REVIEWS

When pathways collide: collaboration

and connivance among signalling
proteins in development
Helen McNeill* and James R. Woodgett‡
Abstract | Signal transduction pathways interact at various levels to define tissue
morphology, size and differentiation during development. Understanding the mechanisms
by which these pathways collude has been greatly enhanced by recent insights into how
shared components are independently regulated and how the activity of one system is
contextualized by others. Traditionally, it has been assumed that the components of
signalling pathways show pathway fidelity and act with a high degree of autonomy.
However, as illustrated by the Wnt and Hippo pathways, there is increasing evidence that
components are often shared between multiple pathways and other components talk to
each other through multiple mechanisms.

A classical example of a signalling pathway with compo- In this Review, we use the canonical and non-canonical
nents that have traditionally been assumed to show path- Wnt pathway and the Hippo growth pathway to exemp-
way fidelity and act with a high degree of autonomy is the lify how better precision, specificity and coordination
cyclic AMP-dependent protein kinase pathway. This sig- can be derived from a minimal parts list. We focus on
nalling system has a core set of molecules that couple the these pathways as they provide examples of several
production of cAMP to the activation of protein kinase A emerging principles of signalling, including the use of
(PKA) and regulation of downstream targets. In some common components, discrimination of signals and
cases this signalling system is linear and relatively insulated, mechanisms for signalling coordination and integration.
such as the adrenaline-induced effects of PKA activation These pathways have been the subject of intense investi-
of phosphorylase kinase leading to glycogen mobilization gation by many laboratories over the past decade and it
in muscle (reviewed in REF. 1). More commonly, however, is thus impossible to adequately review the literature for
these pathways are convoluted and subject to tiered or all of these pathways in this Review. Therefore, we refer
nested levels of input from other pathways. the interested reader to a number of relevant reviews for
Indeed, nature has placed enormous responsibility each pathway (for example, REFS 2–5) and, here, we focus
on relatively few proteins that, together, form the regula- on advances in the past couple of years and on a few
tory skeleton of cellular control. There are only a handful seminal papers that highlight the crosstalk and insula-
of primary signalling pathways that act as guide-wires tion of these signalling cascades. We also note that much
in orchestrating appropriate cellular responses to exter- pathway integration occurs at the level of gene regulation
*‡Samuel Lunenfeld Research
Institute at Mount Sinai
nal and internal cues — in stark contrast to the many through the assembly and disassembly of transcriptional
Hospital, 600 University thousands of proteins and processes they must control. complexes. This topic warrants its own treatise and we
Avenue, Toronto, Although there is a diverse array of receptor proteins that regret that it is only referred to superficially here.
Ontario, Canada M5G 1X5, sense specific ligands, their function is collapsed by the
and *Department of
limited numbers of intracellular pathways to which they The convoluted world of Wnt signalling
Molecular Genetics and
‡
Department of Medical can couple. Moreover, many pathways contain common The Wnt pathway first emerged from elegant classical
Biophysics, University of components, further limiting their apparent potential genetics in Drosophila melanogaster as a key determi-
Toronto, Canada. for specificity. The obvious question, then, is how might nant for segmental and spatial organization of the body
e‑mails: the complexity of a cell be coordinated by such a limited plan. These studies revealed the fundamental architec-
mcneill@[Link];
woodgett@[Link]
vocabulary of commands? A large part of the answer ture of this pathway with double-negative regulation
doi:10.1038/nrm2902 emerges from the multiple ways by which these systems and feedback controls (see below; reviewed in REF. 6).
Published online 12 May 2010 interact, cross-regulate, insulate and collaborate. Geneticists were also the first to realize that the pathway

404 | june 2010 | VoluMe 11 [Link]/reviews/molcellbio

© 2010 Macmillan Publishers Limited. All rights reserved

 focuS on SIgnal IntEg
R ERV
atI EI W
onS

a Cadherin b
LRP5 or LRP5 or
LRP6 Frizzled Plasma LRP6 Frizzled Plasma
membrane Wnt membrane

P
β-catenin β-catenin
Dishevelled P Dishevelled
β-catenin β-catenin
CKI
GSK3 Axin 1
Axin 1 β-catenin β-catenin
APC β-catenin
GSK3 CKI β-catenin
P P β-catenin
APC
β-catenin β-catenin
P P Axin 2
Proteasomal β-catenin
degradation
Nucleus Nucleus
TLE Transcription
TLE β-catenin (for example,
repressor repressor axin 2 and MYC)
TCF TCF

Figure 1 | The canonical Wnt signalling pathway. a | In the absence of a signal, the destruction complex adenomatous
polyposis coli (APC)–axin 1–glycogen synthase kinase 3 (GSK3)– casein kinase 1 (CK1) bindsReviews
Nature and phosphorylates
| Molecular Cellnon-
Biology
cadherin-associated β-catenin, targeting it for destruction by the proteasome. In the nucleus, DNA-binding proteins of the
T cell factor (TCF) and lymphoid enhancer-binding factor 1 (LEF1) family are bound by transcriptional repressors (such as
the transducin-like enhancer proteins (TLEs; homologous to Drosophila melanogaster Groucho)). b | The binding of a Wnt
ligand to its Frizzled receptor and lipoprotein receptor-related protein 5 (LRP5) or LRP6 co-receptor induces a change in
conformation that results in phosphorylation of the co-receptor. This creates a high-affinity binding site for axin 1, causing
disruption of the destruction complex. β-catenin can then accumulate and associate with the TCF or LEF1 proteins,
dislodging the TLE repressors and hence promoting transcriptional activation of a programme of genes, including MYC
and axin 2. Axin 2 feeds back to inhibit the pathway by promoting the assembly of more destruction complexes.

bifurcated, with one collection of components acting in simplifying the analysis of Wnt biology. For example,
on transcriptional regulation through stabilization of Wnt ligands can be classified into acting canonically or
the transcriptional activator β-catenin and the other non-canonically by whether they induce secondary axis
on aspects of planar polarity. each arm of the pathway formation when injected into Xenopus laevis embryos.
shared some components with the other, but certain Conventional descriptions of canonical Wnt signal-
mutations in these shared components had selective ling imply that the pathway is fairly linear, although
effects on one arm or the other, indicating dual and evidence of more complex interactions has been widely
distinct roles. Moreover, some components of each appreciated through proteomic analyses7,8. Adherent cells
arm have the rather disturbing property of lending their contain large amounts of β-catenin, with most of it lining
weight to distinct signalling pathways, thereby raising the plasma membrane in association with cell adhesion
issues of selectivity and specificity. Indeed, this trick molecules of the cadherin family. In the absence of a
of nature has led to considerable confusion with many Wnt signal, a complex of proteins comprising adeno-
researchers as it is generally assumed that the common- matous polyposis coli (APC), the scaffolding protein
ality of a component between two systems translates into axin 1 (which acts to cluster the components), glycogen
that component acting as a functional link. This is not synthase kinase 3 (GSK3), casein kinase 1 (CK1) and
usually the case, but it raises the question of why nature others, acts as a machine to capture ‘surplus’ β-catenin
evolved pathways to employ shared gene products, molecules that are not cadherin-bound and to tag these
given the importance of regulatory pathways. Is it sim- by phosphorylation for ubiquitylation and destruc-
ply parsimony or does this structure offer unappreciated tion by the 26S proteasome (FIG. 1a). This ‘destruction
benefits? Here, we argue the latter, while admitting complex’ is highly efficient and soluble β-catenin levels
Cadherin
that the evidence for this remains sparse. are maintained at very low levels, despite continuous
One of a family of synthesis of β-catenin by the cell.
transmembrane proteins Confusing Wnt terminology The binding of specific Wnt ligands (of which there
that form homotypic, There are distinct branches of Wnt signalling, commonly are 19) to specific Frizzled G protein-coupled receptor
Ca2+-dependent interactions
referred to as canonical and non-canonical pathways (GPCR)-like receptors and low-density lipoprotein recep-
between cells, promoting
adhesion. (FIG. 1; FIG. 2). The former acts largely by controlling lev- tor-related protein (lRP) co-receptors (typically lRP5
els of the non-cadherin-associated pool of β-catenin and or lRP6) occurs through the Dishevelled proteins and
GPCR the latter refers to other actions of Wnt that are β-catenin switches off the destruction complex by stopping phos-
A type of receptor protein that independent. Setting aside the literal inappropriateness phorylation of β-catenin (FIG. 1b). newly synthesized,
traverses the membrane seven
times and is typically coupled
of the term (‘canonical’ is defined in the oxford english unphosphorylated β-catenin, freed from its otherwise
to G proteins (also known as Dictionary as “according to recognized rules or scien- imminent fate of destruction, begins to accumulate
Serpentine receptors). tific laws”), this separation of functions has been useful and to associate with members of the T cell factor (TCF)

nATuRe ReVIeWS | Molecular cell Biology VoluMe 11 | june 2010 | 405

© 2010 Macmillan Publishers Limited. All rights reserved

REVIEWS

a b c d e Plasma
Dachsous membrane
WNT5A
Frizzled Frizzled 7
Vang Gogh Ror and RYK Wnt Wnt RYK
Fat

