Fig. 1. Temporal expression and spatial regulation of vgll3 in Xenopus embryo.
(A) Vgll3 expression detected by RT-PCR starts between stages 11 and 12. (B)
Vgll3 is detected by ISH in r2 during neural tube closure. (C) Vgll3 expression
decreases in stage 18 embryos treated with increasing concentrations of retinoic
acid (RA). FGF8 mRNA injected embryos show an anterior-lateral enlargement of
vgll3 expression domain. Hoxa2 or hoxb2 mRNA injected embryos show a strong
reduction of vgll3 expression. All views are dorsal-anterior. (*), injected side. A dotted line indicates the midline embryo. Quantification of vgll3 regulation results. Three
independent experiments were performed. The number of embryos analysed is
indicated on the top of each bar. (D) Vgll3 is induced in animal caps treated with
noggin + FGF2 (N+F). (E) Vgll3 expression is induced in early but not late animal
cap cells overexpressing wnt8a. (F) Overexpression in animal cap cells of vgll3 in
combination with tead1 stimulates the expression of wnt5a, wnt8b and fgf8. E, non
injected embryo, number indicates the stage; -RT, no cDNA; ni, animal cap from
uninjected embryo; N-tub, N-tubulin. Ornithine decarboxylase (odc) gene expression
is used as control.
Fig. 2. Vgll3 knockdown or overexpression impair trigeminal placode and
nerve formation. (A) Embryos injected with v3MO or v3MOsplicing (v3LMOsp and
v3SMOsp, 20 ng each) exhibit a reduced expression of islet1, neuroD and N-tubulin
in the trigeminal placodes (arrowhead). (B) Overexpression of increasing amounts of
vgll3 mRNA reduces islet1 and N-tubulin expression in stage 19 embryos. Double
ISH shows no prominent overlapping staining between vgll3 (red) and N-tubulin
(blue) (arrow). Injected side (*) was traced by lacZ staining. Gene expression were
assayed by ISH. Arrowhead indicates the trigeminal placodes. (C) Quantification of
results. Images on the top of bars indicated v3LMOsp defects rescued with increasing amounts of vgll3 mRNA injections. Three independent experiments were
performed. The number of embryos analysed is indicated on the top of each bar.
Views are dorsal-anterior excepted for lateral views for stage 28 embryos.
Fig. 3. Vgll3 stimulates pax3 and zic1 expression. (A) Embryos were injected with
v3MO or vgll3 mRNA and analysed at stage 19 for pax3 and zic1 expression.
Defects of CNC lateral streams and trigeminal placode are shown by arrows and
arrowhead respectively. (B) ISH for vgll3 and pax3 shows overlapping expression at
the r2 level (blue arrow). (C) Stage 19 embryos injected with pax3GR/zic1GR
mRNAs show a faint vgll3 and a strong snail2 expression after dexamethasone
treatment (+dex). (D) cMO/LacZ mRNA control and quantification. Three
independent experiments were performed (n=40). (*), injected side. All views are
dorsal-anterior. A dotted line indicates the midline embryo. (E) ISH for vgll3 and zic1 shows overlapping expression at the r2 level (blue arrow). Zic1 expression in stage
Fig. 4. Vgll3 knockdown and overexpression do not affect CNC formation but
block their migration. Embryos were injected with v3MO (40 ng or cMO) or vgll3
mRNA (1ng, or lacZ mRNA) and analysed at different stages for snail2, twist, PCNS
or pcdh18 expression. Pharyngeal arches are indicated (ma, mandibular; hy, hyoid;
br, branchial a for anterior and p for posterior). The mandibular branch of trigeminal
nerve is indicated by an arrowhead. Vgll3 knockdown and overexpression block migration of CNC streams. Overlapping expression of vgll3 and snail2 indicated by
an arrow. The white line indicates the plane of agarose section. Quantification of
results. Three independent experiments were performed. The number of embryos
analysed is indicated on the top of each bar. (*), injected side. The dotted line
indicates the midline embryo. Circle, lateral CNC stream. All views are dorsalanterior
except lateral views for stage 25 and 28 embryos.
