Download Unfolding a chordate developmental program, one cell at a time

Developmental Biology 332 (2009) 48–60
Contents lists available at ScienceDirect
Developmental Biology
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / d eve l o p m e n t a l b i o l o g y
Review
Unfolding a chordate developmental program, one cell at a time: Invariant cell
lineages, short-range inductions and evolutionary plasticity in ascidians
Patrick Lemaire
IBDML, UMR6216, CNRS-Université de la Méditerranée, Case 907, Campus de Luminy, F-13288 Marseille cedex9, France
a r t i c l e
i n f o
Article history:
Received for publication 20 March 2009
Revised 27 April 2009
Accepted 3 May 2009
Available online 8 May 2009
Keywords:
Ascidian
Evolution
Ciona
Halocynthia
Embryo
Induction
Lineage
FGF
Ephrin
Nodal
BMP
ADMP
a b s t r a c t
Ascidians were historically the first metazoans in which experimental embryology was carried out. These
early works by Chabry and Conklin [Chabry, L., 1887. Embryologie normale et tératologique des Ascidie.
Felix Alcan Editeur, Paris; Conklin, E., 1905. The organization and cell lineage of the ascidian egg. J. Acad.,
Nat. Sci. Phila. 13, 1], in particular, led to the idea that the developmental program of these animals was
driven by the cell-autonomous inheritance of localised maternal determinants, rendered precise by the
stereotyped pattern of invariant cell cleavages. Work in the past 20 years indeed identified several
localised maternal determinants of the position of cleavage planes or of some early cell fates. The
overwhelming majority of cells in the three germ layers, however, do not follow a cell-autonomous
differentiation program. Instead, they respond to short-range signals, as described in this review. Careful
analysis of cell–cell contacts suggests that a major function of the invariant position of cleavage plans,
besides segregating competence factors, is to control the relative positions of inducing cells and those
competent to respond. Surprisingly, while the cell lineage is very well conserved between the divergent
species Halocynthia roretzi and Ciona intestinalis, the molecular nature of inducing signals can vary. The
constraints on embryo anatomy thus appear stronger than those on the choice of individual regulatory
molecules.
© 2009 Elsevier Inc. All rights reserved.
Introduction
Looking at an adult ascidian, it is difficult, and slightly degrading,
to imagine that we are close cousins to these creatures: fixed to a
rock, they spend their adult life filtering seawater, protected from
predators by their thick cellulose tunic (Fig. 1). These odd creatures
have, however, retained for over 500 million years an ancestral
chordate tadpole larval form, and, being tunicates, they are indeed
thought to belong to the vertebrates' sister group (Delsuc et al.,
2006; Bourlat et al., 2006).
Historically, these animals had a great conceptual impact on our
understanding of the rules that guide a developmental program.
Laurent Chabry used ascidians to show for the first time, that it
was possible to predict the consequences of interfering with
embryonic development. His experiments on muscle progenitor
ablations led to the theory of mosaic development: cells develop
according to the maternal information they receive during cleavage
(Chabry, 1887). Edward Conklin later showed that ascidian
embryos develop in a stereotyped fashion, with an invariant cell
lineage that he could describe (Conklin, 1905). These major
advances led to the idea that the ascidian developmental program
E-mail address: [email protected].
0012-1606/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.ydbio.2009.05.540
is driven by the precise inheritance of localised maternal determinants made possible by the invariant position of cell cleavage
planes.
Following the excitement of the late XIXth and early XXth centuries, ascidians retreated for a long period to their marine
obscurity, from which they were pulled out once again, by
molecular biology (Kuratani et al., 2006). Initial molecular studies
were mainly carried out on embryos of the Japanese ascidian Halocynthia roretzi. As the number of laboratories interested in
ascidians increased, a more cosmopolitan species was searched
and Ciona intestinalis was chosen. The genomes of C. intestinalis, and
of its relative, Ciona savignyi, have now been sequenced (Dehal
et al., 2002; Small et al., 2007) and over 50 labs worldwide study
ascidian development. This re-emergence of ascidians has been
accompanied by a flurry of recent reviews covering developmental
(Kumano and Nishida, 2007; Nishida, 2005; Satou et al., 2008) or
evolutionary aspects (e.g. Lemaire et al., 2008; Kourakis and Smith,
2005; Jeffery, 2007) of ascidian development. The present review
focuses on the cell fate decisions that take place up to the late
gastrula stage, when most embryonic cells have been assigned a
precise developmental fate. With some exceptions, this review
focuses on the development of Ciona embryos, unless a more
detailed understanding, or a different mechanism, has been
identified in Halocynthia. An interesting discussion of the differences
P. Lemaire / Developmental Biology 332 (2009) 48–60
Fig. 1. Adult Halocynthia roretzi (A) and Ciona intestinalis (B). The scale bar represents
1 cm. The size of the adults can vary according to age and nutritional conditions.
between the developmental programs of these animals can also be
found in Hudson and Yasuo (2008).
Why study ascidians? Simple embryos, simple genomes
C. intestinalis and H. roretzi belong to evolutionary distant taxa,
which probably diverged several hundred million years ago. In spite of
their evolutionary distance, they share a rapid development with few
cells and remarkably similar cell lineages (Kumano and Nishida,
2007). Ciona embryos are smaller (egg diameter of 140 μm) than
those of Halocynthia (280 μm). Two Ciona genomes have been
sequenced but little is known about the Halocynthia genome, except
that the few Halocynthia non-coding sequences that have been
studied cannot be aligned with their Ciona counterparts (Oda-Ishii
et al., 2005). Thus, Ciona embryos are easier to image (Figs. 2A–E) and
reconstruct in 3D (Tassy et al., 2006), and are suitable for genome
wide approaches. Halocynthia embryos are more appropriate for
classical embryological approaches.
The tadpoles of both Ciona and Halocynthia only count around
2600 cells, and less than 20 major tissues (Fig. 2F). Some larval tissues
are differentiated and shared with vertebrates (tail muscle, notochord,
epidermis, neural tissue). Others tissues only terminally differentiate
after metamorphosis and give rise to most of the adult animal. Among
these, the endoderm and heart precursors (Trunk Ventral Cells, TVC)
start differentiating during embryogenesis and have clear vertebrate
homologues. In contrast, the mesenchyme and Trunk Lateral Cells
(TLC) constitute largely undifferentiated populations that cannot
easily be homologised with vertebrate embryonic territories.
Fate restriction occurs early during ascidian development, and the
differentiation potential of each cell in the context of the embryo can
be monitored with precision (Figs. 2G–I, 3. A more detailed fate map
at the 112-cell stage is shown in the Supplemental figure). By the
onset of gastrulation, when embryos count little more than 100 cells,
most (19/22) mesodermal and endodermal pairs of precursors have
been fate restricted to a single tissue type (Fig. 3). Interestingly,
several larval tissues (muscle, posterior endoderm) are produced by
two independent lineages, which follow two very different developmental programs to give rise to indistinguishable tissues (Figs. 2F, 3).
By the mid-gastrula stage, when embryos count around 300 cells, the
neural plate has also been thoroughly patterned (Figs. 3 and 9).
Compared to the simplicity of the mesendoderm, patterning of the
ascidian epidermis is surprisingly sophisticated. While trunk epidermis development has received little attention, tail epidermis
patterning has recently started to be elucidated (Pasini et al., 2006).
49
This tissue forms a very regular structure that only counts 8 cells in
cross section: 1 dorsal and 1 ventral midline cell, 4 medio-lateral
cells, and 2 lateral cells (Fig. 2F). The midline cells have particular
functions: they are required for the morphogenesis of the acellular
median fin of the ascidian tadpole, and also give rise to the epidermal
caudal neurones that form the tail peripheral nervous system (Pasini
et al., 2006).
Ascidian embryos became popular because of their simplicity and
stereotyped development, but also because of the relative ease with
which their transcriptional program can be elucidated. Part of this
success is due to the genomic features of C. intestinalis. Although this
genome has considerably drifted from that of the common ancestor of
vertebrates and ascidians (Dehal et al., 2002; Holland and GibsonBrown, 2003; Hill et al., 2008; Putnam et al., 2008), its simplicity is an
asset for transcriptional analyses. The Ciona genome counts less than
600 candidate transcription factors, most with vertebrate orthologs,
compared with 1500 transcription factors in mammals (Wilson et al.,
2008). The developmental expression patterns for more than 80% of
these factors have been determined (Imai et al., 2004; Miwata et al.,
2006), so that the regulatory state of each cell, defined by the precise
combination of transcription factors it expresses, has been determined
up to the gastrula stages (Imai et al., 2004). Up to this stage, a combinatorial code of only 53 zygotic transcription factors is sufficient to
account for the observed diversity of cell types (Imai et al., 2006).
