microbiology

Developmental Cell

Article

Reciprocal Repression between Sox3 and Snail Transcription Factors Defines Embryonic Territories at Gastrulation Hervé Acloque,1,3 Oscar H. Ocaña,1 Ander Matheu,2 Karine Rizzoti,2 Clare Wise,2 Robin Lovell-Badge,2

and M. Angela Nieto1,* 1Instituto de Neurociencias, CSIC-UMH, San Juan de Alicante 03550, Spain 2National Institute for Medical Research, London NW7 1AA, UK 3Present address: UMR 444, INRA-ENVT, Génétique Cellulaire, Toulouse 31326, France

*Correspondence: [email protected]

DOI 10.1016/j.devcel.2011.07.005

Open access under CC BY license.

SUMMARY

In developing amniote embryos, the first epithelial- to-mesenchymal transition (EMT) occurs at gastrula- tion, when a subset of epiblast cells moves to the primitive streak and undergoes EMT to internalize and generate the mesoderm and the endoderm. We show that in the chick embryo this decision to inter- nalize is mediated by reciprocal transcriptional re- pression of Snail2 and Sox3 factors. We also show that the relationship between Sox3 and Snail is conserved in the mouse embryo and in human cancer cells. In the embryo, Snail-expressing cells ingress at the primitive streak, whereas Sox3-posi- tive cells, which are unable to ingress, ensure the formation of ectodermal derivatives. Thus, the subdi- vision of the early embryo into the two main territo- ries, ectodermal and mesendodermal, is regulated by changes in cell behavior mediated by the antago- nistic relationship between Sox3 and Snail transcrip- tion factors.

INTRODUCTION

The shaping of the early embryo involves the conversion of

a single layer of ectodermal cells (the epiblast) into a multilayered

structure. This complex biological process starts at gastrulation,

when a subset of the initial epiblast cells moves inside the

embryo to become mesoderm and endoderm (Stern, 2004). In

amniotes, gastrulation occurs where Nodal signaling is strongest

(Brennan et al., 2001; Bertocchini and Stern, 2002). Initially, in the

chick embryo, cells accumulate at the posterior part through

epithelial cell intercalation (Voiculescu et al., 2007) in a region

now devoid of the hypoblast, a lower layer that inhibits Nodal

signaling (Bertocchini and Stern, 2002). This accumulation

results in the formation of a midline linear structure called the

primitive streak, from which the presumptive mesendodermal

cells ingress upon undergoing an epithelial-to-mesenchymal

transition (EMT). EMT involves a dramatic change in cell

morphology and behavior that allows cells to break the basal

546 Developmental Cell 21, 546–558, September 13, 2011 ª2011 Els

lamina, internalize, and start migrating toward their destinations

(Harrisson et al., 1991; Acloque et al., 2009). Cells that remain in

the epiblast keep their epithelial character and will contribute to

the ectodermal derivatives, namely the epidermis, the ecto-

dermal placodes, and the anterior central nervous system

(CNS) (Fraser and Stern, 2004). Indeed, much of the CNS will

develop from a subset of the noningressing cells later specified

as neural precursors. Therefore, it is crucial to identify not only

those factors that induce cell ingression at gastrulation but

also those that prevent it, because protection from undergoing

EMT is necessary to ensure the formation of ectodermal deriva-

tives. Indeed, previous studies have shown that committed

neural progenitor cells at the anterior part of the primitive streak

are protected from signals that induce internalization. The zinc

finger transcription factor Churchill and its target SIP-1 are

required to stop ingression movements through the anterior

primitive streak (Sheng et al., 2003). This safeguard mechanism

operates at stages of neural induction and onset of Churchill

expression (from stage HH4+; Hamburger and Hamilton,

1951), but ingression starts at least as early as stage HH2 (Stern

and Canning, 1990), suggesting that a different mechanism must

exist to protect early ectodermal cells from the EMT inducers.

