transcriptional regulation in the early ectodermal lineage of ascidian embryos
TRANSCRIPT
Original Article
Transcriptional regulation in the early ectodermal lineageof ascidian embryos
Yosuke Horikawa, Haruka Matsumoto, Fumika Yamaguchi,† Satomi Ishida andShigeki Fujiwara*
Department of Applied Science, Kochi University, 2-5-1 Akebono-cho, Kochi-shi, Kochi, 780-8520, Japan
In ascidian embryos, ectodermal tissues derive from blastomeres in the animal hemisphere. The animal hemi-sphere-specific gene expression is observed as early as the 16-cell stage. Here, we characterized animal hemi-sphere-specific enhancers of three genes, Ci-ephrin-Ad, Ci-TGFb-NA1 and Ci-Fz4. Deletion analyses identifiedminimal essential elements. Although these elements contained multiple GATA sequences, electrophoreticmobility shift assays revealed that only some of them were strong binding sites for the transcription factorCi-GATAa. On the other hand, the motif-searching software MEME identified an octamer, GA (T/G) AAGGG,shared by these enhancers. In Ci-ephrin-Ad and Ci-TGFb-NA1, the octamer was GATAAGGG, which stronglybound Ci-GATAa. The 397-bp upstream region of Ci-ephrin-Ad contained two strong Ci-GATAa-binding sites,one of which was the octamer motif. Mutation in the octamer motif, but not the other Ci-GATAa-binding site,severely affected the enhancer activity. The 204-bp upstream region of Ci-TGFb-NA1 contained four strongCi-GATAa-binding sites, including the octamer motif. Mutation only in the octamer motif, leaving the otherthree Ci-GATAa-binding sites intact, abolished the enhancer activity. These results suggest a crucial role for theoctamer motif.
Key words: ascidian, ectoderm, enhancer, GATA, transcriptional activation.
Introduction
The ectoderm arises from the animal hemisphere in
chordates, including amphibians and ascidians (Lem-
aire et al. 2008). It is subsequently subdivided into the
epidermis and central nervous system. The mecha-nism of neural induction has been extensively studied
(for reviews see Stern 2005; Schmidt et al. 2013). Fac-
tors promoting epidermal specification and suppress-
ing neural fate have also been identified, which are
activated by bone morphogenetic protein (e.g. Qiao
et al. 2012; Tr�ıbulo et al. 2012). However, few genes
have been identified that were expressed in the
uncommitted general ectoderm. Our knowledge about
the mechanisms of specification of the epidermis and
general ectoderm is fragmentary.
The third cleavage of ascidian embryos divides the
embryo into four animal and four vegetal blastomeres.
The animal blastomeres mainly give rise to the epider-
mis and central nervous system (Conklin 1905; Nishida1987). Nishikata et al. (2001) reported that mRNAs of
a few genes were predominantly localized in the ani-
mal blastomeres of the 8-cell embryo, although it is
not clear whether this localization is achieved by spe-
cific zygotic expression or asymmetrical distribution of
maternal mRNA. The mRNAs localized to the animal
blastomeres include those encoding a WD repeat-con-
taining protein, an F-actin-capping protein, a calcineurininhibitor, and a putative GTP-binding protein (Nishikata
et al. 2001). It is unclear whether these proteins are
involved in the ectodermal genetic regulatory cascade.
At the 16-cell stage, many genes are expressed
exclusively in the animal hemisphere. These include
the genes encoding transcription factors; Ci-SoxF
(Imai et al. 2004) and Ci-fog (Rothb€acher et al. 2007).
Components of cell–cell signaling pathways are alsoexpressed in the animal hemisphere; Ci-ephrin-Aa,
Ci-ephrin-Ad, Ci-TGFb-NA1, Ci-Smad1/5 and Ci-Fz4
*Author to whom all correspondence should be addressed.Email: [email protected]†Present address: Laboratory of Molecular Biology, Medical
Research Center, Kochi Medical School, Kochi, 783-8505,Japan.Received 3 September 2012; revised 9 October 2013;
accepted 10 October 2013.© 2013 The AuthorsDevelopment, Growth & Differentiation © 2013 Japanese
Society of Developmental Biologists
Develop. Growth Differ. (2013) 55, 776–785 doi: 10.1111/dgd.12100
The Japanese Society of Developmental Biologists
(Imai et al. 2004; Hamaguchi et al. 2007; Picco et al.
2007). Cell adhesion molecules (Ci-d1-protocadherin-like and Ci-d-protocadherin-4) are also specifically
expressed (Noda & Satoh 2008).
Transcription of Ci-fog in the animal hemisphere is
activated by the transcription factor Ci-GATAa
(Rothb€acher et al. 2007). A Ci-GATAa-specific mor-
pholino oligo suppresses the enhancer activity of
Ci-fog (Rothb€acher et al. 2007). Transcriptional activa-tion of Ci-otx in the animal hemisphere also requires
Ci-GATAa (Bertrand et al. 2003). The enhancer regions
of these genes contain multiple GATA sequences (Ber-
trand et al. 2003; Rothb€acher et al. 2007), although
binding of the Ci-GATAa protein to these sequences
was not demonstrated. In addition, it is not yet clear
whether only Ci-GATAa is responsible for activation of
animal hemisphere-specific genes.In the present study, we identified animal hemisphere-
specific enhancer elements within the 5′ flanking regions
of three genes (Ci-ephrin-Ad, Ci-TGFb-NA1 and Ci-Fz4).
