Mechanisms of Development 117 (2002) 269–273
www.elsevier.com/locate/modo
Gene expression pattern
Expression and splice variant analysis of the zebrafish tcf4
transcription factor
Rodrigo M. Young, Ariel E. Reyes, Miguel L. Allende*
Millennium Nucleus in Developmental Biology, Departamento de Biologı́a, Facultad de Ciencias, Universidad de Chile, Casilla 653, Santiago, Chile
Received 14 March 2002; received in revised form 9 May 2002; accepted 12 May 2002
Abstract
Wnt signalling has been implicated in antero-posterior patterning of the vertebrate embryonic body axis and in a number of other
developmental processes. One of the downstream effectors of Wnt signalling is the b-catenin protein which complexes with members of
the Lef/tcf transcription factor family. In the zebrafish, specification of the head has been shown to be dependent on the Tcf3 protein which
acts as a repressor of the posteriorizing activity of Wnt (Nature 407 (2000) 913). Here, we report the cloning and expression pattern of the
zebrafish tcf4 gene. In embryos, we find that the tcf4 gene is highly regulated at the level of RNA splicing such that the variant proteins that
are produced contain or lack domains proposed to be essential in repression or activation of transcription. Expression of tcf4 mRNA is first
detected in a graded fashion in the anterior brain and subsequently becomes restricted to the dorsal diencephalon and anterior midbrain. There
is also transient expression in the anterior rhombomeres of the hindbrain and in the developing gut. q 2002 Elsevier Science Ireland Ltd. All
rights reserved.
Keywords: Lef/tcf; Wnt; High mobility group box; Danio rerio development; Diencephalon; Torus semicircularis; Pretectum; Zona limitans
1. Results and discussion
A 308 bp tcf4 polymerase chain reaction (PCR) fragment
(GenBank accession number, AF136455; see Dorsky et al.,
1999) was used to design primers for 5 0 and 3 0 rapid amplification of cDNA ends (RACE). We performed PCR analysis on cDNA prepared from a mixture of 0–48 h postfertilization (hpf) zebrafish embryos. Amplified fragments
were sequenced and corresponded to cDNAs with high
identity to the mammalian tcf4 genes (Fig. 1A). The zebrafish tcf4 gene is expressed as several splice variants as
previously described for the human hTCF-4 gene, which
shows at least 256 different isoforms (Duval et al., 2000).
Since we find that the alternative splicing pattern of the
zebrafish gene is identical to that of the hTCF-4 gene, we
use the same exon numbering convention as that established
by Duval et al. (2000). Also following this nomenclature,
we have named the three classes of splice variants found in
our experiments ‘long’ (L), ‘medium’ (M), and ‘short’ (S),
depending on the length of the COOH-terminal tail. A schematic representation of the five variants cloned and
sequenced in this work is shown in Fig. 1B.
tcf4 isoforms have an amino-terminal stretch of 53 amino
* Corresponding author. Tel.: 156-2-678-7390; fax: 156-2-276-3802.
E-mail address: mallende@machi.med.uchile.cl (M.L. Allende).
acids which shares 93% identity with the b-catenin interaction domain of mouse and human tcf4, and includes
conserved residues shown to be essential for this interaction
(Omer et al., 1999; Poy et al., 2001; Graham et al., 2001). As
described for the mammalian and amphibian genes, we find
alternative splicing in the region corresponding to exon 4 of
tcf4 (amino acids 128–150 of the human sequence; see Fig.
1A). Interestingly, we have cloned variants not encountered
in previous studies which lack both exons 4 and 5 (amino
acids 128–184 of hTCF-4; see Fig. 1B, variants iv and vi).
These sequences correspond roughly to the domain that
interacts with the groucho-like co-repressors and which is
present in both the Tcf-3 and Tcf-4 proteins (Cavallo et al.,
1998; Roose et al., 1998; Brantjes et al., 2001). We speculate that the isoforms lacking this domain may not interact
with the groucho-like proteins and may be unable to function as transcriptional repressors. Though we did not clone a
variant which includes both putative exons 4 and 5, we
performed reverse transcription (RT)-PCR experiments
using primers flanking this region and we detected a band
of the size expected for such a transcript (Fig. 1D, band of
approximately 450 bp). This suggests that there is a possibility of variation in this region producing a protein that
contains the full sequence spanning from the equivalent of
amino acids 128 to 184 of the human sequence (exons 4 and
0925-4773/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved.