Individual loss-of-function analysis of 26 of these by injection of
antisense morpholino oligonucleotides led to the first regulatory
“blueprint”, or network, for a whole metazoan embryo up to gastrulation (Imai et al., 2006). This work has now been extended to the
posterior part of the neural plate (Imai et al., 2009) and complemented by numerous small-scale studies that often included the identification, by phylogenetic footprinting, of the precise cis-regulatory
modules involved in transcriptional regulation (Davidson et al., 2005;
Bertrand et al., 2003; Lamy et al., 2006; Rothbächer et al., 2007).
The challenge ahead is to integrate these functional genomics
studies with the precise embryological framework describing ascidian
embryogenesis. The following sections will give an overview of the
main cellular events that occur during early embryogenesis in ascidians. We will successively explore axis determination by maternal
factors, cell-autonomous zygotic responses to these determinants,
germ layer specification by the FGF pathway up to the 44-cell stage,
regionalisation by nodal/notch signalling up to the early gastrula stage,
and finally neural plate patterning up to the mid-gastrula stages.
Up to the 8-cell stage: determination of primary and secondary
embryonic axes
The embryological studies reported in this section have been
primarily carried out in Halocynthia, because of the larger size of the
eggs of this species. Unless mentioned in the text, it is however
thought that most of what follows also applies to Ciona.
The primary axis
The primary axis, which runs along the animal–vegetal axis of
the egg, is set up during oogenesis (Fig. 4). At fertilisation, the
sperm enters the egg in its animal hemisphere, marked by the polar
bodies. This triggers two successive phases of cytoplasmic rearrangements, also called ooplasmic segregations, during the first cell
cycle (Sardet et al., 2007).
The first phase is actin-dependent and concentrates cortical
maternal determinants for the vegetal-most fate, endoderm, and for
gastrulation at the vegetal pole of the embryo (Nishida, 2005). The
molecular identity of these molecules remains unknown and it is not
clear whether the site of gastrulation and endoderm are specified
independently of each other. The maternal endoderm determinants
act via the β-catenin pathway (Fig. 4, left). Activation of this pathway
50
P. Lemaire / Developmental Biology 332 (2009) 48–60
Fig. 2. Overview of Ciona intestinalis embryogenesis. (A–E) 3D rendering from stacks of confocal images. (A–D) Phalloidin staining of the cortical actin cytoskeleton. Adapted from
FABA (Hotta et al., 2007). (E) Images of a live Ciona larva, electroporated with a ZO-1-GFP fusion protein targeted to the epidermal tight junctions (Roure et al., 2007). Note the small
number of cells covering the embryo. (F) Schemes of an early tailbud stage embryo, with color-coded cell fates as indicated. G–I: Schemes representing the fates of individual
blastomeres at the early gastrula. Drawing based on an early 112-cell embryo, reconstructed in 3D and visualized in 3D Virtual Embryo. Cells are color-coded according to their fate as
in F. G, vegetal view; H, lateral view; I, animal view. The circum-notochord side is up, contra-notochord side is down. Bipotential blastomeres are colored to represent both fates.
When the part of the cell that will contribute to each fate is known, each domain is colored according to its fate. In other cases, the cell is simply hashed with both colors.
by over-expression of activated β-catenin, or treatment with GSK3
inhibitors from early cleavage stages, is sufficient to transform most
embryonic cells into the vegetal-most fate, endoderm. Conversely,
inhibition of the pathway leads to the repression of all vegetal fates
except primary muscle (Imai et al., 2000). A systematic screen for
maternal genes acting in the β-catenin pathway during early
embryogenesis has recently identified 5 novel evolutionary conserved components of this pathway (Wada et al., 2008). The
localised maternal determinant that activates this cascade is
currently not known, but it appears that, in Halocynthia, zygotic
Wnt5a relays this initial signal and is required for the sustained
nuclear accumulation of β-catenin in vegetal blastomeres from the
64-cell stage (Kawai et al., 2007).
On the opposite pole, the ectoderm is specified by the action
of a maternal transcription factor of the GATA family, GATAa, most
similar to vertebrate GATA4/5/6 (Fig. 4, bottom left; Rothbächer
et al., 2007). Although GATAa mRNA and protein are ubiquitously
distributed, the transcriptional activity of GATAa is restricted to
the animal hemisphere by vegetal β-catenin. The molecular
mechanism for this restriction is currently unknown.
Second embryonic axis
The second embryonic axis runs orthogonal to the primary axis
(Fig. 4, right panels). It does not correspond to any precise axis of the
larva, and its initial denomination, “antero-posterior” axis, was
P. Lemaire / Developmental Biology 332 (2009) 48–60
51
Fig. 3. Lineage tree of Ciona intestinalis up to the mid-gastrula stage. As Ciona embryos are bilaterally symmetrical, only half of the embryo, originating from the right-hand AB
blastomere at the 2-cell stage is represented. The 8th division (giving rise to cells whose index starts with a 9, e.g. A9.29) is only represented for a selection of neural plate, secondary
muscle and TVC progenitors that are not yet fate restricted by the early gastrula stage. Colored branches reflect the fate of the corresponding cells as indicated in color legend.
somewhat misleading. For instance, notochord was considered the
“anterior”-most fate although it runs in the center of the larval tail up
to its posterior tip (Fig. 2F, orange). A more accurate, and more
abstract, denomination, is the circum-notochord/contra-notochord
axis (Lemaire et al., 2008; Figs. 2G–I). Throughout this review, these
terms will be used to designate the secondary axis. “Anterior” and
“posterior” will only be used for cells that give rise to tissues located in
an anterior or posterior position in the larvae.
The secondary axis is marked by the adoption of different fates
along the axis (Figs. 2G–I), but also by distinct cleavage patterns
(Fig. 4). In particular, the vegetal contra-notochord cells undergo a
series of markedly unequal cleavages between the 8- and 64-cell
stages, the contra-notochord-most daughter being always smaller
than its sister (Fig. 4, right lowest panels). This axis is set up by the
second phase of cytoplasmic reorganisations, which occurs during
the second half of the first cell cycle. During this process, maternal
axial determinants are translocated along microtubules from the
vegetal pole to the contra-notochord side of the embryo. Cytoplasmic ablations indicate that the circum-notochord fates form by
default (Nishida, 2005). This contrast with the “cortical rotation”
observed in amphibians, which transports axial determinants to the
circum-notochord side (Lemaire et al., 2008). Most identified
ascidian contra-notochord determinants are located in a cortical
cytoplasmic region called PVC, for Posterior Vegetal Cytoplasm,
52
P. Lemaire / Developmental Biology 332 (2009) 48–60
Fig. 4. Specification of the embryonic axes in Halocynthia roretzi. Schemes of eggs and embryos, representing in orange and blue the position of the maternal determinants of
primary (animal–vegetal) and secondary (circum/contra-notochord) axes, respectively, in unfertilized eggs, at the end of the first phase of ooplasmic segregation and at the end
of the second phase of segregation. At the 2-cell stage, the position of the mitotic spindle with respect to the posterior vegetal cytoplasm PVC is indicated. At the 16-cell stage,
the left-hand lateral embryo schemes present the domains of activity of β-catenin (blue) and GATAa (pink) acting in primary axis specification. The right-hand schematic
vegetal views of embryos, revealing the asymmetric nature of cleavage planes in circum vs contra-notochord sides. Arrows placed next to gene names indicate whether the
activity of the gene is increased or decreased. WT: wild type, PVC-: embryos in which the Posterior Vegetal Cytoplasm has been surgically depleted at the end of the 2nd phase
of ooplasmic segregation (the ablation is schematically represented by a red “X”). Eggs and embryos are oriented with their circum-notochord side to the left.
which is inherited by the precursors of tail muscles, endoderm and
germ line (Nishida, 1987; Shirae-Kurabayashi et al., 2006). The PVC
is a complex multilayered cortical structure enriched in cortical
Endoplasmic Reticulum (cER), germ plasm, and anchored localised
mRNAs collectively referred to as Posterior End Mark (PEM) because
of the posterior localisation of the cells that inherit them at the
tailbud stages. By the 8-cell stage, the PVC concentrates into a
subcellular structure, the Centrosome Attracting Body (CAB).