Fig. 5. Vgll3 regulates CNC migration. CNC from neurula embryos injected with
tracer gfp mRNA and v3MO (40 ng or cMO) or vgll3 mRNA (1ng) were (A) grafted on
wild type embryos at stage 17 and migratory phenotype was analysed by GFP
fluorescence 18 hours after transplantation (quantification of results in the right
panel, inside panels showed GFP-positive grafted cells just after transplantation;
magnification has been adjusted to reduce the size of original images) or (B) were
plated on fibronectin and analysed 3h (a,d,g,j) and 18h (b,e,h,k) after plating.
Enlarged views of squares (c,f,i,l). Scale bars: 1 mm (j-k); 250 μm (l). (C) Ratio of
spreading measured by comparing the relative surface area between 18h and 3h of
culture (indicated by the line plot in a-k). #, P value <0.05; error bars represent
mean±s.e.m. (D) MyosinX expression in stage 16 (dorsal-anterior view) or stage 28
(lateral view) embryos. (*), injected side. Dotted line indicates midline embryo.
Fig. 6. Vgll3 regulates specific subset of genes in animal cap explants. (A)
Embryos were injected with pax3GR/zic1GR and v3MO (40 ng or cMO), treated or
not with dexamethasone (±dex) before animal caps dissection and analysis by RTPCR.
(B) Embryos were injected with myc-tead1 (T1), myc-tead2 (T2), HA-vgll3 (V3)
or HA-vgll2 (V2) mRNAs and animal caps were dissected and analysed by RT-PCR.
Vertical white lines indicate spliced images in order to remove results not described
in this paper. E16/E17, control stage 16/17 embryos. -RT, no cDNA.
Fig. 7. Vgll3 interacts with tead proteins and can activate pax3 independently
of tead. (A) Embryos injected with vgll3 (1ng or lacZ) mRNA and T1MO/T2MO or
cMO were fixed at stage 10 or 13 and analysed for pax3. Vgll3 induces ectopic pax3
expression when tead1 and tead2 have been knocked down. Lower right panel (IB
myc): embryos were injected with 50 pg of myc-tead1 or myc-tead2 mRNAs with 15
ng of T1MO or T2MO and analysed by IB. T1MO and T2MO block efficiently tead1
and tead2 expression respectively. Erk2 is used as control. ni: non injected embryo.
(B) Embryos injected with HA-vgll3, myc-tead1 or myc-tead2 mRNAs were processed for immunoprecipitation with HA (IP HA) or myc antibodies (IP myc)
followed by IB with antibodies. Vgll3 interacts with tead1 and tead2 in the embryo.
Lysate, IB of injected embryo before immunoprecipitation. Vertical white lines
indicate spliced images in order to remove results not described in this paper (IP
Fig. 8. Vgll3 interacts with ets1 and activates a tead-dependent luciferase
reporter gene. (A) Pax3 expression in embryos injected with vgll3 or ets1 mRNA.
Ets1 overexpression induces ectopic pax3 expression likewise co-expression of ets1
and vgll3. (*), injected side. (B) Embryos were injected with myc-ets1 or HA-vgll3
mRNAs. Immunoprecipitation with myc (IP myc) or HA antibodies (IP HA) was
followed by IB with antibodies. (C) Schematic representation of the pGL284LUC
reporter plasmid used in transfection assay. The different cis-sequences are
depicted together with the MCAT cis-sequence (black box). (D) Left panel: ets1,
tead1 and vgll3 are expressed in HEK293 cells when analysed by RT-PCR. (-), no
cDNA. Right panel: HEK293 cells were transfected with HA-vgll3, HA-V3ΔTDU, myctead1
or myc-ets1 plasmids and similar amounts of proteins were checked by IB with
myc and HA antibodies. Gapdh and erk2 are used as control. (E) HEK293 cells were
co-transfected with pGL284LUC and HA-vgll3, HA-V3ΔTDU, myc-tead1 and mycets1
plasmids. Relative luciferase activity is expressed as x-fold luciferase activity compared to tead1 or ets1 alone. Bars show mean±s.e.m. of three independent
experiments carried out in duplicate. *P value <0.05.
