A role for N-cadherin in mesodermal morphogenesis during gastrulation.
ABSTRACT Cell adhesion molecules mediate numerous developmental processes necessary for the segregation and organization of tissues. Here we show that the zebrafish biber (bib) mutant encodes a dominant allele at the N-cadherin locus. When knocked down with antisense oligonucleotides, bib mutants phenocopy parachute (pac) null alleles, demonstrating that bib is a gain-of-function mutation. The mutant phenotype disrupts normal cell-cell contacts throughout the mesoderm as well as the ectoderm. During gastrulation stages, cells of the mesodermal germ layer converge slowly; during segmentation stages, the borders between paraxial and axial tissues are irregular and somite borders do not form; later, myotomes are fused. During neurulation, the neural tube is disorganized. Although weaker, all traits present in bib mutants were found in pac mutants. When the distribution of N-cadherin mRNA was analyzed to distinguish mesodermal from neuroectodermal expression, we found that N-cadherin is strongly expressed in the yolk cell and hypoblast in the early gastrula, just preceding the appearance of the bib mesodermal defects. Only later is N-cadherin expressed in the anlage of the CNS, where it is found as a radial gradient in the forming neural plate. Hence, besides a well-established role in neural and somite morphogenesis, N-cadherin is essential for morphogenesis of the mesodermal germ layer during gastrulation.
- [Show abstract] [Hide abstract]
ABSTRACT: Zebrafish gastrulation cell movements occur in the context of dynamic changes in extracellular matrix (ECM) organization and require the concerted action of planar cell polarity (PCP) proteins that regulate cell elongation and mediolateral alignment. Data obtained using Xenopus laevis gastrulae have shown that integrin-fibronectin interactions underlie the formation of polarized cell protrusions necessary for PCP and have implicated PCP proteins themselves as regulators of ECM. By contrast, the relationship between establishment of PCP and ECM assembly/remodeling during zebrafish gastrulation is unclear. We previously showed that zebrafish embryos carrying a null mutation in the four-pass transmembrane PCP protein vang-like 2 (vangl2) exhibit increased matrix metalloproteinase activity and decreased immunolabeling of fibronectin. These data implicated for the first time a core PCP protein in the regulation of pericellular proteolysis of ECM substrates and raised the question of whether other zebrafish PCP proteins also impact ECM organization. In Drosophila melanogaster, the cytoplasmic PCP protein Prickle binds Van Gogh and regulates its function. Here we report that similar to vangl2, loss of zebrafish prickle1a decreases fibronectin protein levels in gastrula embryos. We further show that Prickle1a physically binds Vangl2 and regulates both the subcellular distribution and total protein level of Vangl2. These data suggest that the ability of Prickle1a to impact fibronectin organization is at least partly due to effects on Vangl2. In contrast to loss of either Vangl2 or Prickle1a function, we find that glypican4 (a Wnt co-receptor) and frizzled7 mutant gastrula embryos with disrupted non-canonical Wnt signaling exhibit the opposite phenotype, namely increased fibronectin assembly. Our data show that glypican4 mutants do not have decreased proteolysis of ECM substrates, but instead have increased cell surface cadherin protein expression and increased intercellular adhesion. These data indicate that Wnt/Glypican4/Frizzled signaling regulates ECM assembly through effects on cadherin-mediated cell cohesion. Together, our results demonstrate that zebrafish Vangl2/Prickle1a and non-canonical Wnt/Frizzled signaling have opposing effects on ECM organization underlying PCP and gastrulation cell movements.Developmental Biology 09/2013; · 3.87 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Regulation of gene expression plays a central role in embryonic development. Early stages are controlled by gametic transcripts, which are subsequently substituted with transcripts from the genome of the zygote. Transcriptomic analyses provide an efficient approach to explore the temporal gene expression profiles in embryos and to search for the developmental regulators. We report a study of early Atlantic cod development that used a genome-wide oligonucleotide microarray to examine the composition and putative roles of polyadenylated transcripts.BMC genomics. 07/2014; 15(1):594.
