Genomes & Developmental Control
Spatio-temporal intersection of Lhx3 and Tbx6 defines the cardiac field through
synergistic activation of Mesp
Lionel Christiaen ⁎, Alberto Stolfi, Brad Davidson1, Michael Levine⁎
Department of Molecular & Cell Biology, Division of Genetics, Genomics and Development, Center for Integrative Genomics, University of California Berkeley, CA, 94720-3200, USA
a b s t r a c ta r t i c l ei n f o
Received for publication 18 November 2008
Revised 16 January 2009
Accepted 23 January 2009
Available online 3 February 2009
Mesp encodes a bHLH transcription factor required for specification of the cardiac mesoderm in Ciona
embryos. The activities of Macho-1 and β-catenin, two essential maternal determinants, are required for
Mesp expression in the B7.5 blastomeres, which constitute the heart field. The T-box transcription factor
Tbx6 functions downstream of Macho-1 as a direct activator of Mesp expression. However, Tbx6 cannot
account for the restricted expression of Mesp in the B7.5 lineage since it is expressed throughout the
presumptive tail muscles. Here we present evidence that the LIM-homeobox gene Lhx3, a direct target of β-
catenin, is essential for localized Mesp expression. Lhx3 is expressed throughout the presumptive endoderm
and B7.5 blastomeres. Thus, the B7.5 blastomeres are the only cells to express sustained levels of the Tbx6 and
Lhx3 activators. Like mammalian Lhx3 genes, Ci-Lhx3 encodes two isoforms with distinct N-terminal
peptides. The Lhx3a isoform appears to be expressed both maternally and zygotically, while the Lhx3b
isoform is exclusively zygotic. Misexpression of Lhx3b is sufficient to induce ectopic Mesp activation in cells
expressing Tbx6b. Injection of antisense morpholino oligonucleotides showed that the Lhx3b isoform is
required for endogenous Mesp expression. Mutations in the Lhx3 half-site of Tbx6/Lhx3 composite elements
strongly reduced the activity of a minimal Mesp enhancer. We discuss the delineation of the heart field by the
synergistic action of muscle and gut determinants.
Published by Elsevier Inc.
The reproducibility of metazoan development is ensured by the
tight genetic control of embryogenesis. This genetic control manifests
itself as gene regulatory networks that unfold in space and time in an
exquisitely regulated manner (Levine and Davidson, 2005). In this
framework, spatio-temporal patterns of regulatory gene expression
and activity are established as development proceeds, and ultimately
determine the fate and position of definitive tissues and organs.
Numerous studies have established a range of rules governing
developmental gene expression patterns, which are reflected in the
structure and function of their cis-regulatory DNAs (e.g. Arnone and
Davidson, 1997). Most predominantly, regulatory gene expression
profiles are established by tissue- and stage-specific combinations of
trans-acting activators and/or repressors.
For example, in Drosophila embryos, lateral stripes of rhomboid
and vnd expression are activated in ventral regions of the presumptive
neurogenic ectoderm by broadly distributed Dorsal and Twist acti-
vators, and repressed in the mesoderm by Snail (Ip et al., 1992).
Similarly, Bicoid and Hunchback define a broad domain in the anterior
half of the embryo where the even-skipped stripe 2 enhancer can be
activated. The anterior and posterior borders of the stripe are formed
by the spatially localized Giant and Kruppel repressors, respectively
(Small et al.,1992). In Xenopus embryos, early goosecoid expression is
restricted to the organizer region because it relies in part on over-
lapping activities of Xtwin/siamois homeodomain proteins and Nodal
signaling by the Xnr5 ligand, which are activated in response to
maternal β-catenin and VegT proteins, respectively (reviewed in
Loose and Patient, 2004; Wardle and Smith, 2006). In theses ex-
amples, asymmetric distributions of maternal determinants (e.g.
dorso-ventral gradient of nuclear Dorsal protein) provide initial
positional cues, which areprogressively refinedthrough the activation
of primary (e.g. Twist and Snail) and secondary (e.g. rhomboid and
vnd) zygotic target genes. Because the spatio-temporal expression
patterns of the primary target genes are not perfect read-outs of the
initial maternal cue, they allow for the generation of new expression
patterns of secondary targets.
The expression of the Mesp regulatory gene is restricted to a single
pair of blastomeres, the B7.5 cells, at the 110-cell stage of ascidian
embryogenesis (Imai et al., 2004; Satou et al., 2004). Mesp encodes a
basic helix-loop-helix (bHLH) transcription factor, which defines the
heart field (Davidson et al., 2005, 2006; Satou et al., 2004), and its
expression depends on the activities of two essential maternal deter-
minants, Macho-1 and β-catenin (Satou et al., 2004). The Macho-1
Developmental Biology 328 (2009) 552–560
⁎ Corresponding authors. Fax: +1 510 643 5785.
E-mail addresses: email@example.com (L. Christiaen),
firstname.lastname@example.org (M. Levine).
1Present address: Department of Molecular and Cellular Biology, Molecular
Cardiovascular Research Program, University of Arizona, Tucson, AZ 85724, USA.
0012-1606/$ – see front matter. Published by Elsevier Inc.
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/developmentalbiology
mRNA encodes a Zinc-finger transcription factor that is localized to
the posterior-most vegetal extremity of the embryo (Nishida and
Sawada, 2001). The maternal β-catenin protein is translocated to the
nuclei of vegetal blastomeres at the 16-cell stage (Imai et al., 2000).
Further analyses demonstrated that Tbx6 genes, which mediate the
effects of Macho-1 during primary muscle specification (Yagi et al.,
2005), are direct activators of Mesp (Davidson et al., 2005). However,
Tbx6 genes arenot sufficient to account for the restricted expression of
Mesp in the B7.5 blastomeres since they are broadly distributed
throughout the progenitors of the tail muscles.
