|
|
|
|
|
|
|
||||
| Home Help Feedback Subscriptions Archive Search Table of Contents | ||||
First published online 3 August 2006
doi: 10.1242/dev.02504
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1 Department of Biochemistry, Molecular Biology and Cell Biology, Northwestern
University, Evanston, IL 60208, USA.
2 Robert H. Lurie Comprehensive Cancer Center, Northwestern University,
Evanston, IL 60208, USA.
* Author for correspondence (e-mail: clabonne{at}northwestern.edu)
Accepted 21 June 2006
 |
SUMMARY |
|---|
 TOP SUMMARY INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSupplementary materialREFERENCES |
|---|
Key words: Xenopus, Neural crest, Slug, Snail, Ppa, Ubiquitin
|
INTRODUCTION |
|---|
|
TOPSUMMARYINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSupplementary materialREFERENCES |
|---|
). The precursors of these cells arise from the ectoderm at
the lateral edges of the neural plate during midgastrulation. Following neural
tube closure, these cells undergo an epithelial to mesenchymal transition
(EMT), delaminate from the dorsal aspect of the neural tube, and migrate
extensively to populate distant sites throughout the embryo. Ultimately,
neural crest cells produce most of the neurons and glia of the peripheral
nervous system, melanocytes and nearly all of the elements of the craniofacial
skeleton (Le Douarin and Kalcheim,
1999). Perhaps because of their contribution to such a large
number of derivatives, defects in the neural crest are associated with an
estimated half of all known birth defects
(Hall, 1999). Moreover, a
number of cancers of great clinical significance, including melanomas and
gliomas, are neural crest-derived. Neuroblastomas, one of the most common
pediatric solid tumors, originate from neural crest precursor cells themselves
(Nakagawara and Ohira, 2004).
Because a number of genes that are essential to neural crest development are
misregulated in these tumors (Rosivatz et
al., 2002; Rothhammer et al.,
2004), examining the regulatory programs that control the normal
development of neural crest cells can provide essential insights into how such
programs are inappropriately reactivated during tumorigenesis.
As a population of proliferative, migratory, tissue-invasive stem cells,
the neural crest shares a number of characteristics with metastatic tumor
cells. For instance, the Snail family of transcriptional repressors are key
molecular regulators of both neural crest development and tumor progression.
Members of the Snail family, including the C2H2
zinc-finger transcription factors Slug and Snail, are among
the earliest factors expressed in neural crest-forming regions
(Nieto, 2002). Although Snail
was first characterized in Drosophila, homologs have subsequently
been isolated from many organisms including humans, non-vertebrate chordates,
nematodes, annelids and mollusks
(Manzanares et al., 2001). In
Xenopus, Slug and Snail expression can be detected in neural
crest-forming regions by late gastrula stages, and both factors are expressed
at the lateral edges of the open neural plate in midbrain, hindbrain and
spinal cord regions. Snail is also transiently expressed in the
transverse neural fold, from which neural crest does not arise
(Essex et al., 1993;
Mayor et al., 1995).
The temporal and spatial expression of Snail-related factors prompted
studies to probe their functional roles during neural crest precursor
formation. In Xenopus, overexpression of dominant inhibitory forms of
Slug leads to a loss of neural crest precursor formation, whereas ectopic
wild-type Slug expands early neural crest marker expression
(LaBonne and Bronner-Fraser,
1998; LaBonne and
Bronner-Fraser, 2000). Similar effects have recently been reported
in avian embryos (del Barrio and Nieto,
2002). In both avian and Xenopus embryos, treatment with
Slug antisense oligos inhibits delamination of neural crest cells from the
neural tube, resulting in defective migration
(Carl et al., 1999;
Nieto et al., 1994).
Furthermore, the use of a hormone-inducible dominant negative form of Slug
demonstrated that Snail-related factors play two temporally distinct roles in
neural crest formation (LaBonne and
Bronner-Fraser, 2000). Together, these results indicate that the
neural crest can be categorized into at least two developmental stages
(pre-migratory precursors and maturing, migratory cells), and that
Snail-related factors have essential roles in both of these steps.
One hallmark of the Snail-related factors is their ability to promote EMTs
during embryonic development (Bolos et al.,
2003; Carver et al.,
2001; Savagner,
2001). In addition to their role in triggering neural crest
migration, Slug/Snail expression is also associated with a number of other
developmental EMTs including mesoderm ingression
(Nieto et al., 1994), forming
heart cushions (Romano and Runyan,
2000) and palatal closure
(Martinez-Alvarez et al.,
2004). Significantly, many of the molecular and phenotypic changes
associated with cells undergoing developmental EMTs are also characteristic of
metastatic carcinoma cells (Vernon and
LaBonne, 2004). Overexpression of Snail in epithelial cell lines
is sufficient to induce these cells to undergo an EMT and acquire invasive
properties (Cano et al., 2000;
Thiery, 2002). Furthermore,
whereas little or no Snail expression is detected in
well-differentiated, non-invasive carcinomas, this factor is strongly
expressed in a range of metastatic carcinoma cell lines as well as in many
invasive human carcinomas (Batlle et al.,
2000; Blanco et al.,
2002; Cano et al.,
2000).
