin is a
regulator for cadherin activity (Nagafuchi et al., 1994). Such
mechanisms could serve to enhance neural crest detachment
from neuroepithelial cells.
S. Nakagawa and M. Takeichi
Fig. 4. Schematic representation of full-length or mutant cadherin
constructs of recombinant adenovirus. All constructs have the FLAG
epitope at their carboxyl terminus. TM, transmembrane domain.
Fig. 5. Behavior of neural crest cells ectopically expressing various
cadherin constructs. Embryos were injected with adenoviruses
expressing β-gal (A,B), N-cadherin (C,D), cad7 (E,F), cN390∆
(G,H), or cN/CBR(−) (I,J), and fixed 30 hours after the injection. The
specimens were stained with anti-β-gal (A,B) or anti-FLAG M2
antibody (C-J). (A,C,E,G,I) Dorsal view of whole-mount embryos.
(B,D,F,H,J) The whole-mount samples were sectioned at a thickness
of 100 µm, and photographed under Nomarsky optics. Arrows
indicate cells migrating along the dorsolateral pathway; note the
absence of labeled cells on this pathway in C-F. Arrowheads point to
infected cells localized in the roof plate, some of which are invading
the neural tube lumen.
2969 Cadherin-dependent neural crest emigration
The second mechanism concerns cadherin type switching,
which occurs during neural crest development. Dorsal-most
neuroepithelial cells express N-cadherin and cad6B, while
migrating crest cells express cad7. The expression of cad7 by
neural crest cells seems to begin in advance of their escape
from the tube, but its substantial expression appeared only after
they had left the tube. In these cells, N-cadherin and cad6B
were no longer detected, indicating that N-cadherin and cad6B
are replaced with cad7, when crest cells are leaving the neural
tube. Previous observations showed that heterophilic cad7-
cad6B binding is less stable than homophilic cad6B-cad6B or
cad7-cad7 binding and also that cad7 has little affinity to N-
cadherin (Nakagawa and Takeichi, 1995). Therefore, the
observed replacement of cadherins during neural crest
emergence is expected to facilitate their detachment from
neuroepithelial cells. It should be stressed that the loss of N-
cadherin and cad6B from premigratory crest cells is, in itself,
theoretically sufficient to facilitate this detachment process.
Our results of N-cadherin or cad7 overexpression support
well the idea that the control of cadherin activity or expression
is essential for neural crest emigration. When these cadherins
were overexpressed in the dorsal neural tube, emigration of
melanocyte precursors was prohibited. In considering the
mechanism of this phenomenon, however, we were concerned
that the cadherin overexpression might had affected
melanocyte differentiation, but not adhesion processes. A
melanocyte marker, Mitf, was found to be expressed by the
neural crest cells stuck in the neural tube, suggesting that the
melanocyte differentiation normally proceeded during the
observed events. Another concern along a related line would
be a possible involvement of Wnt signaling. Two Wnt genes,
Wnt-1 and Wnt-3a, are expressed in the midline of the mouse
neural tube; and double mutation for these genes causes
inhibition of neural crest generation and migration (Ikeya et al.,
1997). It is known that cadherin overexpression depletes β-
catenin, a component essential for Wnt signaling, and therefore
blocks this signaling system (Heaseman et al., 1994; Fagotto
et al., 1996; Sanson et al., 1996). However, cN390∆, which can
bind β-catenin as well as intact cadherins, did not suppress the
dorsolateral migration of the crest cells. Moreover, the
expression of cN/CBR(−), which lacks the β-catenin binding
domain, exhibited a partial suppression of crest cell migration.
These findings suggest that the failure of neural crest migration
did not depend on inhibition of the β-catenin-associated Wnt
signaling. The fact that the effect of cadherin overexpression
mimicked that of the Wnt mutation is interesting, as it raises
the possibility that a role of Wnt signals in neural crest
development is to antagonize cadherin function. In this context,
it should be interesting to note that E-cadherin expression is
upregulated in the roof plate of the midbrain in Wnt-1 mutant
mice (Shimamura et al., 1994).