PII: S 0925-477 3(02)00180-6
270
R.M. Young et al. / Mechanisms of Development 117 (2002) 269–273
Fig. 1. (A) Sequence alignment of a subset of the zebrafish tcf4 variants (ztcf-4) cloned in this work with the human (htcf-4) and mouse (mtcf-4) genes. We
show one of the human variants (L2 as defined by Duval et al., 2000) and we refer to splicing landmarks according to the exons present in this gene and to
amino acid positions present in the putative human L2 protein. The b-catenin binding domain, HMG box and CtBP binding domains are indicated as solid lines
above the sequence. Numbers refer to amino acid positions present at the borders of exons 3–4 (128), 4–5 (150) and 5–6 (184) in htcf-4. The solid line above the
gap at position 291 of the human sequence represents a position in which an alternative splice acceptor is used for exon 9 with the possibility of introducing five
additional amino acids. For the zebrafish S variant sequence, we only show the C-terminal Leucine as the rest of the amino acid sequence is identical to the M
variant. Asterisks represent stop codons. (B) Diagrammatic representation of the five tcf4 variants (ii–vi) cloned in this work with the names assigned by us to
each one and their respective GenBank accession numbers. Numbers at the top indicate the putative exons corresponding to each segment of the protein
(boxes). The top-most diagram (i) (labelled PCR*) represents a hypothetical L variant which has been deduced by PCR but has not been cloned. Exons 4 and 5
are alternative, while for exons 9 and 17 alternative splice acceptor sites are used. The hatched pattern in exon 9 represents the alternatively added amino acids
SLLSS and the stippled patterns in exon 17 represent the three different translation products produced (represented by the letters L, M and S in our
nomenclature). (C) Alternative splice acceptor site usage in exon 17. Shown are the three possible deduced amino acid sequences that can be produced at
the C-terminus of the zebrafish tcf4 proteins which classify the variants as L, M or S. (D) Expression of tcf4 during embryogenesis. RT-PCR was carried out
using primers which amplify exons 1 through 6 with total RNA prepared form the indicated embryonic stages. Three different bands are specifically amplified.
The higher molecular weight (MW) band (arrow) corresponds to the variant which contains both presumptive exons 4 and 5 and which is expressed zygotically
at all stages examined. The medium MW band corresponds to variants lacking exon 4 which are expressed maternally and zygotically at all stages examined.
The lower MW band (arrowhead) corresponds to variants lacking both exons 4 and 5 and which can only be detected at 48 hpf. M, Molecular Weight Marker;
hpf, dpf, hours and days post-fertilization, respectively; 50%, 50% epiboly stage; Tb, tailbud stage.
R.M. Young et al. / Mechanisms of Development 117 (2002) 269–273
5; Fig. 1B, variant i). This premise is also supported by the
fact that all of the alternative exon borders found in zebrafish tcf4 coincide with exons mapped in the human genomic
sequence of hTCF-4 (Duval et al., 2000).
Further downstream in the protein sequence, an alternative splice acceptor site in presumptive exon 9 can eliminate
five amino acids (SLLSS, position 291 of the human
sequence), a splice event that has also been described for
hTCF-4 and Xtcf4 (Duval et al., 2000; Pukrop et al., 2001).
The absence of this short stretch is predicted to enhance the
ability of Xtcf4 to form a tertiary complex with DNA and bcatenin, due to an inability of these variants to be phosphorylated at this position (Pukrop et al., 2001).
The High Mobility Group (HMG) box, common to all Tcf
proteins, is present in all variants encountered and shows
100% homology with the human and mouse sequence
within this region.
A third region of splicing complexity is found in the region
corresponding to the carboxy-terminus of the zebrafish tcf4
proteins, as is the case with the mammalian isoforms. After
the amino acid corresponding to position 481 of the human
sequence, three different carboxy-termini are produced. The
different reading frames used arise as a consequence of alternative splice acceptor site usage in exon 17 (Fig. 1C). For this
exon, the S and M isoforms code for 1 and 13 amino acids,
respectively, whereas the zebrafish L form is very similar in
length to the reported human L sequence (145 vs. 137 amino
acids, respectively) though there are only limited stretches of
identity. However, it is noteworthy that two of these regions,
which are absolutely conserved between human and fish (Fig.