Here we investigate how β-catenin contributes to the restricted
expression of Mesp. The β-catenin target gene, Lhx3, is expressed
throughout the presumptive endoderm and B7.5 blastomeres. Thus,
the Tbx6 and Lhx3 activators are co-expressed exclusively in the B7.5
cells. We found that Lhx3 activity is necessary and sufficient, in
sites for the two activators are closely linked within the minimal Mesp
enhancer. Our findings suggest that Mesp expression is governed by a
simple developmental logic – the unique spatio-temporal overlap of
two necessary activators – that is reflected in its regulatory DNA and
allows for a refined read-out of the initial maternal cues that positions
the heart field in the early embryo.
Material and methods
Adult animals and embryo manipulations
Gravid Ciona intestinalis adults were collected at the Pillar Point
harbor (Half Moon Bay, CA) or obtained from M-Rep (San Diego, CA).
Ripe oocytes and sperm were collected surgically and kept separate
untilinvitro fertilization. Fertilized eggsandunfertilized oocyteswere
dechorionated as described (Mita-Miyazawa et al., 1985).
Morpholinos antisense oligonucleotides were dissolved at 0.5 μM
in a mix containing either 10–20% fastgreen or 0.3 μM rhodamine-
dextranas dye. Plasmid DNAwas added at 10–50 ng/μl in the injection
mix. Unfertilized oocytes were injected as described (Bertrand et al.,
2003; Christiaen et al., 2008; Satou et al., 2001a), kept in buffered
artificial sea water, and fertilized. The sequences of the MOs used in
this study were as follows:
Lhx3-E2I3, ACAACAAACATTTCACTTACTTGAA (bold letters highlight
partial exon2 sequence). Standard control MOs from GeneTools
(Oregon, USA) were used as negative control.
Electroporations were performed as described (Corbo et al., 1997)
with 50–100 μg of plasmid DNA per construct. Expression data pre-
sented as histograms were generated from 100–300 embryos ob-
tained in two or more experiments.
Molecular cloning of over-expression and reporter constructs
The Mesp minimal cis-regulatory DNA was described in an
earlier study (Davidson et al., 2005). The EcoO109I-XbaI fragment
from the pCESA plasmid backbone (Corbo et al., 1997; Harafuji et al.,
2002) was removed to reduce non-specific expression of the re-
porter gene in muscle and mesenchyme precursor cells (Davidson
et al., 2005).
The cis-regulatory DNA from the FoxD gene used in this study was a
kind gift from Weiyang Shi (Shi and Levine, 2008). The cis-regulatory
DNA from orphan-TGFβ-a/Bmp2b was cloned by fusing the upstream
non-coding region amplified from genomic DNA using pbmp2b-F1:
AAAGTCGACTGTCAAACATCCGTGTTAGTGA, and pbmp2b-R1: TTTGGAT-
CCCAGTATTCGATGCATCTT, to the first intron amplified using pbmp2d-
F2: AAAGGATCCCGGAGCCTTCGAAACAAA, and pbmp2b-R2: AAAGCGG-
CCGCCCGCCCTGGCAATGTAAT, upstream from the NLS-lacZ coding
sequence present in pCESA. The cis-regulatory DNA from Tbx6b was
amplified using pTbx6b-F: AAAGGATCCCAACGGAGTACGCGTGTCAAGT-
TTAATGG and pTbx6b-R: AAAGCGGCCGCCATATTCGCCATAGTCTTGTCT-
GGTCCAA and cloned upstream of NLS-lacZ (underlined sequences
indicate restriction sites in all primers).
The codingsequences used in this study wereamplified fromcDNA
libraries made from total RNA using the SMART RACE kit (Clontech),
Sensiscript or Omniscript reverse-transcriptase systems (Qiagen). The
following primer sequences were used (underlined sequences indi-
cate restriction sites, brackets indicate start codon and squared brac-
kets indicate STOP codons) : macho-CDS-F: TTGCGGCCGCAACC(ATG)
GCCTTTACTGGTACGATGGGATA, and macho-CDS-R: AACAATTGAGCG-
TGCC[CTA]AAATACTGTCTCG; Tbx6a-CDS-F: TTGCGGCCGCAACC(ATG)
GAGGACAGCAGTTGTTTGAG and Tbx6a-CDS-R: TTGAATTCGAG[TTA]
ATATTCGCTGTTTGTTTTAGA; Tbx6b-CDS-F2: AAGCGGCCGCAACC
(ATG)GAGCTAGCGATTCCAATTCATTGGCAAAAT and Tbx6b-CDS-R:
CAG and beta-cat-CDS-R: AATGCTCAGCTCCAATCCC[TTA]GAGGTCAG-
TATCGAG; Lhx3b-CDS-F2: AAGCGGCCGCAACC(ATG)ATACTAGACACA-
AAGGCACTCGATGAACTTACGAACCTG and Lhx3b-CDS-R: TTGAATTC
[TTA]TTGGAAATGTGTCACGTGGT; Lhx3a-CDS-F: TATGCGGCCGCAACC
(ATG)CAGACCGGAAGTGAGTTTC; and Lhx3a-CDS-R: TATGAATTCA-
Pointmutationsin the Mesp190NlacZ fusion genewereintroduced
using the QuickChange Kit (Stratagene) according to the manufac-
The Lhx3 cis-regulatory DNA used in this study was amplified from
genomic DNA using the following primer sequences:
Lhx3b(−3204)F, TTACTAGTTGCCGTACAGTGGCTTGAATTGTGT and
For RT-PCR analysis, total RNA was extracted from 5 to 20 embryos
using the RNAqueous-micro kit (Ambion) and used for reverse-
transcription using the Sensiscript Reverse-transcriptase (Qiagen).