Although Snail family members are critical for proper neural crest
development and are implicated in the metastatic step of tumorigenesis, the
mechanisms that control the expression or activity of these factors are poorly
understood. For example, Snail and Slug are capable of inducing an EMT when
ectopically expressed in epithelial cell lines, yet in embryos both of these
factors are highly expressed in non-migratory neural crest precursor cells,
and Slug is required for the formation of these precursor cells in
Xenopus. Furthermore, in the non-vertebrate chordates Ciona
and Amphioxus, Snail is expressed in a position analogous to where
neural crest precursors form in vertebrates, yet these cells do not undergo an
EMT or migrate (Corbo et al.,
1997; Langeland et al.,
1998). Taken together, these data point to post-transcriptional
and/or post-translational regulation of Slug and Snail that is essential for
precise control of the activity of these proteins, but is as of yet poorly
understood.
We demonstrate here that the transcriptional repressor Slug is a highly unstable protein. Ubiquitin-mediated proteasomal degradation of this factor is opposed by activation of the neural crest regulatory program, components of which promote Slug stability. We map the portion of the Slug protein that confers instability, and show that this region is sufficient to transfer instability to an unrelated protein, Sox10. Furthermore, we show that Partner of paired (Ppa), an F-box-containing component of a modular E3 ubiquitin ligase, binds to the Slug protein and promotes its degradation. Morpholino-depletion of Ppa stabilizes Slug protein, whereas misexpression of Ppa promotes Slug turnover and inhibits the formation of neural crest precursors. These results shed important new light on the regulation of a protein that plays a central role in neural crest precursor formation as well as developmental and pathological EMTs.
|
MATERIALS AND METHODS |
|---|
|
TOPSUMMARYINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSupplementary materialREFERENCES |
|---|
P38 sense: 5'-ATGCCTGTGTCTGTG-3', H64 sense:
5'-ATGCACCCACCGCTG-3'; Slug P31-H64 sense1:
5'-ATGCCACGATCTTTT-3', antisense1: 5'-TGGGGAGATGATCAC-3,
sense2: 5'-CACCCACCGCTGCCC-3', antisense2:
5'-ATGTGCTACACAGCA-3'. The SlugN-Sox10 fusion was made by
amplifying amino acids 1-63 of Slug (sense: 5'-ATGCCACGATCTTTT-3';
antisense: 5'-AAGCAGTCCTGTCCA-3') and fusing this fragment in the
ClaI site upstream of Sox10 in the StuI site of
pCS2-MycC. All constructs were confirmed by sequencing.
Embryological methods
All results shown are representative of at least three independent
experiments. RNA for injection was produced in vitro from linearized plasmid
templates using the Message Machine kit (Ambion). Collection, injection and in
situ hybridization of Xenopus embryos was as described previously
(LaBonne and Bronner-Fraser,
1998; Bellmeyer et al.,
2003). The in situ probe for Ppa was generated by PCR using the
primers: sense 5'-TCCGCATCACCGACA-3' and antisense
5'-CAATACACCACATCC-3' and then subcloned into the pGEMT vector
(Promega). For 70 µm sections, specimens were post-fixed in 4%
paraformaldehyde (PFA)/0.2% glutaraldehyde, embedded in 3% low melting point
agarose and sectioned using a Leica VT1000M vibratome. The Ppa MO sequence is:
5'-AGACACGAGATGTGGGTCTCCATAG-3' (targeted ATG is
underlined). Where noted, embryos were treated with 10 µg/ml cycloheximide
(Sigma) in 0.1x Marc's modified Ringers (MMR).
Immunoprecipitation and western blot analysis
For immunoprecipitations, embryos were collected at stage 10.5 and lysed in
PBS+1% NP40 containing a protease inhibitor cocktail (Roche). Following
centrifugation, cleared lysates were diluted with RIPA buffer and incubated
with the indicated antibody [0.2 µg 
-Myc (9E10, Santa Cruz) or
-FlagM2 (Sigma)] for 2 hours on ice, followed by a 2-hour incubation
with protein A Sepharose beads. After extensive washing with RIPA buffer, SDS
sample buffer was added to the beads and proteins were resolved by SDS-PAGE.
Immunoblotting was performed using -Flag (1:3000) or -Myc
(1:2000) antibodies as indicated. Labeled proteins were detected using
HRP-conjugated secondary antibodies and enhanced chemiluminescence
(Amersham).
Ubiquitination assays
The ubiquitination assay was modified from
(McGarry, 2005). Fifty
injected embryos were collected in an Eppendorf tube and washed four times
with 1 ml extraction buffer containing protease inhibitors. Washed embryos
were spun at 1000 g and all remaining liquid removed. Embryos
were lysed by pipetting, using 500 µl of extraction buffer plus 20 µg/ml
cytochalasin B, and then spun at 10,000 g at 4°C for 10
minutes. After removing the cytoplasmic layer to a new tube, energy mix (1/20
of volume) and cycloheximide (100 µg/ml) were added. The extract was then
divided in half and treated with either 800 µM MG132 or dimethysulfoxide
vehicle. Samples were taken at regular intervals for western blot
analysis.
|
RESULTS |
|---|
|
TOPSUMMARYINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSupplementary materialREFERENCES |
|---|
). Because
the regulation of Id turnover by the ubiquitin-proteosome pathway is well
documented, and because the half-lives of Id proteins have been previously
calculated (Bounpheng et al.,
1999; Fajerman et al.,
2004), we compared the stability of Slug and Id3 in
Xenopus. Embryos injected at the two-cell stage with mRNA encoding
epitope-tagged forms of Slug or Id3 were allowed to develop to stage 8 and
were then treated with cycloheximide (CHX) to inhibit further protein
synthesis. The embryos were collected at regular time intervals subsequent to
CHX treatment and subjected to western blot analysis. These experiments
confirmed that the Slug protein is highly unstable, and demonstrated that its
half-life is comparable to that of Id3 in early embryos
(Fig. 1A).