1A), correspond to the binding sites to the C-terminal binding
protein (CtBP) co-repressor (Brannon et al., 1999). Therefore, of the five isoforms of tcf4 that we sequenced, only two
(Fig. 1B, variants ii and iii) contain the canonical CtBP binding sites.
We analyzed the expression of tcf4 mRNA during embryogenesis. By RT-PCR, we detect bands of the predicted
sizes for a subset of the splice variants expressed at different
developmental stages (Fig. 1D). The primers used span the
amino-terminal half of the tcf4 predicted protein and therefore we can evaluate alternative splicing of exons 4 and 5 in
this experiment. The variant lacking putative exon 4 is
detected at all stages examined, including fertilized zygotes,
indicating that it is present maternally. The variant containing both exons 4 and 5 is found beginning at gastrula and
during the subsequent stages of embryogenesis examined
(arrow in Fig. 1D). Finally, the variant lacking both exons
4 and 5 (arrowhead in Fig. 1D) was found only at the 48 hpf
stage indicating a very specific temporal regulation of this
splicing event. The appearance of this isoform coincides
with the transient appearance of a novel expression site in
the embryos (see below), which suggests that this isoform
could also be tissue specific. We do not know if a variant
containing exon 4 and lacking exon 5 is produced but we did
not encounter such a clone among the 20 cDNA amplification products spanning this region that we sequenced.
271
The distribution of tcf4 message was analyzed in wholemount embryos by in situ hybridization. Our probe was made
using a 1.3 kb 5 0 RACE product spanning the region coding
for the entire amino-terminal half of tcf4. Therefore, in our in
situ hybridization experiments, we are not able to distinguish
the different isoforms produced by alternative splicing of
exons 4 and 5. Expression was not detected in gastrula
stage embryos (50–70% epiboly) by this method. Beginning
at the one-somite stage, tcf4 mRNA is found to be restricted
to the presumptive forebrain region and the label appears to
be distributed in a graded fashion being highest anteriorly
(Fig. 2A,B), resembling the expression pattern of zebrafish
tcf-3 (Kim et al., 2000). In 18 hpf embryos, we find residual
expression in the ventral telencephalon though expression
appears very strongly in the diencephalon (not shown). At
24 hpf, tcf4 is expressed in a restricted pattern in the brain in a
territory spanning the most dorsal and posterior region of the
diencephalon, though excluding the epithalamic region
(arrowhead in Fig. 2D), extending slightly into the anterior
midbrain (Fig. 2C,D). In the diencephalon, we observe a
higher intensity of staining in the most outer alar region of
Fig. 2. Expression of tcf4 during embryonic development. Whole-mount in
situ hybridizations are shown of the 1–2-somite stage (A,B), 24 hpf (C,D),
and 48 hpf (E,F) stage embryos shown laterally, dorsal up, anterior to the
left (A,C,E) and in dorsal views, anterior to the left (B,D,F). Expression at
the early somite stage (A,B) is restricted to the anterior neural plate. At 24 h
(C,D), expression is only detected in a region including the dorsal diencephalon and anterior midbrain (arrow in D). Note the absence of expression
in the roof of the diencephalon which includes the epiphysis (arrowhead in
D). At 48 h (E,F), expression has extended posteriorly from the diencephalon and includes the anterior tectum and dorsal thalamus. Expression also
appears transiently in two to three rhombomeres in the hindbrain. Lines
shown in (F) indicate the planes of section for the images shown in Fig. 3.
272
R.M. Young et al. / Mechanisms of Development 117 (2002) 269–273
the territory (arrows in Figs. 2D and 3B). Double staining
with a shh probe shows that tcf4 limits anteriorly and
ventrally with the zona limitans (Fig. 3A, arrow), indicating
that it is excluded form the ventral diencephalon.
At 48 hpf, tcf4 is expressed in the dorsal thalamus, the
pretectum, the optic tectum, and the torus semicircularis
(Fig. 3C–E). This distribution of tcf4 transcripts in the anterior brain is similar to what has been reported for the mouse
(Cho and Dressler, 1998; Korinek et al., 1998a; Lee et al.,
1999). Additionally, 3–4 vertical bands of cells in the hindbrain express tcf4 (Fig. 2E,F). The position of these bands
with respect to the otic vesicles indicates that expression is
limited to rhombomeres 4 and 5, and expression is localized
to the border regions of these rhombomeres (Fig. 3G).