Fig. 9. Proposed model showing vgll3 functions in trigeminal placode/nerve
formation and CNC migration. Vgll3 acts via secreted molecules like WNT or FGF
on neural border expressed genes to control the formation of branchial arches and
trigeminal placode. Vgll3 can interact with ets1 and sustain pax3/zic1 expression and
their downstream target genes. Ant, anterior; Post, posterior; ba1, branchial arch1;
CNC, cranial neural crest; RA, retinoic acid.
Figure S1. Vgll3 does not co-localise with trigeminal placode markers. (A) Wholemount
ISH for vgll3 and trigeminal placode genes islet1, neuroD, pax3 and foxi1c. (B)
Whole-mount ISH for trigeminal placodal genes islet1, neuroD, pax3 and foxi1c. Blue
arrow indicates the position of vgll3 expression domain. All staining are visualized in
blue. Analysed embryos are from stage 14 to stage 20 depending on the trigeminal
markers used. All views are dorsal-anterior.
Figure S2. In vivo specificity of vgll3 morpholinos. Embryos were injected with
500pg of vgll3, vgll3 mismatch (mis) or vgll2 mRNAs in the presence (+) or not (-) of
40 ng of vgll3 (v3MO) or control (cMO) morpholino oligonucleotide antisenses. Stage
15 embryos were stopped and analysed by immunoblotting with HA antibody to detect
vgll3. v3MO efficiently blocks the translation of vgll3 mRNA but has no effect on vgll3
mismatch mRNA neither on vgll2 mRNA. Erk2 is used for loading controls. Right panel,
control MO (cMO) has no effect on vgll3 mRNA translation in an in vitro translation
Figure S3. In vivo specificity and efficiency of v3MOsplicing. (A) Schematic
representation of the vgll3 locus. There are two vgll3 homeolog genes, vgll3.L and
vgll3.S, that are expressed in Xenopus laevis. Specific primers for vgll3.L and vgll3.S
(v3Lsp F and v3Ssp F respectively) and common primer for both mRNAs (v3sp R)
were chosen to detect normal and altered splicing for vgll3.L and vgll3.S mRNAs. The
position of the splicing blocking MO specific to vgll3.L pre-mRNA (v3LMOsp for
v3LMOe2i2) or to vgll3.S pre-mRNA (v3SMOsp for v3SMOi1e2) is indicated by a black
box. (B) Embryos were co-injected with 10, 20 or 40 ng of each morpholino splicing
(v3MOsplicing) and gfp mRNA as lineage tracer. Stage 15 embryos were analysed by
RT-PCR to detect vgll3 expression. V3MOsplicing decreases normal spliced vgll3.L
RNA or vgll3.S level compared to non-injected embryos in a dose-dependent manner.
PCR fragments corresponding to the altered splicing mRNAs appeared at expected
size for both vgll3 mRNAs. Odc is used for loading controls.-RT, no DNA for each
Figure S4. Dose dependent effect of v3MO splicing on N-tubulin and snail2
expression domain. (A) Embryos injected with v3LMOsp, specific to vgll3 form L or
v3SMOsp, specific to vgll3 form S or both (v3MOsplicing) exhibit a reduced expression
of N-tubulin in the trigeminal placodes (arrowhead) and an inhibition of CNC migration
as shown snail2 expression. Stronger effect was observed when both splicing MOs
were injected at 10ng each compared to single MO. Injected side (*) was traced by
lacZ staining. Gene expression was assayed by ISH. Arrowhead indicates the
trigeminal placodes. (B) Quantification of results. Three independent experiments were
performed. The number of embryos analysed is indicated on the top of each bar. All
views are dorsal-anterior. sp, splicing.