Prickle PKC CKIε

Dishevelled Atrophin Dishevelled
Diego Dishevelled Canonical signal ?
PI3K
Prickle
SRC
Flamingo
RHOA ? JNK Ca2+

Cytoskeleton CAMKII NFAT Four-jointed Axon guidance

Figure 2 | Non-canonical Frizzled–PcP signalling. Numerous non-canonical Wnt–planar cell polarity (PCP) signalling
pathways have been described. Outlined is a vastly simplified schema highlighting some Nature Reviews
of the | Molecular
key players. Cell Biology
a | The
Frizzled–PCP pathway is enriched asymmetrically at cell boundaries. In Drosophila melanogaster wings, Frizzled,
Dishevelled and Diego accumulate at the distal edge of each cell, and Van Gogh and Prickle accumulate on the proximal
side of each cell. These complexes may be bridged by Flamingo (also known as Stan)–Flamingo interactions across cells, as
well as interactions between Flamingo and Vang Gogh, Flamingo and Frizzled, and Frizzled and Vang Gogh. In addition,
there are mutually repressive interactions in the cell between distal and proximal complexes. Genetically, RHOA has been
placed downstream of the Frizzled–PCP complex. b | WNT5A functions primarily as a non-canonical Wnt ligand, acting
through the Ror and RYK Tyr kinases. Activation of Jun N-terminal kinase (JNK), increases in intracellular Ca2+ leading to
activation of nuclear factor of activated T cells (NFAT) and calmodulin-dependent protein kinase II (CAMKII), and inhibition
of canonical signalling are some of the alterations seen on WNT5A stimulation. c | Non-canonical Wnts may also function
through a subset of Frizzled proteins that bias towards PCP signalling, such as frizzled 7. d | The binding of Dachsous to the
Fat pathway regulates PCP, possibly through the recruitment of casein kinase Iε (CKIε), which has been shown in different
situations to inhibit (or stimulate) PCP signalling. Fat also recruits the transcriptional co-repressor Atrophin (also known as
Grunge), which suppresses transcription of the PCP effector gene four-jointed. e | The atypical Tyr kinase RYK is another
alternative Wnt receptor, which has been shown to function through SRC in axon guidance.

and lymphoid enhancer-binding factor 1 (leF1) family Redirection rather than inhibition
of DnA-binding proteins. Together, these complexes act The treatment of cells with inhibitors of GSK3 results
to induce expression of a cohort of genes including MYC in activation of the canonical pathway, as does genetic
that influence proliferation and apoptosis (reviewed inactivation of this protein kinase, as long as both mam-
in REF. 9). malian GSK3 genes (GSK3A and GSK3B) are silenced14.
The canonical pathway acts in a manner akin to a sta- Genetic inactivation of three of the four GSK3 alleles
tionary car, in which the accelerator is depressed but the does not result in constitutive Wnt pathway activation.
handbrake is also engaged, resulting in no net movement, This is because only a small fraction of the available
albeit at the expense of energy. Wnt ligands have no effect GSK3 molecules (~5%) is physically associated with
on the accelerator but act by relieving the brake. one of the destruction complex. Hence, even with only 25% of
the target genes induced in response to Wnt is axin 2, residual GSK3 molecules in a cell, this kinase saturates
which encodes a protein related to the axin 1 scaffold10. the even lower concentration of the available destruc-
Axin 1 is the least abundant molecule in the destruc- tion complex (as dictated by axin 1 levels11). only the
tion complex and sets the limit on the number of such GSK3 molecules bound to the complex are relevant for
complexes in a cell11, although modulation of APC levels Wnt signalling, and this represents one mechanism that
can also affect the capacity of the pathway 12. Induction leads to signal insulation (see below). As inhibition of
of the axin 2 gene therefore results in an increase in the GSK3 induces canonical signalling and Wnt inactivates
capacity of a cell to process β-catenin for degradation, the destruction complex, then, ipso facto, Wnt must
reducing the activity of the pathway. This feedback loop inactivate GSK3. This supposition is also consonant with
causes an inherent limit to the duration of Wnt signal- the means by which GSK3 is regulated by other pathways
ling. Mutations associated with various cancers (includ- such as mitogens and growth factors that act through
ing most colorectal cancers) inactivate components of phosphatidylinositol 3-kinase (PI3K) or cAMP. These
the destruction complex (APC or axin 1) or remove the pathways induce phosphorylation of Ser21 and Ser9
sites of phosphorylation on β-catenin, effectively cut- on the amino-terminal domains of GSK3α and GSK3β,
ting the brake line (reviewed in REF. 13). Indeed, levels of respectively, which reduces the activity of the kinases15.
β-catenin in such tumours are often far higher than those However, several groups have shown that Wnt disen-
seen during natural activation of the pathway, indicating gages the destruction complex, not through inactivation
the effectiveness of the restraints imposed on β-catenin. of its components, but through a ‘bait and switch’-like