Transverse sections through the hindbrain show that tcf4
RNA is restricted to the alar-most regions mid-way between
pial and ventricular surfaces (Fig. 3F) in regions that coincide with the positions of commissural neurons (Trevarrow
et al., 1990). In the mouse, hindbrain expression of tcf4 was
detected in rhombomeres 1, 3 and 5, and, in a striking
parallel with the fish, expression is also circumscribed
temporally, in this case to a brief interlude between stages
7.5pc and 8.5pc (Cho and Dressler, 1998). It is interesting to
note that the late-appearing hindbrain expression of tcf4—
seen only at 48 hpf in our in situ timecourse—coincides
temporally with the appearance of the 290 bp band in our
RT-PCR experiments (Fig. 1D) and which we assign to the
variant that lacks the groucho interaction domains.
In situ hybridizations done with 3 days post-fertilization
(dpf) embryos show that tcf4 is also expressed in the developing gut (Fig. 3H), which is consistent with the reported
expression of mammalian tcf4 genes and the implication of
hTCF-4 in colorectal cancer pathologies and gut differentiation (Korinek et al., 1997, 1998b; Lee et al., 1999).
2. Experimental procedures
Fig. 3. Expression of tcf4 during late embryogenesis. Double in situ hybridization with tcf4 and shh probes at 24 hpf (A,B) shows that the expression
patterns of these genes do not overlap in the diencephalon. The zona limitans (arrow in A) can be seen between the shh stain (blue) and the tcf4 stain
(red). In a frontal view, it is also possible to observe that the tcf4 stain is
restricted to the alar diencephalon (arrow in B). Parts (C–F) show transverse
sections through the forebrain (C), midbrain (D,E) and hindbrain (F) of 48
hpf embryos hybridized with tcf4 probe (the planes of section are shown in
Fig. 2E). The section through the diencephalon (C) shows that expression is
restricted to the thalamus and is excluded from the dorsal epithalamic
region. At the level of the anterior midbrain (D), expression can be
observed both in the roof of the tectum (arrow) and in the pretectum (arrowhead). In a slightly more posterior section through the anterior midbrain (E),
expression can be seen in the dorsal tectum (arrow) and in the torus semicircularis (arrowhead). A section through the hindbrain (F), shows expression restricted to the alar plates but not the basal plates. A close-up of a
lateral section of the hindbrain (G) shows four vertical bands of cells, the
two central ones in close contact with one another. Analysis of wholemount embryos shows that these bands correspond to the anterior and
posterior regions of rhombomeres 4 and 5. In (H), a close-up image of
the gut region in a 3 dpf embryo shows specific staining in this region
(arrow).
Fish and embryos were maintained according to standard
procedures in our own laboratory. Embryos were raised at
28 8C and fixed for in situ hybridization in 4% paraformaldehyde. Hybridizations were done as previously described
(Jowett and Lettice, 1994). The cloning of zebrafish tcf4
variants was done by 3 0 and 5 0 RACE Kits (Gibco) according to the manufacturer’s instructions. Primers were: for
RT-PCR, Tcf4 F2, 5 0 -TCAAAACAGCTCTTCGGATTCCGAG-3 0 ; Tcf4 R1, 5 0 -CTGTAGGTGATCAGAGGTGTGAG-3 0 . For amplifying full length tcf4 variants, Tcf 1F,
5 0 -CGTCTCCGGCTTCTGTCTCTGC-3 0 , and Tcf R2, 5 0 CGCACAGACGTACAGACAGACAG-3 0 , were used.
Sequences were deposited in GenBank and the accession
numbers are shown in Fig. 1B.
Acknowledgements
The authors are indebted to Steve Wilson for invaluable
help in recognizing the complex expression pattern of tcf4 in
the brain and for helpful comments on the manuscript. The
shh probe was a kind gift of Wen Biao Chen. Technical help
was provided by Claudia d’Alençon and Florencio Espi-
R.M. Young et al. / Mechanisms of Development 117 (2002) 269–273
noza. Grant support was provided by Fondecyt (2010058 to
R.Y., 3010065 to A.R., 1000879 to M.A.) and by the
Millennium Scientific Initiative to M.A. (ICM P99-137-F).
R.Y. received fellowships from Conicyt, The Fulbright
Foundation and The Company of Biologists.
References
Brannon, M., Brown, J.D., Bates, R., Kimelman, D., Moon, R.T., 1999.
XCtBP is a XTcf-3 co-repressor with roles throughout Xenopus development. Development 126, 3159–3170.