Among the key factors that induce EMT at gastrulation and

that are conserved during evolution are the members of the Snail

family (Barrallo-Gimeno and Nieto, 2005). They are fundamental

for EMT at gastrulation in all species analyzed and for additional

developmental EMT processes (Acloque et al., 2009). We previ-

ously found that Snail2 (Slug) downregulation prevented migra-

tion from the primitive streak in the chick embryo (Nieto et al.,

1994), and here we show that Snail2 is sufficient to induce

ectopic delamination in otherwise noningressing epiblast cells,

confirming that these cells need to be protected from Snail2

expression and subsequent ingression. Therefore, we set out

to identify factors that might prevent Snail expression at early

gastrulation, as candidates to play an important role in protecting

epiblast cells from undergoing EMT. We show that Sox3 and

Snail2 are expressed in complementary domains early during

gastrulation; gain- and loss-of-function studies reveal that these

factors antagonize each other to regulate cell ingression. We

show that Sox3-Snail antagonism is implemented through direct

reciprocal transcriptional repression, a relationship that seems

to be conserved in the mouse embryo and in tumor cell lines,

where they also regulate epithelial versus mesenchymal and

evier Inc.

Figure 1. Snail2 Expression in the Epiblast Induces

Ectopic EMT and Cell Delamination

(A–C) Chick embryos were coelectroporated with a vector

encoding GFP and another containing the coding region of

Snail2. The embryos were subsequently allowed to

develop for 15 hr. Endogenous Snail2 expression at the

primitive streak and ectopic expression induced by elec-

troporation could be observed in a dorsal view (B) and in

a transverse section through the epiblast (C). Observe the

delamination of cells from the epiblast. Dotted line in (B)

indicates the level of the section shown in (C). Electro-

poration of a control empty vector (pCX) or that containing

GFP failed to induce cell delamination (not shown).

(D–H) Snail2 electroporation was accompanied by the

induction of RhoB expression, a known Snail2 down-

stream target (D and E), cell delamination, and disruption

of the basement membrane as assessed by the absence

of laminin staining (F and G) and by the repression of

E-cadherin expression in the electroporated cells (H).

Developmental Cell

Sox3 and Snail Define Embryonic Territories

invasive properties. Together, our results show that Snail-Sox3

cross-repression regulates cell ingression at gastrulation in

amniotes and suggest that this antagonistic relationship may

also have important implications in cancer.

Developmental Cell 21, 546–5

RESULTS

Snail2 Induces Ectopic Cell Delamination in the Ectoderm of the Early Chick Gastrula We had previously shown that Snail2 knock-

down in the early chick embryo prevents cell

ingression at the primitive streak and neural

crest delamination from the neural tube (Nieto

et al., 1994). To check whether Snail2 is suffi-

cient to trigger EMT and cell delamination, we

ectopically expressed the coding region of

Snail2 by electroporation of the chick blasto-

derm at stage 2 (Hamburger and Hamilton,

1951) in the ectodermal region corresponding

to the prospective neural plate (see Experi-

mental Procedures). Snail2 ectopic expression

induces cell delamination (Figures 1A–1C), sug-

gesting that Snail2 is sufficient to trigger EMT in

a territory that normally keeps its epithelial

integrity at gastrulation. To confirm that this

induced delamination is due to the activation

of a full EMT program, we examined the expres-

sion of previously described Snail2 target genes

known to be necessary for cell delamination,

such as the small GTPase RhoB (del Barrio

and Nieto, 2002) and E-cadherin (Cano et al.,

2000). As shown in Figures 1D and 1E, Snail2

induces ectopic expression of RhoB, together

with the disruption of basal lamina, as assessed

by the loss of laminin expression (Figures 1F and

1G). Figure 1H shows that E-cadherin is down-

regulated in electroporated cells that undergo

EMT, and cells can be seen delaminating from

the epiblast (compare the electroporated and

the control sides). These data indicate that Snail2 is sufficient

to trigger EMT and cell delamination from the chick epiblast,

suggesting that the latter should be protected from Snail ex-

pression in the early embryo.