|
Given Slug's central role in neural crest development, it is likely that precise control over when and where the protein is present during this process is essential. To gain insight into the importance of protein turnover for the regulation of Slug function, we asked whether instability is a conserved feature of Snail family proteins. Embryos were injected with mRNA encoding Xenopus (Xl) Slug, Xenopus Snail, mouse (Mm) Snail, Drosophila (Dm) Snail, or with Xenopus Sox10 as a control, and samples were collected at blastula, gastrula, neurula and tailbud stages to compare the stability of the various proteins over time. Whereas Sox10 was extremely stable and expressed throughout the developmental time course examined, Xl Slug, Xl Snail, Mm Snail and Dm Snail proteins were all greatly diminished or undetectable by migratory neural crest stages (Fig. 1B), indicating that instability is an evolutionarily conserved feature of these proteins from invertebrates to mammals.
During the course of our studies, two groups reported that human (Hs) Snail
is unstable in tissue culture cells, confirming that at least one member of
this family is regulated by protein turnover in a context other than early
embryonic cells. Moreover, although these two studies contained a number of
conflicting findings, both reported that Hs Snail stability is regulated by
glycogen synthase kinase ß (GSK3ß) phosphorylation and
ßtrcp-mediated ubiquitination (Yook
et al., 2005; Zhou et al.,
2004). Surprisingly, however, examination of the Slug coding
sequence revealed that the well-defined ßtrcp destruction motif
(DSGX2+nS) which is also found in ß-catenin, I
B and Emi
(Fuchs et al., 2004) is absent
in Slug, making it unlikely that Slug is regulated via this mechanism.
Nevertheless, we investigated whether Slug stability is regulated in the
embryo by the GSK3ß/ßtrcp pathway in a manner similar to that
suggested for Hs Snail. Using ß-catenin as a positive control, we
investigated whether inhibiting GSK3ß activity using a well-characterized
dominant negative mutant (DNGSK3ß) would alter Snail or Slug stability in
the early embryo. We co-injected Snail, Slug or ß-catenin along with
DNGSK3ß mRNA into two-cell stage embryos, treated the injected embryos
with CHX at stage 8 and collected samples at the indicated time points.
Whereas ß-catenin was robustly stabilized by the presence of
DNGSK3ß, Slug protein stability was completely unaffected
(Fig. 1D). A weak stabilization
of Snail was noted in some experiments
(Fig. 1D), suggesting that the
GSK3ß pathway may make a minor contribution to Snail regulation. However,
our data indicate that Slug is not regulated by this mechanism.
Slug protein stability is regulated by neural crest transcription factors
Although numerous genes have been implicated in the establishment and
maintenance of neural crest identity, the precise roles of individual factors
and the interplay between these factors remains poorly understood
(Heeg-Truesdell and LaBonne,
2004; Meulemans and
Bronner-Fraser, 2004). As it was clear that Slug is regulated at
the level of protein turnover, we next wished to determine whether the
stability of Slug could be modulated by known neural crest regulatory factors.
This was of particular importance given our experiments demonstrating that
GSK3ß does not regulate Slug stability.
A recent study reported that Slug overexpression was insufficient to
promote neural crest formation in the trunk region of the avian spinal cord,
but that the combined expression of Sox9, FoxD3 and Slug could induce
formation of cells with the phenotypic and morphological characteristics of
definitive neural crest. The mechanisms responsible for this combinatorial
regulation were not investigated, however
(Cheung et al., 2005). To
determine whether FoxD3 and/or Sox9 modulate Slug stability, embryos were
injected with Slug alone or together with FoxD3, Sox9 or both, treated with
CHX at stage 8, and collected at regular time intervals for western blot
analysis. We found that co-expression of either FoxD3 or Sox9 significantly
enhanced Slug stability (Fig.
2A,B), but that co-expression of FoxD3 and Sox9 together did not
result in any further stabilization (data not shown). These data indicate
that, in addition to their individual and independent contributions to neural
crest development, FoxD3 and Sox9 may promote this process through their
ability to stabilize Slug.
The bHLH transcription factor Twist is also closely associated with the
function of Snail family proteins. In Drosophila, Twist is required
for the proper activation and maintenance of snail expression, and
the Twist mutant phenotype can be significantly rescued by expression of Snail
(Ip et al., 1994;
Leptin, 1991). Furthermore,
Twist activity has recently been linked to EMTs during tumor metastasis
(Yang et al., 2004).
Importantly, we found that, like Sox9 and FoxD3, co-expression of Twist was
capable of stabilizing Slug (Fig.
2C), suggesting an additional neural crest regulatory function for
this factor. To determine whether the ability to stabilize Slug was a general
feature of neural crest regulatory factors, we asked whether Opl, a regulator
of neural crest border fates, or Smad1, a downstream mediator of BMP
signaling, were capable of stabilizing Slug. Interestingly, neither Opl nor
Smad1 had any effect on Slug stability in these assays
(Fig. 2D,E), indicating that
the Slug stabilizing activities of Sox9, FoxD3 and Twist are specific.
