Figure 1. Xenopus Follistatin 319 Is a Secreted
(A) Nucleotide and deduced amino acid sequence
of XFS-319. The signal sequence is
underlined; the follistatin modules are boxed.
(B) Comparison of relative positions and numbers
of follistatin modules in testican, osteonectin,
SC1 , and agrin. The open reading frames
are presented as boxes and follistatin modules
as filled boxes. Each box is drawn to scale except
for the long carboxy-terminal box of agrin,
which is denoted by two diagonal lines.
(C) XFS-319 is a secreted protein. Capped synthetic
RNA encoding XFS-319 was used in a
reticulocyte lysate for in vitro translation or injected
in Xenopus oocytes. The conditioned
medium from control uninjected (lane 1) and
injected (lane 2) oocytes as well as the reticulocyte
lysate (lane 3) was fractionated by SDSPAGE.
In both cases, a protein of about 32 kd
was made from the follistatin mRNA.
Figure 2. Expression of Xenopus Follistatin (XF-319) mRNAs during Different Stages of Xenopus Development
(A) Developmental Northern blot using 2 pg of stage-specific poly(A)+ RNA per lane. The blot was probed with full-length XFS-319 cDNA. Two
transcripts of 3.6 and 2.4 kb are detected beginning at late gastrula/early neurula (stage 12). Overexposure of the same blot indicates that the
2.4 kb transcript is present maternally (data not shown). The embryonic stages are as follows: 9, blastula; IO, 11, and 12 are early, middle, and
late gastrula; 13, 14, and 20 are early, middle, and late neurula; 36 is a tadpole stage. The lower panel shows the hybridization of the same blot
with the Xenopus fibronectin gene to test for RNA recovery. The low follistatin and fibronectin signals at stage 11 appear to be due to the fact
that less RNA is loaded in that lane (see whole-mount in situ hybridizations for stage 11 in Figure 38).
(B) XFS-319 transcript is expressed maternally. RT-PCR analysis of total RNA from fertilized eggs demonstrates that the XFS-319 transcript is
encoded maternally. The control lane is a reaction with all the ingredients of the neurula lane taken through the whole procedure, from which,
however, the RT was omitted. The presence of the translational elongation factor, EF-la, is used as positive control.
(C) XFS-319 RNA is present on the dorsal side at the onset of gastrulation. Embryos at stage 10% were dissected into dorsal and ventral parts.
RNA from each part was analyzed by RT-PCR. XFS-319 is present only in the lane containing RNA from the dorsal side. The presence of noggin
and goosecoid in the same fraction controls for the accuracy of the dissections. The assay for EF-la demonstrates again that comparable amounts
of RNA are present in all experimental lanes. The control lane contains all the ingredients of the embryo lanes except for RT.
Figure 3. Whole-Mount In Situ Hybridization of Xenopus Embryoswith
an Antisense XFS-319 RNA
(A) Stage 10 embryo, vegetal view. Follistatin RNAs can be detected
in a few cells of the organizer (top of embryo). (6) Mid-gastrula stages.
RNAs are confined to the prechordal and anterior chordal mesoderm.
(C) Early neurula stage. RNAs are localized to the head mesoderm
and anterior portion of the notochord. (D) Mid-neurula. Broad band of
expression covers the prospective midbrain-hindbrain junction and
hindbrain region. (E) Late neurula. The neural tube and dorsal axis
were isolated from the rest of the embryo, showing that the broad
band of expression is now confined to a few stripes in the midbrain
and hindbrain. (F) Ventral side of the explant in (E), showing theexpression
in the anterior notochord. The arrows indicate the position of the
blastopore. (0) Superficial view of an early tailbud. showing expression
in the pronephros. (H and I) Expression in the eye and pronephros in
tailbud embryo. (J and K) RNAs are expressed abundantly in the head
in swimming tadpole stage embryo. (L) Control using sense XFS-319
RNA as a probe. All embryos shown are albinos.
