Opposing Effects of Wnt and MAPK
on BMP/Smad Signal Duration
Esther M. Verheyen1,*
1Molecular Biology and Biochemistry, Simon Fraser University, Burnaby, BC V5A 1S6, Canada
Signaling downstream of BMP receptors relies on activated nuclear Smad proteins. Recent studies,
one of which is published in the November 30 issue of Cell, shed light on a mechanism that balances
inputs from both activated MAPK and Wnt pathways to regulate the proteasomal degradation of
Smad1 and thus modulate the extent of BMP signaling.
It has long been appreciated that mul-
tiple signaling pathways coexist within
cells during development. Many ex-
amples have been presented in which
these pathways can influence each
other’s behavior through crosstalk.
The puzzle that remains for investiga-
tors is to determine the extent and
relevance of such crosstalk.
The bone morphogenetic proteins
(BMPs) play essential roles during em-
bryogenesis, tissue patterning, prolif-
BMP activity contributes to numerous
and Eng, 2003). With most essential
signaling pathways, multiple levels of
regulation have evolved to ensure
times and the cessation of signaling
when required. Upon BMP ligand
stimulation, the BMP recep-
tor kinases phosphorylate
R-Smads, such as Smad1
(Figure 1). This C-terminal
plex formation with the Co-
Smad Smad4, subsequent
nuclear retention, and DNA
binding, leading to expres-
sion of BMP target genes.
remove the C-terminal re-
ceptor phosphorylation and
inactivate Smads, leading
to them being shuttled out
of the nucleus, yet also al-
lowing possible reactivation
(Chen et al., 2006). The cycle
of phosphorylation and de-
phosphorylation providesa mechanism
by whichSmads monitorBMP receptor
Smads by MAP kinases (MAPKs)
such as Erk, Jnk, and p38 can inhibit
Smad function in vitro and in vivo
et al., 2004). MAPKs can phosphory-
late conserved serine and threonine
sites (pSmad1MAPK) in the linker re-
gions connecting the MH1 and MH2
domains (Figure 1). Linker phosphory-
accumulation of Smads and therefore
inhibits expression of BMP target
genes. Thus, the activity of the BMP
pathway can be tempered by acti-
vated MAPKs downstream of FGF or
Ras signaling. Recently, the MAPK-
like Nemo kinase (Nlk) was also shown
to phosphorylate a Drosophila Smad
(known as Mad), promoting its nuclear
export and inhibiting BMP pathway
activity (Zeng et al., 2007). These stud-
ies reveal that phosphorylation events,
distinct from pSmadCter, can have
a negative impact on BMP signaling.
Smads involves Smurf1, a member of
the E3 ubiquitin ligase family that
targets proteins for ubiquitination and
proteasomal degradation (Zhu et al.,
1999). Smurf1 can target BMP Smads
for degradation in a manner that is
independent of receptor activity. Ear-
lier this year, Sapkota et al. (2007)
found that Smurf1 selectively binds
the protein, leading to its degradation
(Figure 1). Smurf1 was also found to
interfere with nuclear entry of Smads,
attenuation through either
degradation or cytoplasmic
retention of Smads.
Another layer of complex-
vided in a new report in Cell,
in which De Robertis and
phosphorylation but subse-
quent GSK3 phosphoryla-
tion (pSmad1GSK3) in the
linker domain (Fuentealba
et al., 2007) (Figure 1). Both
modifications appear to be
required for Smad1 ubiquiti-
nation and degradation, and
sion of Smad1 in which
Figure 1. Sequential Events in Smad1 Degradation
Activation of BMP receptor kinase activity leads to C-terminal phos-
phorylation of Smad1 (1). Activated MAPKs downstream of multiple in-
puts such as FGF or Ras can phosphorylate the linker domain (2),
which provides a primed substrate for GSK3, which in turn is active
in the absence of Wnt signaling (3). Smurf1 recognizes the doubly
phosphorylated linker domain of Smad1 and polyubiquitinates (Ub) it
(4), targeting Smad1 for degradation in centrosomes (5).
Developmental Cell 13, December 2007 ª2007 Elsevier Inc.
either MAPK or GSK3 target sites were Download full-text
mutated resulted in gain-of-function
phenotypes upon mRNA injections
into Xenopus embryos, and showed
higher gene expression in BMP re-
porter assays. Inhibitors specific to
MAPK, GSK3, or the proteasome all
stabilized pSmadCter. Antibodies that
recognize the pSmadCter, pSmadMAPK,
proved to be particularly useful in fol-
calization of Smad1. Fuentealba et al.
propose that BMP stimulation triggers
events, thereby limiting the duration of
Smad1 activation. While pSmadCter
was detected in the nucleus after
BMP4 treatment, pSmadMAPK
seen in the nucleus and centrosomes,
where the proteasomal machinery is
enriched. pSmadGSK3strongly stained
that converged on the centrosomes.