Thus, a likely mechanism underlying the above phenomenon
is that the overexpressed cadherins directly modulated the
neural crest adhesion properties. Although the endogenous N-
cadherin and cad7 were expressed in different patterns by the
Fig. 6. Ectopic expression of N-cadherin and cad7 proteins in
adenovirus-injected neural tubes. Embryos were injected with AdV-
Ncad (A) or AdV-cad7 (B), fixed after 30 hours, and doubly stained
for N-cadherin (red) and cad7 (green). Arrows indicate cells invading
the lumen of the neural tube. Note that the dorsal-most portion of the
neural tube most strongly expresses these ectopic cadherins.
Fig. 7. Expression of Mitf. Embryos were injected with adenoviruses
expressing β-gal (A,B), N-cadherin (C,D), cad7 (E,F), and cN390∆
(G,H), and fixed after 48 hours. The samples were sectioned, and
doubled-stained for Mitf (A,C,E,G) and β-gal (B) or FLAG (D,F,H).
Small arrows point to cells expressing both Mitf and β-gal or FLAG.
Large arrows indicate Mitf-positive cells clustered in the dorsal
lumen of the neural tube; such cells are never observed in embryos
injected with AdV-β-gal (A) or AdV-cN390∆ (G). Arrowheads in C-
F point to cells expressing Mitf but not the FLAG epitope.
original neural crest, the ectopic expression of them produced
a novel profile of cadherin expression; i.e., a single type of
cadherin, either N-cadherin or cad7, was expressed throughout
from the neural tube to migrating crest cells, abolishing the
cadherin type-switching pattern. This altered cadherin
expression likely caused a persistent sticking of neural crest
precursors to the neuroepithelial layer, interfering with either
or both of the above-proposed mechanisms for neural crest
detachment from the neural tube: that is, the ectopic cadherins
expressed by both neuroepithelial cells and crest precursors
might have prevented them from separating, or an excess
amount of cadherins accumulated at the dorsal neural tube
might have competed with a mechanism to down-regulate
cadherin-mediated intercellular adhesion. One could propose a
third mechanism, i.e., that the overexpression of cadherins
affected cell motility, thereby inhibiting crest emigration. This
is unlikely, because crest cells with a high level of ectopic
cadherins still normally migrated (Fig. 3G, H arrows). In these
cells, the cadherin overexpression was probably induced after
they had escaped from the neural tube (see below).
It remains unsolved as to why crest cells that had failed to
escape from the neural tube eventually accumulated in the
lumen of the tube. Essentially all Mitf-positive cells expressing
ectopic cadherins were excluded from the neuroepithelial layer
into its lumen. Interestingly, the excluded cell clusters
contained a population of cells not expressing the ectopic
cadherins. This finding implies that the exclusion was not
induced by cadherin overexpression in neural crest precursors,
but rather by that in the neuroepithelial cells. A possible
mechanism underlying this phenomenon is the following. The
regulated expression of cadherins in the neuroepithelium may
be required for maintaining its integrity, such as apicobasal
polarity, and this could be essential for the neuroepithelium to
send out neural crest cells in the right direction. Cadherin
overexpression might disorganize the polarized neuroepithelial
structures, leading some of crest cells to migrate in a wrong
direction. If this is the case, this possibility also must be
considered as one of the mechanisms by which melanocyte
precursor emigration was inhibited. It should also be pointed
out that a similar luminal accumulation of crest cells was
observed when their migration onto the extracellular matrix
had been blocked (Bilozur and Hay, 1988). In this case, we
should assume that crest cells were rather passively brought
down into the lumen. For the presence case, such passive
mechanism also should be considered.
Why was behavior of melanocyte precursors selectively
affected in our experiments? The crest cell migration along the
ventral pathway was not dramatically inhibited by the cadherin
overexpression, although some small effects were detectable.
In our experimental protocol, stage 14 embryos were injected.