Figure 4. Transverse Sections of Embryos Stained by Whole-Mount
In Situ Hybridization for XFS-319 RNA Distribution
(A) Transverse section of a late gastrula (stage 11.5). Animal pole is
at the top, dorsal to the left, and ventral to the right. XFS-319 RNA is
exclusively localized to the dorsal side of the embryo. RNA staining
is perinuclear and restricted to cells in the deep layer, prechordal
mesoderm, and anterior chordal mesoderm. The arrow marks the yolk
(B) Transverse section of an early neurula (stage 14) embryo showing
the localization of XFS RNA in the notochord, just underneath the
open neural plate, and the cells of the hypochord.
(C) Transverse section through the forebrain of a swimming tadpole,
showing that XFS RNA is localized as three dorsoventral stripes in
the cells of the ventricular zone (arrows). Staining is absent from
the floor plate and roof plate. For both (B) and (C), dorsal side is at
Figure 5. XFS-319 Inhibits Morphogenetic
Movements and Mesoderm Formation Induced
- + + + + + + Activin
(A) Injection of follistatin RNA blocks the morphogenetic
movement of cells in response to
activin. The top two panels are control caps
injected with globin RNA and show strong elongation
in the presence of activin (plus) and no
elongation in the absence of activin (minus).
The two lower panels are animal caps injected
with follistatin. This treatment completely
blocks the morphogenetic movement induced
(6) Effect of XFS319 protein on animal caps.
The top two panels show that in either buffer
aloneor in the presence of conditioned medium
collected from uninjected oocytes (0) the caps
do not elongate. Incubation of caps in conditioned
medium containing activin results in
elongation, while XFS-319 protein (XFS-319)
does not elicit morphogenetic movements in
the caps. Addition of an equal volume of XFS-
319 to the activinconditioned medium results
in the blockage of the activin-induced morphogenetic
effect. We did not detect any cement
gland under these experimental conditions.
(C) Northern blot analysis of mesoderm induction
in the presence of Xenopus follistatin. Embryos
at the 2cell stage were injected in the
animal pole with different concentrations of
XFS-319 RNA or the control (globin) RNA. Animal
caps were explanted at the blastula stage
(stage 8) and cultured wtth or without activin
until sibling uninjected embryos reached the
early tailbud stage (stage 28). Total RNA from
20 animal caps was analyzed by Northern blots. Control caps injected with 4 ng of globin induce muscle actin transcription in response to activin.
This induction is blocked in a dose-dependent fashion in caps that have received follistatin. The two top bands are uniformly expressed actin
transcripts and thus demonstrate that a comparable amount of RNA was loaded in each lane. The bottom band (arrow) is the muscle-specific
Figure 6. Xenopus Follistatin Induces Neural Markers in the Absence
of the Axial Mesodermal Marker Muscle Actin
Animal caps injected with 2 ng of globin or XFS-319 RNA were cultured
in buffer alone until sibling controls reached earlytailbud stages. These
caps were tested by Northern blot for the expression of two general
neural markers, N-CAM and 64ubulin II, as well as the axial mesodermal
marker muscle actin.
Figure 7. XFS-319 Does Not Induce the Expression of Immediate
Early Mesodermal Markers in Animal Cap Explants
Embryos were injected in the animal pole of both blastomeres with 2
ng of XFS-319 RNA or control globin RNA. Animal caps, from uninjetted
and injected embryos, were explanted at blastula stage (stage
8) and divided into two groups. The first group was allowed to develop
until sibling uninjected embryos reached midgastrula stage (stage
10.5), and RT-PCR assay was used to score for expression of immediate-
early mesodermal markers in uninjected or injected animal caps.