Upon treatment with a proteasome in-
hibitor, all pSmad forms were enriched
in centrosomes, supporting the model
mal proteasome, where it is degraded.
Whether pSmadCteris a required event
for subsequent phosphorylations re-
mains to be seen, as C-terminal mu-
tants resistant to receptor phosphory-
degradation by Smurf1 (Fuentealba
et al., 2007; Sapkota et al., 2007).
Fuentealba et al. (2007) also demon-
strate that activation of Wnt signaling
or inhibition of GSK3 can abrogate
Smad1 inhibition. GSK3 is a constitu-
tively active, multifunctional kinase
that plays roles in Wnt and Hedgehog
signaling, insulin signaling through
PI3K, and numerous other processes
(Forde and Dale, 2007). In the Wnt
pathway it participates in the destruc-
forms of Smad1
tion complex that targets b-catenin
for ubiquitination and degradation.
Wnt signaling inhibits GSK3, allowing
b-catenin stabilization and activation
of target genes with Tcf/Lef transcrip-
tion factors. Treatment of cells with
Wnt3a can inhibit GSK3 and lead to
in pSmad1GSK3. In Xenopus embryos
neural tissue differentiates when BMP
signaling is inhibited, while epidermis
forms upon high BMP signaling. Ex-
pression of the Wnt pathway inhibitor
Dickkopf1 (Dkk1) causes neuralized
phenotypes like those caused upon
inhibition of BMP (Niehrs et al., 2001).
Fuentealba et al. showed that injection
of either Wnt morpholinos (MO) or
Dkk1 mRNA resulted in expansion of
neural tissues, and that this effect
required intact GSK3 phosphorylation
sites in Smad1. Furthermore, elevated
Wnt or Dkk1 MO could induce epider-
mis, and this effect required active
(2007) and Sapkota et al. (2007) report
comparable findings, there are some
solved questions. Fuentealba et al.
propose that GSK3 phosphorylation
is a critical step toward degradation,
while Sapkota et al. find that GSK3
phosphorylation is not essential, but
acts to enhance Smurf1 binding and
appears that pSmad1Cterphosphory-
lation is not essential for subsequent
MAPK and GSK3 phosphorylation,
the nuclear enrichment of Smad1
seen upon pSmad1Cterphosphoryla-
nuclear MAPKs, facilitating subse-
quent degradation. Interestingly, both
groups find that total amounts of
Smad1 areunchangedin experimental
conditions, showing that the pools of
a small proportion of total protein. It
remains to be seen how the MAPK/
GSK3 induced degradation integrates
regulation of Smads.
These findings shed light on a mys-
tery that has puzzled Xenopus embry-
ologists—namely, how loss of either
BMP or Wnt can trigger the neural
program, even though they have dis-
tinct requirements in patterning. The
integration of negative regulation by
MAPK and GSK3 with the positive
effects of BMP and Wnt illustrates the
interplay of multiple pathways regulat-
ing Smad1 during development.
Aubin, J., Davy, A., and Soriano, P. (2004).
Genes Dev. 18, 1482–1494.
Chen, H.B., Shen, J., Ip, Y.T., and Xu, L. (2006).
Genes Dev. 20, 648–653.
Forde, J.E., and Dale, T.C. (2007). Cell. Mol.
Life Sci. 64, 1930–1944.
Fuentealba, L., Eivers, E., Ikeda, A., Hurtado,
C., Kuroda, H., Pera, E., and Robertis, E.M.D.
(2007). Cell 131, 980–993.
Kretzschmar, M., Doody, J., and Massague, J.
(1997). Nature 389, 618–622.
Niehrs, C., Kazanskaya, O., Wu, W., and
Glinka, A. (2001). Int. J. Dev. Biol. 45, 237–240.
Sapkota, G., Alarcon, C., Spagnoli, F.M.,
Brivanlou, A.H., and Massague, J. (2007).
Mol. Cell 25, 441–454.
Waite, K.A., and Eng, C. (2003). Nat. Rev.
Genet. 4, 763–773.
Zeng, Y.A., Rahnama, M., Wang, S., Sosu-
Sedzorme, W., and Verheyen, E.M. (2007).
Development 134, 2061–2071.
Zhu, H., Kavsak, P., Abdollah, S., Wrana, J.L.,
and Thomsen, G.H. (1999). Nature 400, 687–
Developmental Cell 13, December 2007 ª2007 Elsevier Inc.