58, September 13, 2011 ª2011 Elsevier Inc. 547

Figure 2. Snail2 and Sox3 Genes Are Complemen-

tarily Expressed in the Chick Gastrula

(A–C) Snail2 expression starts at the posterior part of the

blastula at stage EGXIV (A), and its expression is associ-

ated with the primitive streak as it progresses in the gas-

trulating embryo (B and C).

(D–F) Sox3 expression is detected in the epiblast and it

gets progressively downregulated at the primitive streak

as it forms.

(G and H) Double labeling for Snail2 (green) and Sox3 (red)

transcripts at the full primitive streak stage (HH4) shows

the complementary expression pattern between Snail2

and Sox3. A transverse vibratome section taken at the

level shown by the dotted line in (G) allows a better

assessment of the mutually exclusive expression pattern

and the continued Snail2 expression in migratory mesen-

dodermal cells emanating from the streak (H).

Developmental Cell

Sox3 and Snail Define Embryonic Territories

Snail2 and Sox3 Expression Domains Delineate Ingressing versus Noningressing Embryonic Cell Populations Because the process of cell ingression has to be tightly regulated

to maintain the balance between ectodermal (noningressing) and

mesendodermal (ingressing) progenitors, Snail2 is restricted

around the primitive streak, where cells are internalized, from

the first stages of streak formation (Nieto et al., 1994; Figures

2A–2C). To search for genes functionally equivalent to Churchill

but at the early primitive streak stages, we focused on the Sox3

gene for several reasons: (1) it is highly expressed at epiblast of

early-stage embryos, and its expression disappears in the terri-

tory surrounding the primitive streak (Rex et al., 1997; Figures

2D–2F); (2) Sox3 and Snail2 show mutually exclusive expression

patterns (Figures 2G and 2H); and (3) Sox2, a closely related

gene, has been shown to prevent the induction of Snail2 by

BMP in the dorsal neural fold (Wakamatsu et al., 2004), but it is

not expressed at the early gastrulation stages in chick. There-

fore, we decided to manipulate either Snail2 or Sox3 expression

548 Developmental Cell 21, 546–558, September 13, 2011 ª2011 Els

in the chick blastoderm by electroporation and assess the

effects on their respective expression. We first induced ectopic

expression of Snail2 by electroporating in regions such as the

anterior epiblast. We found that Sox3 expression was repressed

in 75% of the embryos (Figures 3F–3H; black arrow in H; n = 16),

which is best seen in transverse sections (Figures 3I–3K). GFP

electroporation alone does not have the same effect (Figures

3A–3E; n = 13). We then electroporated a Snail2 dominant-nega-

tive form lacking the transactivation domain (Morales et al., 2007;

DN-Snail2). This led to an extension of Sox3 expression to the

primitive streak up to the midline, a region that normally

expresses Snail2 (Figures 3L–3Q; black arrow in N; 53%, n =

17). Conversely, ectopic Sox3 expression at the primitive streak

strongly represses endogenous Snail2 expression (Figures 3R–

3X; 74%; n = 19), indicating that Sox3 and Snail2 act in a mutually

antagonistic way.

Next, we examined whether the observed changes in the ex-

pression domains of Snail2 and Sox3 correlate with cell behavior.

With this aim, we followed cell movements near the primitive

Figure 3. Snail2 and Sox3 Behave as Mutual

Transcriptional Repressors in the Chick Embryo (A–E) Control embryo electroporated with a vector encod-

ing GFP (A) shows the normal pattern of Sox3 expression

in a dorsal view (B) and in a transverse section (C). (D) and

(E) are high power images of the areas demarcated in (C).

(F–K) Similar images taken from an embryo electro-

porated with Snail2 in the anterior epiblast shows the

repression of Sox3 expression in the electroporated

region (arrow in H and sections in I). High power images

comparing the electroporated with the control side

confirm Sox3 downregulation (J and K).

(L–Q) Electroporation of a dominant-negative form of

Snail2 (DN-Snail2) in the primitive streak extended Sox3

expression to the embryonic midline (arrow in N), as better

assessed in the images shown in (O–Q).