406 | june 2010 | VoluMe 11 [Link]/reviews/molcellbio

© 2010 Macmillan Publishers Limited. All rights reserved

 focuS on SIgnal IntEg
R ERV
atI EI W
onS

mechanism14. Binding of Wnt to the Frizzled–lRP5 phosphorylation of its n-terminal domain, the effect
or lRP6 receptor complex induces a conformational on β-catenin phosphorylation will be marginal and it
change in lRP5 or lRP6, such that it becomes an will not accumulate. It is also possible that participation
attractive substrate for GSK3 and CK1 (REFS 16,17). in the destruction complex physically shields GSK3
Phosphorylation of the co-receptor by these kinases cre- from the protein kinases that phosphorylate and inacti-
ates a high-affinity binding site for axin 1 and this leads vate most of the GSK3 molecules that are not associated
to dissolution of the destruction complex, although the with the complex.
precise mechanism remains unclear 18. Therefore, Wnt There is also genetic evidence in D. melanogaster
increases lRP5 or lRP6 phosphorylation by GSK3 and that the dual roles of β-catenin in cell adhesion and
CK1, which leads to lower β-catenin phosphorylation. nuclear signalling are insulated. Certain Armadillo (the
Importantly, Wnt signalling does not change the activ- fly homologue of β-catenin) mutants have been isolated
ity state of these two protein kinases, rather it redirects that affect only cell adhesion or transcriptional proper-
their attention to a distinct substrate or substrates (lRP5 ties (as assessed by tissue integrity and segmental polar-
and/or lRP6). A dual role in Wnt signalling has also ity defects)22. The advantages of this duality of protein
been proposed for APC in D. melanogaster, whereby function may relate to higher-order coordination, a
this tumour suppressor not only promotes β-catenin possibility we return to below.
processing in the absence of a Wnt signal but, in the
presence of Wnt, stimulates Axin degradation, hence Non-canonical Wnt signalling
accelerating the signal19. Despite the complexities, the canonical Wnt pathway
is still far better understood than the non-canonical
Insulating the wiring Wnt pathway (FIG. 2). originally, the non-canonical Wnt
The proteins that comprise the canonical Wnt pathway pathway was synonymous for the planar cell polarity
are remarkable in that several also have important roles (PCP) pathway — a pathway that regulates tissue
in other cellular functions (for example, β-catenin has a morphogenesis and the synchronous polarity of sheets
dual role in cell adhesion and signalling, APC has a dual of cells (reviewed in REF. 23). This was based on the dis-
role in protein degradation and microtubule organiza- covery in D. melanogaster that mutations in Frizzled
tion, and GSK3 and CK1 are pleiotropic). Teleologically, and Dishevelled, as well as the other proteins encoded
it makes little sense for such an essential and influential by ‘core PCP genes’ (Van Gogh, Flamingo (also known
pathway to be comprised of elements that seem to be as Starry night), Diego and Prickle) disrupt tissue
borrowed from the inventory of other cellular processes. organization in an Armadillo-independent manner.
There are important questions raised by this organiza- Given the presence of Frizzled and Dishevelled in the
tional strategy. First, are there effective mechanisms to PCP pathway, it was long thought that a non-canonical
insulate the shared components such that the cell can Wnt would activate Frizzled to provide spatial informa-
maintain signal discrimination? If so, are there advan- tion to the PCP pathway. However, extensive genetic
tages in such organization compared to having entirely analysis failed to identify any PCP function for the
distinct molecules for each pathway? D. melanogaster Wnt ligands. loss of all five of the Wnt
As described for GSK3 above, only the small frac- ligands expressed in the wing 24 and removal of most
tion of molecules of GSK3 that are physically associ- Wnt ligands in the abdomen failed to disrupt their
ated with the destruction complex are relevant to Wnt PCP25, showing that Wnt ligands do not control PCP
signalling. There is a considerable amount of literature in these tissues. Therefore, rather than calling Frizzled-
that associates, for example, polypeptide growth factor- mediated PCP signalling the non-canonical Wnt path-
induced inactivation of GSK3 through n-terminal way, it is probably more appropriate to refer to it as the
domain (Ser21 or Ser9) phosphorylation with direct Frizzled–PCP pathway.
activation of the Wnt pathway, as judged by β-catenin A search for alternative upstream regulators of the
stabilization. Despite this literature, cellular stimuli that Frizzled–PCP pathway in D. melanogaster identified
lead to inhibition of GSK3 through n-terminal domain Fat and Dachsous as important PCP regulators in all
phosphorylation (for example, mitogens that activate tissues (FIG. 2; FIG. 3). Fat and Dachsous and their mam-
AKT1 (also known as PKB)) do not cause stabilization malian homologues are large, atypical cadherins that
of β-catenin. Moreover, mutants of GSK3 that cannot control PCP in D. melanogaster and mammals26–29. In
be inactivated by n-terminal domain phosphorylation D. melanogaster, Fat and Dachsous also function in
are fully competent in mediating Wnt activation of growth regulation. Although they were first proposed
β-catenin14,20,21. These findings indicate that the small to act upstream of the Frizzled core PCP genes28, more
number of GSK3 molecules bound to axin 1 is insulated recent studies have suggested that they may act in a par-
from the bulk of the protein kinase that is not associated allel pathway for PCP control30 (reviewed in REF. 31).
with the destruction complex. The isolation mechanism Current models suggest that Dachsous binds to Fat to
Planar cell polarity responsible for this signalling selectivity is not com- inhibit its activity. The cytoplasmic domain of Fat binds
A mechanism of cellular pletely understood. However, axin 1 acts to chaperone a transcriptional co-repressor, Atrophin (also known
organization, distinct from both β-catenin and GSK3, rendering the phosphoryl- as Grunge)32, which regulates PCP target genes, such
apical–basal polarity, that is
important in providing a higher
ation reaction, in essence, first-order, reducing sensi- as four-jointed. Fat activity is controlled by gradients of
order of arrangements in flat tivity to the specific activity of GSK3. Hence, even if Dachsous and Four-jointed33,34. How Fat activity then
sheets of epithelial cells. the axin 1-associated GSK3 is partially inactivated by regulates tissue organization is still unclear.

nATuRe ReVIeWS | Molecular cell Biology VoluMe 11 | june 2010 | 407

© 2010 Macmillan Publishers Limited. All rights reserved

REVIEWS

canonical pathway (that is, through β-catenin), whereas

WnT5A, WnT4 and WnT11 have been classed as
non-canonical Wnt ligands that do not affect β-catenin
Dachsous levels but cause convergent extension movements
Plasma during development. A plethora of downstream media-
Fat membrane
tors respond to these non-canonical Wnt ligands
(reviewed in REF. 38), suggesting that there are in fact
Atrophin Merlin
CKIε multiple non-canonical Wnt pathways.
Expanded one class of downstream mediators have been called
PCP Salvador the Wnt–Ca2+ pathway and are associated with the acti-
Hippo vation of a phospholipase C (PlC)-mediated increase in
intracellular Ca2+ levels and the activation of calmodulin-
Dachs MATS dependent protein kinase II (CAMKII), protein kinase C
Warts
(PKC) and the nuclear factor of activated T cells (nFAT)
transcription factor. Another set of responses involves
Nucleus the Rac, Rho and Rap small G proteins, and is associ-
Yorkie
ated with activation of the jun n-terminal kinase (jnK)
? Scalloped Cyclin E, bantam and diap1 pathway and alterations in the cytoskeleton. However,
the boundaries between these classes are still unclear and
need further clarification. A larger question that is still
Figure 3 | The Fat growth and PcP pathway. The large
Nature Reviews | Molecular Cell Biology
cadherin Dachsous acts as an inhibitory ligand for the unanswered is why it is generally accepted that Wnt pro-
large cadherin Fat. Fat functions to regulate planar cell teins regulate PCP in vertebrates, despite careful genetic
polarity (PCP) signalling by recruiting the transcriptional analysis in D. melanogaster excluding a role for Wnts in
co-repressor Atrophin (also known as Grunge). Atrophin PCP. one possibility, of course, is that the use of Wnts
represses transcription of the PCP gene four-jointed. in PCP is a vertebrate-specific adaptation. However, an
The identity of the transcription factor that Atrophin binds alternative possibility is that the role of Wnts in PCP
to regulate four-jointed transcription is at present unclear. in higher organisms may also be indirect, as has been
Fat also regulates growth by controlling the Hippo kinase shown in D. melanogaster, in which Wingless regulates
pathway. Current models suggest that Fat regulates Hippo the expression of PCP regulators such as Dachsous and
activity through either the FERM domain protein Expanded
Four-jointed. In any case, more studies are needed to
or the small myosin-like protein Dachs. Dachs has been
shown to regulate the stability of Warts; in turn, Warts define the exact role of the non-canonical Wnts in PCP.
phosphorylates Yorkie, leading to its export from the
nucleus. In the absence of Fat or the other Hippo Crosstalk between various Wnt pathways
components, Yorkie promotes the transcription of A notable characteristic of many forms of non-canonical
Cyclin E, the microRNA bantam and the anti-apoptotic Wnt signalling seems to be an inhibition of canonical
gene Drosophila inhibitor of apoptosis 1 (diap1; also known signalling. early studies showed that overexpressed
as thread). Salvador and MATS (mob as tumour-suppressor WnT5A can block the stabilization of β-catenin induced
protein 1) are adaptor proteins essential for Hippo and by WnT1 (REF. 39). expression of WnT5A can also acti-
Warts activity. Merlin is another FERM domain protein that vate neMo-like kinase (nlK), which phosphorylates
functions in parallel to Expanded to control activity of the
TCF transcription factors, thus inhibiting canonical
Hippo pathway.
Wnt signalling 40,41. In fact, it seems that simply inhibit-
ing canonical signalling can, in some conditions, lead to
The Frizzled–PCP and Fat–Dachsous pathways are PCP effects. For example, overexpression of the canonical
conserved in vertebrates, in which they function to Wnt feedback inhibitor, the cytoplasmic eF-hand protein
regulate tissue organization, most notably the polar- naked, produces strong PCP effects in D. melanogaster 42
ized movements of convergent extension during develop- and convergent extension defects in X. laevis43.
ment, the orientation of hair cells in the inner ear and Recent work has identified additional pathways
neural tube closure (reviewed in REF. 35). In a surprising that mediate crosstalk between the non-canonical and
twist, the loss of function of several Wnt ligands that canonical Wnt pathways. lee and co-workers44,45 found
are largely non-canonical Wnt ligands gives rise to simi- that stimulation of WnT5A leads to increased PKC
Convergent extension lar PCP phenotypes in vertebrates. This has led to the activity. Activated PKC phosphorylates RoRα (retinoic
A non-mitotic developmental proposal that some Wnts may function in a vertebrate acid-related orphan nuclear receptor α), which in turn
process that involves non-canonical Wnt–PCP pathway. It is still unclear how binds with high affinity to β-catenin at promoters and
elongation in one axis of a
the non-canonical Wnt–PCP pathways relate to the core represses transcription, thus inhibiting β-catenin-
band (or bands) of cells, usually
resulting in the coverage of a PCP pathway or the Fat–Dachsous–PCP cassette. dependent activity. This obviously opens up many
structure. Wnt ligands have been traditionally subdivided into avenues for modifying β-catenin activity, through the
canonical and non-canonical groups, based on their abil- many pathways that can activate PKC. In yet another
Neural tube ity to induce a secondary body axis in X. laevis and to mechanism for crosstalk between canonical and non-
The precursor structure of the
vertebrate nervous system that
transform C57MG epithelial cells, indicating β-catenin canonical Wnts, Sato et al.46 have recently shown that
develops into the brain and signalling 36,37. WnT1, WnT3A, WnT8A and WnT8B WnT5A can also suppress canonical Wnt signalling by
spinal cord. have therefore been suggested to act through the competing for Frizzled receptors and also by inducing