Inheritance of this structure is responsible for the series of unequal
cleavages observed in the contra-notochord-most vegetal cells
during cleavage stages (Nishida, 2005). Surgical removal of the
PVC leads to a radialization of both fates and cleavage patterns of
the embryo (Fig. 4 bottom right). Muscle is lost, while the notochord is expanded.
Many of the activities of the PVC can be ascribed to the function of
individual PEM mRNAs (Fig. 4, right). Some of them act as classical
fate determinants. The most famous of these, Macho1, coding for a
zic-family transcription factor, has a dual activity. It acts throughout
the vegetal contra-notochord territories, as a determinant for the
primary muscle lineage (Nishida and Sawada, 2001) and as a
competence factor for mesenchyme (Kobayashi et al., 2003). Another
Halocynthia maternal PEM RNAs, Wnt5, has also been shown to act in
contra-notochord (mesenchyme, muscle) fate specification (Nakamura et al., 2006). Interestingly, genes that produce PEM RNAs can
have additional zygotic function: for instance, zygotic Wnt5 transcripts subsequently play a role in notochord specification in Halocynthia (Nakamura et al., 2006). Besides these classical fate
determinants, other determinants act in the anchoring or translation
of PEM RNAs or in the function of the CAB. The POPK-1 kinase is
required for the positioning and concentration of the cER, germ
plasm and PEM RNAs to the CAB (Nakamura et al., 2005). Its
depletion only rarely perturbs the cleavage pattern, but it leads to a
specific loss of muscle and mesenchyme, probably by an inhibition of
Macho1 function. The founding member of the PEM RNAs, PEM-1, is
required for the proper positioning of the mitotic spindle during the
early unequal cleavages at the contra-notochord pole of embryos and
is thus required for the contra-notochord-specific cleavage pattern.
PEM-1 is thought to act as an adaptor pulling on the distal tip of
astral microtubules to attract the centrosome to the PVC or CAB
(Negishi et al., 2007; Munro, 2007).
Up to the early 32-cell stage: cell-autonomous regionalisation of
the embryo by maternal determinants
The initial experiments by Chabry and Conklin suggested that
much of ascidian development is controlled locally by the simple
inheritance of localised maternal determinants. Indeed, by the 16cell stage, combinations of regulatory genes define only 6 territories
in the embryo, each comprising no more than 2 pairs of cells
(reviewed in Satou et al., 2008; Fig. 5B). One of these territories,
the B5.2 vegetal contra-notochord cell pair, inherits the germ plasm
and is distinguished by its transcriptional silence (Tomioka et al.,
2002). Some of these regulatory genes are validated direct targets
of the maternal determinants. For instance, Tbx6 paralogues are
targets of Macho1 (Yagi et al., 2004), FOG is a target of GATA4/5/6
(Rothbächer et al., 2007), and FoxD is a target of β-catenin (Imai et
al., 2002). The simultaneity of onset of expression of early zygotic
genes within each territory suggests that most animal-specific
genes are targets of GATA4/5/6, most vegetal genes respond to βcatenin, and most genes expressed in B5.1 are Macho1 targets (as
indicated tentatively on Fig. 5B).
With the exception of b5.4, which only contributes to epidermis,
none of the territories defined at the 16-cell stage gives rise to a single
larval tissue type. The presence, alongside zygotically transcribed
regulatory genes, of 6 secreted or trans-membrane signalling ligands
(Fig. 5B) suggests that, during this initial regionalisation, combina-
P. Lemaire / Developmental Biology 332 (2009) 48–60
53
indicate that they play major, and antagonistic, roles in germ layer
specification (Fig. 6B).
FGF9/16/20 induces mesoderm (notochord, mesenchyme), animal neural tissue, anterior endoderm and, in Halocynthia at least,
posterior endoderm (Fig. 6B). Over-activation of the FGF pathway
increases the number of induced cells in most cases, revealing that
only some of the competent cells are induced in vivo. Because all
vegetal cells express FGF and all animal cells touch at least one vegetal
cell, it is puzzling that not all competent cells are induced. The precise
mechanisms restricting the response to the inducer in the ectoderm
and the mesendoderm differ, but they involve in each case a
combination of short-range signalling and precise positioning of the
cleavage plane of a bipotential precursor.
Polarisation of bipotential mesenchyme/muscle and notochord/nerve
cord precursors
Fig. 5. Onset of the Ciona zygotic program. (A) Animal (right) and vegetal (left)
schemes of embryos, on which cells on one side have been labeled (A-line in black, Bline in orange, a-line in red and b-line in green). (B) representation of the expression
patterns of the 18 regulatory genes zygotically expressed at the 16-cell stage (adapted
from Satou et al., 2008). Genes in violet are transcriptional regulators, genes in green
code for signaling ligands. These genes combinatorially define 6 domains: B5.2
(transcriptionally silent), b-line, a-line, B5.1, A5.2 and A5.1. Gataa, β-catenin and
Macho1 and likely direct regulators of the indicated co-expressed genes.
tions of transcriptional regulators establish territories of competence,
while the signalling ligands define signalling sources. As cleavage
proceeds, cells furthest from these sources (primary muscle, early
epidermis) form autonomously. In the case of the tail muscle, a cellautonomous regulatory network has emerged. Activates directly Tbx6
genes (Yagi et al., 2004), which in turn activate the unique bHLH
muscle regulatory factor (MRF) found in the ascidian genome (Meedel
et al., 1997; Imai et al., 2006), which is required for muscle formation
(Meedel et al., 2007). Most cells, however, are affected by the
signalling sources established by the 16-cell stage. The following
sections will focus on the successive signalling events that regionalise
each of the 6 initial territories.
The 32-cell stage: germ layer inductions through the antagonistic
actions of the FGF and Ephrin pathways
Of the six zygotically-expressed ligands from the 16-cell stage, two,
FGF9/16/20 and EphrinAd, account for most of the known cell communication events up to the late 32-cell stage (Fig. 6). Although FGFs
are used prior to gastrulation in vertebrate and ascidian embryos,
their precise functions differ largely in the two groups of animals
(Lemaire et al., 2008). The instructive function of Eph/Ephrin in early
cell fate decisions is so far unique to ascidians.
FGF9/16/20 is transcribed in most cells of the vegetal hemisphere,
except in the transcriptionally silent B5.2 cell pair (Fig. 6A). In a
complementary manner, EphrinAd is transcribed throughout the
animal hemisphere during the 16- and 32-cell stages (Fig. 6A)
(Bertrand et al., 2003; Imai et al., 2004; Picco et al., 2007; Shi and
Levine, 2008). Gain and loss of function for FGF9/16/20 and Ephrin Ad
At the 32-cell stage, primary notochord (A6.2, A6.4) and
mesenchyme (B6.4, B6.2) progenitors are still multipotential. During
the next division, A6.2/A6.4 each gives birth to a notochord and a
nerve cord precursor, B6.4 to a mesenchyme and a muscle precursor,
and B6.2 to a bipotential mesenchyme/notochord and a muscle
precursor (Fig. 6B). When FGF signalling is inhibited, the notochord
and mesenchyme precursors adopt the fate of their sister cell, nerve
cord and muscle, respectively (Minokawa et al., 2001; Kim et al., 2007;
Picco et al., 2007). Conversely, when the pathway is ubiquitously
activated, nerve cord and muscle adopt a notochord and mesenchyme
identity (Kim et al., 2000; Minokawa et al., 2001). Isolation of the
notochord/nerve cord or mesenchyme/muscle mother cell at the
beginning or at the end of the 32-cell stage reveals that during this cell
cycle, the mother cell become polarised, so that only the most vegetal
daughter adopts the induced fate (Kim et al., 2007; Picco et al., 2007).