(R–X) Ectopic Sox3 expression by electroporation at the

primitive streak inhibited Snail2 expression (black arrow in

T and a section at different magnifications in (U–X).

evier Inc.

D

BF(t=0) GFP (t=0)

A

Cell tracking

(8h)

CB 100

H

G F

P

ps

2

E

i

GF

Ingressed 0

50

100

non ingressed

( %

)

G F

P +

S n

a i l 2

Ingressed non ingressed

0

50

100

( %

)

L

G F

P +

S

o x

3

KJ

Ingressed non ingressed

0

50

100

( %

)

PM ON

Ingressed non 0

50

100

( %

)

G F

P +

D

N - S

n a

i 2

TQ SR

G F

P +

D

N S

o x

3

ingressed

0

50

100

( %

)

XU WV

Ingressed non ingressed

0

50

100

( %

)

P +

S n

a i l 2

+ S

o x

3

Ingressed non ingressed

0

G F

P

Figure 4. Sox3 Overexpression or Snail2 Inhibition

Block Ingression at the Primitive Streak

(A–X) Time-lapse confocal analysis of cell movements in

cultured embryos electroporated with GFP alone (A–D),

GFP plus Snail2 (E–H), GFP plus Sox3 (I–L), GFP plus

a dominant-negative form of Snail2 (M–P), GFP plus

a dominant-negative form of Sox3 (Q–T), and GFP, Sox3,

and Snail2 (U–X). BF (t = 0) and GFP (t = 0) show trans-

mitted light and fluorescent images obtained 6 hr after

electroporation, just before the start of time lapse anal-

yses. Colored dots in (B, F, J, N, R, and V) represent the

initial position of cells that were subsequently followed.

Dots in (C, G, K, O, S, and W) represent the final position of

those cells and the lines their tracks. (D), (H), (L), (P), (T),

and (X) show the percentages of ingressing and non-

ingressing cells in each condition. The white dotted lines

indicate the relative position of the midline; black arrows

and ps indicate primitive streak. Control GFP-expressing

embryos showed the normal movements of epiblasts cells

toward the primitive streak and subsequent ingression (C).

Snail2 overexpression increases the percentage of in-

gressing cells (E–H). Ectopic Sox3 expression blocked cell

ingression through the primitive streak without affecting

the convergence movement toward the midline (compare

C and D with K and L). A dominant-negative form of Snail2

also blocked cell ingression (compare C and D with O and

P), whereas a dominant-negative form of Sox3 increased

ingression (compare T with D). Overexpression of both

Snail2 and Sox3 could rescue the ingression behavior of

epiblast cells (compare C and D with W and X).

Developmental Cell

Sox3 and Snail Define Embryonic Territories

streak in embryos electroporated with GFP-containing control

vectors or with GFP plus different Snail2 and Sox3 constructs.

Cells expressing GFP converge at the primitive steak, ingress,

and migrate away in the mesendoderm as expected for a normal

embryo (Figures 4A–4D and Movies S1 and S2 available online).