408 | june 2010 | VoluMe 11 [Link]/reviews/molcellbio

© 2010 Macmillan Publishers Limited. All rights reserved

 focuS on SIgnal IntEg
R ERV
atI EI W
onS

the internalization of frizzled 2. Recently, WnT5A has Drosophila melanogaster Mammals

been shown to use Dishevelled and APC to modulate its
effects on focal adhesion dynamics during cell move-
ment47. APC has long been recognized to have important Fat FAT4?
roles in microtubule organization and dynamics48. This
provides an apt example of how the lines between the Plasma
canonical and non-canonical elements can be blurred. membrane
There is still a great deal of debate as to how non-
canonical Wnt ligands function. For example, WnT5A
has been shown to regulate Ca2+, Rac, Src kinases, Expanded FRMD6
CAMKII, jnK and β-catenin44–46,49–52. Part of the answer Kibra Kibra
Merlin NF2
to this complexity may lie in the cellular context and the
specific Frizzled receptors these cells express. For exam- Salvador WW45
ple, although WnT5A is needed for convergent extension Hippo MST1
processes in embryonic development in a β-catenin- or MST2
independent manner, co-expression of frizzled 4 and
lRP5 can convert the WnT5A response to stabilizing MATS Mob
Warts LATS1
β-catenin, as can expression of frizzled 5 (REFS 52,53). or LATS2

Non-Frizzled Wnt receptors YAP or TAZ

Although it is clear that some of the diversity of Wnt Yorkie
signalling can be explained by the Frizzled receptor
complement present on different cells, excitingly, recent
data have suggested that some of the answer may also
lie in newly described non-Frizzled receptors for some ↑ Proliferation, ↓ apoptosis and
↑ expression of Cyclin E, diap1 and bantam
Wnt ligands: the RYK and Ror Tyr kinases51 (reviewed
in REFS 38,45,54). Figure 4 | The core Hippo pathway. The core Hippo
Ror and RYK kinases have been shown to function pathway was defined inNature Reviews
Drosophila | Molecular as
melanogaster Cell
a Biology
as Wnt receptors in various systems. Ror kinases have cassette composed of the sterile 20-like kinase Hippo, which
a CRD domain, similar to that of Frizzled, and WnT5A forms a complex with the WW domain adaptor protein
binds to the CRD domain of RoR2 with high affinity 52. Salvador. Hippo phosphorylates and activates the DBF family
overexpressed RoR2 leads to convergent extension kinase Warts, which forms a complex with the adaptor
defects, suggesting that it functions in a PCP pathway protein MATS (mob as tumour suppressor protein 1). Warts
phosphorylates the transcriptional co-activator Yorkie.
like WnT5A51. The addition of WnT5A to purified
Phosphorylated Yorkie is bound by 14-3-3 proteins and
RoR2 (REF. 52) can inhibit or activate β-catenin signalling, excluded from the nucleus. When in the nucleus, Yorkie
depending on the co-receptor context. Aside from the promotes the transcription of growth-promoting and
presence or absence of RYK and RoR2, other factors can apoptosis-inhibiting genes, such as Cyclin E and Drosophilia
bias WnT5A to act through a canonical or non-canonical inhibitor of apoptosis 1 (diap1; also known as thread).
pathway. For example, a secreted collagen protein, colla- Upstream of Hippo lies the cadherin Fat and the FERM
gen triple helix repeat-containing protein 1 (CTHRC1), domain proteins Expanded and Merlin. Kibra is a WW
can potentiate the Wnt–RoR2–Frizzled interaction, domain protein that promotes Expanded–Merlin
enhancing PCP signalling of WnT5A55. Conversely, there interactions and enhances Hippo pathway activity. The
are also non-Wnt ligands for the Frizzled receptors, such analogous pathway in mammals is outlined on the right.
YAP (also known as YAP65) and transcriptional co-activator
as norrin56, that can activate the canonical pathway. This
with PDZ-binding motif (TAZ; also known as WWTR1) are
exciting finding suggests that there are many other, as yet homologous to Yorkie, MST1 and MST2 are homologous to
unidentified, ligands that can affect the selection of the Hippo, LATS1 and LATS2 are homologous to Warts, Mob
canonical and non-canonical pathways. proteins are homologous to MATS, WW45 is homologous to
Salvador, neurofibromin 2 (NF2) is homologous to Merlin and
Growth control by Fat PCP regulators FERM domain-containing protein 6 (FRMD6) is homologous
Surprisingly, the large cadherins Fat and Dachsous are not to Expanded. The vertebrate homologue of Fat is FAT4,
dedicated solely to the control of PCP but have another, but this protein has not yet been shown to function in the
CRD domain apparently independent, role as important growth regu- vertebrate Hippo pathway.
A Cys-rich region located on lators. loss of fat leads to dramatic tissue overgrowth,
the extracellular portion of
leading to its original classification as a D. melanogaster
GPCR receptors that is
essential for binding ligands. tumour suppressor gene. Subsequently, genetic studies conserved in vertebrates. The Sterile 20-like kinase Hippo
indicated that fat regulates growth through a conserved forms a complex with the WW-repeat scaffolding protein
WW repeat kinase cassette, known as the Hippo kinase pathway Salvador to phosphorylate and activate the DBF family
A protein motif that binds (reviewed in REF. 31). kinase Warts. Activated Warts, in association with adap-
certain polyPro peptides
and/or phosphorylated
The Hippo pathway was first discovered and charac- tor protein MATS (mob as tumour-suppressor protein 1),
peptides (usually phosphoSer terized in D. melanogaster (reviewed in REF. 57). The core phosphorylates and inhibits Yorkie. Yorkie is a non-DnA-
or phosphoThr peptides). of the Hippo pathway is relatively simple (FIG. 4) and well binding co-activator that enhances the transcription of