These genes were chosen because their specific
expression was clear and strong (Imai et al. 2004;
Picco et al. 2007). We found multiple GATA
sequences in the upstream regions of these genes.
Electrophoretic mobility shift assays revealed thatsome, but not all, of them strongly bind the
Ci-GATAa protein. However, disruption of only one of
them, whose GATA sequence was followed by AGGG,
severely affected the enhancer activity. Based on these
results, the possible involvement of transcription factor
(s) other than Ci-GATAa is discussed.
Materials and methods
Animals
Juvenile adults of Ciona intestinalis were purchased
from the National Bio-Resource Project of MEXT,
Japan. The animals were cultured in Tosa Bay near
the Usa Marine Biological Institute of Kochi University.
Eggs and sperm were obtained from the gonoducts ofmature adults. Eggs were inseminated with non-self
sperm. Fertilized eggs were dechorionated with 0.05%
actinase E (Kaken Pharmaceutical) and 1% sodium
thioglycolate. Embryos were reared in artificial seawa-
ter Super Marine Art SF1 (Tomita Pharmaceutical).
Construction of plasmids
Genomic DNA fragments corresponding to the pro-
moter region of Ci-ephrin-Ad, Ci-TGFb-NA1 and Ci-
Fz4 were amplified by polymerase chain reaction
(PCR), using the primers ephrin1.7F and ephrin1.7R
for Ci-ephrin-Ad, tgfna3.0F and tgfna3.0R for
Ci-TGFb-NA1, and fz2.0F and fz2.0R for Ci-Fz4 (TableS1). The PCR products were inserted into pGEM-T
(Promega). The Ci-ephrin-Ad fragment was excised
with PstI and BamHI from pGEM-T, while the
Ci-TGFb-NA1 and Ci-Fz4 fragments were excised with
XhoI and BamHI. These fragments were inserted into
pSP72-1.27 (Corbo et al. 1997). The resultant plas-
mids were named Ephrin1648Z, Tgfna2940Z and
Fz1964Z, respectively.For construction of Ephrin545Z, Ephrin497Z, Eph-
rin447Z, Ephrin397Z and Ephrin333Z, PCR was
carried out using a lacZ-specific primer (gal-A2) and a
Ci-ephrin-Ad-specific primer (ephrin545F, ephrin497F,
ephrin447F, ephrin397F and ephrin333F, respectively;
Table S1). Ephrin1648Z was used as a template. For
construction of Tgfna799Z, Tgfna599Z, Tgfna399Z,
Tgfna204Z and Tgfna66Z, PCR was carried out usingtgfna3.0R and one of the following primers (tgfna799F,
tgfna599F, tgfna399F, tgfna204F and tgfna66F,
respectively; Table S1). Tgfna2940Z was used as a
template. For construction of Fz1425Z, Fz1294Z,
Fz975Z, Fz784Z and Fz387Z, PCR was carried out
using fz2.0R and one of the following primers
(fz1425F, fz1294F, fz975F, fz784F and fz387F, respec-
tively; Table S1). Fz1964Z was used as a template.PCR products were once inserted into pGEM-T,
excised with XhoI and BamHI, and inserted into
pSP72-1.27.
Site-directed mutagenesis was performed according
to the protocol described by Fujiwara et al. (1998). To
create point mutations within the Ci-ephrin-Ad enhan-
cer, the XhoI-BamHI fragment of Ephrin397Z was
inserted into pBluescript II SK+ (Stratagene). Uracil-containing single-stranded DNA was prepared using
the Escherichia coli strain CJ236 (Takara). Mutagenic
oligonucleotides used were E-S1 m, E-S2 m, E-G1 m,
E-G2 m, E-S1G1 m and E-Mm (Table S1). For point
mutations in the Ci-TGFb-NA1 enhancer, we trans-
formed CJ236 with pGEM-T containing the PCR prod-
uct using primers tgfna204F and tgfna3.0R (described
above). Mutagenic oligonucleotides used were T-Gmand T-Mm (Table S1).
Electroporation and detection of transgene expression
Plasmid DNA was prepared using QIAGEN tip-100
(Qiagen). Transgenes were introduced into dechorion-
ated embryos by electroporation, according to the pro-
tocol described by Corbo et al. (1997). Expression oflacZ was detected by in situ hybridization as described
by Kanda et al. (2009). In each experiment, 100–400electroporated embryos were observed under an
AZ100 microscope (Nikon). Embryos expressing lacZ
in one or more blastomeres were counted as positive.
ª 2013 The Authors
Development, Growth & Differentiation ª 2013 Japanese Society of Developmental Biologists
Early ectodermal enhancers in ascidians 777
The percentage of positive embryos among morpho-logically normal embryos was calculated.