- 07/2014; , ISBN: 978-953-51-1590-8
A role for N-cadherin in mesodermal morphogenesis during gastrulation
Rachel M. Warga, Donald A. Kane⁎
Department of Biological Sciences, Western Michigan University, Kalamazoo, MI 49008, USA
Received for publication 14 March 2007; revised 26 June 2007; accepted 28 June 2007
Available online 6 July 2007
Cell adhesion molecules mediate numerous developmental processes necessary for the segregation and organization of tissues. Here we show
that the zebrafish biber (bib) mutant encodes a dominant allele at the N-cadherin locus. When knocked down with antisense oligonucleotides, bib
mutants phenocopy parachute (pac) null alleles, demonstrating that bib is a gain-of-function mutation. The mutant phenotype disrupts normal
cell–cell contacts throughout the mesoderm as well as the ectoderm. During gastrulation stages, cells of the mesodermal germ layer converge
slowly; during segmentation stages, the borders between paraxial and axial tissues are irregular and somite borders do not form; later, myotomes
are fused. During neurulation, the neural tube is disorganized. Although weaker, all traits present in bib mutants were found in pac mutants. When
the distribution of N-cadherin mRNA was analyzed to distinguish mesodermal from neuroectodermal expression, we found that N-cadherin is
strongly expressed in the yolk cell and hypoblast in the early gastrula, just preceding the appearance of the bib mesodermal defects. Only later is
N-cadherin expressed in the anlage of the CNS, where it is found as a radial gradient in the forming neural plate. Hence, besides a well-established
role in neural and somite morphogenesis, N-cadherin is essential for morphogenesis of the mesodermal germ layer during gastrulation.
© 2007 Elsevier Inc. All rights reserved.
Keywords: Mesoderm; Muscle; Notochord; Cell adhesion; N-Cadherin; Gain-of-function; Gastrulation
the biber (bib) mutant identified in the Tübingen large-scale
screen for mutations that affect morphogenesis in the zebrafish,
carried out at the Max Planck Institut für Entwicklungsbiologie
it was placed into the gastrulation and tail class, where it was
described briefly as a simple homozygous recessive lethal,
having indistinct somites and a disorganized tail (Hammersch-
midt et al., 1996). Phenotypically unique amongst the mutants
found in the screen, bib was never successfully placed in a
In the first portion of this work, we demonstrate that the bib
mutation is in a single gene that encodes the zebrafish homolog
of N-cadherin. One of the earliest calcium-dependent cell
adhesion molecules identified, N-cadherin was first described in
the CNS of the mouse and chicken (Hatta and Takeichi, 1986),
of cadherins known as the classic cadherins (Takeichi et al.,
1988). In its most well-known role, N-cadherin is strongly
et al., 1997; Radice et al., 1997; Redies et al., 1993; Redies and
Takeichi, 1993), where its expression pattern is complementary
to the expression of E-cadherin in the epidermal ectoderm
(Takeichi, 1995). In general, the pattern of expression during
motility, suggesting a potential role in the epithelial to mesen-
chymal transition (Duband et al., 1988; Hatta et al., 1987; Oda
et al., 1998). N-Cadherin tends to be expressed in mesenchymal
cell types whereas E-cadherin is expressed in epithelial cell
often E-cadherin expression is lost in metastastic cells, while the
expression of N-cadherin is elevated (Derycke and Fracke,
The N-cadherin protein is characterized by five extracellular
cadherin (EC) repeats, a single pass transmembrane domain and
a cytoplasmic domain. The EC repeats are extremely conserved
among theirhomologs inother species, and arenecessary forthe
specific adhesion properties of the cadherins (Blaschuk et al.,
Developmental Biology 310 (2007) 211–225
E-mail address: firstname.lastname@example.org (D.A. Kane).
0012-1606/$ - see front matter © 2007 Elsevier Inc. All rights reserved.