Figure S5. Vgll3 impairs trigeminal nerve formation. (A) Embryos injected with 40
ng of v3MO exhibit a decrease of HNK-1 staining (arrowhead) and a quivering
phenotype for 3A10 staining (bottom panel) in the trigeminal ophthalmic and maxillomandibular
branches (50%, n=30 for both staining). (B) Embryos injected with vgll3
mRNA (1ng) exhibit a similar decrease of staining and a quivering phenotype in the
trigeminal branches for HNK-1 (50%, n=25) and 3A10 staining (80%, n=25). Major
branch connected to trigeminal ganglia disappeared (arrowhead) (C) Upper row:
Embryos injected with cMO or lacZ mRNA show no difference in HNK-1 staining
between non injected side (ni) or injected side (*). Lower row: Stage 14 embryos
injected with cMO, v3MO, gfp or vgll3 mRNAs show no apoptotic cells when analysed
by TUNEL. The median line of the embryo is indicated with a dotted line. The injected
side was traced by lacZ staining. HNK1 (specifically labelling axonal outgrowth
antibody) and 3A10 (neurofilament specific antibody) expression was assayed by
immunohistochemistry on stage 28 and stage 42 respectively. (D) TUNEL apoptosis
analysis of stage 17 embryos. V, trigeminal nerve; VII, facial nerve; ey, eye; ov, otic
Figure S6. A second vgll3 morpholino (v3MO2) phenocopies v3MO effects. (A)
Embryos were injected with 10 ng of a second non overlapping vgll3 (v3MO2)
morpholino oligonucleotide antisense, then fixed at stage 16 or 19 and stained for
snail2 expression. No significant effect was detected at stage 16 but defects in neural
crest migration were observed at stage 19 (brackets). (*) injected side. (B) Specificity
of v3MO2 in in vitro translation experiments. (C) Immunohistochemistry analysis of
HNK-1 and 3A10 (stage 28 and stage 35 respectively). White and black arrowheads
show defects of the ophthalmic and maxillo-mandibular branches of the trigeminal
nerve respectively. (*) injected side.
Figure S7. Vgll3 is required for the formation of neural crest derivatives. Embryos
were injected with 40 ng of v3MO (or cMO) or 1 ng of vgll3 mRNA (or gfp mRNA) and
fixed at stage 47 before alcian blue staining (A-I). (A, B, C, D) The formation of cranial
cartilage is impaired when vgll3 is knocked-down when compared to control MO
injected embryo [A compared to D (embryos injected with cMO show no alteration)].
This is more obvious on dissected embryos that revealed on the injected side a high
rate of cranial cartilage hypoplasia including severe loss of Meckel’s (MC), ceratohyal
(CH) and branchial cartilages (Br) (100%, n=15, Fig. 6B,C). (E, F) Transversal sections
of paraffin embedded tissues showed that the ceratohyal cartilage on the injected side
was shorter than on the uninjected control side (Fig. 6E) and displayed numerous
smaller chondrocytes (Fig. 6F, enlarged view of E indicated in the square line. Scale
bars represent 500 μm (E) and 130 μm (F). (G, H, I, J) Embryos injected with vgll3
mRNA showed an impaired cartilage development with the ceratohyal cartilage being
severely disorganized (100%, n=15, Fig. 6G, H). In control experiments, embryos
injected with gfp mRNA showed no change in cartilage head morphology (Fig. 6I, J).
ey, eye. Ventral views with anterior to the top (A, D, G, I) and dorsal views (J). (*)
injected side. Midline embryo is indicated by a dotted line.
Figure S8. The histidine repeat in vgll3 is required for pax3 activation. (A)
Sequence alignment of histidine (H) tract in Vgll3 proteins from different vertebrate
species. Numbers indicate the relative position of the histidine tract. Xenlae, Xenopus
laevis, Xentrop, Xenopus tropicalis. (B) Embryos were injected with 1 ng of vgll3 (vgll3
wt) or vgll3Dhis, fixed at stage 11 and analysed for pax3 expression. Vgll3Dhis cannot
activate pax3 (100%, n=30). (*) injected side. (C) HEK293 cells, recently tested for
contamination, were transfected with 1μg of plasmid encoding HA-vgll3wt or HAvgll3Dhis,
fixed and analysed by immunocytochemistry using HA antibody. The
deletion of histidine tract from vgll3 does not impair its nuclear localization. White arrow
indicates nuclear staining. Scale bars, 10 μm.