PCR was performed using the FlexiTaq DNA polymerase (Promega)
and the following primer pairs: Lhx3-E1aF: ACATCCCGACCTCAAC-
GACT; Lhx3-E1bF: GAAGTCGAAACGCACGTCAC; Lhx3-E3R: TGTCC-
TGTGCCCTCCTAACC; CA7-F: CTGCCAGCAGCAGCTCACTC and CA7-R:
In situ hybridization and β-galactosidase detection
X-gal staining and in situ hybridization were performed essentially
as described (Christiaen et al., 2008). The digoxigenin-labeled
antisense RNA probe for Mesp was synthesized by in vitro transcrip-
tion from linearized plasmid DNA containing a 3′RACE fragment
amplified using the Mesp3′F primer sequence: GCGGTACCGACAG-
TAACGATTTCCA, and the Universal Primer Mix from the SMARTcDNA
kit (Clontech) and cloned into the pCRII-TOPO vector (Invitrogen).
Tbx6 and Lhx3 probes were synthesized from clones cicl007g16 and
cilv006c13, respectively (note that the Lhx3 probe recognized both
isoforms but could not detect the weak Lhx3a maternal expression;
Imai et al., 2004).
Ectopic Tbx6b efficiently trans-activates Mesp in endoderm precursor
Mesp expression is first detected in the B7.5 blastomeres of 110-cell
embryos (Fig.1A;Imai et al.,2004; Satou et al.,2004). Previous studies
L. Christiaen et al. / Developmental Biology 328 (2009) 552–560
have shown that Mesp expression requires the activities of two
essential maternal determinants: the zinc-finger transcription factor
Macho-1, and β-catenin, which determine primary muscle and endo-
derm fates, respectively (Imai et al., 2000; Nishida and Sawada, 2001;
Satou et al., 2004; Yagi et al., 2004). In addition, the Macho-1 target
gene Tbx6 was previously shown to directly activate Mesp (Davidson
et al., 2005). However, there are at least three distinct Tbx6 genes in C.
intestinalis (Ci-Tbx6a, -b and -c), which are broadly expressed in B7.5
and muscle precursor cells (Takatori et al., 2004). Therefore,
additional molecular inputs must be invoked to explain the restricted
expression of Mesp in B7.5 cells.
We used a mis-expression strategy to determine which blasto-
meres are competent to activate Mesp in response to ectopic expres-
sion of Macho-1 and Tbx6 proteins in the early embryo. Towards this
goal, cis-regulatory DNA from FoxD was used to drive ectopic
expression of Macho-1, Tbx6a, -b, or -c in early vegetal blastomeres
(Fig. 1B; Shi and Levine, 2008), where nuclear β-catenin was pre-
viously observed (Imai et al., 2000). Embryos were co-electroporated
with each mis-expression transgene and a minimal Mesp190NlacZ
reporter gene, followed by staining for endogenous Mesp expression
(Fig.1C) or β-galactosidase activity (Fig.1D) at the earlygastrula stage
(Hotta et al., 2007).
Using X-gal staining, we observed that mis-expression of
Macho-1, Tbx6a, Tbx6b or Tbx6c led to variable ectopic induction
of the Mesp190NlacZ reporter gene (Fig. S1). The strongest
staining was obtained with the FoxDNTbx6b transgene, which
was the only one to produce ectopic expression of endogenous
Mesp transcripts (Figs. 1C and S1A–C). Ectopic activation of both
the endogenous gene and reporter transgene expression was
particularly strong in A- and B-line endoderm precursors (Fig.
1E). Notably, ectopic expression was never observed in primary
muscle precursors, which express endogenous Tbx6b transcripts.
These observations indicate that Tbx6b transactivates Mesp more
efficiently than Tbx6a or Tbx6c. Moreover, endoderm precursors
are more competent than other vegetal blastomeres to express
Mesp in response to Tbx6b.
β-catenin activity confers competence to activate Mesp in response to
β-catenin restricts competence for Mesp induction in the vegetal
hemisphere (Satou et al., 2004). Therefore, we predicted that mis-
expression of Tbx6b in the animal hemisphere would not be sufficient
to transactivate Mesp. To test this possibility, we used the cis-
regulatory DNA from the orphan-TGFβ-a gene, referred to as bmp2b
in the present study, which is only expressed in the animal blas-
tomeres starting at the 8-cell stage (Figs. 2A, S1; Imai et al., 2004). As
expected, mis-expression of Tbx6b in the animal blastomeres was
not sufficient to transactivate Mesp (Figs. 2B, C). To test whether β-
catenin might work together with Tbx6b for Mesp activation, we
over-expressed both Tbx6b and β-catenin using the bmp2b cis-
regulatory DNA. This combination appeared sufficient to transactivate
Mesp in a- and b-neural precursor cells in 30 to 40% of the embryos
(Figs. 2B, D).
Since muscle precursor cells express Tbx6b but not Mesp, we asked
whether this is due to insufficient β-catenin activity. To examine this
possibility, the cis-regulatory DNA from Tbx6b was used to over-
express β-catenin, which proved sufficient to induce ectopic expres-
sion of Mesp in the primary muscle precursors of more than 20% of the
embryos (Figs. 2B, E).
Notably, ectopic expression of β-catenin in animal blastomeres
(bmp2bNβcat) led to ectopic Mesp activation in B-line muscle and
mesenchyme precursor blastomeres (Figs. 2B, D). These non cell-
autonomous effects indicate that signaling molecules might be
involved in Mesp activation downstream of β-catenin (see Discussion;
and Imai et al., 2006).
Fig. 1. Ectopic Tbx6b activates Mesp in endoderm precursors. (A) Wild-type, early gastrula stage embryo hybridized with a Mesp-specific probe, vegetal view. (B) Schematic
representation of the spatial activity of the FoxD driver from 16-cell to early gastrula stage embryos. (C, D) Ectopic expression of Tbx6b using the FoxD driver produced ectopic
activation of endogenous Mesp (C) and a minimal Mesp190 reporter transgene. Note ectopic expression in endodermal cells. Inset in (C) shows a lateral view of a different embryo.
(E) Schematic representation of the early gastrula stage embryo, vegetal view, blastomeres are recognized by their lineage (a-, b-, A- and B-lines) and specific fate (color coded,
modified from Nishida, 2005). Proportion of embryos showing ectopic Mesp or Mesp190ΔNlacZ expression upon FoxDNTbx6b electroporation in indicated blastomeres (n=total
number of embryos scored in at least three independent experiments).