In the mesenchyme lineage, the contacts established by the
bipotential progenitor with FGF-expressing cells are intrinsically
polarised (Fig. 7A) and the asymmetric reception of FGF is sufficient
to polarise the cell (Kim et al., 2007). Polarisation of the notochord/
nerve cord precursor is more sophisticated as this cell expresses
FGF9/16/20 itself (Fig. 7B). Its polarisation requires a contact with
animal cells and involves animally expressed ephrinAd (Kim et al.,
2007; Picco et al., 2007). When the function of EphrinAd or its
receptor Eph3, is blocked, polarisation of the mother cell is prevented
and both daughters form notochord, conversely if EphrinAd is
overexpressed, twice as many nerve cord cells form (Picco et al.,
2007) (Figs. 6B, 7B). At the molecular level, the action of FGF9/16/20
and EphrinAd converge on the MAP kinase ERK, which FGF9/16/20
activates, and EphrinAd represses (Picco et al., 2007; Shi and Levine,
2008). Interestingly, while it is likely that polarisation by FGF/Ephrin
is a mechanism shared by Halocynthia and Ciona, notochord
induction in Halocynthia, but not in Ciona, requires additional Bmp
signalling (Darras and Nishida, 2001, Hudson and Yasuo, personal
communication).
How the initial polarisation of these mother cells can lead to two
daughter cells with distinct fates remains mysterious. A possible
mechanism is that the polarised activation of the FGF receptor or of
the Eph receptor leads to the sequestration/activation of the
corresponding signalling components at the membrane facing the
inducer, and their inheritance by only one of the two daughters.
Future work will probably explore this hypothesis.
Induction of endoderm
In contrast to vertebrates, in which the FGF pathway is involved in
mesoderm but not endoderm induction, FGF signalling is involved in
ascidian endoderm induction during the 32-cell stage.
Shi and Levine (2008) showed that at this stage, all A-line cells that
contribute to the anterior trunk endoderm are in fact bipotential
54
P. Lemaire / Developmental Biology 332 (2009) 48–60
Fig. 6. Germ layer specification at the 32-cell stage: the FGF/EphrinAd antagonism. A) Expression profiles of FGF9/16/20 (pink) and EphrinAd (brown) at the 16- and 32-cells stages.
B) Effect on fate specification of interfering between the 16 and late 32-cell stage with FGF or Ephrin signaling. Animal (top) or vegetal (bottom) schemes of embryos on which the
left-hand cells have been color-coded for cell fates and the right-hand side indicates cell identities. Fates colored as indicated.
endoderm/TLC precursors. The level of activation of the FGF pathway
in each precursor determines the final fate it achieves (Fig. 6B). Only
one of these bipotential progenitors, A6.3, will give daughters of both
fates. This progenitor is located closest to the equator, in contact with
EphrinAd-expressing animal cells (Fig. 7C). It becomes polarised by a
similar mechanism as the notochord/nerve cord precursor (Shi and
Levine, 2008).
Formation of the posterior B-line endoderm (B6.1, B6.3) has not
been carefully analysed in Ciona, though loss of function of FGF9/16/
20 is not sufficient to affect this cell fate in whole embryos (Fig. 6B). In
Halocynthia, formation of this tissue requires either FGF or Bmp
signalling (Kondoh et al., 2003). Explant experiments revealed that
FGF signals, originating from the B-line endoderm precursors
themselves, act between the 16- and 32-cell stages to suppress the
muscle fate and promote posterior endoderm. FGFs signals are
however not strictly required. BMP signals act redundantly around
the same time, but originate from the A-line vegetal cells, explaining
why inhibition of FGF signalling prevents posterior endoderm
formation in B-line explants but not in the context of the embryo.
The precise ligands involved are not known but FGF9/16/20 is suitably
expressed in both Ciona and Halocynthia (Bertrand et al., 2003;
Kumano et al., 2006). Bmp2/4 is expressed in anterior endoderm
precursors from the 44-cell stage in Halocynthia, but this expression is
not conserved with Ciona, thus raising doubts as to the conservation of
the role of Bmp in posterior endoderm induction in this species.
Selection of induced animal neural precursors
FGF9/16/20 also induces neural fate in 2 pairs of animal cells, a6.5
and b6.5 (Fig. 6B). In this case, inhibition of FGF/ERK signalling at
different time points with pharmacological inhibitors revealed that
the induction of neural precursors occurs after the division of the
bipotential epidermis/neural precursors, a5.3 and b5.3 (Hudson and
Lemaire, 2001; Hudson et al., 2003). Analysis of 3D digital models of
P. Lemaire / Developmental Biology 332 (2009) 48–60
55
Fig. 7. Polarisation or induction by short-range cell interactions at the 32-cell stage. Left, snapshot from 32-cell stage embryos analysed in 3D Virtual embryo, showing cell
identity and cell fates (dark red: muscle; bright red: anterior and dorsal neural plates; orange: notochord/posterior nerve cord; dark green: mesenchyme/muscle; bright green:
TLC/anterior endoderm; light blue: posterior endoderm; dark blue: anterior endoderm; light pink: trunk epidermis; pink: tail epidermis). Right, schematic drawings of the same
embryos showing the expression pattern of FGF9/16/20 (violet) and EphrinAd (brown). The blue arrows indicate the orientation of the next mitotic spindle. Surfaces of contact
in A–C relate to the contact between individual signaling cells and the polarized progenitor (A: B6.4; B: A6.2; C: A6.3). Surfaces of contact in D are between each indicated
animal cell and all vegetal cells expressing FGF9/16/20. The measures of surfaces of contact indicated correspond to a representative embryo, and show little variability between
embryos (Tassy et al., 2006).
embryos indicated that the cleavage plane of a5.3 and b5.3 partitions
the surface of contact with FGF-expressing cells in a very asymmetric
manner as the neural progenitor inherits more than 80% of the
contacts with the inducing cells (Fig. 7D). Further, it was shown that
the measure of the surface of contact between inducing and
responding cells is a critical parameter of the selection of induced
cells among the pool of competent ectodermal cells (Tassy et al.,
2006). Interestingly the response to the inducer is not a linear
function of the surface of contact: above a certain threshold of contact,
cells are fully induced, while no induction occurs below this threshold
(Tassy et al., 2006). As animal cells express EphrinAd, it is tempting,
though not experimentally tested, to propose that the threshold of
response to FGF is defined by the level of ERK inhibition provided in
animal cells by EphrinAd.
Induction of the dorsal midline tail epidermis
The epidermis of ascidians is very precisely patterned into midline,
medio-lateral and lateral cells (Fig. 2I). When cultured in isolation
from the 32-cell stage, whole posterior animal explants only give rise
to lateral epidermis, indicating that signals originating from the
vegetal hemisphere induce the midline and paraxial fates. This
process starts at the 32-cell stage with the induction of the dorsal
midline epidermis by FGF9/16/20 (Pasini et al., 2006).
56
P. Lemaire / Developmental Biology 332 (2009) 48–60
44-cell to early gastrula stage: the nodal/delta2 relay
By the late 32-cell stage, most of the induced mesendodermal
and ectodermal cells have received an initial inductive clue, but
several of these tissues require a stabilisation or a refinement of
their fates. In the cases of the primary notochord and the anterior
neural tissue, this stabilisation involves continued FGF signalling up
to the 64-cell and early gastrula stage, respectively (Yasuo and
Hudson, 2007; Hudson and Lemaire, 2001). The FGF involved in
the stabilisation of the neural fate is unknown, while a combination of FGF8/17/18 and FGF9/16/20 acts in the notochord lineage
(Yasuo and Hudson, 2007).
In the other lineages, Nodal activated by FGF9/16/20 in the
posterior ectodermal b6.5 lineage plays a major role. It acts either
directly or via a divergent Notch ligand, Delta2, which Nodal activates
by the early gastrula stage in the descendents of b6.5, as well as in the
A7.6 (TLC) and A7.8 (lateral nerve cord) cells (Fig. 8A).
Secondary notochord induction
At the 44-cell stage, the progenitor of the secondary notochord
lineage, B7.3, is still bipotential (Fig. 3). Between the 64- and 76-cell
stages, this cell undergoes a very unequal division to give a large
mesenchyme progenitor, B8.5, and a small secondary notochord
Fig. 8. Inductions between the 44-cell and early gastrula stages. (A) Expression of nodal and delta2 at the 64- and 112-cell stages. Expression, normally bilateral, is only represented
on the left side of schemes. The identity of cells expressing the genes is indicated on the right-hand side of schemes. (B) Effect on cell fate specification of interfering with Nodal and
Notch signaling from the 44-cell stage. Cells are color-coded according to their fate on the left of each scheme. The names of cells whose fates are affected are indicated on the right.