nATuRe ReVIeWS | Molecular cell Biology VoluMe 11 | june 2010 | 409

© 2010 Macmillan Publishers Limited. All rights reserved

REVIEWS

Contact inhibition genes that promote cell proliferation (such as Cyclin E and a cytoplasmic protein with two WW domains and a C2
The growth-suppressive effect the microRnA bantam) and prevent apoptosis (such as domain that is essential for maximal Hippo pathway acti-
that occurs when epithelial Drosophila inhibitor of apoptosis 1 (diap1; also known as vation71–73. In addition, Kibra can bind Hippo, promoting
cells are in physical contact. thread)). The phosphorylation of Yorkie by Warts inhibits its association with membranes, which may contribute
Yorkie nuclear localization, preventing it from acting on to activation of the pathway. expanded can also regu-
its growth-promoting and apoptosis-inhibiting targets. late Yorkie in a Hippo-independent manner, sequester-
Yorkie functions with the transcription factor Scalloped ing Yorkie at apical junctions57,74. The small myosin-like
in the regulation of growth control58–60. protein Dachs functions downstream of Fat to modulate
Genetic studies clearly indicate that Fat regulates the Hippo pathway activity, possibly by controlling the stabil-
Hippo pathway31,61–64, but these and other studies also sug- ity of Warts62. More work is clearly needed to understand
gest that there are other, as yet unidentified, cell surface how Fat and other cell surface receptors regulate growth
receptors65–67. How the Hippo pathway connects to the through the Hippo pathway.
Fat and Dachsous cadherins and other cell surface recep- The study of Hippo pathway components in vertebrates
tors is still unclear and is an area of much research. Some has revealed conservation of the general organization seen
studies suggest that Fat regulates Hippo activity through in D. melanogaster, although, as is often the case, there
the FeRM domain protein expanded, which functions are multiple family members in the vertebrate pathway
redundantly with another FeRM domain protein Merlin (reviewed in REF. 57). The vertebrate orthologues of Hippo
(also known as neurofibromin 2 (nF2) in mammals) to are MST1 (also known as STK4) and MST2 (also known
control Hippo activity 68. Transcription of fat, dachsous, as STK3). like Hippo, MST1 and MST2 phosphorylate
four-jointed, expanded and Merlin are enhanced by loss of and activate the Warts orthologues, lATS1 and lATS2.
Hippo pathway activity, providing negative feedback and lATS1 and lATS2 phosphorylate the two Yorkie homol-
also allowing changes in Hippo activity to be transmitted ogues, YAP (also known as YAP65) and transcriptional
to adjacent cells. Fat also binds CK1ε, which promotes co-activator with PDZ-binding motif (TAZ; also known
Hippo pathway activity 69,70. There are multiple interac- as WWTR1), leading to the cytoplasmic retention of these
tions between the upstream components. expanded can growth-promoting transcriptional co-activators. YAP and
bind Hippo, and Merlin can bind Salvador71. Interactions TAZ partner with four TeA domain transcription factors
between expanded and Merlin are promoted by Kibra, (TeAD1–TeAD4), Runx, eRBB4 and p73. Interestingly,
loss of MST1 and MST2, loss of WW45 (Salvador homol-
ogue 1), lATS1 and lATS2, and amplification of YAP, are
Cell contact
involved in various cancers. In addition, the long-studied
? FRMD6
but poorly understood phenomena of contact inhibition
NF2 were shown to require YAP and the Hippo pathway.
There are four Fat-like genes in mammals, but it is not
WW45
MST1
yet known if they regulate Hippo activity, or if there are
or MST2 other regulators in higher organisms.
Mob
LATS1
Bountiful crosstalk in the Hippo pathway
or LATS2 As we learn more about the Hippo pathway, the initially
↑ EGFR ↑ BMP signalling
simple linear pathway described from genetic studies
YAP or TAZ ↑ SHH signalling begins to unravel into a much more complicated and
↑ Amphiregulin Binds Dishevelled; ↓ Wnt signalling
intertwined picture, revealing many points of crosstalk
with multiple signalling pathways (FIG. 5). Work in D. mel-
Cell cycle; anti-apoptosis
anogaster has shown that Wingless and Decapentaplegic
(DPP) regulate expression of Dachsous and Four-
Figure 5 | crosstalk in the Hippo pathway. Recent studies have uncovered jointed28,75, suggesting that activity of Wingless and DPP
unsuspected crosstalk between the Hippo pathway and other
Nature growth
Reviews regulators.
| Molecular CellCell
Biology may affect Hippo signalling. Conversely, expression of
contact is still detected in mouse embryonic fibroblasts (MEFs) lacking both MST1 (also
known as STK4) and MST2 (also known as STK3; homologues of Drosophila melanogaster
Wingless and Serrate (a notch ligand) are regulated by
Hippo), indicating an Mts-independent input (possibly a kinase) that can phosphorylate Yorkie62. Yorkie also regulates diverse morphogen signal-
and activate LATS1 and LATS2 (homologues of D. melanogaster Warts). LATS1 or LATS2 ling by controlling expression of the heparan sulphate
then inhibit YAP (also known as YAP65) or transcriptional co-activator with PDZ-binding proteoglycans Dally and Dally-like76. Interestingly, the
motif (TAZ; also known as WWTR1), which are homologues of D. melanogaster Yorkie, to microRnA bantam is not only regulated by the Hippo
control growth. YAP65 can function as a transcriptional co-activator of Smads, enhancing pathway, but also by notch, Wingless and insulin–target
bone morphogenetic protein (BMP) signalling. YAP can also stimulate the transcription of of rapamycin (ToR) signalling 77. The complexity of
Gli proteins, enhancing Sonic hedgehog (SHH) signalling. AKT1 can inhibit MST1 kinase, bantam control also highlights differential transcrip-
providing another avenue for the growth promotion and anti-apoptotic activity of tion factor interactions with Yorkie in growth control.
phosphoinositide 3-kinase (PI3K)–AKT1 signalling pathways. TAZ can bind Dishevelled
Whereas the TeA domain family transcription factor
proteins, inhibiting β-catenin signalling in the cytoplasm. When TAZ translocates to the
nucleus, it frees Dishevelled to activate β-catenin, enhancing Wnt signalling. Nuclear YAP
Scalloped functions with Yorkie to regulate growth in the
also regulates the expression of amphiregulin, an epidermal growth factor receptor wing, recent studies have shown that the TAle family
(EGFR) ligand. Therefore, loss of Mts or Lats in one cell can drive proliferation in that cell transcription factor Homothorax, together with the zinc
while simultaneously activating the proliferation of nearby cells through the EGFR finger transcription factor Teashirt, function with Yorkie
pathway. FRMD6, FERM domain-containing protein 6; NF2, neurofibromin 2. to regulate growth in the D. melanogaster eye disc78.

410 | june 2010 | VoluMe 11 [Link]/reviews/molcellbio

© 2010 Macmillan Publishers Limited. All rights reserved

 focuS on SIgnal IntEg
R ERV
atI EI W
onS

Although the transcriptional co-activator YAP was The benefits of double duty
thought to be committed to the Hippo pathway, sev- even if cells are capable of effectively segregating signals
eral exciting studies have uncovered roles for YAP in that seem to pass through common elements, why take the
regulating the signalling strength of numerous other risk? one can only speculate, but part of the answer may
well-characterized signalling pathways. A recent study stem from the fact that the number of signalling pathways
revealed that YAP can act as a transcriptional activa- discovered to date is remarkably limited, given biological
tor in the bone morphogenic protein (BMP) pathway complexity. Although there is tremendous diversity at the
with SMAD1 and is needed for the BMP suppression of level of ligands and receptors, once the receptor Tyr kinase
neural differentiation of mouse embryonic stem cells79. or GPCR has been engaged, there are a much smaller
Therefore, when Hippo pathway activity is high, YAP is number of options available for the transmission of sig-
phosphorylated and excluded from the nucleus, reducing nals in the cell. Moreover, despite the practical scientific
activity of the BMP pathway. This provides an integra- need to probe cellular responses to individual signals, this
tion of Hippo and BMP signalling at the transcriptional scenario is hardly physiological. Cells are exposed to a ver-
level. This integration is conserved: in D. melanogaster, itable cacophony of signals that is constantly changing.
Yorkie is needed for maximal signalling by the BMP In this reality, it is not so much that signals turn on or off,
orthologue DPP79. Recent studies have revealed that rather how they interact and integrate in the context of
YAP also provides an integration point in Hedgehog any given cell. Perhaps this is a clue to the preponderance
signalling 80, acting as a transcriptional co-activator for of shared components in signalling? Yet, given the estab-
Gli transcription factors. lished roles of Fat and Dachsous in regulating the Hippo
There is also integration between Hippo signalling pathway and separately controlling PCP, one has to won-
and Wnt signalling. The Yorkie homologue TAZ binds to der if this separation is only apparent. It seems wasteful to
Dishevelled proteins in the cytoplasm81. This interaction position Fat as a key mediator of tissue growth and tissue
inhibits Dishevelled function, leading to reduced β-catenin patterning and not to take advantage of that position to
signalling on Wnt stimulation. Conversely, loss of TAZ integrate those signals.
enhances β-catenin signalling. This crosstalk is conserved evidence for the advantages of shared molecules is subtle,
in D. melanogaster, in which loss of Yorkie enhances the but it may be argued that it is obscured by the insulation
expression of Wingless target genes in vivo81. Similarly, systems discussed above — systems that are essential to
some of the pro-growth, anti-apoptotic functions of the avoid the chaos that would occur if there was indiscrimi-
PI3K–AKT1 pathway were recently found to be due to nate cross-contamination of information. Despite this,
phosphorylation of MST1, providing another point of there are examples that hint at the existence of hidden
integration between known growth factor pathways underpinnings of signal transduction. The most obvious
and Hippo activity 82. theoretical advantage of common components is cross-
until recently, the pro-growth and anti-apoptotic func- talk and coordination. There are many well-documented
tions of YAP were thought to be purely cell autonomous, biochemical links between signalling pathways, such as
acting on cell cycle regulators and inhibitors of apoptosis. the transcriptional product of one signal being a negative
However, this has changed, as a recent study identified regulatory protein of another (for example, the mitogen-
the epidermal growth factor receptor (eGFR) ligand activated protein kinase phosphatases that are induced at
amphiregulin as a transcriptional target of the Hippo the transcriptional level by stresses that act through jnK
pathway in mammals83 and showed that activation of and p38 to dephosphorylate and inactive the extracellular
YAP leads to proliferation of neighbouring cells in an signal-regulated kinases85). The Wnt pathway offers simi-
eGFR-dependent manner. This study also showed that lar links, such as the role of tankyrase-mediated poly-ADP
Yorkie interacts with eGFR in D. melanogaster, suggest- ribosylation in destabilizing axin 1, thereby providing a new
ing that alterations in Hippo pathway activity will have input for regulators of tankyrase on Wnt pathway sensi-
non-autonomous effects in both normal development and tivity 86. These links serve as direct and rapid coordinators
in cancers. that tune the activity of signals at an intensive and detailed
Finally, there are hints that there are as yet unidenti- level. The importance of these mechanisms has been par-
fied complexities even at the core of the Hippo pathway. ticularly recognized through systems biology efforts to
A recent study 84 examined mice lacking both MST1 and model short-term pathway responses (in minutes) to spe-
MST2. As expected, targeted loss of MST1 and MST2 in cific triggers, such as tumour necrosis factor-α87. There are
the liver resulted in massive overgrowth and eventual also advantages in terms of multi-purposing. The destruc-
hepatocellular carcinoma. Surprisingly, however, they tion complex, for example, processes not only β-catenin
found that even in the absence of MST1 and MST2, con- for degradation but also [Link] mechanism of MYC
tact inhibition of mouse embryonic fibroblasts (MeFs) degradation is, seemingly, not regulated by Wnt but,
occurs normally. In these MeFs, phosphorylation of given that MYC is an important transcriptional target of
lATS1, lATS2 and YAP occurs normally; therefore, β-catenin, the common machinery at the least provides
ADP ribosylation there must be lATS1 and lATS2 activators other than opportunity for coordination.
A post-translational MST1 and MST2 in MeFs. Similarly, when they examined less obvious inter-pathway influences occur on
modification involving the the liver of MST1 and MST2 double mutants, YAP phos- longer timescales (hours rather than seconds or minutes).
transfer of an ADP ribose
moiety from nicotinamide
phorylation occurred, but in the absence of lATS1 and This evidence was derived initially from developmental
adenine dinucleotide to Arg, lATS2 activation. The identity of the lats activator biology studies of model organisms that noted recipro-
Glu or Asp acid side chains. and YAP kinase are currently unknown. city of phenotypes in processes governing fundamental