L. Christiaen et al. / Developmental Biology 328 (2009) 552–560
Taken together, our observations support the notion that both
Tbx6b and sustained levels of β-catenin activity are sufficient to acti-
vate Mesp expression in the early embryo.
Ectopic expression of the β-catenin target gene Lhx3 transactivates
Several lines of evidence suggest that β-catenin activates Mesp
indirectly. First, over-expression of β-catenin induced only modest
ectopic activation of Mesp, similar to that observed upon over-
expression of Macho-1. Second, β-catenin is active in all vegetal blas-
tomeres at the 32-cell stage (Imai et al., 2000), and Tbx6b efficiently
induces Mesp in endoderm precursors, thereby raising the possibility
that β-catenin induces the expression of a co-activator in the pre-
sumptive endoderm and B7.5 cells.
Previous studies have shown that the LIM-homeobox gene Lhx3 is
activated in direct response to nuclear β-catenin and is more highly
expressed in endoderm precursors at the 110-cell stage (Fig. 3B, Satou
et al., 2001b). Notably, Lhx3 is also expressed in the B7.5 blastomeres
(arrows, Fig. 3B), which are therefore the only cells to strongly ex-
press both Lhx3 and Tbx6 (Fig. 3A; compare with B). In addition, we
observed that Lhx3 is ectopically activated following electroporation
of bmp2bNβ-catenin or Tbx6Nβ-catenin transgenes (Figs. 3C, D).
Thus, Lhx3 is expressed at the right time and place to function down-
stream of β-catenin as a co-activator of Mesp expression in the B7.5
Previous gene models suggest that Lhx3 encodes distinct isoforms,
which differ by 5′ exons encoding distinct N-terminal peptides that
are poorly conserved among chordate orthologs of Lhx3 proteins
(Fig. 3E, and unpublished observations). Conventional RT-PCR assays
were done to determine when each form is expressed during Ciona
embryogenesis (Fig. 3F). The Lhx3a isoform was observed throughout
embryogenesis, including unfertilized eggs, with peak levels at the
64–76 cell stages. In contrast, Lhx3b expression is strictly zygotic, with
expression detected from the 32-cell to the early neurula stage, when
Mesp transcripts are no longer detected by in situ hybridization. Either
or both forms may contribute to Mesp regulation, although Lhx3b
seems to be a better candidate since it is zygotically expressed at the
110-cell stage when Mesp is first activated.
Both Lhx3a and Lhx3b were ectopically expressed in the tail
muscles using Tbx6b regulatory DNA, but only Lhx3b induced strong
ectopic expression of Mesp and the Mesp190ΔNlacZ transgene in B-
line mesenchyme and muscle precursor cells (Figs. 3G–I, S2 and data
not shown). In an attempt to understand why Lhx3a appeared less
efficient for Mesp induction, we expressed Lhx3a-GFP and Lhx3b-GFP
fusion proteins in B-line cells, using the Tbx6 regulatory DNA. Both of
these constructs were able to induce ectopic expression of a Mesp
reporter construct, although the Lhx3b-GFP fusion protein was still
more efficient (A.S., data not shown). In these experiments, Lhx3b-
GFP localized strictly to the nuclei, while Lhx3a-GFP distributed more
broadly in the cytoplasm (Fig. S2). This suggests that the isoform-
specific N-terminal peptide may interfere with nuclear localization,
which could account for the reduced trans-activating efficiency of
Lhx3a. Taken together, our observations suggest that the zygotic
Lhx3b isoform is a likely partner of Tbx6b, which functions down-
stream of β-catenin to activate Mesp.
Tbx6b and Lhx3b synergy is sufficient to activate Mesp
The preceding results suggest that Tbx6b synergizes preferentially
with Lhx3b during Mesp activation. Further evidence was obtained by
mis-expressing both proteins in animal blastomeres using the bmp2b
enhancer (see Fig. 2A). Neither Tbx6b nor Lhx3b alone was sufficient
to induce Mesp expression in animal blastomeres (Fig. 4E). In
contrast, the combination of Tbx6b and Lhx3b (Bmp2NTbx6b+
Bmp2NLhx3b) resulted in intense ectopic induction of both the
endogenous Mesp gene (Figs. 4A, B, E, F) and the Mesp190NlacZ
reporter gene (Figs. 4C, D, F) in the animal hemisphere. Notably,
ectopic Mesp activation was observed in all animal cells, but the a-
and b-neural precursors expressed Mesp in a larger fraction of
Fig. 2. Ectopic β-catenin and Tbx6b activate Mesp in neural precursor cells. (A) Schematic representation of the spatial activity of the bmp2b and Tbx6b drivers. Vegetal views are
shown, note that the bmp2b driver is active in the whole animal hemisphere but only the a- and b-neural precursors are depicted. (B) Proportions of embryos showing endogenous
Mesp expression in indicated tissues and conditions. P-values refer to student's t-tests. (n=total number of embryos scored in at least two independent experiments). (C–E)
Embryos electroporated with indicated constructs and hybridized with a Mesp-specific probe. (inset in (E): posterior view). Arrowheads point to ectopic Mesp expression in a-neural
(D) and muscle precursor (E) cells.
L. Christiaen et al. / Developmental Biology 328 (2009) 552–560
embryos (Fig. 4E, see Discussion). These observations strongly sup-
port the notion that Tbx6b and Lhx3b function synergistically to
activate Mesp expression.