P. Lemaire / Developmental Biology 332 (2009) 48–60
57
progenitor, B8.6 (Darras and Nishida, 2001). In Ciona, this fate
decision requires Nodal signalling from the b6.5 line up to the 64cell stage, followed by Delta2 signals, from the A7.6 blastomere (Fig.
8B, Hudson and Yasuo, 2006). The adjacent positions of inducing and
induced cells suggest that Nodal, like FGF, acts at short range.
Interestingly, in embryos compromised for Notch or Nodal signalling,
the notochord progenitor does not adopt its sister fate (mesenchyme).
This suggests either that the geometrically asymmetric division is
important for fate specification, or that additional signals are acting to
discriminate between mesenchyme and notochord fates. It is
currently not known whether the unequal cleavage of B7.3 is affected
by inhibition of Nodal or Notch signalling.
Surprisingly, the mechanisms of secondary notochord induction
appear to differ markedly in Halocynthia and Ciona. BMP2/4 was
shown to act as secondary notochord inducer in Halocynthia and to
control the asymmetry of cleavage of B7.3 (Darras et al., 2001). This
gene is, however, not expressed in a suitable location in Ciona, and
BMP pathway inhibition has no effect on secondary notochord
induction in this species (Hudson and Yasuo, 2006, Hudson and
Yasuo, personal communication).
Mesenchyme and TLC induction
In addition to notochord, inhibition of Nodal signalling affects
other mesodermal cell types. It is required for the activation of Twistlike 1 in the B8.5 mesenchyme precursor and twist-like 1, Hand and
Delta2 in the A7.6 TLC precursor (Fig. 8B, Hudson and Yasuo, 2006;
Imai et al., 2006). Unlike the situation observed in the secondary
notochord, the nodal signals that activate A7.6 markers do not originate from the b6.5 progeny, but rather from the vegetal cells that
expressed Nodal at the 32-cell stage (Shi and Levine, 2008). As the
B8.5 precursor is located away from the b6.5 precursors, a similar
situation may apply in this mesenchyme lineage. Finally, also in
contrast with the secondary notochord case, interference with Notch
signalling does not prevent expression of Twist-like 1 in TLC or
mesenchyme precursors, suggesting specification of these cells is
Delta2-independent (Fig. 8B; Hudson and Yasuo, 2006).
Fig. 9. Regionalisation of the Ciona neural plate between the 64-cell and mid-gastrula
stages. Schematic representations of the geometrically organized Ciona neural plate at
the indicated stages, showing the correlation between the activation of the Nodal, Notch
and RK pathways and successive fate restriction events. The black arrows indicate the
orientation of cell cleavages that occur between successive stages. The cell names are
color-coded according to their lineages and fate restriction. b-line cells are not
represented at the mid-gastrula stage as their position is not precisely known. Cells that
do not give rise to a neural (CNS, PNS) fate are boxed with a hashed line. 64-cell stage:
the lateral identity (beige) of neural cells is established by nodal signalling. 76-cell
stage: Delta2-mediated Notch signalling differentiates between column 3 (beige) and 4
(light blue) in the A-line neural plate. Early gastrula stage, a second phase of Delta2
signalling distinguishes columns 1 (white) and 2 (light green) in the A-line neural plate.
Mid-gastrula stage: ERK is activated throughout rows I and III. Activation in Row I is
required for the distinction between Rows I and II.
Medio-lateral regionalisation of the neural plate
In addition to patterning the mesoderm, Nodal also patterns the
whole neural plate, and adjacent epidermis, along its medio-lateral
axis (Figs. 8B, 9) (Hudson and Yasuo, 2005; Hudson et al., 2007).
Between the 32- and 64-cell stages, the Nodal signalling pathway is
activated in the lateral precursors of the neural plate (a7.13; A7.8), and
confers to these cells their lateral neural plate identity (Figs. 8B, 9A;
Hudson and Yasuo, 2005; Hudson et al., 2007). Interestingly, although
Nodal is known to affect neural plate patterning in vertebrates, it does
so in an opposite way: Nodal signalling induces medial, rather than
lateral, neural fates (reviewed in Strähle et al., 2004).
By the 64-cell stage, Nodal has turned on Delta2 in cells (b7.9/10,
A7.6), which contact the side of A7.8 that will be inherited by its
lateral-most daughter, A8.16 (Fig. 8A; Hudson et al., 2007). Polarisation of A7.8 before its cleavage at the 76-cell stage, confers its distinct
lateral-most identity to A8.16 (Figs. 8B, 9B; Hudson et al., 2007).
By the early gastrula stage, Delta2 is expressed in both A8.16
(Column 4) and A8.15 (Column 3) and signals to the next A-line neural
cell, A8.8, to which it confers a column 2 identity (Fig. 9C). Thus,
between the 64-cell and early gastrula stage, two temporally
separable, short-range, Delta2/notch signalling events confer to the
A-line neural precursors their four distinct column identity (Hudson
et al., 2007).
Interestingly, while FGF is also involved up to the 44-cell stage in
the specification of A8.16 in Halocynthia, an unknown, Nodal independent signal relays this initial FGF signal in this species (Tokuoka
et al., 2007).
Dorsal midline epidermis induction
As in other tissues, Nodal acts in relay to FGF9/16/20 during the
induction of the dorsal midline epidermis in the b8.18, b8.20 and b8.21
daughters of b7.9, b7.10 and b7.11, respectively (Fig. 8B, center; Pasini
et al., 2006). The precise timing of the induction and a potential role
for delta2 in the process have not been investigated. As Nodal and
Delta2 are expressed in all daughters of b7.9 and b7.10, these ligands
are unlikely to explain why a single daughter of each of these
precursors adopts a midline epidermal fate.
Early to mid-gastrula stages: regionalisation of the neural plate,
induction of secondary muscle fates, ventral midline
epidermis induction
This period sees the fate restriction of two of the three remaining
bipotential mesendoderm precursors: b8.17 (secondary muscle, posterior endoderm) and A8.16 (secondary muscle, posterior lateral
nerve cord). In parallel, the neural plate becomes regionalised along
its anterior–posterior axis, and the ventral midline epidermis is
induced by the underlying posterior endoderm.
By the end of the period, the only mesendoderm progenitors still
bipotential are the equicompetent daughters of B7.5 (primary
muscle and TVC), which divide asymmetrically during the late
neurula stage to give rise to one TVC and one anterior muscle cell
(Davidson et al., 2006).
58
P. Lemaire / Developmental Biology 332 (2009) 48–60
Antero posterior regionalisation of the neural plate
During the gastrula stages, every neural plate cell divides once,
along the A–P axis of the embryo, to give rise to 6 rows of cells (Fig. 9,
bottom right). By the mid-gastrula stage, specific markers characterise the distinct identities of each of these cells (e.g. Hudson et al.,
2007; Imai et al., 2009). The MAP kinase ERK is specifically activated
by phosphorylation in all row I and III cells, a feature conserved in
Halocynthia (Nishida, 2003). Inhibition of ERK activity from the
early gastrula stage led to the conversion of all row I cells to their
corresponding row II sister (Hudson et al., 2007). Thus, ERK
activation is responsible for the distinction between Row I and II
identities without interfering with medio-lateral column identities.
The ligand responsible for ERK activation is currently unknown, but
FGF9/16/20 is a strong candidate as it is expressed in the right
territories. The effect of ERK inhibition on row III identity has not yet
been studied.
Secondary muscle induction
The two secondary muscle lineages b8.17 and A8.16 give rise to
the muscle cells located at the tip of the tail. In contrast to the
primary muscle lineage which forms cell-autonomously, the
secondary lineages are obtained in response to a cascade of signalling events starting at the 32-cell stages with successive
inductions by FGF, nodal and delta2 as indicated above. During
the gastrula stages, each of these precursors undergoes one division, which segregates the muscle fate from either posterior lateral
neural tissue (A8.16-line) or endodermal strand (b9.17-line).
Although experiments carried out in Ciona and Halocynthia are
not always easy to compare, specification mechanisms in these
lineages appear to differ between species.