Electrophoretic mobility shift assay
The entire translated region of Ci-GATAa was amplified
by PCR using the cDNA clone “cien229100” as a tem-
plate (Tassy et al. 2010; http://www.aniseed.cnrs.fr/
index.php). The PCR was carried out using primersGATA-F and GATA-Nhe-R (Table S1), and Tks Gflex
DNA polymerase (Takara), according to the protocol
supplied by the manufacturer. The product of the PCR
was briefly treated with Taq DNA polymerase (Ampli-
qon) to create the 3′ overhang of a single adenine
nucleotide, and inserted into pGEM-T. The Ci-GATAa
cDNA was again amplified using Tks Gflex DNA poly-
merase and the primers GATA-F and M13-Forward-40(5′-GTTTTCCCAGTCACGAC-3′). The product of the
PCR was excised with NheI. The pCMX-hRARa plas-
mid (Umesono et al. 1991) was digested with MscI
and NheI to remove the cDNA encoding human reti-
noic acid receptor a (hRARa). The Ci-GATAa cDNA
was inserted into the pCMX vector. The Ci-GATAa
protein was produced using the TNT coupled reticulo-
cyte lysate system (Promega). Double-stranded DNA
probes were labeled with [c-32P] adenosine triphos-phate (Perkin Elmer Japan), according to the protocol
described by Kanda et al. (2009). Electrophoretic
mobility shift assays (EMSAs) were carried out using a
Gel-shift assay system (Promega). Autoradiograms
were obtained using BAS2500 (Fuji Film).
Results
Identification of early animal-hemisphere enhancers
A fragment of the C. intestinalis genomic DNA was
obtained, which contained the promoter region of Ci-
ephrin-Ad, including putative transcription and transla-
tion initiation sites (Fig. 1a). The fragment was fused
in-frame with lacZ and introduced into fertilized eggs
by electroporation (Fig. 1b). The resultant transgene,named Ephrin1648Z, contained 1648 bp of the 5′flanking region, according to the gene model
KH.C3.716.v1.A.SL1-1 in the C. intestinalis genome
database (Satou et al. 2005; http://ghost.zool.kyoto-u.
ac.jp/SearchGenomekh.html). Ephrin1648Z was intro-
duced into one-cell embryos. The embryos were fixed
at the 16-cell stage and the lacZ mRNA was detected
by in situ hybridization. Ephrin1648Z was expressed in
(a)
(b)
(c) (d)
(e) (f)
(g) (h)
Fig. 1. The animal hemisphere enhancer of Ci-ephrin-Ad. (a) The genomic structure of the Ci-ephrin-Ad gene, predicted in the Ciona
intestinalis genome database (Satou et al. 2005; http://ghost.zool.kyoto-u.ac.jp/SearchGenomekh.html). Exons are indicated by boxes.
The open reading frame is in green. The 1.6-kb upstream region examined in the present study is indicated. (b) Diagrams showing the
different 5´ flanking regions used for the analysis in (c–h). The E1 enhancer is indicated by a red box. Percentage of embryos expressing
lacZ among normally developing embryos is indicated. (c–h) The expression pattern of transgenes shown in (b). All panels show the
animal side of the 16-cell embryos. The anterior side is up. Names of the blastomeres are indicated in (c).
ª 2013 The Authors
Development, Growth & Differentiation ª 2013 Japanese Society of Developmental Biologists
778 Y. Horikawa et al.
all of the blastomeres (a5.3, a5.4, b5.3 and b5.4 blas-tomere pairs) in the animal hemisphere (Fig. 1c).
Transgenes containing 545, 497, 447, or 397 bp of
the upstream sequence gave essentially similar pat-
terns of expression, although some embryos did not
express lacZ in all of the eight blastomeres (Fig. 1d–g).No embryo carrying Ephrin333Z expressed lacZ in any
blastomere (Fig. 1h). These results indicate that 397
bp of the 5′ flanking region of Ci-ephrin-Ad is suffi-cient, and the region between �397 and �334,
named E1, is necessary for transcriptional activation in
the animal hemisphere.
We also isolated the promoter regions of Ci-TGFb-NA1and Ci-Fz4, and placed them upstream of lacZ. (Figs 2
and 3). A transgene named Tgfna2940Z contained
2940 bp of the 5′ flanking region of Ci-TGFb-NA1,according to the gene model KH.C4.547.v1.A.nonSL1-1(Fig. 2a,b). Tgfna2940Z was expressed in all of the
eight blastomeres in the animal hemisphere of the
16-cell embryo (Fig. 2c). Deletion analyses revealed
that 204 bp of the upstream sequence of Ci-TGFb-NA1 was sufficient, and the region between �204 and
�66, named T1, was necessary for transcriptional acti-
vation in the animal hemisphere (Fig. 2d–h). Fz1964Z
contained 1964 bp of the 5′ flanking region of Ci-Fz4,according to the gene model KH.C6.162.v2.A.
SL2-1 (Fig. 3a,b). Fz1964Z recapitulated the animal
hemisphere-specific expression of Ci-Fz4 (Fig. 3c).Deletion analyses revealed that 975 bp of the upstream
sequence of Ci-Fz4 was sufficient, and the region
between �975 and �784, named F1, was necessary
for transcriptional activation in the animal hemisphere
(Fig. 3d–h). The region between �783 and �67 was
deleted from Fz975Z. The resultant transgene was
not expressed, suggesting that the proximal region
(downstream to �784) also contains enhancer element(s) necessary for transcriptional activation (data not
shown).