Finally, gene knock-down assay using antisense morpholino
oligonucleotides (MOs) showed that Lhx3b is required for Mesp
expression in B7.5 blastomeres. Three different MOs were used: two
isoform-specific translation inhibitor MOs targeting the initiation
codon of either Lhx3a or Lhx3b (Lhx3a-ATG and Lhx3b-ATG, res-
pectively) and a splicing-inhibitor MO targeting the exon2–intron3
junction of zygotic transcripts (E2I3 MO, note that this morpholino
can target zygotic transcripts of both isoforms). The Lhx3b-ATG MO
was found to inhibit translation of the reporter gene from an
Lhx3bNmCherry transgene (Fig. S3). RT-PCR analysis following injec-
tion of the E2I3 MO showed that zygotic Lhx3b transcripts, but not
maternal Lhx3a mRNAs, were disrupted (Fig. S3). Injection of each of
the three MOs inhibited Mesp expression (Figs. 5, S4) and altered early
development, even though injection of Lhx3a-ATG and E2I3 MOs
showed more severe phenotypes, suggesting a more pleiotropic role
for the Lhx3a isoform (Fig. S4, see Discussion). Notably, the loss of
Mesp expression upon injection of the Lhx3b-ATG MO indicates
Fig. 3. Lhx3 is a β-catenin downstream target that can transactivate Mesp. (A, B) 110 cells stage embryos hybridized with Tbx6b (A) and Lhx3 (B) probes. Vegetal views, note that
Tbx6 and Lhx3 expression patterns overlap only in B7.5 cells (arrows). (C, D) The bmp2bNβ-catenin and Tbx6bNβ-catenin transgenes produce ectopic activation of Lhx3 in the
animal hemisphere (C) and B-line muscle precursors (D), respectively (32-cell stageembryos areshown in animal (C) or vegetal (D) views).(E) Schematic representation of the Lhx3
locus showing the two transcript models and isoform-specific PCR primers (E1aF, E1bF and E3R). (F) RT-PCR analysis of isoform-specific Lhx3 expression during early Ciona
intestinalis development. Asterisks indicate the earliest expression observed. (Ca7: cytoplasmic actin; st.: stages; hpf: hours post-fertilization; UF: unfertilized egg; 4c to 110c: 4 cells
to 110 cells stages; eG: early gastrula; eN: early neurula; N: neurula; LN: late neurula; eTB: early tailbud; mTB: mid-tailbud). (G, H) Embryos electroporated with the Tbx6bNLhx3b
transgene, showing ectopic expression of endogenous Mesp (G) and reporter gene (H). (I) Proportions of embryos showing ectopic Mesp and Mesp190ΔNlacZ in B-line tissues upon
Tbx6bNLhx3b electroporation. (Note the reduced expression in B7.5 blastomeres using X-gal staining, which is due to a delay between transcription activation and accumulation of
active beta-galactosidase enzyme). (n=total number of embryos scored in at least three independent experiments).
L. Christiaen et al. / Developmental Biology 328 (2009) 552–560
that the Lhx3b isoform is required for proper Mesp activation in B7.5
Putative Lhx3 binding sites are required for activation of the Mesp
To determine whether Lhx3b interacts directly with Mesp regu-
latory sequences, we re-examined the minimal 190 bp Mesp enhancer
from C. intestinalis and compared it with the upstream non-coding
region of the Mesp gene in Ciona savignyi (Fig. 6A, note that this
element is sufficient to activate reporter gene expression in B7.5 cells,
e.g. Fig. 6C). Alignment of the orthologous sequences identified
several potential binding sites for homeodomain proteins in addition
tothe previouslycharacterized Tbx6 sites(sites Ato D; Davidson et al.,
2005), which resembled the consensus sequence GWTCACACCT (Yagi
et al., 2005). Among these sites, one perfectlyconserved sequencewas
most similar to known Lhx3 binding sites identified in mammals (Fig.
6A, site 3; Lhx3 consensus: aaTTAATTAA, (Berger et al., 2008; Yaden
et al., 2005). Notably, sites “2” and “3” overlap two of the previously
identified Tbx6 binding sites (sites “B” and “D”; Davidson et al., 2005).
Fig. 5. Lhx3b function is required for Mesp expression in B7.5 blastomeres. (A–F) Embryos injected with indicated antisense morpholino oligonucleotides and hybridized with a Mesp
probe. Arrows indicate the B7.5 cells. Co-injected rhodamine-dextran was used as injection marker (C, F). Note the absence of staining following injection of the Lhx3b-ATG MO. (A,
D: n=number of embryos showing Mesp expression/total number of analyzed embryos).
Fig. 4. Lhx3b can synergize with Tbx6b in ectopic animal hemisphere expression assays. (A–D) Early gastrula embryos electroporated with combinations of Mesp190ΔNlacZ,
bmp2bNTbx6b and/or bmp2bNLhx3b transgenes were stained for endogenous Mesp (A, B) or β-galactosidase activity (C, D). Note the strong ectopic activation of both Mesp and
reporter transgene in animal hemisphere cells. (A, C) animal views; (B, D) vegetal views. (E, F) Proportion of embryos showing ectopic Mesp (E, F) and Mesp190ΔNlacZ (F)
expression in indicated tissues and conditions. (n=total number of embryos scored in at least three independent experiments). (Note reduced X-gal staining in B7.5 cells, due to the
above-mentioned delay in β-galactosidase protein production, Fig. 3).
L. Christiaen et al. / Developmental Biology 328 (2009) 552–560
Point mutations were introduced in each one of the six candidate
homeodomain binding sites (Fig. 6A). For the two “composite
elements”, point mutations were designed to leave the Tbx6 site
unaffected (Fig. 6A, “m3”) or to convert it into an optimal Tbx6 site
(“m2”). Point mutations in the Lhx3 moiety of the composite sites
strongly reduced B7.5 expression of the reporter gene (summarized in
Fig. 5B), but had no effect on sporadic ectopic activation in mesen-
chyme and muscle precursors (Figs. 6B, D, E). In contrast, mutations in
any of the other four putative Lhx3 sites do not significantly alter
reporter gene activity. Thus, the putative Lhx3 moieties in the two
composite Tbx6/Lhx3 binding sites are crucial for Mesp activation in
B7.5 blastomeres (see Discussion).