Finally, because many neural crest regulatory factors are involved in highly
complex feedback loops, we investigated whether Snail was capable of
stabilizing Slug. However, as with Opl and Smad1, we found that Snail
overexpression had no effect on Slug stability
(Fig. 2F).
|
|
;
Finley et al., 1994;
Hochstrasser et al., 1991). We
co-injected Slug, the P38 mutant or Slug-C with UbK48R and
performed western blots to detect ubiquitin incorporation. Whereas full-length
Slug and the unstable P38 mutant were both clearly ubiquitinated in
this assay, the stable C-terminal zinc finger mutant was not
(Fig. 3F). To confirm and
extend these findings, we examined whether Slug was polyubiquitinated and
degraded in a proteasome-dependent fashion. Lysates from embryos injected with
mRNA encoding Slug or the Slug-C mutant were treated with either DMSO or the
proteasome inhibitor MG132. Accumulation of polyubiquitinated species was
dramatically increased in the Slug sample treated with MG132 as compared to
both the control DMSO-treated samples and the MG132 treated Slug-C sample
(Fig. 3G). These data confirm
that Slug is targeted for proteasomal degradation via poly-ubiquitination of
the N-terminal portion of the protein.
The F-box protein Ppa is expressed in the neural crest and binds to Slug
We next sought to identify E3 ubiquitin ligases that might regulate Slug
turnover. Interestingly, although most E3 ligases have been reported to be
broadly expressed, the F-box protein Ppa shows developmentally restricted
expression in both Drosophila and zebrafish embryos
(Das et al., 2002;
Raj et al., 2000). The
Xenopus homolog of Ppa had been previously isolated in our laboratory
to examine a putative role in regulating neural plate border cell types.
Similar to Drosophila and zebrafish Ppa, the Xenopus Ppa
protein contains an F-box together with 11 leucine rich repeats (LRRs) and is
95% identical to the previously identified zebrafish ortholog (see Fig. S1 in
the supplementary material). F-box-containing proteins function as substrate
recognition subunits for SCF (Skp1-Cullin-F-box) ubiquitin ligases, and
consistent with this we found that Xenopus Ppa binds to Skp1 through
its F-box (Fig. S1 in the supplementary material). Whereas the N-terminal
region of Drosophila Ppa contains a putative
alanine/histidine/proline-rich repressor domain and a PEST domain
(Raj et al., 2000) these
motifs are not conserved in any other known Ppa homologs, including
Xenopus Ppa.
Ppa is not expressed in neural crest-forming regions at late gastrula and early neural plate stages, a time when Slug function is known to be essential for the establishment of the progenitor cell population (Fig. 4Aa). However, Ppa becomes highly expressed in neural crest cells as the neural folds are closing (Fig. 4Ab-d), and is expressed in at least some migrating neural crest cells (Fig. 4Ae,f). Because expression of Ppa appeared to be dynamic, and it was not expressed in neural crest precursors at all stages, we next directly compared its expression to that of Slug using double whole mount in situ hybridization. Again, these experiments clearly indicated that Ppa expression entirely co-localizes with that of Slug as the neural folds are closing (Fig. 4Ba,b). Interestingly, however, by stage 20, just prior to the onset of neural crest migration in cranial regions, Ppa expression is downregulated/lost in a subset of neural crest precursors (Fig. 4Bc). Similarly, at early migratory stages it is clear that many neural crest cells proximal to the neural tube (e.g. cells that have just begun migration) do not express Ppa (Fig. 4Bd). To obtain better resolution, we examined the co-localization of Ppa and Slug expression in sections. Interestingly, we noted that in rostral regions of stage 15 embryos, where neural crest precursors are already specified and where cranial neural folds are closing, expression of Slug is contained completely within the Ppa expression domain (Fig. 4Ca). However, neural crest precursors are specified in a rostral-to-caudal progression, and when we examined more caudal sections from the same embryo, at axial levels where neural crest precursors are likely to be newly induced, we noted that the domains of Slug staining lay largely outside the Ppa expressing region (Fig. 4Cb). Sections through embryos in which neural crest cells are initiating migration in cranial (Fig. 4Cc) or spinal cord (Fig. 4Cd) regions also show loss of Ppa expression in some Slug-expressing cells. Together, these findings are consistent with a role for Ppa in controlling Slug protein levels such that they are maximized at times when Slug is known to play critical and essential roles in neural crest development. Interestingly, we also found that in embryos injected with either FoxD3 or Sox9, expression of Ppa was significantly reduced (Fig. S1 in the supplementary material). By contrast, Ppa protein levels are unaffected by exogenously coexpressed FoxD3 or Sox9 (data not shown). Regulation of Ppa at the level of transcription provides one potential mechanism via which these neural crest regulatory factors may act to stabilize Slug protein.
|
|
Given that Ppa regulates the turnover of a key neural crest regulatory factor, overexpression of Ppa might be expected to have deleterious consequences for neural crest precursor formation. To test this possibility, embryos were injected in one of two cells with Ppa mRNA and the lineage tracer, nß-gal, and the expression of neural crest markers was examined at neural plate stages by in situ hybridization. Ppa injection caused the downregulation or loss of a wide variety of early neural crest markers including Slug (100%, n=18), FoxD3 (100%, n=23), Twist (100%, n=19) and Sox10 (100%, n=20) and this disruption persisted to migratory stages (Fig. 5Ea,b and data not shown). We found that the ability of Ppa to reduce the expression of neural crest markers was dependent on the presence of an intact F-box (data not shown). Importantly, Ppa overexpression did not perturb the expression of muscle actin, demonstrating that the observed loss of neural crest precursors is not due to an underlying defect in the mesodermal cell population (data not shown). Furthermore, Ppa injection led to expanded expression of the neural plate marker Sox3 (90%, n=20), suggesting that cells overexpressing Ppa adopt a neural fate instead of becoming neural crest precursors (Fig. 5Ec).