Possible binding sites for transcription factors
We searched for sequence motifs shared by Ci-eph-
rin-Ad, Ci-TGFb-NA1 and Ci-Fz4, using MEME (Bailey
& Elkan 1994; http://meme.sdsc.edu/meme/intro.html).MEME identified an octamer sequence, GA (T/G)
AAGGG. The E1 and T1 enhancers contained a single
octamer motif each (GATAAGGG), named E-M1 and
T-M1, respectively (Fig. 4). MEME did not find this
motif within the 975-bp proximal region of Ci-Fz4.
However, we found a similar sequence, GAAAAGGG,
at nucleotide position �73 of Ci-Fz4 (data not shown).
We also searched for putative binding sites for knowntranscription factors. Ci-GATAa is necessary for activa-
tion of genes in the animal hemisphere of the Ciona
(a)
(b)
(c) (d)
(e) (f)
(g) (h)
Fig. 2. The animal hemisphere enhancer of Ci-TGFb-NA1. (a) The genomic structure of the Ci-TGFb-NA1 gene, predicted in the Ciona
intestinalis genome database. Exons are indicated by boxes. The open reading frame is in green. The 2.94-kb upstream region exam-
ined in the present study is indicated. (b) Diagrams showing the different 5´ flanking regions used for the analysis in (c–h). The T1 enhan-
cer is indicated by a red box. Percentage of the embryos expressing lacZ among normally developing embryos is indicated. (c–h) The
expression pattern of transgenes shown in (b). All panels show the animal side of the 16-cell embryos. The anterior side is up.
ª 2013 The Authors
Development, Growth & Differentiation ª 2013 Japanese Society of Developmental Biologists
Early ectodermal enhancers in ascidians 779
embryo (Rothb€acher et al. 2007). The binding
sequence for Ci-GATAa is 5′-GATA-3′, and it had no
strong constraint on surrounding sequences (K. Nitta
& P. Lemaire, pers. comm., 2012). Ci-SoxF is the only
transcription factor whose mRNA is exclusively
detected in the animal hemisphere at the 16-cell stage(Imai et al. 2004). The core recognition sequence for
the Sox group transcription factors is AACAAT (Mertin
et al. 1999). In vitro binding experiments revealed that
the consensus sequence for C. intestinalis Sox pro-
teins is (A/G) (A/C) CAA (T/A) (K. Nitta & P. Lemaire,
pers. comm., 2012). We found six GATA-binding sites
and two Sox-binding sites within the 397-bp proximal
region of Ci-ephrin-Ad (Fig. 4a). The E1 enhancer con-tained two GATA-binding sites (named E-G1 and E-
G2) and one Sox-binding site (named E-S1; Fig. 4a,b).
The 204-bp upstream region of Ci-TGFb-NA1 con-
tained no Sox-binding site, but four GATA-binding
sites (Fig. 4a). The T1 enhancer contained three of
them (named T-G1–T-G3; Fig. 4c). E-M1 and T-M1
included the GATA sequence (E-G2 and T-G3, respec-
tively, Fig. 4).
The octamer motif is necessary for enhancer activity
Since the E1 enhancer is necessary for transcriptional
activation in the animal hemisphere, we tested the
requirement of putative binding sites for transcription
factors by site-directed mutagenesis. First, a point
mutation was generated at the Sox-binding site within
the E1 enhancer (E-S1; Fig. 5a,b). The mutant trans-
gene, named Ephrin (S1 m) Z, was expressed normally
in the animal hemisphere at the 16-cell stage (Fig. 5b).Another possible Sox-binding sequence was found
downstream of nucleotide position �333 (E-S2;
Fig. 5a). Disruption of E-S2, or a double mutation in
E-S1 and E-S2, did not affect the expression in the
animal hemisphere (Fig. 5c,d). One of the two GATA-
binding sites within the E1 enhancer (E-G1) was
mutagenized (Fig. 5a,e). The mutant transgene, named
Ephrin (G1 m) Z, was expressed in the animal hemi-sphere (Fig. 5e). In contrast, a mutation in the other
GATA sequence (E-G2) caused complete silencing of
the reporter gene (Fig. 5f). Note that E-G2 overlaps
with E-M1 (GATAAGGG) (Figs. 4b and 5a), and that a
point mutation in E-G2 completely inactivated the
enhancer that still contained five additional GATA
sequences (see Fig. 4a). We then introduced a point
mutation into E-M1 without disrupting E-G2 (GAT-AAGGG to GATActtt). The resultant transgene, named
Ephrin (M1 m) Z, was not expressed in any blastomere
(Fig. 5g). About 83% of the embryos carrying Eph-
rin397Z expressed lacZ, while only 14% of the
embryos carrying Ephrin (M1 m) Z expressed it.