Fig. 6. Putative Lhx3 binding sites in composite Lhx3-Tbx6 elements are required for enhancer activity in B7.5 cells. (A) Sequence of the minimal Mesp190 cis-regulatory element,
showing conservation between Ciona intestinalis (upper sequence) and Ciona savignyi (lower sequence). Previously identified Tbx6 sites are highlighted (A to D sites; blue boxes).
Candidate Lhx3 and other homeodomain core binding sequences are shown (green boxes). Nucleotides substitutions introduced for mutational analysis are indicated above the C.
intestinalis sequence. (B) Mutational analysis of the candidate Lhx3 sites. Note that only point mutations of the Lhx3 sites that overlap Tbx6 sites strongly reduced reporter gene
expression. (n=number of embryos scored). (C–E) Selected embryos hybridized with a lacZ-specific probe after electroporation of the wild-type (wt) or m3 mutant enhancer
constructs (m3). Arrows point tothe B7.5 cells. Note the absence of B7.5 staining inpanel E. (D) and (E) illustrate spontaneous ectopic expression in B7.7 (open arrowheads) and B8.5
(solid arrowheads) mesenchyme precursors.
L. Christiaen et al. / Developmental Biology 328 (2009) 552–560
Definition of the cardiac field through localized Mesp activation
In ascidian embryos, the trunk ventral cells (TVCs) constitute the
heart progenitors (Hirano and Nishida, 1997; Satou et al., 2004).
Lineage studies demonstrated that they originate from the B7.5
blastomeres, the only cells that express the bHLH transcription factor
Mesp in 110-cell and gastrula stage embryos (Satou et al., 2004). Mesp
function is required for early heart specification; however, an
additional FGF signal induces the cardiac fate in the anterior-most
B7.5 grand daughter cells at the neurula stage (i.e. the TVCs; Davidson
et al., 2006). Therefore, Mesp activity defines the competence to form
cardiac mesoderm and its restricted expression in the B7.5 lineage
delineates the heart field.
We have presented evidence that Tbx6b and Lhx3 establish the
cardiac field through the selective activation of Mesp in the B7.5
blastomeres. Tbx6b is broadly expressed throughout the presumptive
tail muscles, while Lhx3 is expressed throughout the endoderm. Their
expression profiles overlap only in the B7.5 blastomeres. Thus, Tbx6-
Lhx3 synergy is sufficient to account for the restricted activation of
Mesp in the B7.5 lineage.
Tbx6b is activated in the posterior vegetal blastomeres by Macho-1
(Yagi et al., 2004, 2005), which is encoded by a maternal mRNA that
becomes localized to the posterior-vegetal cortex upon fertilization
(e.g. Sardet et al., 2003). During early cleavage stages, Macho-1 mRNA
segregates to the posterior-most vegetal blastomeres. Notably, it co-
segregates with the germ plasm, which presumably contains
unknown inhibitors of zygotic transcription (e.g. Shirae-Kurabayashi
et al., 2006). Therefore, the B7.5 progenitor cells (namely B6.3, B5.2
and B4.2 blastomeres at the 32,16 and 8-cell stages, respectively) are
transcriptionally silent (e.g. Imai et al., 2006). Hence, Tbx6 expression
is activated by the Macho-1 protein in B7.5 cells at the 64-cell stage,
while their sister B7.6 cells retain the germ plasm and Macho-1 mRNA.
In parallel, the β-catenin protein becomes localized within the nuclei
of vegetal blastomeres at the 16-cell stage (Imai et al., 2000).
However, because of the aforementioned global transcriptional
inhibition, Lhx3 expression is only initiated at the 64-cell stage in
B7.5 blastomeres. Thus, at the 64-cell stage, a gene regulatory cascade
is triggered in the newly born B7.5 blastomeres, which induces zygotic
expression of Tbx6 and Lhx3 in response to maternal determinants,
thus leading to Mesp activation at the 110-cell stage (Fig. 7).
Previous studies by Imai et al. (2006) identified Fgf9/16/20 as a β-
catenin target that functions together with Tbx6 to activate Mesp.
Indeed, knock-down of FGF9/16/20 using MOs led to a significant
reduction of Mesp expression in early gastrula embryos, as assessed
by quantitative RT-PCR (Imai et al., 2006). Some of our observations
support this notion since ectopic expression of Mesp and/or
Mesp190NlacZ was more often observed in a- or b-neural precursors
(Figs. 4E, F) and in B-line mesenchyme (Fig. 6B), which all receive an
FGF9/16/20 signal by the 32-cell stage as compared to their sister
epidermal and primary muscle cells, respectively (Bertrand et al.,
2003; Imai et al., 2003). Fgf9/16/20 transcripts were detected in the
precursors of the B7.2 endoderm progenitors, which contact B7.5 and
thereby produce the most likely source of FGF9/16/20 involved in
Mesp activation (Bertrand et al., 2003; Imai et al., 2002). However,
Fgf9/16/20 is also expressed in the B8.7 and B8.8 primary muscle
precursors, and induces mesenchyme fate in B7.7 blastomeres (Imai
et al., 2003), which express sustained levels of Tbx6b but do not
activate Mesp. Therefore, the FGF9/16/20 expression profile is not
sufficient to explain the spatio-temporal restriction of Mesp activa-
tion. Instead, we propose that FGF9/16/20 enhances the levels of
Mesp expression in B7.5 blastomeres. Further studies will be required
to determine whether FGF/MAPK signaling directly activates Mesp
(for example, through putative Ets1/2 binding sites in the Mesp
enhancer, Fig. 6A).
Lhx3 isoforms and the molecular basis for transcriptional synergy
In an attempt to elucidate the contribution of each Lhx3 isoform to
Mesp activation, we used gain and loss of function experiments. All
morpholino antisense oligonucleotides used in this study strongly
reduced Mesp expression, suggesting that both maternal Lhx3a and
zygotic Lhx3a and Lhx3b transcripts are required for proper Mesp
activation. Each MO also disrupted early and late development, with
MOs targeting the Lhx3a isoform having the strongest effects (Fig. S4
and data not shown). Conversely, mis-expression of either Lhx3a or
Lhx3b in Tbx6b-expressing cells led to ectopic activation of Mesp
reporter constructs. Taken together, these observations indicate that
both isoforms contribute to Mesp activation, eitherdirectlyor indirectly
(e.g. Lhx3a may have pleiotropic functions in early development).