Overexpression of wild-type Snail2 increases the proportion of

cells that ingress (Figures 4E–4H and Movies S3 and S4). After

ectopic expression of Sox3, electroporated cells still converge

at the primitive streak but are unable to ingress (Figures 4I–4L

and Movies S5 and S6). This observation confirms that Sox3

needs to be downregulated to allow cell ingression at the primi-

tive streak, but does not impair the convergence of epiblastic

Developmental Cell 21, 546–

cells at the primitive streak. If the defect in cell

ingression is mediated by the observed Snail2

downregulation by Sox3 (Figures 3R–3X), pre-

venting Snail2 function should elicit a similar

defect. Indeed, overexpression of DN-Snail2 in

the primitive streak also inhibits cell ingression

(Figures 4M–4P and Movies S7 and S8), which

is compatible with the observed activation of

Sox3 expression (Figures 3L–3Q). In contrast,

loss of Sox3 function by ectopic expression of

DN-Sox3 close to the primitive streak favors

cell ingression (Figures 4Q–4T and Movies S9

and S10), as also observed after Snail2 overex-

pression (Figures 4E–4H). On the other hand,

overexpression of Snail2 together with Sox3

and GFP in the region of endogenous ingression

rescued the phenotype of Sox3 overexpression,

as the cell movements observed resembled

those in the control embryo (Figures 4U–4X and Movies S11

and S12). Altogether, these results strongly suggest that the

decision to ingress at the primitive streak depends on the inter-

actions between Snail2 and Sox3 transcription factors.

Snail2 and Sox3 Are Reciprocal Direct Transcriptional Repressors Next, we investigated whether Snail2 and Sox3 can bind to each

other’s promoter to directly repress expression. We found one

Snail-binding site in sequences located 50 to the Sox3 coding region, which is conserved in chick, mouse, and human (Fig-

ure S1). First, we used a luciferase reporter driven by the 50

558, September 13, 2011 ª2011 Elsevier Inc. 549

A

B

C

D

E

F

Figure 5. Snail2 and Sox3 Are Direct Mutual Tran-

scriptional Repressors

(A) Conserved Snail-binding site in the Sox3 promoter (see

also Figure S1).

(B) Snail2 overexpression in cultured embryos repressed

the activity of the Sox3 promoter, whereas a deletion that

removes the Snail binding site significantly restores

promoter activity.

(C) Chromatin immunoprecipitation (ChIP) analysis

confirms that Snail2 can directly bind to the Sox3

promoter. ChIP assays were carried out with anti-myc

antibodies on embryo extracts that were previously elec-

troporated with either GFP or myc-Snail2. PCR fragments

were obtained after amplification of the indicated promoter

region from the input (1), histone H3 positive control (2),

IgG negative control (3), and myc immunoprecipitated

fractions (4).

(D) Conserved sequence for Sox transcription factors

binding in the Snail2 promoter (see also Figure S1).

(E) Sox3 overexpression in cultured embryos repressed

Snail2 promoter activity, and the deletion of the Sox-

binding site restored most of the promoter activity.

(F) Chromatin immunoprecipitation (ChIP) analyses con-

firmed that Sox3 can directly bind to the Snail2 promoter.

ChIP assays were carried out as described above.

Developmental Cell

Sox3 and Snail Define Embryonic Territories

Sox3 sequences (Figure 5A), to test whether Snail2 can modu-

late the activity of the Sox3 promoter by coelectroporating Snail2

with the luciferase construct in the epiblast of preprimitive streak

stage embryos (see Experimental Procedures). Snail2 reduced

the activity of the Sox3 promoter construct. This repression is

dependent on the presence of the conserved Snail2-binding

box (Figure 5B), because electroporation with a construct

lacking this box significantly restores activity (Figure 5B). To

determine whether Snail2 can bind directly to the Snail box

in vivo, we performed Chromatin immunoprecipitation (ChIP)

experiments after electroporating embryos with GFP and a

myc-tagged version of Snail2. After dissecting the GFP-positive

tissues, we carried out ChIP with an anti-myc antibody. A frag-

ment of genomic DNA corresponding to the Snail-box containing

region of the Sox3 promoter was amplified in embryos electropo-

rated with the myc-tagged Snail2 protein (Figure 5C, compare

line 4 for GFP and myc-Snail2 experiments). We then performed

similar experiments to determine whether Sox3 acts on a similar

way on the Snail2 promoter and found that this is indeed the case

(Figures 5D–5F). The conserved Sox DNA binding sequence

located in the chick Snail2 promoter is shown in Figure 5D, and

its conservation in other species is shown in Figure S1. Thus,

Snail2 and Sox3 can bind directly to each other’s promoter in vivo

at conserved elements, providing a molecular mechanism for

their antagonistic role in the cell decision to ingress at gastrula-

tion. This interaction defines two main territories within the

early embryos: the noningressing ectoderm and the ingressing

mesendoderm.