(a)
(b)
(c) (d)
(e) (f)
(g) (h)
Fig. 3. The animal hemisphere enhancer of Ci-Fz4. (a) The genomic structure of the Ci-Fz4 gene, predicted in the Ciona intestinalis
genome database. Exons are indicated by boxes. The open reading frame is in green. The 1.96-kb upstream region examined in the
present study is indicated. (b) Diagrams showing the different 5´ flanking regions used for the analysis in (c–h). The F1 enhancer is indi-
cated by a red box. Percentage of the embryos expressing lacZ among normally developing embryos is indicated. (c–h) The expression
pattern of transgenes shown in (b). All panels show the animal side of the 16-cell embryos. The anterior side is up.
ª 2013 The Authors
Development, Growth & Differentiation ª 2013 Japanese Society of Developmental Biologists
780 Y. Horikawa et al.
The T1 enhancer contains three GATA sequences,one of which (T-G3) overlaps with T-M1 (GATAAGGG)
(Figs 4c and 6a). Disruption of T-G3 (GATAAGGG to
GtaAAGGG) completely silenced the expression of
lacZ (Fig. 6b). A point mutation of T-M1 that did not
disrupt T-G3 (GATAAGGG to GATActtt) also caused
severe inactivation of the enhancer (Fig. 6c). About
81% of the embryos carrying Tgfna204Z expressed
lacZ, while 3.6% of the embryos carrying Tgfna(M1 m) Z did so.
Binding of the Ci-GATAa protein to wild-type and
mutagenized DNA sequences
Electrophoretic mobility shift assays were carried out
using a full-length Ci-GAGAa protein synthesized in
vitro. An oligonucleotide probe containing the E-M1(GATAAGGG) and its flanking sequences formed a
DNA/protein complex (Fig. 7a, lane 2). We prepared amutant probe (E-M1 m), in which AGGG of E-M1 was
disrupted without affecting GATA. E-M1m did not form
a complex with Ci-GATAa (Fig. 7a, lane 3). A probe
containing E-G1 formed a complex with Ci-GATAa
(Fig. 7a, lane 4). E-G3, E-G5 and E-G6 formed a weak
band (Fig. 7a, lanes 5, 7, 8). No specific shift was
observed with E-G4 (Fig. 7a, lanes 6). A probe con-
taining T-M1 (overlapping with T-G3) and T-G2 formeda complex with Ci-GATAa (Fig. 7b, lane 2). The shifted
band became weak when mutation was created in
either T-G2 or T-G3 (Fig. 7b, lanes 3, 5). When the
AGGG moiety of T-M1 was disrupted without affecting
GATA, the complex’s formation was also affected
(Fig. 7b, lane 4). No shifted band was observed when
both T-G2 and T-G3 were mutated (Fig. 7b, lane 6).
T-G1 and T-G4 also strongly formed a complex withCi-GATAa (Fig. 7b, lanes 7, 8).
Discussion
Binding of Ci-GATAa is strongly influenced by flanking
sequences
Direct binding assays of the ascidian GATA proteinwere first reported in the present study. Our EMSA
revealed cases in which the affinity of the Ci-GATAa
binding to DNA was strongly affected by the
sequences surrounding the core GATA sequence. The
397-bp upstream region of Ci-ephrin-Ad contains six
GATA sequences. However, only two of them (E-G1
and E-G2) can strongly bind Ci-GATAa. These results
seem inconsistent with unpublished results obtainedby K. Nitta and P. Lemaire (pers. comm., 2012).
They searched for a consensus Ci-GATAa-binding
sequence by the systematic evolution of ligands using
exponential enrichment (SELEX). They did not see a
strong preference outside of the GATA sequence,
although TGATAA was a little better (K. Nitta & P.
Lemaire, pers. comm., 2012). In the present study,
Ci-GATAa seemed to prefer GATAAGGG. The under-lined nucleotides may not have been recognized by
SELEX, possibly because only two or three of the
underlined nucleotides are conserved in many cases.
Bertrand et al. (2003) and Rothb€acher et al. (2007)
carried out mutagenesis of many GATA sequences
within the enhancer region of Ci-otx and Ci-fog, and
reported strong suppression of reporter genes. Similar
results were obtained by antisense-mediated knock-down of Ci-GATAa (Bertrand et al. 2003; Rothb€acher
et al. 2007). Two critical GATA sequences within the
Ci-fog enhancer are GATAAAGT and GATACTTT
(Rothb€acher et al. 2007). The former is similar to the
octamer motif, but the latter looks like the mutant
(a)
(b)
(c)
Fig. 4. Putative transcription factor-binding sites in the E1 and
T1 enhancers (a) Schematic diagrams showing the 5´ flanking
regions of Ci-ephrin-Ad and Ci-TGFb-NA1. Putative binding
sites for the GATA and Sox transcription factors are indicated
by blue and yellow boxes, respectively. The octamer motif
(GATAAGGG) is indicated by red boxes. (b) Nucleotide
sequence of the E1 enhancer. (c) Nucleotide sequence of the
T1 enhancer.
ª 2013 The Authors
Development, Growth & Differentiation ª 2013 Japanese Society of Developmental Biologists
Early ectodermal enhancers in ascidians 781
sequences that did not bind Ci-GATAa in the present
study. None of the seven GATA sequences within the
Ci-otx enhancer look similar to the strong binding sites
(Bertrand et al. 2003). Considering the present results,
direct binding assays will be necessary to confirm that
these GATA sequences are genuine Ci-GATAa-binding
sites.