nATuRe ReVIeWS | Molecular cell Biology VoluMe 11 | june 2010 | 411

© 2010 Macmillan Publishers Limited. All rights reserved

REVIEWS

embryonic patterning processes such as segmental polar- and other pathways, the result being an effective block
ity, in which, for example, mutants in the Hedgehog to differentiation (but not cellular viability) as visualized
pathway caused effects on Wingless-regulated pro- in embryonic stem cells and neuronal precursors14,93.
cesses, and vice versa. These influences probably reflect However, inactivation of 75% of this promiscuous regula-
programming events that attempt to compensate for tor results in minimal, if any, developmental effects. Such
reduced activities in certain pathways through sporadic a ‘non-outcome’ must reflect a considerable correction of
loss of cells and form the basis of biological robustness. signalling receptivity in the background. Although such
Although a loss-of-function mutation that permeates changes in GSK3 are unlikely to be physiologically rel-
all cells will usually defeat such compensatory systems, evant, this common protein may act as a tether, helping
development succeeds in the face of frequent accidents to align and adjust pathway activities.
as a consequence of intrinsic recalibration of cellular
sensitivities to pathways. using the canonical Wnt path- Concluding remarks
way as an example, the absolute levels of β-catenin seem The fact that key regulatory pathways often share com-
less important than fold differences89, although this may ponents complicates their analysis and the understand-
be tissue dependent as suggested by an allelic series of ing of their functions. use of small molecule inhibitors or
APC mutations90,91. This phenomenon has advantages in RnA interference will break down the natural barriers and
organismal adaptation and allows the resetting of global mechanisms of insulation, resulting in these tools having
cellular responses to a signal. The mechanisms by which an impact on multiple systems. Development of therapeu-
this is achieved are being uncovered. tic strategies should take the linkages into consideration in
The simple axin 2 negative feedback loop is one exam- estimating therapeutic windows. Although a small mol-
ple of resetting, but it is selective for the canonical Wnt ecule might show exquisite selectivity for its target protein,
pathway. More complex loops, whereby a regulator con- that specificity is lost if its target has a natural promiscu-
trols both positive and negative elements of a system, can ity that is normally held in check, for example through
lead to changes in multiple parameters including signal sequestration. Targeting selective insulating systems may
amplitude, duration and sensitivity92. Such wider impacts provide new opportunities such that only a specific role
may be mediated by regulators like GSK3, which has roles of a target might be affected, leaving its other functions
in not only the Wnt, growth factor and insulin signal- intact. Finally, the core adaptability of signalling mol-
ling pathways but also regulates the Hedgehog and notch ecules may provide natural buffering against undesired
pathways. Complete inactivation of GSK3 leads to activa- effects if the differential sensitivities of pathways are taken
tion and dysregulation of β-catenin, notch, Hedgehog into account.

1. Cohen, P. The role of protein phosphorylation in neural 11. Lee, E., Salic, A., Kruger, R., Heinrich, R. & 21. Ng, S. S. et al. Phosphatidylinositol 3-kinase signaling
and hormonal control of cellular activity. Nature 296, Kirschner, M. W. The roles of APC and Axin derived does not activate the Wnt cascade. J. Biol. Chem. 284,
613–620 (1982). from experimental and theoretical analysis of the 35308–35313 (2009).
2. van Amerongen, R. & Nusse, R. Towards an integrated Wnt pathway. PLoS Biol. 1, E10 (2003). Provides compelling evidence to debunk the link
view of Wnt signaling in development. Development 12. Benchabane, H., Hughes, E. G., Takacs, C. M., between PI3K signals and Wnt responses,
136, 3205–3214 (2009). Baird, J. R. & Ahmed, Y. Adenomatous polyposis coli is demonstrating signal authenticity.
3. MacDonald, B. T., Tamai, K. & He, X. Wnt/β-catenin present near the minimal level required for accurate 22. Orsulic, S. & Peifer, M. An in vivo structure-function
signaling: components, mechanisms, and diseases. graded responses to the Wingless morphogen. study of Armadillo, the β-catenin homologue, reveals
Dev. Cell 17, 9–26 (2009). Development 135, 963–971 (2008). both separate and overlapping regions of the protein
References 2 and 3 provide excellent and 13. Polakis, P. The many ways of Wnt in cancer. Curr. Opin. required for cell adhesion and for Wingless signaling.
up‑to‑date reviews of Wnt signalling. Genet. Dev. 17, 45–51 (2007). J. Cell Biol. 134, 1283–300 (1996).
4. Zhao, B., Lei, Q. Y. & Guan, K. L. The Hippo-YAP 14. Doble, B. W., Patel, S., Wood, G. A., Kockeritz, L. K. & 23. Simons, M. & Mlodzik, M. Planar cell polarity
pathway: new connections between regulation of Woodgett, J. R. Functional redundancy of GSK-3α and signaling: from fly development to human disease.
organ size and cancer. Curr. Opin. Cell Biol. 20, GSK-3β in Wnt/β-catenin signaling shown by using an Annu. Rev. Genet. 42, 517–540 (2008).
638–646 (2008). allelic series of embryonic stem cell lines. Dev. Cell 12, 24. Chen, W. S. et al. Asymmetric homotypic interactions
5. Zeng, Q. & Hong, W. The emerging role of the Hippo 957–971 (2007). of the atypical cadherin Flamingo mediate intercellular
pathway in cell contact inhibition, organ size control, Describes the insensitivity to the loss of GSK3 polarity signaling. Cell 133, 1093–1105 (2008).
and cancer development in mammals. Cancer Cell 13, alleles in embryonic stem cells and reveals a 25. Lawrence, P. A., Casal, J. & Struhl, G. Towards a model
188–192 (2008). requirement for GSK3 in differentiation. of the organisation of planar polarity and pattern in
References 4 and 5 provide excellent and 15. Cohen, P. & Frame, S. The renaissance of GSK3. the Drosophila abdomen. Development 129,
up‑to‑date reviews of Hippo signalling. Nature Rev. Mol. Cell Biol. 2, 769–776 (2001). 2749–2760 (2002).
6. Klingensmith, J. & Nusse, R. Signaling by Wingless in 16. Davidson, G. et al. Casein kinase 1γ couples Wnt 26. Casal, J., Struhl, G. & Lawrence, P. A. Developmental
Drosophila. Dev. Biol. 166, 396–414 (1994). receptor activation to cytoplasmic signal transduction. compartments and planar polarity in Drosophila. Curr.
7. Major, M. B. et al. Wilms tumor suppressor WTX Nature 438, 867–872 (2005). Biol. 12, 1189–1198 (2002).
negatively regulates WNT/β-catenin signaling. Science 17. Zeng, X. et al. A dual-kinase mechanism for Wnt 27. Rawls, A. S., Guinto, J. B. & Wolff, T. The cadherins
316, 1043–1046 (2007). co-receptor phosphorylation and activation. Nature Fat and Dachsous regulate dorsal/ventral signaling in
8. Miller, B. W. et al. Application of an integrated physical 438, 873–877 (2005). the Drosophila eye. Curr. Biol. 12, 1021–1026
and functional screening approach to identify 18. Zeng, X. et al. Initiation of Wnt signaling: control of (2002).
inhibitors of the Wnt pathway. Mol. Syst. Biol. 5, 315 Wnt coreceptor LRP6 phosphorylation/activation via 28. Yang, C. H., Axelrod, J. D. & Simon, M. A. Regulation
(2009). Frizzled, Dishevelled and Axin functions. Development of Frizzled by Fat-like cadherins during planar polarity
References 7 and 8 provide excellent examples 135, 367–375 (2008). signaling in the Drosophila compound eye. Cell 108,
of the power of proteomic approaches for the References 16–18 describe the positive roles 675–688 (2002).
unbiased assessment of pathway topology and of two protein kinases in Wnt signalling that 29. Saburi, S. et al. Loss of Fat4 disrupts PCP signaling
interactions. had previously been ascribed negative and oriented cell division and leads to cystic kidney
9. Mosimann, C., Hausmann, G. & Basler, K. β-catenin functions. disease. Nature Genet. 40, 1010–1015 (2008).
hits chromatin: regulation of Wnt target gene 19. Takacs, C. M. et al. Dual positive and negative Provides evidence for conserved roles of
activation. Nature Rev. Mol. Cell Biol. 10, 276–286 regulation of Wingless signaling by Adenomatous D. melanogaster Fat and mammalian FAT4 in PCP
(2009). polyposis coli. Science 319, 333–336 (2008). pathways.
10. Jho, E. H. et al. Wnt/β-catenin/TCF signaling induces 20. McManus, E. J. et al. Role that phosphorylation of 30. Casal, J., Lawrence, P. A. & Struhl, G. Two separate
the transcription of Axin2, a negative regulator of the GSK3 plays in insulin and Wnt signalling defined molecular systems, Dachsous/Fat and Starry night/
signaling pathway. Mol. Cell. Biol. 22, 1172–1183 by knockin analysis. EMBO J. 24, 1571–1583 Frizzled, act independently to confer planar cell
(2002). (2005). polarity. Development 133, 4561–4572 (2006).

