Fig. 1. Wild-type and mutant type I activin receptors. (A) Protein sequence alignment between Xenopus ALK4 (XALK4) and three human type I receptors: ALK4, ActRI and ALK5. The signal sequence is overlined and the transmembrane domain is boxed. The cytoplasmic domain starts with amino acid (aa) 143, the GS domain contains the GS core sequence at aa 182-187. (B) XALK4 expression pattern. Top, RT-PCR analysis of XALK4 at different developmental stages; stage 2, 2-cell stage; stages 9 to 11, gastrula stages; stages 12 to 20, neurula stages; stage 28 tailbud stage; stage 35, tadpole. XALK4 is expressed maternally and persists during development. Bottom: in situ hybridization of XALK4 at gastrulation (left, vegetal view with dorsal lip at the top) and tailbud (right, anterior at the left) stages, showing uniform expression in dorsal and ventral sides at early stage and expression in many tissues in late embryos. (C) Schematic presentation of the wild-type and mutant receptors used in this study. tXALK4 contains the first 164 aa of XALK4, which includes the extracellular and transmembrane domains, but excludes GS and kinase domains. GS domain of ALK4 contains Thr at positions 202 and 206, which have been shown to be important for its function in cell culture assays. ALK4-T206E mutant induces transcriptional responses in a ligand-independent manner in cell culture (Willis et al., 1996).
Fig. 2. Mesoderm induction assay by wild-type and mutant ALK4 in
ectodermal explants. (A) Morphology of animal caps expressing the
different forms of ALK4. Embryos were injected with RNAs
encoding wild-type ALK4 or a constitutively active point mutant
ALK4-T206E. Caps dissected at blastula stages 8 to 9 were allowed
to grow in the absence (left panels) or presence (right panels) of
activin until sibling control embryos reached neurula stage 18.
(B) Expression of mesoderm-specific markers in animal caps injected
with different receptor RNAs. Xbra is an early pan-mesodermal
marker, muscle actin is a marker of paraxial mesoderm. Lanes 1-3,
caps incubated in buffer alone; lanes 4-6, caps incubated in the
presence of activin. Lanes 1 and 4, uninjected control caps; lanes 2
and 5, caps injected with wild-type ALK4 RNA; lanes 3 and 6, caps
injected with ALK4-T206E RNA; lanes 7 and 8, whole-embryo
controls, in the absence (lane 7) or presence (lane 8) of reverse
transcriptase in the RT-PCR reaction. While wild-type ALK4 is a
weak mesoderm inducer, ALK4-T206E induces mesoderm in animal
caps in the absence of added activin.
Fig. 3. Truncated XALK4 inhibits mesoderm induction by activin
and BMP4. (A) tXALK4 inhibits mesoderm induction by activin.
Embryos were injected with tXALK4, tXActRIIB or tBR RNAs at 2-
cell stage and animal cap explants were dissected at blastula stages 8
to 9. The caps were either incubated alone (lanes 1 to 4), or with
activin (lanes 5 to 8). Total RNA was assayed at either gastrula stage
11 (top panels, ‘Early’) or tailbud stage 28 (bottom panels, ‘Late’) by
RT-PCR for expression of mesodermal-specific markers. Lanes 1 and
5, uninjected controls; lanes 2 and 6, injected with tXALK4; lanes 3
and 7, injected with tBR; lanes 4 and 8, injected with tXActRIIB.
Lane 9 is a negative control without reverse transcriptase in the RTPCR
reaction, and lane 10 is a positive control using RNA extracted
from whole embryos at either gastrula (top panel) or tailbud (lower
panel) stages. (B) tXALK4 inhibits mesoderm induction by BMP4.
BMP4 RNA was injected alone or coinjected with the truncated
receptor RNAs. Explants were assayed as described above. Xwnt8
and Xhox3 are early markers of ventroposterior mesoderm. Globin is
a marker of ventral mesoderm. (C) tXALK4 does not inhibit
mesoderm induction by bFGF. Late blastula explants were incubated
in the absence (lanes 1 to 4) or presence (lanes 5 to 8) of 100 ng/ml
bFGF protein, and assayed when sibling controls reached late
neurula stages as described above.
Fig. 4. Wild-type XALK4 receptor rescues tXALK4 blocked activin,
but not BMP4, mesoderm induction. Ectodermal explants derived
from embryos injected with tXALK4 RNA alone or coinjected with
tXALK4 and XALK4 or XActRIIB RNAs were assayed for
expression of mesodermal markers. Lane 1, uninjected control
explants in buffer alone; lanes 2-5, explants incubated with activin
(A) or injected with 0.5 ng BMP4 RNA (B). Lane 3 is animal caps
injected with 2 ng tXALK4 RNA, lane 4 is caps coinjected with 2 ng
tXALK4 and 0.5 ng XALK4 RNA, and lane 5 is caps coinjected with
2 ng tXALK4 and 0.5 ng XActRIIB RNA. (A) Induction of muscle
actin by activin, assayed when control siblings reached tailbud stage
28, was rescued by coinjection of wild-type XALK4 RNA, but not
by XActRIIB. (B) Expression of BMP4-induced early mesodermal
marker Xbra at control sibling gastrula stage 11 and the late marker
globin at stage 28 could not be rescued by co-expression of XALK4,
but was rescued by wild-type XActRIIB receptor.