Possible role of the octamer motif shared by the
animal hemisphere enhancers
The MEME software identified the octamer motif
shared by the E1 and T1 enhancers. Point mutations
of the motif resulted in severe reduction of enhancer
activity. There are at least three possible explanations
of how the octamer motif contributes to transcriptionalactivation. First, Ci-GATAa binds to the motif and acti-
vates transcription. The affinity of the GATA sequence
for Ci-GATAa is affected by the adjacent sequence,
where AGGG facilitates the binding of Ci-GATAa to the
GATA sequence. Second, the octamer motif is a bind-
ing site for a transcription factor other than Ci-GATAa.
Although E-M1 and T-M1 can bind Ci-GATAa in vitro,
these sequences may be occupied in vivo by another
transcription factor that recognizes the octamer motif
as a whole. Third, Ci-GATAa binds to the GATA part
of the octamer motif, while another transcription factorsynergistically binds to the AGGG part. The combina-
torial binding of two transcription factors may acceler-
ate transcriptional activation.
Mutations in E-M1 and T-M1 (GtaAAGGG or GAT-
Acttt) indeed affected the binding of Ci-GATAa, sup-
porting the first possibility. However, this does not
exclude the second and third explanations. The 397-bp
upstream region of Ci-ephrin-Ad contains two strongCi-GATAa-binding sites (E-G1 and E-G2). However,
only E-G2 is necessary, while E-G1 is dispensable for
transcriptional activation. Disruption of E-G1 revealed
that only a single strong Ci-GATAa-binding site (and a
few weak sites?) was sufficient for transcriptional
(a)
(b)
(c)
(d)
(e)
(f)
(g)
Fig. 5. Effect of point mutations in putative transcription factor-binding sites in the 5´ flanking region of Ci-ephrin-Ad. Diagrams of
transgenes and their pattern of expression are shown side-by-side. Percentage of the embryos expressing lacZ among normally devel-
oping embryos is also indicated. (a) Ephrin397Z, carrying no mutation. (b) Ephrin (S1 m) Z, carrying a mutation in the Sox-binding site E-
S1. (c) Ephrin (S2 m) Z, carrying a mutation in the other Sox-binding site E-S2. (d) Ephrin (S1S2 m) Z, carrying mutations in both E-S1
and E-S2. (e) Ephrin (G1 m) Z, carrying a mutation in the GATA sequence E-G1. (f) Ephrin (G2 m) Z, carrying a mutation in the GATA
sequence E-G2, which overlaps with the octamer motif E-M1. (g) Ephrin (M1 m) Z, carrying a mutation in E-M1. The mutation was
created to convert GATAAGGG into GATATCCC, not to disrupt E-G2.
ª 2013 The Authors
Development, Growth & Differentiation ª 2013 Japanese Society of Developmental Biologists
782 Y. Horikawa et al.
activation. In contrast, the Ci-TGFb-NA1 upstreamregion contained four binding sites for Ci-GATAa. Dis-
ruption of only one of them (T-G3 or T-M1) abolished
the enhancer activity, although three strong binding
sites remained. These observations suggest that theoctamer motif is not just one of many Ci-GATAa-bind-
ing sites. In vitro binding of Ci-GATAa to the octamer
sequence does not necessarily mean that the octamer
is also legitimately occupied in vivo by Ci-GATAa.
Maternal Ci-GATAa mRNA is ubiquitous, and seems to
be translated in both animal and vegetal blastomeres
(Bertrand et al. 2003). Disruption of the b-cateninfunction caused ectopic expression of the Ci-fog repor-ter gene in the vegetal hemisphere, suggesting that
b-catenin suppresses Ci-GATAa at the post-transla-
tional level (Rothb€acher et al. 2007). As described
above, one of the critical GATA sequences within the
Ci-fog enhancer is similar to the octamer motif. It is
therefore possible that the octamer motif is also involved
in the animal-specific and b-catenin-sensitive transcrip-
tional activation. Extensive binding analyses usingnuclear extracts and identification of binding proteins will
help elucidate what binds to the octamer motif in vivo.
Other transcription factors
In sea urchin embryos, maternally provided SoxB1 is
responsible for gene expression in the presumptive
ectodermal region in the animal hemisphere (Kennyet al. 1999). Sox transcription factors are expressed
mainly in the central nervous system in vertebrates
(a)
(b)
(c)
Fig. 6. Effect of point mutations in putative transcription factor-
binding sites in the 5´ flanking region of Ci-TGFb-NA1. Diagrams
of trangenes containing mutations and their pattern of expression
are shown side-by-side. Percentage of the embryos expressing
lacZ among normally developing embryos is indicated. (a)
Tgfna204Z, carrying no mutation. (b) Tgfna (G3 m) Z, carrying a
mutation in the GATA sequence T-G3, which overlaps with the
octamer motif T-M1. (c) Tgfna (M1 m) Z, carrying a mutation in
T-M1. The mutation was created to convert GATAAGGG into
GATATCCC, not to disrupt T-G3.