We next sought to determine the extent to which the loss of neural crest
cell formation in Ppa-injected embryos was due to effects on Slug. Since
individual E3s are known to have multiple targets
(Pickart, 2001), the most
effective means of specifically disrupting the Ppa-Slug interaction required
identifying and eliminating the sequence in Slug responsible for Ppa binding.
Consistent with their relative stability, we found that the N terminus of Slug
bound to Ppa in co-immunoprecipitation (IP) experiments, but the C terminus
did not (Fig. 6A). Because
transfer of the N-terminal 63 amino acids of Slug to Sox10 rendered Sox10
unstable (Fig. 3D), we tested
whether the ability to bind Ppa had also been conferred on the SlugN-Sox10
fusion. Embryos were co-injected with wild-type Slug, Sox10 or SlugN-Sox10
mRNA along with Ppa, and were harvested at gastrula stages for IP/western blot
analysis. These experiments demonstrated that Slug-N was sufficient to confer
Ppa binding ability on Sox10 (Fig.
6B), providing a potential mechanism to explain the instability of
the SlugN-Sox10 fusion protein.
To more precisely map the Ppa interacting domain of Slug, we utilized the
N-terminal deletion mutant series previously generated to examine Slug
instability. We found that Ppa bound well to the P31 mutant, but that
deletion of an additional seven amino acids (P38) caused this
interaction to weaken significantly (Fig.
6C). Together with our previous experiments showing that the
addition of amino acids M1-H64 was sufficient to transfer both Ppa binding and
instability to Sox10, these data indicated that Ppa binding was mediated by a
region spanning amino acids P31 to H64. To confirm these results, we
constructed an internal deletion mutant that removed these amino acids.
Co-immunoprecipitation experiments demonstrated that the deletion mutant
P31-H64 was no longer capable of binding to Ppa, indicating that this
sequence was necessary for the E3 to interact with Slug
(Fig. 6D). Consistent with
these findings, deletion of this region also abolished the accumulation of
poly-ubiquitinated Slug proteins following MG132 treatment
(Fig. 3G), further confirming
the role of Ppa as a bona fide E3 subunit for Slug.
Disruption of the Ppa binding domain stabilizes Slug
Examination of the region between P31 and H64 of Slug did not reveal any
defined protein interaction motifs. This was not particularly surprising since
few sequence elements or structural features constituting such a destabilizing
signal or degron have been well characterized to date. Nevertheless, whereas
some proteins, such as ß-catenin, are targeted for degradation by short,
highly conserved peptide sequences, other proteins seem to depend on an
extended protein surface or prefolded structure to trigger their proteolysis
(Laney and Hochstrasser,
1999). Ppa contains 11 leucine-rich repeats, which generally
mediate protein-protein interactions and can interact with extended
hydrophobic regions (Kobe and Kajava,
2001). Interestingly, we noted that the region between P31-H64 of
Slug was rich in hydrophobic amino acids
(Fig. 7A).
To test the hypothesis that this hydrophobic region might mediate Slug's
interaction with Ppa, we introduced point mutations designed to disrupt such
an interaction. In Slug LY33,34AA (Slug
1) and Slug VW58,59AA
(Slug2), target hydrophobic amino acids were mutated to
alanines (Fig. 7A). The ability
of these Slug mutants to bind Ppa was then examined in co-immunoprecipitation
assays. Although neither of these mutations alone was sufficient to disrupt
the Slug-Ppa interaction, in combination they abolished Ppa binding
(Fig. 7B). These data support a
model in which Ppa interacts with Slug using extended contacts over a stretch
of hydrophobic residues. Snail family orthologs all contain a comparable
hydrophobic amino acid-rich region, further suggesting that these proteins
interact with Ppa using a similar motif.
|
1,2 were collected at progressively later stages of
development to compare the stability of these two proteins by western blot. We
found that Slug1,2 protein is more stable than wild-type Slug,
indicating that elimination of Ppa binding protects Slug from degradation
(Fig. 7C). To determine if
stabilizing Slug had consequences for normal neural crest development, Slug or
Slug1,2 mRNA was injected together with the lineage tracer
nß-gal into one cell of two-cell stage embryos. In situ hybridization of
embryos at neural plate stages revealed that Slug and Slug1,2
both potently expanded the neural crest precursor population
(Fig. 8Aa,e), indicating that
the stabilizing mutations do not have antimorphic effects on Slug function.
When sibling embryos were examined at stage 24, Slug-injected embryos showed
an increased number of migratory neural crest cells, consistent with the
expansion seen in the progenitor pool at earlier stages (compare
Fig. 8Ab with 8Ac).