Snail2 Represses Ectodermal Markers without Inducing Mesodermal Fate In addition to a strong repression of Sox3 expression in the

epiblast, Snail2 is also able to repress other ectodermal markers,

such as the epidermal marker Dlx5 (Figure 6A; 100%; n = 3) and

the neuroectodermal marker Otx2, albeit in a limited territory

550 Developmental Cell 21, 546–558, September 13, 2011 ª2011 Els

(Figure 6B; 62%; n = 8). This is consistent with the known role

of Snail as an epithelial repressor (Nieto, 2002). In contrast,

Sox3 induces Dlx5, although it is unable to activate Otx2 expres-

sion (Figure 6A; 80%; n = 5; Figure 6B; 100%; n = 11).

We had observed that, in addition to repressing epithelial

markers, Snail2 induces delamination of epiblast cells by trig-

gering EMT (Figure 1). We wondered whether this morphological

change is accompanied by the induction of mesendodermal

fate, because primitive streak cells are Snail2 positive and give

rise to mesoderm and endoderm. However, Snail2 electropora-

tion did not induce the expression of the primitive streak and

mesoderm markers Brachyury or Tbx6L in any of the embryos

analyzed (Figures 6C and 6D; n = 5 and n = 8, respectively). Simi-

larly, Snail2 could not induce the expression of the endodermal

marker Sox17 (Figure 6E; 100%; n = 12). In turn, ectopic expres-

sion of Sox3 in the primitive streak dramatically reduces the

expression of mesodermal markers (Figures 6C and 6D; 100%;

n = 8 and n = 9, respectively), as expected from the reduction

observed in cell ingression (Figures 4I–4L). Sox3 is unable to

induce Sox17 (Figure 6E; 100%; n = 9), in agreement with the

endodermal fate being determined in cells after ingression (Ta-

katori et al., 2010). Altogether, our data suggest that Snail and

Sox3 do not behave as mesodermal-endodermal or neural

inducers, respectively, but rather as regulators of cell behavior

and movement. This is in agreement with the phenotype of Snail

mutant mouse embryos, which still form mesoderm but are

unable to migrate because of a defective EMT at the primitive

streak (Carver et al., 2001). It is also compatible with Sox3 being

an early neural marker and expressing cells becoming definitive

neural only later when they express Sox2 (Linker and Stern,

2004). Furthermore, as expected from the antagonistic relation-

ship between Snail2 and Sox3, the effects of Snail2 on Dlx5 and

Otx2 expression depend on Sox3 suppression, because overex-

pression of Snail2 together with Sox3 in the presumptive neural

ectoderm rescued their inhibition (Figure 6F; 67%; n = 6).

evier Inc.

Figure 6. Snail2 Represses Neural and Nonneural

Ectodermal Markers without Inducing Meso-

dermal or Endodermal Fate

(A–E) Snail2 or Sox3 were electroporated together with

GFP in the epiblast of stage HH2 embryos. Snail2

repressed the expression of Dlx5, a nonneural ectodermal

marker (A) and Otx2, an anterior neural marker (B),

whereas Sox3 extended Dlx5 expression (A). Snail2 was

unable to induce the expression of mesodermal markers,

as assessed by examining the expression of the T-box

genes Brachyury and Tbx6L (C and D) or the endodermal

marker Sox17 (E), whereas Sox3 strongly repressed

mesodermal (C and D) but not endodermal markers (E).

Overexpression of both Snail2 and Sox3 did not affect the

expression of the ectodermal markers Dlx5 and Otx2 (F),

suggesting that their repression by Snail2 acts through

Sox3 downregulation.

Developmental Cell

Sox3 and Snail Define Embryonic Territories

These results indicate that, even though the antagonistic rela-

tionship between Sox3 and Snail2 divides the early embryo in

two main territories from which either the ectoderm or the mes-

endoderm will form, it is cell adhesion and behavior rather than

the induction of a change of cell fates that drives the segregation.