412 | june 2010 | VoluMe 11 [Link]/reviews/molcellbio

© 2010 Macmillan Publishers Limited. All rights reserved

 focuS on SIgnal IntEg
R ERV
atI EI W
onS

31. Sopko, R. & McNeill, H. The skinny on Fat: an 54. Green, J. L., Kuntz, S. G. & Sternberg, P. W. 78. Peng, H. W., Slattery, M. & Mann, R. S. Transcription
enormous cadherin that regulates cell adhesion, tissue Ror receptor tyrosine kinases: orphans no more. factor choice in the Hippo signaling pathway:
growth, and planar cell polarity. Curr. Opin. Cell Biol. Trends Cell Biol. 18, 536–544 (2008). Homothorax and Yorkie regulation of the microRNA
21, 717–723 (2009). 55. Yamamoto, S. et al. Cthrc1 selectively activates the bantam in the progenitor domain of the Drosophila
32. Fanto, M. et al. The tumor-suppressor and cell planar cell polarity pathway of Wnt signaling by eye imaginal disc. Genes Dev. 23, 2307–2319
adhesion molecule Fat controls planar polarity via stabilizing the Wnt-receptor complex. Dev. Cell 15, (2009).
physical interactions with Atrophin, a transcriptional 23–36 (2008). 79. Alarcon, C. et al. Nuclear CDKs drive Smad
co-repressor. Development 130, 763–774 (2003). 56. Clevers, H. Wnt signaling: Ig-norrin the dogma. transcriptional activation and turnover in BMP and
33. Matakatsu, H. & Blair, S. S. Interactions between Fat Curr. Biol. 14, R436–R437 (2004). TGF-β pathways. Cell 139, 757–769 (2009).
and Dachsous and the regulation of planar cell 57. Badouel, C. et al. The FERM-domain protein 80. Fernandez, L. A. et al. YAP1 is amplified and
polarity in the Drosophila wing. Development 131, Expanded regulates Hippo pathway activity via direct up-regulated in Hedgehog-associated
3785–3794 (2004). interactions with the transcriptional activator Yorkie. medulloblastomas and mediates Sonic hedgehog-
34. Simon, M. A. Planar cell polarity in the Drosophila Dev. Cell 16, 411–420 (2009). driven neural precursor proliferation. Genes Dev. 23,
eye is directed by graded Four-jointed and 58. Goulev, Y. et al. Scalloped interacts with Yorkie, the 2729–2741 (2009).
Dachsous expression. Development 131, 6175–6184 nuclear effector of the Hippo tumor-suppressor 81. Varelas, X. et al. The Hippo pathway regulates
(2004). pathway in Drosophila. Curr. Biol. 18, 435–441 Wnt/β-catenin signalling. Dev. Cell 18, 579–591
35. Yin, C., Ciruna, B. & Solnica-Krezel, L. Convergence (2008). (2010).
and extension movements during vertebrate 59. Wu, S., Liu, Y., Zheng, Y., Dong, J. & Pan, D. The 82. Yuan, Z. et al. Phosphoinositide 3-kinase/Akt inhibits
gastrulation. Curr. Top. Dev. Biol. 89, 163–192 TEAD/TEF family protein Scalloped mediates MST1-mediated pro-apoptotic signaling through
(2009). transcriptional output of the Hippo growth-regulatory phosphorylation of threonine 120. J. Biol. Chem. 285,
36. Du, S. J., Purcell, S. M., Christian, J. L., McGrew, L. L. pathway. Dev. Cell 14, 388–398 (2008). 3815–3824 (2010).
& Moon, R. T. Identification of distinct classes and 60. Zhang, L. et al. The TEAD/TEF family of transcription 83. Zhang, J. et al. YAP-dependent induction of
functional domains of Wnts through expression of factor Scalloped mediates Hippo signaling in organ amphiregulin identifies a non-cell-autonomous
wild-type and chimeric proteins in Xenopus embryos. size control. Dev. Cell 14, 377–387 (2008). component of the Hippo pathway. Nature Cell Biol. 11,
Mol. Cell. Biol. 15, 2625–2634 (1995). 61. Bennett, F. C. & Harvey, K. F. Fat cadherin 1444–1450 (2009).
This analysis provides an effective classification modulates organ size in Drosophila via the 84. Zhou, D. et al. Mst1 and Mst2 maintain hepatocyte
schema and structure–function analysis of the Wnt Salvador/Warts/Hippo signaling pathway. Curr. Biol. quiescence and suppress hepatocellular carcinoma
family of ligands. 16, 2101–2110 (2006). development through inactivation of the Yap1
37. Wong, G. T., Gavin, B. J. & McMahon, A. P. 62. Cho, E. et al. Delineation of a Fat tumor suppressor oncogene. Cancer Cell 16, 425–438 (2009).
Differential transformation of mammary epithelial cells pathway. Nature Genet. 38, 1142–1150 (2006). 85. Owens, D. M. & Keyse, S. M. Differential regulation
by Wnt genes. Mol. Cell. Biol. 14, 6278–6286 63. Silva, E., Tsatskis, Y., Gardano, L., Tapon, N. & of MAP kinase signalling by dual-specificity protein
(1994). McNeill, H. The tumor-suppressor gene fat controls phosphatases. Oncogene 26, 3203–3213 (2007).
38. Angers, S. & Moon, R. T. Proximal events in Wnt signal tissue growth upstream of Expanded in the Hippo 86. Huang, S. M. et al. Tankyrase inhibition stabilizes
transduction. Nature Rev. Mol. Cell Biol. 10, signaling pathway. Curr. Biol. 16, 2081–2089 axin and antagonizes Wnt signalling. Nature 461,
468–477 (2009). (2006). 614–620 (2009).
39. Torres, M. A. et al. Activities of the Wnt-1 class of 64. Willecke, M. et al. The Fat cadherin acts through the Reveals a new level of regulation of Wnt signalling
secreted signaling factors are antagonized by the Hippo tumor-suppressor pathway to regulate tissue through the tankyrase‑mediated destabilization of
Wnt-5A class and by a dominant negative cadherin in size. Curr. Biol. 16, 2090–2100 (2006). axin 1.
early Xenopus development. J. Cell Biol. 133, References 62–64 position and characterize the 87. Albeck, J. G. et al. Quantitative analysis of pathways
1123–37 (1996). key molecules associated with Fat signalling into a controlling extrinsic apoptosis in single cells. Mol. Cell
40. Ishitani, T. et al. The TAK1-NLK-MAPK-related pathway. 30, 11–25 (2008).
pathway antagonizes signalling between β-catenin and 65. Yu, J., Poulton, J., Huang, Y. C. & Deng, W. M. The 88. Arnold, H. K. et al. The Axin1 scaffold protein
transcription factor TCF. Nature 399, 798–802 Hippo pathway promotes Notch signaling in regulation promotes formation of a degradation complex for