However, only Lhx3b expression is strictly zygotic and mis-expression
assayssuggested thatit is moreefficient than Lhx3a for Mesp activation.
Therefore, we cannot exclude the possibility that both Lhx3a and Lhx3b
isoforms contribute to Mesp activation through direct synergy with
Tbx6b, but our observations suggest that Lhx3b plays a predominant
role downstream of β-catenin during heart field specification.
Mammalian Lhx3 genes encode distinct isoforms with different
terminal trans-activation domains (Sloop et al.,1999, 2001). The Ciona
Lhx3 gene possesses an analogous genomic organization and produces
similar isoforms to those seen in mammals, although the specific N-
terminal peptides are poorly conserved (data not shown). Human
LHX3a and LHX3b isoforms have different transactivation and DNA
binding capabilities resulting from distinct auto-inhibitory effects
mediated by the different N-terminal peptides. In addition, LHX3a
synergizes more efficiently than LHX3b with the Pit-1 POU home-
odomain protein to transactivate the prolactin and TSHβ promoters in
cell culture (Sloop et al., 1999, 2001). In light of these and our obser-
vations, it seems possible that the N-terminal peptide of the Ciona
Lhx3b isoform exerts a reduced auto-inhibitory effect allowing for
more efficient synergy with Tbx6b during Mesp activation.
What is the mechanism of Tbx6-Lhx3 synergy? The tight linkage of
critical Tbx6 and Lhx3 binding sites raises the possibility of direct
protein–protein interactions. Suchinteractions might be importantfor
Fig. 7. Summary model: Mesp expression requires transactivation by Tbx6b and Lhx3,
which overlap only in B7.5 cells. Schematic 64- and 110-cell stages embryos, vegetal
in the B7.5 cells, where Mesp is specifically activated at the 110-cell stage (red dots). The
regulatorycascade leading to Mesp activation is initiated at the 64-cell stage in B7.5 cells.
L. Christiaen et al. / Developmental Biology 328 (2009) 552–560
efficient occupancy of the operator sites in cells that contain both Download full-text
activators. Indeed, it is of note that the linked Tbx6/Lhx3 binding sites
in the two composite elements within the minimal Mesp enhancer are
suboptimal sites that deviate from the consensus recognition se-
quences. Low affinity sites might help ensure that Mesp is activated
only in cells that contain both Tbx6 and Lhx3. In principle, optimal
sites could be occupied without cooperative protein–protein interac-
tions, resulting in expanded expression in the tail muscles or endo-
derm by Tbx6 or Lhx3 alone. Thus, enhancer structure, the organiza-
tion of binding sites, might determine whether target genes are
expressed in both the endoderm and muscles, or solely within the
cardiac lineage. The arrangement of Tbx6 and Lhx3 binding sites in
the Mesp enhancer might represent an optimal design for ensuring
stringent regulation in the cardiac field.
Mesp orthologs are involved in early heart specification in both
Ciona and vertebrate embryos (Davidson, 2007). Much less is known
about the regulation of Mesp genes in the early mesoderm of mouse
embryos, except for the fact that separate cis-regulatory modules
control Mesp1 and Mesp2 expression in the early mesoderm and pre-
somitic mesoderm (PSM), respectively (e.g. Haraguchi et al., 2001). It
is of note that Tbx6 is required for Mesp2 expression in the PSM and
cis-regulatory analyses demonstrated that this interaction is direct
(Yasuhikoet al.,2006). However, it remains tobe determinedwhether
a Tbx6-Mesp regulatory connection also exists in the early mesoderm.
Lhx3 is a critical determinant of the endoderm in Ciona, but not in
vertebrate embryos, where Lhx3 genes are best known for there roles
in motoneuron and adenohypophyseal gland development. However,
Lhx3 is expressed in the early endoderm of Amphioxus (Wang et al.,
2002). Given the revised phylogenetic position of Cephalochordates
as an out-group to vertebrates and tunicates (Delsuc et al., 2006), this
raises the possibility that an early role of Lhx3 in endoderm and
possibly heart development was specifically lost in vertebrates.
Studies in Hemichordate, Echinoderm and more basal protostomes
species might prove particularly informative and shed light on a pos-
sible dual, gut (Lhx3)–muscle (Tbx6), origin of the heart.
We thank Weiyang Shi for the FoxD enhancer. We are indebted to
Yutaka Satou, who called our attention to the existence of Lhx3
isoforms in Ciona species. This work was supported by a grant from
the NSF to M.L.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.ydbio.2009.01.033.
Arnone, M.I., Davidson, E.H., 1997. The hardwiring of development: organization and
function of genomic regulatory systems. Development 124, 1851–1864.
Berger, M.F., Badis, G., Gehrke, A.R., Talukder, S., Philippakis, A.A., Pena-Castillo, L.,
Alleyne, T.M., Mnaimneh, S., Botvinnik, O.B., Chan, E.T., et al., 2008. Variation in
homeodomain DNA binding revealed by high-resolution analysis of sequence
preferences. Cell 133, 1266–1276.
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.
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.
Corbo, J.C., Levine, M., Zeller, R.W., 1997. Characterization of a notochord-specific
enhancer from the Brachyury promoter region of the ascidian, Ciona intestinalis.
Development 124, 589–602.
Davidson, B., 2007. Ciona intestinalis as a model for cardiac development. Semin. Cell
Dev. Biol. 18, 16–26.
Davidson, B., Shi, W., Levine, M., 2005. Uncoupling heart cell specification and migration
in the simple chordate Ciona intestinalis. Development 132, 4811–4818.
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,
Delsuc, F., Brinkmann, H., Chourrout, D., Philippe, H., 2006. Tunicates and not cepha-
lochordates are the closest living relatives of vertebrates. Nature 439, 965–968.