(a) (b)
Fig. 7. Electrophoretic mobility shift
assay. Double-stranded radioactive
probes were incubated with in vitro
translated Ci-GATAa protein. For lane
1, TC14-3 was used as a control pro-
tein. TC14-3 is a calcium-dependent
galactose-binding lectin, whose cDNA
was obtained from the budding ascid-
ian Polyandrocarpa misakiensis (Mat-
sumoto et al. 2001). DNA/protein
complexes were fractionated on a poly-
acrylamide gel. The black arrowhead
indicates specific signals of the com-
plex’s formation. The open arrowhead
indicates non-specific signals. Probes
used are indicated below the gel
image. Lower case letters (in red) indi-
cate mutagenized nucleotides. (a) The
GATA sequences within the Ci-ephrin-
Ad upstream region were examined. (b)
The GATA sequences within the Ci-
TGFb-NA1 upstream region were
examined.
ª 2013 The Authors
Development, Growth & Differentiation ª 2013 Japanese Society of Developmental Biologists
Early ectodermal enhancers in ascidians 783
(Pevny & Lovell-Badge 1997). In the Ciona embryo,SoxF is specifically expressed in the animal hemi-
sphere from the 16-cell stage (Imai et al. 2004). SoxB1
is also expressed in all of the eight animal blasto-
meres, although it is expressed in anterior vegetal
blastomeres (Imai et al. 2004). Therefore, these Sox
factors may play a role in the gene regulatory network
in the animal hemisphere. In the present study, we
found two putative Sox-binding sequences in the 5′flanking region of Ci-ephrin-Ad. However, these sites
were unnecessary for transcriptional activation, at least
at the 16-cell stage. Although animal hemisphere-spe-
cific expression of many transcription factors, signaling
proteins and receptors is observed at the 16-cell
stage, these proteins require a time-lag to exhibit their
function.
Since zygotic expression at the 16-cell stage is theearliest, maternal transcription factors likely play impor-
tant roles. A large amount of expressed sequence tag
(EST) and in situ data have been accumulated for
maternal transcripts (e.g. Nishikata et al. 2001). In con-
trast, information about maternal proteins is still limited
(Nomura et al. 2009; Endo et al. 2011). In addition to
analyzing enhancers, it is important to characterize
maternal proteins involved in activation of the gene reg-ulatory network at the earliest stage of embryogenesis.
Acknowledgments
This work was supported in part by Grants-in-aid from
the Japan Society for the Promotion of Science
(#22570214 and #25440192). We thank Kazuko Hiray-
ama, Chikako Imaizumi, Shota Chiba, Nori Satoh, Yu-taka Satou and the Maizuru Fisheries Research Station
of Kyoto University for providing us with animals
through the National Bio-Resource Project, MEXT,
Japan. We are particularly grateful to the late Ms. Ka-
zuko Hirayama who helped us set up the culture sys-
tem in Kochi University. We also thank Kaz Nitta for
sharing his unpublished data.
References
Bailey, T. L. & Elkan, C. 1994. Fitting a Mixture Model by Expec-
tation Maximization to Discover Motifs in Biopolymers. Proc.2nd Int. Conf. Intell. Syst. Mol. Biol. Menlo Park, CA: AAAIPress, pp. 28–36.
Bertrand, V., Hudson, C., Caillol, D., Popovici, C. & Lemaire, P.2003. Neural tissue in ascidian embryos is induced byFGF9/16/20, acting via a combination of maternal GATA andEts transcription factors. Cell 115, 615–627.
Conklin, E. G. 1905. The organization and cell lineage of ascidianegg. J. Acad. Nat. Sci. (Philadelphia) 13, 1–119.
Corbo, J. C., Levine, M. & Zeller, R. W. 1997. Characterization ofa notochord-specific enhancer from the Brachyury promoter
region of the ascidian, Ciona intestinalis. Development 124,589–602.
Endo, T., Ueno, K., Yonezawa, K., Mineta, K., Hotta, K., Satou,Y., Yamada, L., Ogasawara, M., Takahashi, H., Nakajima,A., Nakachi, M., Nomura, M., Yaguchi, J., Sasakura, Y.,Yamazaki, C., Sera, M., Yoshizawa, A. C., Imanishi, T., Tan-iguchi, H. & Inaba, K. 2011. CIPRO 2.5: Ciona intestinalis
protein database, a unique integrated repository of large-scale omics data, bioinformatic analyses and curated anno-tation, with user rating and reviewing functionality. Nucleic
Acids Res. 39 (Database Issue), D807–D814.Fujiwara, S., Corbo, J. C. & Levine, M. 1998. The Snail repressor
establishes a muscle/notochord boundary in the Ciona
embryo. Development 125, 2511–2520.Hamaguchi, M., Fujie, M., Noda, T. & Satoh, N. 2007. Microarray
analysis of zygotic expression of transcription factor genesand cell signaling molecule genes in early Ciona intestinalis
embryos. Dev. Growth Differ. 49, 27–37.Imai, K. S., Hino, K., Yagi, K., Satoh, N. & Satou, Y. 2004. Gene
expression profiles of transcription factors and signaling mol-ecules in the ascidian embryo: towards a comprehensiveunderstanding of gene networks. Development 131, 4047–4058.