We saw in the previous section that in Ciona, ERK activation during
the early gastrula stage distinguishes between row I, including A8.32
(muscle), and row II, including A9.31 (lateral posterior nerve cord). In
Halocynthia, explant experiments revealed that A8.16 becomes
polarised during the 110-cell stage and that this polarisation is
required for muscle formation (Nishida, 1990). Pharmacological
inhibition of FGF/ERK signalling, however, reveals that FGF/ERK
signalling is not required past the 64-cell stage for this polarisation. In
contrast, Wnt5a, emitted by A7.6 is required between the 44-cell and
early gastrula stages (Tokuoka et al., 2007). The function of Wnt5 has
not been assayed in Ciona secondary muscle formation.
Little is known in Ciona about the distinction between the
secondary muscle (b9.33) and endodermal stand (b9.34). In Halocynthia, secondary muscle specification occurs later than in the A-line,
around the 130-cell stage (Nishida, 1990) and requires FGF signalling
up to this stage. Wnt5a signals of either maternal or zygotic origin are
additionally required.
Ventral midline epidermis induction
The ventral midline epidermis originates from the b8.27, b8.28,
b8.31 and b8.32 cells. Among these cells, b8.27 is the only one to
give rise exclusively to midline epidermis. The other cells are
bipotential midline/medio-lateral epidermis precursors (Figs. 2G–I;
Supplemental figure). The two fates segregate during the next,
medio-lateral, cell division (Pasini et al., 2006). Ventral midline
induction surprisingly involves a different inducer than dorsal midline induction. Gain-of-function experiments indicate that activation
of the Bmp pathway during gastrulation is sufficient to convert most
B-line epidermal cells into midline fates. Conversely, inhibition of
ADMP function by morpholino injection leads to a specific loss of the
ventral, but not dorsal, epidermis midline (Pasini et al., 2006).
Consistent with a direct inducing role, ADMP is expressed at the
mid-gastrula stage in the B-line ventral midline vegetal cells, in
direct contact with epidermal midline precursors (Pasini et al.,
2006).
Conclusions
Work in the 19th and early 20th century led to the suggestion that
ascidian embryos form in a largely mosaic (i.e. cell-autonomous)
fashion, making use of their invariant lineage to precisely segregate
localised maternal determinants to the progenitors of the various cell
fates. The results presented in this review give a strikingly different
view of ascidian embryogenesis.
Invariant lineage and mosaic development are not synonymous
The first surprise is the sheer importance of cell–cell communication during ascidian early embryogenesis. By the mid-gastrula
stage, only around 10/300 cells form cell-autonomously: the
primary muscle and lateral tail epidermis. Additionally, some lateral
trunk epidermal cells may also form autonomously, but this will not
change the message that the overwhelming majority of ascidian
cells are induced during early development. In most cases, these
inductions occur at very short range, between contacting cells. Some
inductions, like that of the a6.5 neural progenitor by FGF9/16/20
even require a precise partitioning of the surface of contact with
inducing cells between induced and uninduced sisters. As we
generate 3D models of embryos for progressively later stages of
development, it will be interesting to test whether the precise
surface of contact is used generally to select induced cells among a
competent domain, or whether this is only critical for early FGF
signalling. In any case, the high prevalence of short-range signalling
probably imposes constraints on the lineage and may explain why it
has been so well conserved between distantly related ascidian
species.
The second surprise of recent ascidian work is that very few
signalling pathways are involved in early fate decisions. Some of these,
like FGF, nodal, Bmp or Notch, are “classical” ligands implicated in
numerous cell fate decisions across evolution. In contrast the role of
Ephrin as an antagonist of FGF signalling appears to be a peculiarity of
ascidians. Strikingly, because few pathways are involved in early fate
decisions, each of them can induce several fates in different cells. This
requires sophisticated mechanisms to control the competence of cells
to respond to inducers. For instance, the fate induced in various cells
by FGF between the 16- and late 32-cell stage is specified by a
combinatorial code involving a very limited number of transcription
factors. The ubiquitous maternal ETS factor ETS1/2 is required in all
lineages, but it cooperates with distinct regionally active factors in
each lineage. ETS1/2, FoxAa, and ZicL activate Brachyury in the
notochord (Kumano et al., 2006; Yagi et al., 2004), ETS1/2 and GATAa
activate Otx in the animal neural lineages (Bertrand et al., 2003),
while the factors cooperating with ETS1/2 to activate the early
mesenchyme marker Twist-like 1 are unknown. Likewise, in the B7.5
lineage, Mesp activation is specified by the combination of two
transcription factors, Lhx and Tbx6, whose genes are respectively
activated by β-catenin and Macho1, and define a domain of competence to respond to FGF during the late gastrula and neurula stages
(Christiaen et al., 2009). Interestingly in these cases, the control of the
domains of expression/activity of the competence factors occurs cellautonomously. The fixed ascidian lineage, rather than being the
hallmark of “mosaic” development, thus appears to regulate cell
communication in two complementary ways: it precisely partitions
maternal information to define initial competence domains, and it
controls the contacts between these competent blastomeres and the
short-range signalling sources.
The third surprise is that on several occasions Halocynthia and
Ciona, while sharing very similar or identical early lineages, and a
global logic of inductions, use different inducers. This is particular
P. Lemaire / Developmental Biology 332 (2009) 48–60
59
true for the secondary mesoderm lineages. For instance, the B-line
secondary notochord lineage is induced in two steps in both species.
The first step is FGF9/16/10-dependent in both species, but the
second steps differ. In Ciona, Nodal from the b6.5 line and Delta2
from A7.6 are acting in relay. In contrast in Halocynthia, although
the second step also involves contacts with the b6.5-line and A7.6,
the unidentified ligands operating act via pathways distinct from
Nodal and Notch. Thus, the constraints on cell contacts and the
position of cleavage planes appear to be stronger than those on the
use of particular signalling pathways. A similar finding has
previously been reported during vulval development in nematodes
(Félix, 2005).
Acknowledgments
How much conservation with other chordates?
Appendix A. Supplementary data
Ascidians thus sport an odd adult body plan, derived genomes and
surprisingly stereotyped early mode of development. Will the
emerging detailed understanding of the ascidian developmental
program teach us something about vertebrates?
Considering that ascidian and vertebrate embryos have very
similar fate maps from the onset of gastrulation (see Lemaire et al.,
2008 for a discussion of ascidian and vertebrate fate maps) and that
both make extensive usage of inductive processes, this appears likely.
If, however, one delves into the details, as exemplified in this review,
inducers and their targets often differ in the two taxa and our
current understanding of the ascidian and vertebrate programs
points to only a few conserved islands (e.g. the heart specification
network, Davidson, 2007) in a sea of differences (reviewed in
Lemaire et al., 2008; Lemaire, 2006). Interestingly the similarity of
ascidian and vertebrate programs does not increase significantly as
development proceeds and the body forms become increasingly
alike. A large-scale comparison of the in situ hybridisation patterns of
Ciona and zebrafish orthologous genes reveals that a surprising level
of divergence of individual expression patterns of both regulatory
and non-regulatory genes extends throughout embryogenesis
(Sobral et al., submitted).
In spite of these puzzling results, it is difficult to imagine that
totally unrelated developmental programs can lead to such similar
tadpole larval forms. The comparison of the Ciona and Halocynthia
programs presented here suggests that the cellular logic may be
more constrained than the molecular identity of individual players.
But the cellular logic itself is under transcriptional control, and
there must be a level at which commonalities can be detected.
Because of their simpler gene content, suitability to functional
genetics and imaging approaches, ascidians are among the few
multicellular organisms for which an integrated systems-level
understanding of the developmental program, from transcriptional
processes to the mechanics of cell behaviour may be obtained (for
an example of an integrated view of the transcriptional control of
cell behaviour, see Christiaen et al., 2008). Yet, although large
transcriptional networks (e.g. Imai et al., 2006) are starting to
emerge, these networks have so far only modestly contributed to
our understanding of Ciona embryogenesis, and its evolutionary
conservation between chordates.
This may be because the networks are still very incomplete.
Systematic chromatin immunoprecipitation experiments for major
nodes of the current network are currently underway in Ciona (Y.