1990; Yagi and Takeichi, 2000). The cytoplasmic domain of the
protein binds alpha- and beta-catenin, and indirectly actin. This
anchors the protein to the cytoskeleton of the cell and connects
the molecule to the WNT and other signaling pathways
(Christofori, 2003; Derycke and Fracke, 2004; Gumbiner and
McCrea, 1993; Herrenknecht et al., 1991; Kintner, 1992;
McCrea et al., 1991; Sanson et al., 1996). The protein is thought
to function as a dimer, and the EC repeats are necessary for this
dimerization (Brieher et al., 1996; Nagar et al., 1996; Shapiro et
al., 1995). Based on EC domain switching studies, the adhesive
specificity of cadherins is primarily conferred by the first EC
domain (Blaschuk et al., 1990; Nose et al., 1990), a specificity
that is mediated through a twofold symmetric interaction
between EC1 domains on opposing cells (Boggon et al., 2002;
Harrison et al., 2005).
In the second portion of this work, we examine the role of
N-cadherin in the development of the mesoderm during
gastrulation and early segmentation. Although N-cadherin is
expressed in the mesoderm, targeted knockout studies in the
mouse suggests that it is not essential for mesodermal
morphogenesis during gastrulation; defects are seen only later
during somitogenesis where they are limited to malformations
of somitic and cardiac structures (Radice et al., 1997; Horikawa
et al., 1999). When N-cadherin function is rescued in the heart,
allowing the mutant embryos to develop further, dramatic
defects occur in brain development (Luo et al., 2001). Even
more subtle are targeted N-cadherin mutations in Drosophila.
Null mutant embryos are partly viable with no defects in
mesodermal structures and, when examined more closely, the
mutants show axonal pathfinding errors in the central nervous
system (Iwai et al., 1997).
Recently N-cadherin mutations have been identified in the
al., 2002; Masai et al., 2003; Wiellette et al., 2004). AlthoughN-
cadherin expression is reported as fairly ubiquitous throughout
the zebrafish embryo during early development (Bitzer et al.,
1994; Lele et al., 2002), the phenotypes observed in pac alleles
where it is necessary for the convergence of the ectoderm during
neurulation, and later, the cellular organization of the brain and
retina (Lele et al., 2002; Malicki et al., 2003; Masai et al., 2003).
Outside the nervous system, N-cadherin seems only required
during late somitogenesis for the radial migration of a select
population of slow muscle fibers to the surface of the myotome
(Cortes et al., 2003). Hence, to date, the studies in zebrafish and
other animals support the idea that N-cadherin has no role in
mesodermal morphogenesis during gastrulation.
In this report, we demonstrate that the zebrafish mutant bib is
a dominant gain-of-function mutation in the N-cadherin gene.
We focus primarily on cellular defects that are observed in the
bib mutant mesoderm, especially during gastrulation, and
show that pac mutants exhibit similar albeit more subtle defects
in the same cell types. Analysis of RNA expression during the
blastula and gastrula stages correlates the presence of N-
cadherin RNA in the hypoblast with the time of appearance and
location of the bib mutant phenotypes, showing a surprisingly
The gain-of-function nature of the bib allele reveals that
N-cadherin is playing a far greater role in morphogenesis of
the mesoderm during gastrulation then previously thought.
The bibtb8allele was isolated in a large-scale mutagenesis screen (Haffter
type fish for mapping. Subsequent generations were out-crossed to various wild-
type strains of fish that were polymorphic for the appropriate microsatellite
markerssurroundingthe biblocus.Thepactm101alleleof N-cadherin(Leleet al.,
2002)was usedfor complementationstudies.The trim209alleleof Van Goghlike
were used for double mutant studies.
Mapping of the bib locus
bib was initially mapped to one arm of Linkage Group 20 by half tetrad
analysis (Johnson et al., 1995). Fine-resolution mapping was obtained using a
larger panel of 353 EP-diploid embryos to identify two closely linked
microsatellite markers (Knapik et al., 1998). An independent panel of 657
related diploid embryos was used to further characterize recombination
frequencies between closely linked microsatellite markers and the bib locus.
Microsatellite markers were mapped as simple sequence length polymorphisms
and genes were mapped as single-stranded conformation polymorphisms as in
McFarland et al. (2005).