(1999). of cell differentiation, proliferation, and oocyte c-Myc. EMBO J. 28, 500–512 (2009).
41. Meneghini, M. D. et al. MAP kinase and Wnt polarity. PLoS One 3, e1761 (2008). 89. Goentoro, L. & Kirschner, M. W. Evidence that fold-
pathways converge to downregulate an HMG-domain 66. Meignin, C., Alvarez-Garcia, I., Davis, I. & change, and not absolute level, of β-catenin dictates
repressor in Caenorhabditis elegans. Nature 399, Palacios, I. M. The Salvador-Warts-Hippo pathway Wnt signaling. Mol. Cell 36, 872–884 (2009).
793–797 (1999). is required for epithelial proliferation and axis 90. Buchert, M. et al. Genetic dissection of
42. Rousset, R. et al. Naked cuticle targets Dishevelled to specification in Drosophila. Curr. Biol. 17, differential signaling threshold requirements for the
antagonize Wnt signal transduction. Genes Dev. 15, 1871–1878 (2007). Wnt/β-catenin pathway in vivo. PLoS Genet. 6,
658–671 (2001). 67. Polesello, C. & Tapon, N. Salvador-Warts-Hippo e1000816 (2010).
43. Yan, D. et al. Cell autonomous regulation of multiple signaling promotes Drosophila posterior follicle cell 91. Yokota, Y. et al. The adenomatous polyposis coli
Dishevelled-dependent pathways by mammalian Nkd. maturation downstream of Notch. Curr. Biol. 17, protein is an essential regulator of radial glial polarity
Proc. Natl Acad. Sci. USA 98, 3802–3807 (2001). 1864–1870 (2007). and construction of the cerebral cortex. Neuron 61,
44. Lee, J. M. et al. RORα attenuates Wnt/β-catenin 68. Hamaratoglu, F. et al. The tumour-suppressor genes 42–56 (2009).
signaling by PKCα-dependent phosphorylation in colon NF2/Merlin and Expanded act through Hippo 92. Goentoro, L., Shoval, O., Kirschner, M. W. & Alon, U.
cancer. Mol. Cell 37, 183–195 (2010). signalling to regulate cell proliferation and apoptosis. The incoherent feedforward loop can provide fold-
45. Minami, Y., Oishi, I., Endo, M. & Nishita, M. Ror-family Nature Cell Biol. 8, 27–36 (2006). change detection in gene regulation. Mol. Cell 36,
receptor tyrosine kinases in noncanonical Wnt 69. Sopko, R. et al. Phosphorylation of the tumor 894–899 (2009).
signaling: their implications in developmental suppressor fat is regulated by its ligand Dachsous 93. Kim, W. Y. et al. GSK-3 is a master regulator of neural
morphogenesis and human diseases. Dev. Dyn. 239, and the kinase Discs overgrown. Curr. Biol. 19, progenitor homeostasis. Nature Neurosci. 12,
1–15 (2010). 1112–1117 (2009). 1390–1397 (2009).
46. Sato, A., Yamamoto, H., Sakane, H., Koyama, H. & 70. Feng, Y. & Irvine, K. D. Processing and
Kikuchi, A. Wnt5a regulates distinct signalling phosphorylation of the Fat receptor. Proc. Natl Acad. Acknowledgements
pathways by binding to Frizzled2. EMBO J. 29, Sci. USA 106, 11989–11994 (2009). The authors acknowledge support from the Canadian
41–54 (2010). 71. Yu, J. et al. Kibra functions as a tumor suppressor Institutes of Health Research.
47. Matsumoto, S., Fumoto, K., Okamoto, T., Kaibuchi, K. protein that regulates Hippo signaling in conjunction
& Kikuchi, A. Binding of APC and dishevelled mediates with Merlin and Expanded. Dev. Cell 18, 288–299 Competing interests statement
Wnt5a-regulated focal adhesion dynamics in migrating (2010). The authors declare no competing financial interests.
cells. EMBO J. 29, 1192–1204 (2010). 72. Baumgartner, R., Poernbacher, I., Buser, N., Hafen, E.
48. Rusan, N. M. & Peifer, M. Original CIN: reviewing & Stocker, H. The WW domain protein Kibra acts
roles for APC in chromosome instability. J. Cell Biol. upstream of Hippo in Drosophila. Dev. Cell 18, DATABASES
181, 719–726 (2008). 309–316 (2010). Entrez gene: [Link]
49. O’Connell, M. P. et al. The orphan tyrosine kinase 73. Genevet, A., Wehr, M. C., Brain, R., Thompson, B. J. & axin 2 | four-jointed | GSK3A | GSK3B | MYC
receptor, ROR2, mediates Wnt5A signaling in Tapon, N. Kibra is a regulator of the Salvador/Warts/ uniProtKB: [Link]
metastatic melanoma. Oncogene 29, 34–44 Hippo signaling network. Dev. Cell 18, 300–308 AKT1 | APC | Armadillo | axin 1 | Dachsous | Dishevelled | DPP |
(2010). (2010). Expanded | Fat | Frizzled | Hippo | Kibra | LRP5 | LRP6 | MATS |
50. Roman-Gomez, J. et al. WNT5A, a putative tumour 74. Oh, H., Reddy, B. V. & Irvine, K. D. Phosphorylation- Merlin | ROR2 | RYK | Salvador | TAZ | Warts | Wingless |
suppressor of lymphoid malignancies, is inactivated independent repression of Yorkie in Fat-Hippo WNT1 | WNT3A | WNT4 | WNT5A | WNT8A | WNT8B | WNT11 |
by aberrant methylation in acute lymphoblastic signaling. Dev. Biol. 335, 188–197 (2009). YAP | Yorkie
leukaemia. Eur. J. Cancer 43, 2736–46 (2007). 75. Rogulja, D., Rauskolb, C. & Irvine, K. D. Morphogen
51. Oishi, I. et al. The receptor tyrosine kinase Ror2 is control of wing growth through the Fat signaling FURTHER INFORMATION
involved in non-canonical Wnt5a/JNK signalling pathway. Dev. Cell 15, 309–321 (2008). Helen Mcneill’s homepage:
pathway. Genes Cells 8, 645–654 (2003). 76. Baena-Lopez, L. A., Rodriguez, I. & Baonza, A. The [Link]
52. Mikels, A. J. & Nusse, R. Purified Wnt5a protein tumor suppressor genes dachsous and fat modulate James R. Woodgett’s homepage:
activates or inhibits β-catenin-TCF signaling different signalling pathways by regulating Dally and [Link]
depending on receptor context. PLoS Biol. 4, e115 Dally-like. Proc. Natl Acad. Sci. USA 105, 9645–9650 Pathway Interaction Database:
(2006). (2008). Canonical Wnt pathway; Non-canonical Wnt pathway
53. He, X. et al. A member of the Frizzled protein family 77. Herranz, H. & Milan, M. Signalling molecules, growth ucSD–nature Signaling gateway: MST1; MST2
mediating axis induction by Wnt-5A. Science 275, regulators and cell cycle control in Drosophila. Cell all liNks are acTive iN THe oNliNe PdF
1652–4 (1997). Cycle 7, 3335–3337 (2008).

nATuRe ReVIeWS | Molecular cell Biology VoluMe 11 | june 2010 | 413

© 2010 Macmillan Publishers Limited. All rights reserved