Conservation of the Sox3/Snail Antagonistic Relationship in Mouse Embryos Next, we wondered whether our observations in the chick could

be extended to other systems, and we compared the expression

of Sox3 and Snail genes in the mouse gastrula. Snail2 is not ex-

pressed in the primitive streak or early mesoderm in mouse

embryos because of a reshuffling of Snail1 and Snail2 expression

domains during evolution (Sefton et al., 1998; Locascio et al.,

2002). Thus, we compared the expression patterns of Snail1

and Sox3. As previously described, Sox3 is strongly expressed

in the epiblast (Wood and Episkopou, 1999; Figure 7A), and

Snail1 is expressed at the primitive streak and in the mesodermal

cells delaminating from it, but not in the epiblast (Sefton et al.,

1998; Figure 7A). This complementary expression is compatible

with the idea that, in the mouse, Snail1 and Sox3 repress each

other’s transcription. This is also reinforced by the conservation

of the Snail-binding site in the mouse Sox3 promoter and the

Developmental Cell 21, 546–5

presence of two consensus Sox DNA-binding

sequences located in conserved regions of the

mouse Snail1 promoter (Figure S1).

To directly examine the influence of Sox3

on Snail1 expression, we used wild-type cells

(CCE cells; Robertson et al., 1986; Keller et al.,

1993) or Sox3 null mouse ES cells (M. Parsons,

C.W., and R.L.-B., unpublished data) that were

analyzed after 5 or 8 days in the presence or in

the absence of LIF, the latter allowing the

generation of embryoid bodies. Snail1 expres-

sion is activated in Sox3-deficient embryoid

bodies (Figure 7B), suggesting that the absence

of Sox3 leads to a derepression of Snail1

expression. Concomitant with Snail1 activation,

E-cadherin was downregulated (Figure 7B),

indicating the presence of aberrant expression

of Snail1 and downregulation of E-cadherin,

as confirmed by immunohistochemistry (Figure 7C). Further-

more, although embryoid bodies derived from wild-type ES cells

look round and show a compact morphology, those derived from

Sox3 mutant ES cells are irregular. We excluded both the possi-

bility of these cells dying and the existence of differences in the

rate of cell proliferation (Figure S2). Rather, they are healthy cells

that appear to disaggregate, resembling a process of EMT (Fig-

ure 7C). These results pointed to a conserved antagonistic rela-

tionship between Sox3 and Snail1 in the mouse.

Further support for the conservation comes from the observa-

tion of gastrulation defects in chimeras obtained after injection

of Sox3 null ES cells in mouse blastocysts (M. Parsons, C.W.,

and R.L.-B., unpublished data). Sox3 is located on the X chro-

mosome; therefore, targeted XY ES cells are null for Sox3.

However, this gastrulation defect was observed in the context

of chimeric embryos, because ubiquitous deletion of a floxed

allele of Sox3 by bactinCre results in generation of live animals

(Rizzoti et al., 2004). This suggests that the presence of Sox3

null cells in a wild-type host embryo is incompatible with

a compensation mechanism that allows normal gastrulation in

an entirely Sox3 null embryo. The latter is likely to involve

Sox2, which, in contrast to the situation in the chick, shows

similar expression at these stages in the mouse (Avilion et al.,

58, September 13, 2011 ª2011 Elsevier Inc. 551

Figure 7. The Antagonistic Relationship between

Sox3 and Snail Factors Is Conserved in the Mouse

(A) Expression of Snail1 and Sox3 in 7,5 d.p.c. mouse

embryos. Snail1 is the family member expressed in

the primitive streak and early mesendodermal cells in the

mouse (Sefton et al., 1998). Note the similarities in the

expression patterns between chick and mouse embryos.

A mutually exclusive expression pattern is observed in the

transverse sections obtained at the level of the primitive

streak.

(B) Snail1 expression is increased in embryoid bodies (EB)

obtained from Sox3 knockout embryonic stem cells

compared to wild-type stem cells (CCE line). Consistent

with the increase in Snail1 expression, E-cadherin tran-

scripts are downregulated in Sox3 null EBs (T-test statis-

tical analysis).