Isolation of the bib mutation and genotypic characterization
To locate the site of the bib mutation we used the RT–PCR approach
the published cDNA. We confirmed the mutation by isolating genomic DNA
from individual mutant and wild-type embryos to generate sequence for the site
CCT CAT C 3′).
Characterization of individual embryos to determine genotype in subsequent
experiments was performed using the above reverse primer and one of the two
following forward primers (bib: F-5′ CCC ATT GAC ATC ATC AA 3′, 55 °C
annealing; wild type: F-5′ CCC ATC GAC ATC ATC AT 3′, 60 °C annealing)
which took advantage of a polymorphism based on a silent nucleotide change in
the mutant background as well as the mutant lesion, see examples in Fig. 2C.
Alternatively, embryos were genotyped by the closely linked markers z6425 or
z78809 (distal) and z3964, z20582, z61014, z59566 or z9343 (proximal).
Time-lapse and data analysis
For in vivo observations, embryos derived from bib heterozygotes were
mounted and recorded in multi-plane as previously described (Warga and Kane,
between 30° and 60° lateral from the shield. Afterwards, the recordings were
analyzed using Cytos Software, looping half hour segments from the time lapse
recording. In the case of Fig. 6, black and white images from one plane of the
time-lapse video recording were imported into Adobe Photoshop and pseudo-
colored to aid in presentation.
Embryo manipulation, immunohistochemistry and RNA in situ
Morpholino antisense oligonucleotides (5′-TCT GTA TAA AGA AAC
CGA TAG AGT T 3′) corresponding to −40 to −16 of the N-cadherin cDNA
(Lele et al., 2002) were injected into 1-cell stage embryos at a final con-
centration of 50 μM. At the 18-somite stage embryos were photographed and
212R.M. Warga, D.A. Kane / Developmental Biology 310 (2007) 211–225
Transplantation was carried out between doming and 30% epiboly stage as
described elsewhere (Ho and Kane, 1990) with the following modification:
donor embryos were injected with either fixable FITC–dextran or a combination
of neutral rhodamine–dextran and fixable biotinylated–dextran (Molecular
Probes). Hosts were examined at 30 h and recorded live as described (Warga and
Kane, 2003), or fixed in 4% paraformaldehyde and processed for the fixable
To detect the FITC–tracer we used a peroxidase conjugated anti-FITC
Detection of the biotin tracer was performed using the Vectastain Elite ABC
(peroxidase) kit and Vector SG enzyme substrate (Vector laboratories), giving a
blue reaction product.
RNA in situ hybridization was carriedout as describedin Thisse et al. (1994)
with modifications from a more up to date protocol by the same authors at http://
zfin.org/zf_info/monitor/vol5.1/vol5.1.html. Embryos were cleared in 70%
glycerol,photographed ona ZeissAxioskopequippedwitha SonyF-707Digital
Camera, and in most instances, genotyped.
Cell contact is perturbed in the bib mesoderm
The bib phenotype first becomes apparent near the end of
gastrulation, when the anlage of the notochord begins to
characteristically kink in the mutant (Figs. 1A, A′) and cells in
the paraxial mesoderm appear to loose cell contact, phenotypes
easily observed on thestereomicroscope.By the3-somite stage,
the paraxial mesoderm of the mutant has disaggregated into a
loose association of spherical cells (Figs. 1B, B′). Curiously,
where paraxial cells contacted the notochord they appeared to
adhere to it, accentuating its knobby and uneven appearance. In
the notochord region itself, cell adhesion defects did not appear
as severe, although cells were not as closely packed as usual and
did not organize into the uniform rows as seen in normal
embryos (Figs. 1B, B′). All these presumptive adhesion defects
were particularly severe in the tailbud of the mutant (which can
mesodermal tissues remained disorganized, somite boundaries
were fused (Figs. 1C–E′), and the tail appeared vacuolated,
flatter, and often curved up, the basis for the moniker “biber”,
German for “beaver”.