Harafuji, N., Keys, D.N., Levine, M., 2002. Genome-wide identification of tissue-specific
enhancers in the Ciona tadpole. Proc. Natl. Acad. Sci. U. S. A. 99, 6802–6805.
Haraguchi, S., Kitajima, S., Takagi, A., Takeda, H., Inoue, T., Saga, Y., 2001. Transcriptional
regulation of Mesp1 and Mesp2 genes: differential usage of enhancers during
development. Mech. Dev. 108, 59–69.
Hirano, T., Nishida, H., 1997. Developmental fates of larval tissues after metamorphosis
in ascidian Halocynthia roretzi. I. Origin of mesodermal tissues of the juvenile. Dev.
Biol. 192, 199–210.
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.
Imai, K., Takada, N., Satoh, N., Satou, Y., 2000. (beta)-catenin mediates the specification
of endoderm cells in ascidian embryos. Development 127, 3009–3020.
Imai, K.S., Satoh, N., Satou, Y., 2002. Early embryonic expression of FGF4/6/9 gene and
its role in the induction of mesenchyme and notochord in Ciona savignyi embryos.
Development 129, 1729–1738.
Imai, K.S., Satoh, N., Satou, Y., 2003. ATwist-like bHLH gene is a downstream factor of an
endogenous FGF and determines mesenchymal fate in the ascidian embryos.
Development 130, 4461–4472.
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.
Ip, Y.T., Park, R.E., Kosman, D., Bier, E., Levine, M.,1992. The dorsal gradient morphogen
regulates stripes of rhomboid expression in the presumptive neuroectoderm of the
Drosophila embryo. Genes Dev. 6, 1728–1739.
Levine, M., Davidson, E.H., 2005. Gene regulatory networks for development. Proc. Natl.
Acad. Sci. U. S. A. 102, 4936–4942.
Loose, M., Patient, R., 2004. A genetic regulatory network for Xenopus mesendoderm
formation. Dev. Biol. 271, 467–478.
Mita-Miyazawa, I., Ikegami, S., Satoh, N., 1985. Histospecific acetylcholinesterase
development in the presumptive muscle cells isolated from 16-cell-stage ascidian
embryos with respect to the number of DNA replications. J. Embryol. Exp. Morphol.
Nishida, H., 2005. Specification of embryonic axes and mosaic development in
ascidians. Dev. Dyn. 233, 1177–1193.
Nishida, H., Sawada, K., 2001. macho-1 encodes a localized mRNA in ascidian eggs that
specifies muscle fate during embryogenesis. Nature 409, 724–729.
Sardet, C., Nishida, H., Prodon, F., Sawada, K., 2003. Maternal mRNAs of PEM and macho
1, the ascidian muscle determinant, associate and move with a rough endoplasmic
reticulum network in the egg cortex. Development 130, 5839–5849.
Satou, Y., Imai, K.S., Satoh, N., 2001a. Action of morpholinos in Ciona embryos. Genesis
Satou, Y., Imai, K.S., Satoh, N., 2001b. Early embryonic expression of a LIM-homeobox
gene Cs-lhx3 is downstream of beta-catenin and responsible for the endoderm
differentiation in Ciona savignyi embryos. Development 128, 3559–3570.
Satou, Y., Imai, K.S., Satoh, N., 2004. The ascidian Mesp gene specifies heart precursor
cells. Development 131, 2533–2541.
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.
Sloop, K.W., Meier, B.C., Bridwell, J.L., Parker, G.E., Schiller, A.M., Rhodes, S.J., 1999.
Differential activation of pituitary hormone genes by human Lhx3 isoforms with
distinct DNA binding properties. Mol. Endocrinol. 13, 2212–2225.
Sloop, K.W., Dwyer, C.J., Rhodes, S.J., 2001. An isoform-specific inhibitory domain
regulates the LHX3 LIM homeodomain factor holoprotein and the production of a
functional alternate translation form. J. Biol. Chem. 276, 36311–36319.
Small, S., Blair, A., Levine, M.,1992. Regulation of even-skipped stripe 2 in the Drosophila
embryo. EMBO J. 11, 4047–4057.
Takatori, N., Hotta, K., Mochizuki, Y., Satoh, G., Mitani, Y., Satoh, N., Satou, Y., Takahashi,
H., 2004. T-box genes in the ascidian Ciona intestinalis: characterization of cDNAs
and spatial expression. Dev. Dyn. 230, 743–753.
Wang, Y., Zhang, P.J., Yasui, K., Saiga, H., 2002. Expression of Bblhx3, a LIM-homeobox
gene, in the development of amphioxus Branchiostoma belcheri tsingtauense. Mech.
Dev. 117, 315–319.
Wardle, F.C., Smith, J.C., 2006. Transcriptional regulation of mesendoderm formation in
Xenopus. Semin. Cell Dev. Biol. 17, 99–109.
Yaden, B.C., Savage, J.J., Hunter, C.S., Rhodes, S.J., 2005. DNA recognitionproperties of the
LHX3b LIM homeodomain transcription factor. Mol. Biol. Rep. 32, 1–6.
Yagi, K., Satoh, N., Satou, Y., 2004. Identification of downstream genes of the ascidian
muscle determinant gene Ci-macho1. Dev. Biol. 274, 478–489.
Yagi, K., Takatori, N., Satou, Y., Satoh, N., 2005. Ci-Tbx6b and Ci-Tbx6c are key mediators
of the maternal effect gene Ci-macho1 in muscle cell differentiation in Ciona
intestinalis embryos. Dev. Biol. 282, 535–549.
Yasuhiko, Y., Haraguchi, S., Kitajima, S., Takahashi, Y., Kanno, J., Saga, Y., 2006. Tbx6-
mediated Notch signaling controls somite-specific Mesp2 expression. Proc. Natl.
Acad. Sci. U. S. A. 103, 3651–3656.
L. Christiaen et al. / Developmental Biology 328 (2009) 552–560