Satou, personal communication) and for some orthologous nodes in
vertebrates (e.g.. Brachyury: Morley et al., 2009). These approaches
may help identify conserved chordate linkages in regulatory networks. In parallel, networks act in a precise cellular context, and it will
be important to integrate the transcriptional networks with the
position and behaviour of cells. This will certainly benefit from the
small number of transparent cells in ascidian early embryos, whose
precise individual morphology can be quantified in 3D reconstructed
embryos (Tassy et al., 2006).
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.ydbio.2009.05.540.
I am grateful for discussions, comments and personal communications from Hiroki Nishida (Osaka, Japan), Clare Hudson and
Hitoyoshi Yasuo (Villefranche-sur-Mer, France), Lionel Christiaen
(Berkeley, USA) and my colleagues at IBDML (Marseilles, France).
The picture of an adult H. roretzi shown on Fig. 1A is a kind gift from
Hiroki Nishida. Work in the laboratory is supported by CNRS and by
grants from the ANR (Chor-reg-net and Chor-evo-net) and from the
European Union (Transcode, Plurigenes, Myores, Marine Genomics
Europe).
References
Bertrand, V., Hudson, C., Caillol, D., Popovici, C., Lemaire, P., 2003. Neural tissue in
ascidian embryos is induced by FGF9/16/20, acting via a combination of maternal
GATA and Ets transcription factors. Cell 115, 615–627.
Bourlat, S.J., et al., 2006. Deuterostome phylogeny reveals monophyletic chordates and
the new phylum Xenoturbellida. Nature 444, 85–88.
Chabry, L., 1887. Embryologie normale et tératologique des Ascidie. Felix Alcan Editeur,
Paris.
Christiaen, L., Stolfi, A., Davidson, B., Levine, M., 2009. Spatio-temporal intersection of
Lhx3 and Tbx6 defines the cardiac field through synergistic activation of Mesp. Dev.
Biol. 328, 552–560.
Christiaen, L., Davidson, B., Kawashima, T., Powell, W., Nolla, H., Vranizan, K., Levine, M.,
2008. The transcription/migration interface in heart precursors of Ciona intestinalis.
Science 320, 1349–1352.
Conklin, E., 1905. The organization and cell lineage of the ascidian egg. J. Acad., Nat. Sci.
Phila 13, 1.
Darras, S., Nishida, H., 2001. The BMP signaling pathway is required together with the
FGF pathway for notochord induction in the ascidian embryo. Development 128,
2629–2638.
Davidson, B., 2007. Ciona intestinalis as a model for cardiac development. Semin. Cell
Dev. Biol. 18, 16–26.
Davidson, B., Shi, W., Beh, J., Christiaen, L., Levine, M., 2006. FGF signaling delineates the
cardiac progenitor field in the simple chordate, Ciona intestinalis. Genes Dev. 20,
2728–2738.
Davidson, B., Shi, W., Levine, M., 2005. Uncoupling heart cell specification and migration
in the simple chordate Ciona intestinalis. Development 132, 4811–4818.
Dehal, P., et al., 2002. The draft genome of Ciona intestinalis: insights into chordate and
vertebrate origins. Science 298, 2157–2167.
Delsuc, F., Brinkmann, H., Chourrout, D., Philippe, H., 2006. Tunicates and not
cephalochordates are the closest living relatives of vertebrates. Nature 439,
965–968.
Félix, M., 2005. An inversion in the wiring of an intercellular signal: evolution of Wnt
signaling in the nematode vulva. BioEssays 27, 765–769.
Hill, M.M., Broman, K.W., Stupka, E., Smith, W.C., Jiang, D., Sidow, A., 2008. The
C. savignyi genetic map and its integration with the reference sequence
facilitates insights into chordate genome evolution. Genome Res. 18,
1369–1379.
Holland, L.Z., Gibson-Brown, J.J., 2003. The Ciona intestinalis genome: when the
constraints are off. BioEssays 25, 529–532.
Hotta, K., Mitsuhara, K., Takahashi, H., Inaba, K., Oka, K., Gojobori, T., Ikeo, K., 2007.
A web-based interactive developmental table for the ascidian Ciona intestinalis,
including 3D real-image embryo reconstructions: I. From fertilized egg to
hatching larva. Dev. Dyn. 236, 1790–1805.
Hudson, C., Lemaire, P., 2001. Induction of anterior neural fates in the ascidian
Ciona intestinalis. Mech. Dev. 100, 189–203.
Hudson, C., Darras, S., Caillol, D., Yasuo, H., Lemaire, P., 2003. A conserved role for the
MEK signalling pathway in neural tissue specification and posteriorisation in the
invertebrate chordate, the ascidian Ciona intestinalis. Development 130, 147–159.
Hudson, C., Lotito, S., Yasuo, H., 2007. Sequential and combinatorial inputs from Nodal,
Delta2/Notch and FGF/MEK/ERK signalling pathways establish a grid-like
organisation of distinct cell identities in the ascidian neural plate. Development
134, 3527–3537.
Hudson, C., Yasuo, H., 2006. A signalling relay involving Nodal and Delta ligands acts
during secondary notochord induction in Ciona embryos. Development 133,
2855–2864.
Hudson, C., Yasuo, H., 2005. Patterning across the ascidian neural plate by lateral Nodal
signalling sources. Development 132, 1199–1210.
Hudson, C., Yasuo, H., 2008. Similarity and diversity in mechanisms of muscle fate
induction between ascidian species. Biol. Cell 100, 265–277.
Imai, K., Takada, N., Satoh, N., Satou, Y., 2000. (beta)-catenin mediates the specification
of endoderm cells in ascidian embryos. Development 127, 3009–3020.
60
P. Lemaire / Developmental Biology 332 (2009) 48–60
Imai, K.S., Hino, K., Yagi, K., Satoh, N., Satou, Y., 2004. Gene expression profiles of
transcription factors and signaling molecules in the ascidian embryo: towards a
comprehensive understanding of gene networks. Development 131, 4047–4058.
Imai, K.S., Levine, M., Satoh, N., Satou, Y., 2006. Regulatory blueprint for a chordate
embryo. Science 312, 1183–1187.
Imai, K.S., Satoh, N., Satou, Y., 2002. An essential role of a FoxD gene in notochord
induction in Ciona embryos. Development 129, 3441–3453.
Imai, K.S., Stolfi, A., Levine, M., Satou, Y., 2009. Gene regulatory networks underlying the
compartmentalization of the Ciona central nervous system. Development 136,
285–293.
Jeffery, W.R., 2007. Chordate ancestry of the neural crest: new insights from ascidians.
Semin. Cell Dev. Biol. 18, 481–491.
Kawai, N., Iida, Y., Kumano, G., Nishida, H., 2007. Nuclear accumulation of beta-catenin
and transcription of downstream genes are regulated by zygotic Wnt5alpha and
maternal Dsh in ascidian embryos. Dev. Dyn. 236, 1570–1582.
Kim, G.J., Yamada, A., Nishida, H., 2000. An FGF signal from endoderm and localized
factors in the posterior–vegetal egg cytoplasm pattern the mesodermal tissues in
the ascidian embryo. Development 127, 2853–2862.
Kim, G.J., Kumano, G., Nishida, H., 2007. Cell fate polarization in ascidian
mesenchyme/muscle precursors by directed FGF signaling and role for an
additional ectodermal FGF antagonizing signal in notochord/nerve cord
precursors. Development 134, 1509–1518.
Kobayashi, K., Sawada, K., Yamamoto, H., Wada, S., Saiga, H., Nishida, H., 2003.
Maternal macho-1 is an intrinsic factor that makes cell response to the same FGF
signal differ between mesenchyme and notochord induction in ascidian embryos.
Development 130, 5179–5190.
Kondoh, K., Kobayashi, K., Nishida, H., 2003. Suppression of macho-1-directed muscle
fate by FGF and BMP is required for formation of posterior endoderm in ascidian
embryos. Development 130, 3205–3216.
Kourakis, M.J., Smith, W.C., 2005. Did the first chordates organize without the
organizer? Trends Genet. 21, 506–510.
Kumano, G., Nishida, H., 2007. Ascidian embryonic development: an emerging model
system for the study of cell fate specification in chordates. Dev. Dyn. 236,
1732–1747.