Although cellular defects were first detected within meso-
the mutant became disorganized and it was difficult to define
Eventually, by 30 h, both the brain and spinal cord were filled
with rounded cells and both structures acquired an uneven
appearance (Figs. 1E, E′). Despite these morphological abnorm-
alities, the basic body plan was normal at the gross level, and, at
the cellular level, mutant tissues differentiated into histotypes
muscle formed in the myotomal regions (Figs. 1G, G′) and
differentiated neurons formed in the CNS (data not shown).
bib is a mutation in N-cadherin
Using microsatellite markers and half-tetrad analysis, we
mapped bib to one arm of Linkage Group 20 (Fig. 2A) close to
several centromeric markers (Fig. 2B; Table 1). This placed bib
in the vicinity of the gene N-cadherin (also termed cdh2).
Ironically, this gene was initially rejected as a candidate because
the bib phenotype was distinct from the phenotype of alleles of
parachute (pac) (compare Figs. 1E′ to 1F), mutations in the
zebrafish N-cadherin locus (Birely et al., 2005; Lele et al., 2002;
Malicki et al., 2003; Masai et al., 2003; Wiellette et al., 2004).
Nevertheless, fine mapping placed bib very close to N-cadherin
(0 recombinants in 192 meioses), and subsequently, matings
between bib/+ and pactm101/+ adults revealed that bib and pac
failed to complement. Interestingly, bib/pac transheterozygotes
resembled bib homozygous mutants, the more extreme pheno-
type. Normally a transheterozygote phenotype resembles an
intermediate phenotype close to the weaker of the two
The bib allele was sequenced from cDNA generated by RT–
PCR. The mutation is a transversion in a nucleotide within the
coding region of the first EC domain of the N-cadherin protein,
changing the well-conserved isoleucine242to an asparagine
(Figs. 2D–F). To confirm that the bib mutation was in N-
Fig. 1. Morphology of the bib phenotype. (A) Dorsal view; arrows indicate
kinked mutant notochord, inset shows high magnification of rounded cells in
paraxial mesoderm, dotted line outlines ventral portion of anterior neural plate.
(B) High magnification of notochord (nc) and paraxial mesoderm (pm). (C) Side
view: arrows indicate position of somite furrows; arrowhead indicates flatter
mutant tailbud. (D) Side view: arrowhead indicates defects in brain subdivi-
sions; arrow indicates defects in muscle segmentation. Side views of (E) wild-
type and bib siblings and (F) pac mutant; arrowheads indicate uneven surface of
neural tube, arrows indicate vacuoles in disorganized tail. (G) High magni-
fication of wild type and bib striated muscle.
213 R.M. Warga, D.A. Kane / Developmental Biology 310 (2007) 211–225
cadherin, we sequenced individual embryos and designed and
tested primers specific for either the wild-type or the bib N-
cadherin variants. In all cases, the mutation always segregated
with the bib phenotype (Fig. 2C). To determine if the bib
phenotype was the result of a co-segregating second mutation
that was exasperating a pac phenotype, we tested linkage of the
bib phenotype to several markers closely linked to N-cadherin
(Table 1), looking for changes in the bib phenotype. Recombi-
nation events were found on either side of the bib locus between
closelylinked proximal and distal markers (see mutant 6in Figs.
2B and C; also, and Table 1), and in these cases, the bib
phenotype was unchanged. Hence, it is unlikely that the bib
phenotype is the result of an additional mutation.
bib is a dominant gain-of-function mutation
The complementation analysis between bib and pac sug-
gested that bib was not acting as a simple recessive mutation:
pac is thought to be a null phenotype and yet the bib phenotype
was more extreme. The analysis of mutations in E-cadherin
(Kane et al., 2005) showed that non-synonymous changes in
conserved amino acids of the EC1 domain often result in
Examination of bib clutches in the early somite stages
identified a class of embryos that exhibited morphological
defects in the tailbud which was intermediate in phenotype
between bib mutants and completely normal embryos (Figs.
3A–C). In these intermediate individuals, somites form but their
epithelial arrangement was mildly disorganized (see Fig. 5C′.3).