During the initial 20 hours following the injection, only cells
that would migrate through the ventral pathway are generated;
and after about 24 hours (stage 21), melanocyte precursors first
appear in the trunk areas (Erickson et al., 1992), therefore
leaving the neural tube relatively late (Henion and Weston,
1997). As the virus-derived cadherin expression gradually
increased during 24 hours, the maximum overexpression of
these cadherins happened to coincide with the onset of the
generation of melanocyte precursors; i.e., among various
neural crest precursors being generated, the future melanocytes
and surrounding neuroepithelial cells must have undergone the
strongest exogenous cadherin expression while the former
were still localized in the neural tube. These considerations
explain why the melanocyte pathway was most severely
affected. In many of the crest cells that migrated along the
ventral pathway, ectopic cadherin expression would have been
induced only after their escape from the neural tube, as there
was a lag period in the viral vector expression.
In the present studies, we used two control N-cadherin
constructs, cN390∆ and cN/CBR(−). Although the former
showed no effect on neural crest emigration processes, the
latter one exhibited a partial effect. The adhesive function of
this cadherin construct was not extensively analyzed. Since this
construct lacks the catenin-binding region, its activity as a
cadherin adhesion molecule should be not as good as that of
the intact N-cadherin. Nevertheless, the juxtamembrane region
of the cadherin cytoplasmic domain, left intact in the
cN/CBR(−), was shown to have some biological activity
(Kintner, 1992; Riehl et al., 1996), implying that this construct
still possesses certain functions to regulate cell adhesion. The
effect of its expression on neural crest emigration is possibly
based on such partial adhesion-related functions. However,
cN390∆ did not show any inhibition of neural crest emigration.
This result is consistent with the fact that cN390∆ cannot
function as an adhesion molecule, because of the deletion in the
extracellular domain. cN390∆ is known to exhibit a dominant
negative effect on endogenous cadherin-mediated adhesion in
certain epithelial cell lines (Fujimori and Takeichi, 1993). Such
inhibitory effect was not clearly detected in the present system.
Perhaps, the expression level of cN390∆ was not sufficient to
inhibit endogenous cadherin activity, or the neural tube system
was not sensitive enough to detect adhesion defects.
The present study did not focus on the role of cadherins in
the later migration processes of crest cells including homing,
mainly because of the technical reason that the virus-derived
cadherin expression was transient, diminishing during the
migration. We previously hypothesized that cadherins
expressed by migrating crest cells may play roles in the
regulation of their homing behavior (Nakagawa and Takeichi,
1995), but this idea remains to be tested in future studies.
We thank Dr S. Nagata for plasmid pEF-Fc, and helpful advice on
fusion protein production; Drs K. Moriyoshi and S. Nakanishi for
their kind advice on adenovirus experiments and for the plasmids
necessary for generating recombinant adenoviruses; Dr Y. Takahashi
for advice on manipulation of chicken embryos; Drs Y. Wakamatsu
and N. Funayama for advice on FLAG tag and data analysis; Dr M.
Mochii for anti-Mitf antibody; and Drs S. Tsukita and H. Oda for
allowing us to use their confocal microscope. This work was
supported by a grant from the program Grants-in-Aid for Creative
Fundamental Research from the Ministry of Education, Science,
Sports, and Culture of Japan, and by a grant from the Human Frontier
Science Program. S.N. is a recipient of a Fellowship of the Japan
Society for the Promotion of Science for Junior Scientists.
Aberle, H., Schwartz, H. and Kemler, R. (1996). Cadherin-catenin complex:
protein interactions and their implications for cadherin function. J. Cell
Biochem. 61, 514-523.
Akitaya, T. and Bronner-Fraser, M. (1992). Expression of cell adhesion
molecules during initiation and cessation of neural crest cell migration. Dev.
Dyn. 194, 12-20.
S. Nakagawa and M. Takeichi
2971 Cadherin-dependent neural crest emigration
Bilozur, M. E. and Hay, E. D. (1988). Neural crest migration in 3D
extracellular matrix utilizes laminin, fibronectin, or collagen. Dev. Biol. 125,
Barth, A. I. M., Näthke, I. S. and Nelson, W. J. (1997). Cadherins, Catenins
and APC protein: interplay between cytoskeletal complexes and signaling
pathways. Curr. Opin. Cell Biol. 9, 683-690.