Kumano, G., Yamaguchi, S., Nishida, H., 2006. Overlapping expression of FoxA and Zic
confers responsiveness to FGF signaling to specify notochord in ascidian embryos.
Dev. Biol. 300, 770–784.
Kuratani, S., Wada, H., Kusakabe, R., Agata, K., 2006. Evolutionary embryology
resurrected in Japan with a new molecular basis: Nori Satoh and the history of
ascidian studies originating in Kyoto during the 20th century. Int. J. Dev. Biol. 50,
451–454.
Lamy, C., Rothbächer, U., Caillol, D., Lemaire, P., 2006. Ci-FoxA-a is the earliest zygotic
determinant of the ascidian anterior ectoderm and directly activates Ci-sFRP1/5.
Development 133, 2835–2844.
Lemaire, P., 2006. Developmental biology. How many ways to make a chordate? Science
312, 1145–1146.
Lemaire, P., Smith, W.C., Nishida, H., 2008. Ascidians and the plasticity of the chordate
developmental program. Curr. Biol. 18, R620–R631.
Meedel, T.H., Farmer, S.C., Lee, J.J., 1997. The single MyoD family gene of Ciona intestinalis
encodes two differentially expressed proteins: implications for the evolution of
chordate muscle gene regulation. Development 124, 1711–1721.
Meedel, T.H., Chang, P., Yasuo, H., 2007. Muscle development in Ciona intestinalis
requires the b-HLH myogenic regulatory factor gene Ci-MRF. Dev. Biol. 302,
333–344.
Minokawa, T., Yagi, K., Makabe, K.W., Nishida, H., 2001. Binary specification of nerve
cord and notochord cell fates in ascidian embryos. Development 128,
2007–2017.
Miwata, K., Chiba, T., Horii, R., Yamada, L., Kubo, A., Miyamura, D., Satoh, N., Satou, Y.,
2006. Systematic analysis of embryonic expression profiles of zinc finger genes in
Ciona intestinalis. Dev. Biol. 292, 546–554.
Morley, R.H., Lachani, K., Keefe, D., Gilchrist, M.J., Flicek, P., Smith, J.C., Wardle, F.C., 2009.
A gene regulatory network directed by zebrafish No tail accounts for its roles in
mesoderm formation. Proc. Natl. Acad. Sci. U. S. A. 106, 3829–3834.
Munro, E., 2007. Asymmetric cell division: a CAB driver for spindle movements. Curr.
Biol. 17, R639–R641.
Nakamura, Y., Makabe, K.W., Nishida, H., 2005. POPK-1/Sad-1 kinase is required for the
proper translocation of maternal mRNAs and putative germ plasm at the posterior
pole of the ascidian embryo. Development 132, 4731–4742.
Nakamura, Y., Makabe, K.W., Nishida, H., 2006. The functional analysis of Type I
postplasmic/PEM mRNAs in embryos of the ascidian Halocynthia roretzi. Dev.
Genes Evol. 216, 69–80.
Negishi, T., Takada, T., Kawai, N., Nishida, H., 2007. Localized PEM mRNA and protein are
involved in cleavage-plane orientation and unequal cell divisions in ascidians. Curr.
Biol. 17, 1014–1025.
Nishida, H., 1987. Cell lineage analysis in ascidian embryos by intracellular injection of a
tracer enzyme. III. Up to the tissue restricted stage. Dev. Biol. 121, 526–541.
Nishida, H., 1990. Determinative mechanisms in secondary muscle lineages of ascidian
embryos: development of muscle-specific features in isolated muscle progenitor
cells. Development 108, 559–568.
Nishida, H., Sawada, K., 2001. macho-1 encodes a localized mRNA in ascidian eggs that
specifies muscle fate during embryogenesis. Nature 409, 724–729.
Nishida, H., 2003. Spatio-temporal pattern of MAP kinase activation in embryos of the
ascidian Halocynthia roretzi. Dev. Growth Differ. 45, 27–37.
Nishida, H., 2005. Specification of embryonic axis and mosaic development in ascidians.
Dev. Dyn. 233, 1177–1193.
Oda-Ishii, I., Bertrand, V., Matsuo, I., Lemaire, P., Saiga, H., 2005. Making very similar
embryos with divergent genomes: conservation of regulatory mechanisms of Otx
between the ascidians Halocynthia roretzi and Ciona intestinalis. Development 132,
1663–1674.
Pasini, A., Amiel, A., Rothbächer, U., Roure, A., Lemaire, P., Darras, S., 2006. Formation of
the ascidian epidermal sensory neurons: insights into the origin of the chordate
peripheral nervous system. PLoS Biol. 4, e225.
Picco, V., Hudson, C., Yasuo, H., 2007. Ephrin–Eph signalling drives the asymmetric division
of notochord/neural precursors in Ciona embryos. Development 134, 1491–1497.
Putnam, N.H., et al., 2008. The amphioxus genome and the evolution of the chordate
karyotype. Nature 453, 1064–1071.
Rothbächer, U., Bertrand, V., Lamy, C., Lemaire, P., 2007. A combinatorial code of
maternal GATA, Ets and beta-catenin-TCF transcription factors specifies and
patterns the early ascidian ectoderm. Development 134, 4023–4032 (the first
two authors contributed equally).
Roure, A., Rothbächer, U., Robin, F., Kalmar, E., Ferone, G., Lamy, C., Missero, C., Mueller,
F., Lemaire, P., 2007. A multicassette Gateway vector set for high throughput and
comparative analyses in Ciona and vertebrate embryos. PLoS ONE 2, e916.
Sardet, C., Paix, A., Prodon, F., Dru, P., Chenevert, J., 2007. From oocyte to 16-cell
stage: cytoplasmic and cortical reorganizations that pattern the ascidian embryo.
Dev. Dyn. 236, 1716–1731.
Satou, Y., Satoh, N., Imai, K.S., 2008. Gene Regulatory Networks in the Early Ascidian
Embryo. Biochim. Biophys. Acta 1789, 268–273.
Shi, W., Levine, M., 2008. Ephrin signaling establishes asymmetric cell fates in an
endomesoderm lineage of the Ciona embryo. Development 135, 931–940.
Shirae-Kurabayashi, M., Nishikata, T., Takamura, K., Tanaka, K.J., Nakamoto, C.,
Nakamura, A., 2006. Dynamic redistribution of vasa homolog and exclusion of
somatic cell determinants during germ cell specification in Ciona intestinalis.
Development 133, 2683–2693.
Small, K.S., Brudno, M., Hill, M.M., Sidow, A., 2007. A haplome alignment and reference
sequence of the highly polymorphic Ciona savignyi genome. Genome Biol. 8, R41.
Strähle, U., Lam, C.S., Ertzer, R., Rastegar, S., 2004. Vertebrate floor-plate specification:
variations on common themes. Trends Genet. 20, 155–162.
Tassy, O., Daian, F., Hudson, C., Bertrand, V., Lemaire, P., 2006. A quantitative approach
to the study of cell shapes and interactions during early chordate embryogenesis.
Curr. Biol. 16, 345–358.
Tokuoka, M., Kumano, G., Nishida, H., 2007. FGF9/16/20 and Wnt-5alpha signals are
involved in specification of secondary muscle fate in embryos of the ascidian,
Halocynthia roretzi. Dev. Genes Evol. 217, 515–527.
Tomioka, M., Miya, T., Nishida, H., 2002. Repression of zygotic gene expression in the
putative germline cells in ascidian embryos. Zoolog. Sci. 19, 49–55.
Wada, S., Hamada, M., Kobayashi, K., Satoh, N., 2008. Novel genes involved in
canonical Wnt/beta-catenin signaling pathway in early Ciona intestinalis
embryos. Dev Growth Differ. 50, 215–227.
Wilson, D., Charoensawan, V., Kummerfeld, S.K., Teichmann, S.A., 2008. DBD—
taxonomically broad transcription factor predictions: new content and
functionality. Nucleic Acids Res. 36, D88–D92.
Yagi, K., Satoh, N., Satou, Y., 2004. Identification of downstream genes of the ascidian
muscle determinant gene Ci-macho1. Dev. Biol. 274, 478–489.
Yasuo, H., Hudson, C., 2007. FGF8/17/18 functions together with FGF9/16/20 during
formation of the notochord in Ciona embryos. Dev. Biol. 302, 92–103.