After sorting 190 embryos into normal, intermediate and
strong classes, we found on genotyping that the intermediate
phenotype correlated with the heterozygous genotype (with the
exception of two heterozygous individuals misidentified as
normal). Analysis of expression of papc in the paraxial
mesoderm (Yamamoto et al., 1998) followed by genotyping
showed that the tailbud of the heterozygote is wider than
normal, but not as wide as the homozygous mutants (Figs.
3D–F). Hence, the bib phenotype segregated as a semi
dominant trait (Fig. 3G). Comparison of pac clutches revealed
a similar cellular disorganization in the tailbud of pac homo-
zygotes (Figs. 3H, I). However, analysis of papc expression
(Figs. 3J, K) followed by genotyping revealed that the pac
phenotype segregated as a simple recessive trait (Fig. 3L). Note
the striking similarity of the bib heterozygous and pac homo-
Fig. 2. Identification of bib as N-cadherin. (A) Genetic map of Linkage Group 20. Shown are the positions of microsatellite (Z) markers used in mapping and the genes
N-cadherin(ncad) and desmoglein2(dsg2). Markers in green were used to calculate recombination frequencies in Table 1. (B) Examples of productsobtained for three
closely linked markers. Arrows indicate product that segregates with mutant allele; black asterisk denotes a homozygous mutant individual possessing both products
indicating a recombination event between the marker and the N-cadherin locus. (C) Products obtained from same embryos shown in panel B using the wild-type or
mutant N-cadherin-specific primers. The expected product (arrow) is 200 bp. (D) Domain structure of the N-cadherin protein. Cyt, cytoplasmic domain; EC1–5,
extracellular domains; Pro, prodomain; S, signal sequence; TM, transmembrane domain. (E) The zebrafish amino acid sequence for the carboxyl-end of EC1
containing the bib mutation and the corresponding amino acid sequence alignments for human, mouse and chick and more distantly related cadherins. Identical
residues, black; conserved residues, grey. (F) Sequence trace data showing site of mutation.
Recombination rates, based on half tetrad analysis, showing closely linked
Intervalbib/bib bib/++/+ cM
aCentromere is in this interval.
214 R.M. Warga, D.A. Kane / Developmental Biology 310 (2007) 211–225
To directly test if the bib gene product was producing the
extreme phenotype of bib mutants, we blocked the bib product
using morpholino antisense oligonucleotides (Fig. 4). The
oligonucleotides were originally designed to knock down the
normal N-cadherin gene product to phenocopy the pac mutant
in the study by Lele et al. (2002). We injected the oligonucleo-
tides into one-cell stage embryos produced from a cross of two
bib heterozygotes, a cross that would normally yield 25% bib
mutants. However, 100% of the injected embryos exhibited the
pac phenotype (Figs. 4C, D). In particular, none of the embryos
hadfused somites,an obviousphenotypictrait that distinguishes
bib from pac mutants (Figs. 4A, B). To verify that bib homo-
Fig. 3. bib has a semi-dominant phenotype. (A–F and H–K) Vegetal views, showing cellular morphology and width of tailbud in (A–F) bib siblings and (H–K) pac
siblings. (A–C and H–I) Live embryos and (D–F and J–K) papc expression. Arrows indicate width of tailbud and arrowheads indicate patches of round disaggregated
cells. (G and L) Comparison of width of tailbud using papc expression based on genotype of embryos.
Fig. 4. Antisense knockdown of the bib mutant product phenocopies the pac mutant. (A) Homozygous pac mutant, (B) uninjected homozygous bib mutant, and (C)
heterozygous and (D) homozygous bib mutant MO-injected siblings showing phenocopy of pac homozygote. Examples of PCR products from the closely linked
Z6425 marker used for genotyping (E) uninjected controls and (F) MO-injected experiments; individuals B–D are indicated. (G) The ratio of genotypes in the MO
experiment is not statistically different from the expected 1:2:1 ratio.
215R.M. Warga, D.A. Kane / Developmental Biology 310 (2007) 211–225