Bronner-Fraser, M. (1986). Analysis of the early stages of trunk neural crest
migration in avian embryos using monoclonal antibody HNK1. Dev. Biol.
Duband, J. L., Volberg, T., Sabanay, I., Thiery, J. P. and Geiger, B. (1988).
Spatial and temporal distribution of the adherens-junction-associated
adhesion molecule A-CAM during avian embryogenesis. Development 103,
Erickson, C. A. and Weston, J. A. (1983). An SEM analysis of neural crest
migration in the mouse. J. Embryol. exp. Morph. 74, 97-118.
Erickson, C. A., Duong, T. D. and Tosney, K. W. (1992). Descriptive and
experimental analysis of the dispersion of neural crest cells along the
dorsolateral path and their entry into ectoderm in the chick embryo. Dev.
Biol. 151, 251-272.
Erickson, C. A. and Perris, R. (1993). The role of cell-cell and cell-matrix
interactions in the morphogenesis of the neural crest. Dev. Biol. 159, 60-74.
Erickson, C. A. and Goins, T. L. (1995). Avian neural crest cells can migrate
in the dorsolateral path only if they are specified as melanocytes.
Development 121, 915-924.
Fagotto, F., Funayama, N., Gluck, U. and Gumbiner, B. M. (1996). Binding
to cadherins antagonizes the signaling activity of β-catenin during axis
formation in Xenopus. J. Cell Biol. 132, 1105-1114.
Fujimori, T. and Takeichi, M. (1993). Disruption of epithelial cell-cell
adhesion by exogenous expression of a mutated nonfunctional N-cadherin.
Mol. Biol. Cell 4, 37-47.
Hamburger, V. and Hamilton, H. L. (1951). A series of normal stages in the
development of the chick embryo. J. Morphol. 88, 49-92.
Hatta, K., Nose, A., Nagafuchi, A. and Takeichi, M. (1988). Cloning and
expression of cDNA encoding a neural calcium-dependent cell adhesion
molecule: its identity in the cadherin gene family. J. Cell Biol. 106, 873-
Hatta, K. and Takeichi, M. (1986). Expression of N-cadherin adhesion
molecules associated with early morphogenetic events in chick
development. Nature 320, 447-449.
Heasman, J., Crawford, A., Goldstone, K., Garner-Hamrick, P.,
Gumbiner, B., McCrea, P., Kintner, C., Noro, C. Y. and Wylie, C. (1994).
Overexpression of cadherins and underexpression of β-catenin inhibit dorsal
mesoderm induction in early Xenopus embryos. Cell 79, 791-803.
Henion, P. D. and Weston, J. A. (1997). Timing and pattern of cell fate
restrictions in the neural crest lineage. Development 124, 4351-4359
Hopp, T. P., Prickett, K. S., Priice, V. L., Libby, R. T., March, C. J.,
Cerretti, D. P., Urdal, D. L. and Conlon, P. J. (1988). A short polypeptide
marker sequence useful for recombinant protein identification and
purification. Biotechnology 6, 1204-1210.
Ikeya, M., Lee, S. M. K., Johnson, J. E., McMahon, A. P. and Takada, S.
(1997). Wnt signaling required for expansion of neural crest and CNS
progenitors. Nature 389, 966-970.
Kanegae, Y., Makimura, M. and Saito, I. (1994). A simple and efficient
method for purification of infectious recombinant adenovirus. Jpn. J. Med.
Sci. Biol. 47, 157-166.
Kintner, C. (1992). Regulation of embryonic cell adhesion by the cadherin
cytoplasmic domain. Cell 69, 225-236.
LeDouarin (1982). The Neural Crest. Cambridge: Cambridge Univ. Press.
Mochii, M., Mazaki, Y., Mizuno, N., Hayashi, H. and Eguchi, G. (1998).
Role of Mitf in differentiation and transdifferentiation of chicken pigmented
epithelial cell. Dev. Biol. 193, 47-62.
Moriyoshi, K., Richards, L. J., Akazawa, C., O’Leary, D. D. and
Nakanishi, S. (1996). Labeling neural cells using adenoviral gene transfer
of membrane-targeted GFP. Neuron 16, 255-260.