Interestingly, the cranial region of many embryos injected with
Slug1,2 had areas that were cleared of migrating
Sox10-expressing neural crest (compare
Fig. 8Af with 8Ag), suggesting
that prolonged expression of high levels of Slug protein after the onset of
neural crest migration is incompatible with the normal development of these
cells. Importantly, however, these effects do not represent antimorphic
activities of the mutant protein, as expression of high levels of wild-type
Slug can also have similar consequences (data not shown). Another noteworthy
phenotypic consequence of the stabilized Slug protein was the presence of
premature, ectopic migratory neural crest cells in spinal cord regions
(Fig. 8Ah), further supporting
a role for Snail family proteins in promoting EMTs and migratory behavior.
|
|
1,2
mutant, which is resistant to Ppa-mediated degradation, was sufficient to
restore neural crest precursor formation in Ppa-injected embryos. We injected
mRNA encoding Ppa, Slug1,2, or both together with nß-gal
into one cell of two-cell stage embryos. Injected embryos were cultured to
neural plate stages when the expression of neural crest precursor markers was
examined by in situ hybridization. Expression of Ppa alone dramatically
reduced the expression of Slug (100%, n=14)
(Fig. 8Ba), whereas expression
of Slug1,2 expanded the neural crest precursor cell population
(93%, n=29) (Fig.
8Bb). Importantly, we found that co-expression of
Slug1,2 could rescue the Ppa-mediated loss of neural crest
precursors (82%, n=23; Fig.
8Bc), suggesting that Slug is the major target of Ppa regulation
in these cells. Whereas wild-type Slug is also capable of rescuing the
Ppa-overexpression neural crest defect when expressed at high levels, the
Slug1,2 mutant that does not bind to Ppa rescues this
phenotype much more effectively, and this is particularly evident at lower
doses (data not shown). |
DISCUSSION |
|---|
|
TOPSUMMARYINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSupplementary materialREFERENCES |
|---|
), Smads (Izzi and
Attisano, 2004) and MyoD
(Tintignac et al., 2005), are
known to be regulated at the level of protein turnover. Slug and Snail are
excellent examples of proteins for which such precise and accurate control is
necessary, given their critical roles in both neural crest precursor cell
formation and migration, and their ability to induce EMTs when overexpressed
in epithelial cells. Despite their developmental and clinical importance,
however, the mechanisms via which Snail family members control these processes
are poorly understood, and even less is known about the ways in which these
transcriptional repressors are themselves regulated. In this paper, we
demonstrate that Slug is a labile protein, the expression of which is
dynamically regulated by the opposing forces of neural crest regulatory
factors and the ubiquitin-mediated proteasomal degradation pathway.
Several recent studies have investigated the functional roles of
Snail-related factors during neural crest development and their involvement in
the metastatic progression of tumors. However, although much emphasis has been
placed on the ability of Snail family members to induce both physiological and
pathological EMTs, the expression of these genes is not always sufficient to
induce this dramatic morphological transformation. For example,
Xenopus Slug is highly expressed in neural crest precursors, and its
activity is required for their formation, but these cells are epithelial and
do not undergo an EMT or migrate until much later in development. Furthermore,
Snail mRNA expression is generally inversely correlated with E-cadherin, but
some cell lines show high levels of both genes
(Dominguez et al., 2003).
Indeed, recent studies have suggested that only mutant forms of Snail are
capable of downregulating E-cadherin activity or inducing an EMT when
overexpressed in certain epithelial cell lines
(Yook et al., 2005;
Zhou et al., 2004). Together,
these data point to a dynamic, context-dependent regulation of Slug and Snail
that allows them to regulate essential targets without always inducing an
EMT.
The non-vertebrate chordate Amphioxus provides a model for
considering this aspect of Slug regulation. Despite the presence of seemingly
appropriate signaling conditions, the induction of early patterning genes such
as Pax3/7 (Holland et al.,
1999) and Opl
(Gostling and Shimeld, 2003),
and the expression of Snail
(Langeland et al., 1998) at
the neural plate border, Amphioxus lacks definitive neural crest
(Holland and Holland, 2001).
This deficiency has been hypothesized to be due to the absence of other neural
crest specifiers such as Sox9, FoxD3 and Twist
(Yu et al., 2002). Although
these transcription factors have been shown to confer different aspects of the
neural crest phenotype, their concerted activity is important for the proper
development of these cells. Here we demonstrate that the Slug protein is
stabilized by co-expression of Sox9, FoxD3 or Twist, while overexpression of
other neural crest regulatory factors including Opl, Smad1 or Snail has no
effect on Slug stability. We suggest that one way in which Sox9, FoxD3 and
Twist, which are not expressed at the neural plate border in non-vertebrate
chordates, may contribute to neural crest formation is by stabilizing Snail
family proteins, thus altering their functional capacity to promote neural
crest development. We are currently investigating the mechanism(s) underlying
the ability of these neural crest transcription factors to influence Slug
protein stability, and one means by which they appear to do so is by
regulating the expression of Ppa. We found that expression of either FoxD3 or
Sox9 reduced the expression of Ppa as detected by in situ
hybridization (Fig. S1 in the supplementary material).
We show here that Slug is a target of the ubiquitin-mediated proteasomal
degradation pathway. Despite providing most of the substrate binding
specificity to this pathway (Hershko and
Ciechanover, 1998), most E3 ligases examined thus far have broad,
diffuse expression patterns in early embryos. A noteworthy exception to this
is Ppa, an F-box protein first characterized in Drosophila, where it
was one of only two E3s found to have specifically localized mRNA transcripts
(Das et al., 2002;
Raj et al., 2000). We report
here that Xenopus Ppa is expressed at the neural plate border and in
migrating neural crest cells, and that Ppa directly binds to Slug and promotes
its degradation. Significantly, we find that ectopic Ppa expression leads to
an almost complete loss of neural crest precursor cells. Morpholino-mediated
depletion of Ppa leads to the stabilization of Slug protein
(Fig. 5C) and also leads to
aberrant neural crest development (A.E.V. and C.L., unpublished). The latter
finding is of interest, but it must nevertheless be interpreted with caution
since the range of substrates regulated by Ppa remains unknown. Instead, to
specifically address the significance of Ppa regulation for Slug stability and
function, we mapped the Ppa interaction site to a region in the Slug N
terminus that is rich in hydrophobic amino acids and characterized minimal
mutations that disrupt the interaction between these two factors. These
mutations significantly stabilize the Slug protein, and can rescue neural
crest precursor formation in Ppa-injected embryos, suggesting that Slug is the
major target of Ppa regulation in neural crest cells. Finally, we demonstrate
that the persistence of too much Slug protein in migratory neural crest cells
severely disrupts their development, consistent with a need for tight control
over the levels and activity of Slug protein in early embryos.