Fig. 5. Truncated XALK4 blocks mesoderm formation in vivo.
Albino embryos were injected with 2 ng of tXALK4 or constitutively
active ALK4-T206E RNAs in the marginal zone of one blastomere at
the 2-cell stage. The embryos were fixed at gastrula stage 10 and
whole-mount in situ hybridization was performed using an Xbra
antisense probe to assess the effect of these constructs on mesoderm
formation in vivo. (A,D) Uninjected control embryos, showing an
intact ring of Xbra expression during early gastrula stages.
(B,E) Embryos injected with tXALK4 RNA. Consistent with
tXALK4 inhibition of mesoderm induction, Xbra expression is only
detected in half of the marginal zone. (C,F) Embryos injected with
ALK4-T206E RNA. These embryos display an expanded Xbra
expression that includes the marginal zone and the animal cap on one
side of the embryo.
Fig. 6. Expression of tXALK4 in embryos leads to the elimination of the body axis and a severe reduction of mesodermal tissue.
(A-D) Embryos at the top of each panels are uninjected controls, while the embryos at the bottom of each panels are injected with tXALK4.
(A) Phenotype of embryos injected with 1 ng of tXALK4 RNA in the marginal zone of all four blastomeres at the 4-cell stage, display no
obvious axis and resemble the ‘bubble’ phenotype described for tXActRIIB. (B-D) Analysis of tissue-specific molecular markers by wholemount
immunohistochemistry. (B) Staining with a muscle-specific antibody shows reduced muscle tissue in injected embryo. (C) Staining for a
notochord antigen shows the same reduction in the injected embryo as above. (D) Staining for a neural-specific antigen demonstrates that neural
tissue, though not reduced as severely, is disorganized. (E-G) The phenotype imposed by tXALK4 in the embryo can be rescued by coinjection
of wild-type activin type I receptor. (E) Control uninjected tadpole. (F) Phenotype of embryo injected with 2 ng of tXALK4 into the two dorsal
blastomeres in the marginal zone. (G) Coinjection of tXALK4 with 100 pg of the wild-type XALK4 RNA can rescue the phenotype and restore
body axis in embryos.
Fig. 7. tXALK4 does not induce neuralization in intact explants and
does not inhibit epidermal induction by BMP4. (A) Ectodermal
explants derived from embryos injected with tXALK4, tXActRIIB or
tBR RNAs were assayed by RT-PCR for the expression of
ectodermal-specific markers when sibling stages reached late neurula
(stage 20). NRP-1 and NCAM are neural-specific markers. Lane 1 is
uninjected animal cap control, lanes 2 to 4 are caps injected with
RNA encoding tXALK4, tBR and tXActRIIB, respectively. Lane 5 is
a negative control without reverse transcriptase, and lane 6 is a
positive control with RNA extracted from whole embryos. tXALK4
does not neuralize intact ectodermal explants. (B) Ectodermal
explants derived from embryos injected with tXALK4 or tXActRIIB
were dissociated in Ca2+/Mg2+-free medium. They were
reaggregated either immediately (lanes 1 to 3) or following 4 hours
dissociation (lanes 4 to 9). As shown previously (Wilson and
Hemmati-Brivanlou, 1995), 4 hours of dissociation changed animal
caps from epidermal to a neural fate (lanes 4-6). In lanes 7 to 9,
BMP4 protein was added at 50 ng/ml concentration during
dissociated culture. Unlike tXActRIIB, expression of tXALK cannot
block BMP4-dependent epidermal induction. Lanes 1, 4 and 7 are
uninjected controls; lanes 2, 5 and 8 are embryos injected with
tXALK4 RNA; lanes 3, 6 and 9 are injected with tXActRIIB RNA.
Total RNA was extracted from explants and assayed by RT-PCR
when control siblings reached late neurula stages 17 to 18.
Fig. 8. Working model for differential involvement of XALK4 and
XActRIIB in signal transduction by activin and BMP4. Different
receptor complexes are involved in mesoderm and epidermis
induction by activin and BMP4. For details, see Discussion.
acvr1b () gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF stage 10, vegetal view with dorsal lip up.
alk4 () gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF stage 28, lateral view, anterior left, dorsal up.