Neither of the two general immediate early markers of mesoderm,
X&a and MM, nor the dorsal specific markers, noggin and goosecoid,
nor the ventral marker Xwnf-8 are induced by XFS-319. Addition of
activin to the uninjected caps induced the expression of all of these
markers in the explants, demonstrating that these caps are responsive
to induction. The embryo lane represents RNA extracted from stage
10.5 embryos and provides a positive control for the primers and the
size of the expected product. The negative control lane contains all
the ingredients of the embryo lane except for RT. The second group
was cultured until early tailbud stages and analyzed by RT-PCR for
expression of neural markers (as in Figure 9). This provides a positive
control for the neural inducing activity of XFS-319 RNA.
Figure 8. Dose Response Effect of XFS-319 RNA and Neural Induction
Embryos were injected in the animal pole at the 2-cell stage, in both
blastomeres, with different concentrations of XFS-319 RNA. Animal
caps were explanted from blastulae (stage 8) and cultured until early
tailbud stage (stage 28) at which point total RNA was harvested from
all explants and analyzed by Northern blots for the expression of the
general neural marker, f&tubulin isotype II, as well as the cement gland
marker, XAG-7. While the uninjected control animal cap does not show
expression of either marker, as little as 50 pg of XFS-319 induces the
general neural marker. Moreover, the cement gland-specific marker
requires a minimum of 250 pg to be induced. The same blot was
stripped of its signals and reprobed with muscle actin. This control
demonstrates that neuralization is direct and was not accompanied
by mesoderm induction. The two top ubiquitous bands of actin demonstrate
that comparable amounts of RNA were loaded in each lane.
Figure 9. Anteroposterior Character of the Neural Tissue Induced by
Embryos were injected with 2 ng of XFS-319 or control globin RNA
at the 2-cell stage, and, at the blastula stage, animal caps were removed,
and the expression of a set of neural anteroposterior markers
in tailbud stage explants was assessed by RT-PCR. While uninjected
caps or caps injected with the same amount of the control globin RNA
do not induce the expression of any of these markers, animal caps
expressing XFS-319 specifically induce the expression of the general
neural marker N-CAM and anterior markers opsin, En-2, and tanabin.
Krox-20 and Xlhbox-6 expression is absent in these caps. For opsin,
the arrow points to the band of predicted size for this marker. The
axial mesodermal marker, muscle actin, is not induced in any explant.
The EF-la control demonstrates that a comparable amount of RNA
was assayed in each set.
Figure 10. Activin, but Not Noggin, Induces
the Expression of Xenopus Follistatin
(A) Animal cap exolants were isolated from
btastulae (stage 8) and incubated in the presence
or absence of activin. When sibling controls
reached the gastrula stage (stage 10.5)
total RNA was assayed by RT-PCR for XFS-
expression. Explants incubated in the
presence of activin induce the expression of
follistatin. The embryo lane contains RNA from
sibling controls (stage 10.5). The control lane
contains all the ingredients of the embryo lane
except for RT. The EF-1 o signal demonstrates
that a comparable amount of RNA was assayed
in each lane.
(6) Left panel, 2-41 stage embryos were injected
in the animal pole of both blastomeres
with 1 ng of noggin RNAor control globin RNA.
The explants were cut, harvested, cultured,
and scored for XFS-319 RNA under the same
conditions described for (A). Noggin does not
induce the expression of follistatin RNA. Noggin
in these experiments is active, since it was
able directly to neuralize animal cap explants
from the same experiment harvested at tailbud
(C) Activin induces the expression of follistatin
in the context of the whole embryo. Embryos
were injected at the &to 16-cell stage in a single
vegetal blastomere with 2 pg of either Xenopus
activin Bb or globin control RNA. Embryos
were allowed to develop until early neurula
stage (stage 21) and ware then assayed by
whole-mount in situ hybridization for the expression
of follistatin. While the control injected
embryo at the right demonstrates a single primary
axis (I), the embryo on the left injected
with activin contains an ectopic axis (II) that
stains positively for follistatin transcript. The in
situ probe used in this study and the sense
control probe (not shown) are the same as
those in Figure 3.