Importantly, although two recent studies using mammalian cell cultures implicated the GSK3ß/ßtrcp pathway in the regulation of Hs Snail stability, we find that this pathway does not regulate Slug stability in Xenopus embryos. In marked contrast to the robust stabilization of ß-catenin, we found that blocking GSK3ß activity had no effect on Slug stability. This is consistent with the absence in Slug of the ßtrcp binding site present in human and mouse Snail proteins. Interestingly, however, the GSK3ß/ßtrcp pathway also appears to play, at most, a minimal role in regulating Snail stability in Xenopus embryos. Because Snail family proteins fulfill many essential functions in organisms ranging from arthropods to vertebrates, the existence of multiple, context-dependent mechanisms for controlling the activity and stability of these proteins makes considerable sense. Ppa, the F-box protein characterized here, can promote the degradation of Snail family proteins from Drosophila through to mammals (Fig. 5A; A.E.V. and C.L., unpublished), suggesting that this is a highly conserved mechanism for regulating Slug/Snail activity, and therefore is likely to be of significance beyond its role in neural crest development. Although additional E3s that modulate the levels, and thus the activity, of these critical regulatory factors may yet be described, determining the contribution of Ppa to the control of Slug/Snail-mediated EMTs during tumor progression will be important.
|
Supplementary material |
|---|
|
TOPSUMMARYINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSupplementary materialREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPSUMMARYINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSupplementary materialREFERENCES |
|---|
Batlle, E., Sancho, E., Franci, C., Dominguez, D., Monfar, M., Baulida, J. and Garcia De Herreros, A. (2000). The transcription factor snail is a repressor of E-cadherin gene expression in epithelial tumour cells. Nat. Cell Biol. 2, 84-89.[CrossRef][Medline]
Bellmeyer, A., Krase, J., Lindgren, J. and LaBonne, C. (2003). The protooncogene c-myc is an essential regulator of neural crest formation in Xenopus. Dev. Cell 4, 827-839.[CrossRef][Medline]
Blanco, M. J., Moreno-Bueno, G., Sarrio, D., Locascio, A., Cano, A., Palacios, J. and Nieto, M. A. (2002). Correlation of Snail expression with histological grade and lymph node status in breast carcinomas. Oncogene 21,3241 -3246.[CrossRef][Medline]
Bolos, V., Peinado, H., Perez-Moreno, M. A., Fraga, M. F.,
Esteller, M. and Cano, A. (2003). The transcription factor
Slug represses E-cadherin expression and induces epithelial to mesenchymal
transitions: a comparison with Snail and E47 repressors. J. Cell
Sci. 116,499
-511.
Bounpheng, M. A., Dimas, J. J., Dodds, S. G. and Christy, B.
A. (1999). Degradation of Id proteins by the
ubiquitin-proteasome pathway. FASEB J.
13,2257
-2264.
Cano, A., Perez-Moreno, M. A., Rodrigo, I., Locascio, A., Blanco, M. J., del Barrio, M. G., Portillo, F. and Nieto, M. A. (2000). The transcription factor snail controls epithelial-mesenchymal transitions by repressing E-cadherin expression. Nat. Cell Biol. 2,76 -83.[CrossRef][Medline]
Carl, T. F., Dufton, C., Hanken, J. and Klymkowsky, M. W. (1999). Inhibition of neural crest migration in Xenopus using antisense slug RNA. Dev. Biol. 213,101 -115.[CrossRef][Medline]
Carver, E. A., Jiang, R., Lan, Y., Oram, K. F. and Gridley,
T. (2001). The mouse snail gene encodes a key regulator of
the epithelial-mesenchymal transition. Mol. Cell.
Biol. 21,8184
-8188.
Chau, V., Tobias, J. W., Bachmair, A., Marriott, D., Ecker, D.
J., Gonda, D. K. and Varshavsky, A. (1989). A multiubiquitin
chain is confined to specific lysine in a targeted short-lived protein.
Science 243,1576
-1583.
Cheung, M., Chaboissier, M. C., Mynett, A., Hirst, E., Schedl, A. and Briscoe, J. (2005). The transcriptional control of trunk neural crest induction, survival, and delamination. Dev. Cell 8,179 -192.[CrossRef][Medline]
Corbo, J. C., Erives, A., Di Gregorio, A., Chang, A. and Levine, M. (1997). Dorsoventral patterning of the vertebrate neural tube is conserved in a protochordate. Development 124,2335 -2344.[Abstract]
Das, T., Purkayastha-Mukherjee, C., D'Angelo, J. and Weir, M. (2002). A conserved F-box gene with unusual transcript localization. Dev. Genes Evol. 212,134 -140.[CrossRef][Medline]
del Barrio, M. G. and Nieto, M. A. (2002).