Nakagawa, S. and Takeichi, M. (1995). Neural crest cell-cell adhesion
controlled by sequential and subpopulation-specific expression of novel
cadherins. Development 121, 1321-1332.
Nagafuchi, A., Ishihara, S. and Tsukita, S. (1994). The roles of catenins in
the cadherin-mediated cell-cell adhesion: Functional analysis of E-cadherin-
α-catenin fusion molecules. J. Cell Biol. 127, 235-245.
Nieto, M. A., Sargent, M. G., Wilkinson, D. G. and Cooke, J. (1994).
Control of cell behavior during vertebrate development by slug, a zinc finger
gene. Science 264, 835-839.
Niwa, H., Yamamura, K. and Miyazaki, J. (1991). Efficient selection for
high-expression transfectants with a novel eukaryotic vector. Gene 108, 193-
Opdecamp, K., Nakayama, A., Nguyen, M. T., Hodgkinson, C. A., Pavan,
W. J. and Arnheiter, H. (1997). Melanocyte development in vivo and in
neural crest cell cultures: crucial dependence on the Mitf basic-helix-loop-
helix-zipper transcription factor. Development 124, 2377-2386.
Riehl, R., Johnson, K., Bradley, R., Grunwald, G. B., Cornel, E.,
Lilienbaum, A. and Holt, C. E. (1996). Cadherin function is required for
axon outgrowth in retinal ganglion cells in vivo. Neuron 17, 837-848.
Sanson, B., White, P. and Vincent, J. P. (1996). Uncoupling cadherin-based
adhesion from wingless signaling in Drosophila. Nature 383, 627-630.
Schorle, H., Meier, P., Buchert, M., Jaenisch, R. and Mitchell, P. J. (1996).
Transcription factor AP-2 essential for cranial closure and craniofacial
development. Nature 381, 235-238.
Shimamura, K., McMahon, A. P. and Takeichi, M. (1994). Wnt-1-
dependent regulation of local E-cadherin and αN-catenin expression in the
embryonic mouse brain. Development 120, 2225-2234.
Suda, T. and Nagata, S. (1994). Purification and characterization of the Fas-
ligand that induces apoptosis. J. Exp. Med. 179, 873-879.
Takeichi, M. (1993). Cadherins in cancer: Implication for invation and
metastasis. Curr. Opin. Cell Biol. 5, 806-811.
Takeichi, M. (1995). Morphogenetic roles of classic cadherins. Curr. Opin.
Cell Biol. 7, 619-627.
Takeichi, M., Hirano, S., Matsuyoshi, N. and Fujimori, T. (1993).
Cytoplasmic control of cadherin-mediated cell-cell adhesion. Cold Spring
Harbor Quant. Biol. LVII, 327-334.
Thiery, J. P., Delouvee, A., Gallin, W. J., Cunningham, B. A. and Edelman,
G. M. (1984). Ontogenic expression of cell adhesion molecules: L-CAM is
found in epithelia derived from the three primary germ layers. Dev. Biol.
Tosney, K. W. (1978). The early migration of neural crest cells in the trunk
region of the avian embryo: an electron microscopic study. J. Embryol. exp.
Morph. 62, 317-333.
Tremblay, P., Kessel, M. and Gruss, P. (1995). A transgenic neuroanatomical
marker identifies cranial neural crest deficiencies associated with the Pax3
mutant splotch. Dev. Biol. 171, 317-329.
Tucker, G. C., Aoyama, H., Lipinski, M., Tursz, T. and Thiery, J. P. (1984).
Identical reactivity of monoclonal antibodies HNK-1 and NC-1:
conservation in vertebrates on cells derived from the neural primordium and
on some leukocytes. Cell Differ.14, 223-230
Zhang, J., Hagopian-Donaldson, S., Serbedzija, G., Elsemore, J., Plehn-
Dujowich, D., McMahon, A. P., Flavell, R. A. and Williams, T. (1996).
Neural tube, skeletal and body wall defects in mice lacking transcription
factor AP-2. Nature 381, 238-241.