Kanda, M., Wada, H. & Fujiwara, S. 2009. Epidermal expressionof Hox1 is directly activated by retinoic acid in the Ciona in-
testinalis embryo. Dev. Biol. 335, 454–463.Kenny, A. P., Kozlowski, D. J., Oleksyn, D. W., Angerer, L. M. &
Angerer, R. C. 1999. SpSoxB1, a maternally encoded tran-scription factor asymmetrically distributed among early seaurchin blastomeres. Development 126, 5473–5483.
Lemaire, P., Smith, W. C. & Nishida, H. 2008. Ascidians and theplasticity of the chordate developmental program. Curr. Biol.18, R620–R630.
Matsumoto, J., Nakamoto, C., Fujiwara, S., Yubisui, T. & Ka-wamura, K. 2001. A novel C-type lectin regulating cellgrowth, cell adhesion and cell differentiation of the multipo-tent epithelium in budding tunicates. Development 128,3339–3347.
Mertin, S., McDowall, S. G. & Harley, V. R. 1999. The DNA-bind-ing specificity of SOX9 and other SOX proteins. Nucleic
Acids Res. 27, 1359–1364.Nishida, H. 1987. Cell lineage analysis in ascidian embryos by
intracellular injection of a tracer enzyme. III. Up to the tissuerestricted stage. Dev. Biol. 121, 526–541.
Nishikata, T., Yamada, L., Mochizuki, Y., Satou, Y., Shin-i, T.,Kohara, Y. & Satoh, N. 2001. Profiles of maternallyexpressed genes in fertilized eggs of Ciona intestinalis. Dev.Biol. 238, 315–331.
Noda, T. & Satoh, N. 2008. A comprehensive survey of cadherinsuperfamily gene expression patterns in Ciona intestinalis.Gene Expr. Patterns 8, 349–356.
Nomura, M., Nakajima, A. & Inaba, K. 2009. Proteomic profilesof embryonic development in the ascidian Ciona intestinalis.Dev. Biol. 325, 468–481.
Pevny, L. H. & Lovell-Badge, R. 1997. Sox genes find their feet.Curr. Opin. Genet. Dev. 7, 338–344.
Picco, V., Hudson, C. & Yasuo, H. 2007. Ephrin-Eph signallingdrives the asymmetric division of notochord/neural precur-sors in Ciona embryos. Development 134, 1491–1497.
Qiao, Y., Zhu, Y., Sheng, N., Chen, J., Tao, R., Zhu, Q., Zhang, T.,Qian, C. & Jing, N. 2012. AP2c regulates neural and epider-mal development downstream of the BMP pathway at earlystages of ectodermal patterning. Cell Res. 22, 1546–1561.
ª 2013 The Authors
Development, Growth & Differentiation ª 2013 Japanese Society of Developmental Biologists
784 Y. Horikawa et al.
Rothb€acher, U., Bertrand, V., Lamy, C. & Lemaire, P. 2007. Acombinatorial code of maternal GATA, Ets and b-catenin-TCF transcription factors specifies and patterns the earlyascidian ectoderm. Development 134, 4023–4032.
Satou, Y., Kawashima, T., Shoguchi, E., Nakayama, A. & Satoh,N. 2005. An integrated database of the ascidian, Ciona in-
testinalis: towards functional genomics. Zool. Sci. 22, 837–843.
Schmidt, R., Str€ahle, U. & Scholpp, S. 2013. Neurogenesis in ze-brafish - from embryo to adult. Neural Dev. 8, 3.
Stern, C. D. 2005. Neural induction: 10 years on since the‘default model’. Curr. Opin. Cell Biol. 18, 692–697.
Tassy, O., Dauga, D., Daian, F., Sobral, D., Robin, F., Khoueiry,P., Salgado, D., Fox, V., Caillol, D., Schiappa, R., Laporte,B., Rios, A., Luxardi, G., Kusakabe, T., Joly, J. S., Darras,S., Christiaen, L., Contensin, M., Auger, H., Lamy, C., Hud-son, C., Rothb€acher, U., Gilchrist, M., Makabe, K. W., Hotta,K., Fujiwara, S., Satoh, N., Satou, Y. & Lemaire, P. 2010.The ANISEED database: digital representation, formalization
and elucidation of a chordate developmental program. Gen-ome Res. 20, 1459–1468.
Tr�ıbulo, C., Barrionuevo, M. G., Ag€uero, T. H., S�anchez, S. S.,Calcaterra, N. B. & Aybar, M. J. 2012. DNp63 is regulatedby BMP4 signaling and is required for early epidermal devel-opment in Xenopus. Dev. Dyn. 241, 257–269.
Umesono, K., Murakami, K. K., Thompson, C. C. & Evans, R. M.1991. Direct repeats as selective response elements for thethyroid hormone, retinoic acid, and vitamin D3 receptors.Cell 65, 1255–1266.
Supporting Information
Additional supporting information may be found in the
online version of this article at the publisher’s web-site:
Table S1. Primers used for construction of plasmids.
ª 2013 The Authors
Development, Growth & Differentiation ª 2013 Japanese Society of Developmental Biologists
Early ectodermal enhancers in ascidians 785