Overexpression of Snail family members highlights their ability to promote
chick neural crest formation. Development
129,1583
-1593.
Dominguez, D., Montserrat-Sentis, B., Virgos-Soler, A., Guaita,
S., Grueso, J., Porta, M., Puig, I., Baulida, J., Franci, C. and Garcia de
Herreros, A. (2003). Phosphorylation regulates the
subcellular location and activity of the snail transcriptional repressor.
Mol. Cell. Biol. 23,5078
-5089.
Essex, L. J., Mayor, R. and Sargent, M. G. (1993). Expression of Xenopus snail in mesoderm and prospective neural fold ectoderm. Dev. Dyn. 198,108 -122.[Medline]
Fajerman, I., Schwartz, A. L. and Ciechanover, A. (2004). Degradation of the Id2 developmental regulator: targeting via N-terminal ubiquitination. Biochem. Biophys. Res. Commun. 314,505 -512.[CrossRef][Medline]
Finley, D., Sadis, S., Monia, B. P., Boucher, P., Ecker, D. J.,
Crooke, S. T. and Chau, V. (1994). Inhibition of proteolysis
and cell cycle progression in a multiubiquitination-deficient yeast mutant.
Mol. Cell. Biol. 14,5501
-5509.
Fuchs, S. Y., Spiegelman, V. S. and Kumar, K. G. (2004). The many faces of beta-TrCP E3 ubiquitin ligases: reflections in the magic mirror of cancer. Oncogene 23,2028 -2036.[CrossRef][Medline]
Gostling, N. J. and Shimeld, S. M. (2003). Protochordate Zic genes define primitive somite compartments and highlight molecular changes underlying neural crest evolution. Evol. Dev. 5,136 -144.[CrossRef][Medline]
Hall, B. K. (1999). The Neural Crest in Development and Evolution. New York: Springer-Verlag Press.
Heeg-Truesdell, E. and LaBonne, C. (2004). A slug, a fox, a pair of sox: transcriptional responses to neural crest inducing signals. Birth Defects Res. Part C Embryo Today 72,124 -139.[CrossRef][Medline]
Hershko, A. and Ciechanover, A. (1998). The ubiquitin system. Annu. Rev. Biochem. 67,425 -479.[CrossRef][Medline]
Hochstrasser, M., Ellison, M. J., Chau, V. and Varshavsky,
A. (1991). The short-lived MAT alpha 2 transcriptional
regulator is ubiquitinated in vivo. Proc. Natl. Acad. Sci.
USA 88,4606
-4610.
Holland, L. Z. and Holland, N. D. (2001). Evolution of neural crest and placodes: amphioxus as a model for the ancestral vertebrate? J. Anat. 199, 85-98.[Medline]
Holland, L. Z., Schubert, M., Kozmik, Z. and Holland, N. D. (1999). AmphiPax3/7, an amphioxus paired box gene: insights into chordate myogenesis, neurogenesis, and the possible evolutionary precursor of definitive vertebrate neural crest. Evol. Dev.1,153 -165.[CrossRef][Medline]
Ip, Y. T., Maggert, K. and Levine, M. (1994). Uncoupling gastrulation and mesoderm differentiation in the Drosophila embryo. EMBO J. 13,5826 -5834.[Medline]
Izzi, L. and Attisano, L. (2004). Regulation of the TGFbeta signalling pathway by ubiquitin-mediated degradation. Oncogene 23,2071 -2078.[CrossRef][Medline]
Kobe, B. and Kajava, A. V. (2001). The leucine-rich repeat as a protein recognition motif. Curr. Opin. Struct. Biol. 11,725 -732.[CrossRef][Medline]
LaBonne, C. and Bronner-Fraser, M. (1998). Neural crest induction in Xenopus: evidence for a two-signal model. Development 125,2403 -2414.[Abstract]
LaBonne, C. and Bronner-Fraser, M. (2000). Snail-related transcriptional repressors are required in Xenopus for both the induction of the neural crest and its subsequent migration. Dev. Biol. 221,195 -205.[CrossRef][Medline]
Laney, J. D. and Hochstrasser, M. (1999). Substrate targeting in the ubiquitin system. Cell 97,427 -430.[CrossRef][Medline]
Langeland, J. A., Tomsa, J. M., Jackman, W. R., Jr and Kimmel, C. B. (1998). An amphioxus snail gene: expression in paraxial mesoderm and neural plate suggests a conserved role in patterning the chordate embryo. Dev. Genes Evol. 208,569 -577.[CrossRef][Medline]
Le Douarin, N. M. and Kalcheim, C. (1999). The Neural Crest. Cambridge: Cambridge University Press.
Leptin, M. (1991). twist and snail as positive
and negative regulators during Drosophila mesoderm development.
Genes Dev. 5,1568
-1576.
Light, W., Vernon, A. E., Lasorella, A., Iavarone, A. and
LaBonne, C. (2005). Xenopus Id3 is required downstream of Myc
for the formation of multipotent neural crest progenitor cells.
Development 132,1831
-1841.
Manzanares, M., Locascio, A. and Nieto, M. A. (2001). The increasing complexity of the Snail gene superfamily in metazoan evolution. Trends Genet. 17,178 -181.[CrossRef][Medline]
Martinez-Alvarez, C., Blanco, M. J., Perez, R., Rabadan, M. A., Aparicio, M., Resel, E., Martinez, T. and Nieto, M. A. (2004). Snail family members and cell survival in physiological and pathological cleft palates. Dev. Biol. 265,207 -218.