Brantjes, H., Roose, J., van de Wetering, M., Clevers, H., 2001. All tcf
HMG box transcription factors interact with groucho-related co-repressors. Nucleic Acids Res. 29, 1410–1419.
Cavallo, R.A., Cox, R.T., Moline, M.M., Roose, J., Polevoy, G.A., Clevers,
H., Peifer, M., Bejsovic, A., 1998. Drosophila tcf and groucho interact
to repress wingless signalling activity. Nature 395, 604–608.
Cho, E.A., Dressler, G.R., 1998. Tcf4 binds b-catenin and is expressed in
distinct regions of the embryonic brain and limbs. Mech. Dev. 77, 9–18.
Dorsky, R.I., Snyder, A., Cretekos, C.J., Grunwald, D.J., Geisler, R., Haffter, P., Moon, R.T., Raible, D.W., 1999. Maternal and embryonic
expression of zebrafish lef1. Mech. Dev. 86, 147–150.
Duval, A., Rolland, S., Tubacher, E., Bui, H., Thomas, G., Hamelin, R.,
2000. The human T-cell transcription factor-4 gene: structure, extensive
characterization of alternative splicings, and mutational analysis in
colorectal cancer cell lines. Cancer Res. 60, 3872–3879.
Graham, T.A., Ferkey, D.M., Mao, F., Kimelman, D., Xu, W., 2001. Tcf4
can specifically recognize b-catenin using alternative conformations.
Nat. Struct. Biol. 8, 1048–1052.
Jowett, T., Lettice, L., 1994. Whole-mount in situ hybridizations on zebrafish embryos using a mixture of digoxigenin- and fluorescein-labelled
probes. Trends Genet. 10, 73–74.
273
Kim, C-H., Oda, T., Itoh, M., Jiang, D., Artinger, K.B., Chandrasekharappa,
S.C., Driever, W., Chitnis, A.B., 2000. Repressor activity of headless/
tcf-3 is essential for vertebrate head formation. Nature 407, 913–
916.
Korinek, V., Barker, N., Morin, P.J., van Wichen, D., de Weger, R.,
Kinzler, K.W., Vogelstein, B., Clevers, H., 1997. Constitutive transcriptional activation by a b-catenin–Tcf complex in APC-/- colon carcinoma. Science 275, 1784–1787.
Korinek, V., Barker, N., Willert, K., Molenaar, M., Roose, J., Wagenaar,
G., Markman, M., Lamers, W., Destree, O., Clevers, H., 1998a. Two
members of the Tcf family implicated in Wnt/b-catenin signalling
during embryogenesis in the mouse. Mol. Cell. Biol. 18, 1248–1256.
Korinek, V., Barker, N., Moerer, P., van Donselaar, E., Huls, G., Peters,
P.J., Clevers, H., 1998b. Depletion of epithelial stem-cell compartments
in the small intestine of mice lacking Tcf4. Nat. Genet. 19, 379–383.
Lee, Y-J., Swenki, B., Shoichet, S., Shivdasani, R., 1999. A possible role
for the High Mobility Group box transcription factor tcf4 in vertebrate
gut epithelial cell differentiation. J. Biol. Chem. 274, 1566–1572.
Omer, C.A., Miller, P., Diehl, R.E., Kral, A., 1999. Identification of Tcf4
residues involved in high-affinity b-catenin binding. Biochem. Biophys.
Res. Commun. 256, 584–590.
Poy, F., Lepourcelet, M., Shivdasani, R.A., Eck, M.J., 2001. Structure of a
human Tcf4–b-catenin complex. Nat. Struct. Biol. 8, 1053–1057.
Pukrop, T., Gradl, D., Henningfeld, K.A., Knöchel, W., Wedlich, D., Kühl,
M., 2001. Identification of two regulatory elements within the High
Mobility Group transcription factor XTCF4. J. Biol. Chem. 276,
8968–8978.
Roose, J., Molenaar, M., Peterson, J., Hurenkamp, J., Brantjes, H., Moerer,
P., van de Wetering, M., Destree, O., Clevers, H., 1998. The Xenopus
wnt effector XTcf-3 interacts with groucho-related transcriptional
repressors. Nature 395, 608–612.
Trevarrow, B., Marks, D., Kimmel, C.B., 1990. Organization of hindbrain
segments in the zebrafish embryo. Neuron 4, 669–679.