(C) EBs derived from the CCE line are round and smooth,

whereas those derived from Sox3 null ES cells show

irregular edges with budding cells delaminating from the

EBs, leading to abundant isolated cells in the culture

medium (not shown). E-cadherin is strongly down-

regulated in Sox3�/� compared to wild-type (WT) EBs. Conversely, Snail1 mRNA and protein are detected only in

Sox3�/� EBs, mainly in cells at the edges. (D) Expression of Snail1 in WT and Sox3�/�; Sox2+/�

mouse embryos. Ectopic Snail1 expression can be

observed in the mutant embryos leading to an extended

region of EMT and cell ingression. The dotted lines in the

whole mounted embryos indicate the level of the sections

shown in the lower panels. Red brackets indicate the

areas of cell delamination in the enlarged pictures and in

the diagrams, which schematically show the defective

phenotype in the gastrulating embryos. Dotted lines in the

diagram indicate the midline.

Developmental Cell

Sox3 and Snail Define Embryonic Territories

2003). Therefore, if the antagonistic relationship between Sox3

and Snail is conserved in the mouse as the EB experiments

indicate, then gastrulation defects similar to those we described

here in the chick should be observed after lowering the Sox2

dose in the Sox3 mutant. Because Sox2 null embryos die at

peri-implantation stages, we analyzed Sox3 null; Sox2 heterozy-

gous embryos and indeed, we found ectopic Snail1 expression

and deformed embryos with an extended area of cell delamina-

tion at the primitive streak (Figure 7D; n = 3). These results

support and extend those found in the Sox3 null EBs, indicative

that the antagonistic relationship with Snail is conserved in the

mouse.

Conservation of the Sox3/Snail Antagonistic Relationship in Human Cancer Cells Because Snail factors are also implicated in the repression of

the epithelial phenotype and the induction of EMT during tumor

progression (Thiery et al., 2009), we checked whether human

cancer cell lines also show an antagonistic relationship between

Snail and Sox3 transcription factors. We examined one epithelial

cell line derived from breast carcinoma (MCF7) and three inde-

pendent mesenchymal and invasive lines, two of them also

derived from breast tumors (MDA231 and MDA435) and one

552 Developmental Cell 21, 546–558, September 13, 2011 ª2011 Elsevier Inc.

from a melanoma (A375P). Sox3 is strongly ex-

pressed in MCF7 cells and it is absent from

A375P, MDA231, and MDA435. Conversely,

Snail1 expression is high in MDA231, MDA435, and A375P and

very low in MCF7 cells (Figure 8A). To address whether the

expression of Snail1 and Sox3 is interdependent, we first inter-

fered with Sox3 expression in MCF7 cells by transfecting specific

siRNAs (see Experimental Procedures). We found an efficient

Sox3 downregulation concomitant with an increase in Snail1

expression (Figure 8B) accompanied by the decrease in the

epithelial marker Claudin1, and an increase in the mesenchymal

markers Adam12 and Fibronectin (Figure 8B). We did not

observe a reduction in E-cadherin levels in this transient downre-

gulation of Sox3 expression. Importantly, because MCF7 cells

do not express significant levels of Snail1, we added 2 ng/ml

of its potent inducer TGFb to the cultures. This leads to three-

fold induction of Snail1 expression, which correlates with a

50% decrease of Sox3 transcripts (Figure 8C). Interestingly,

the changes in Sox3 expression were dependent on Snail1,

because Sox3 transcript levels remained unaffected in the

presence of a Snail1-specific siRNA, which prevented Snail1

induction by TGFb (Figure 8C). To examine whether the interde-

pendent changes in gene expression had an impact on cell

morphology and behavior, we infected the mesenchymal

MDA435 cells with a retrovirus containing the Sox3 coding

sequence to generate MDA435-Sox3 cells stably expressing

A B C

D E F

G

H

Figure 8. The Antagonistic Relationship between Sox3 and Snail Factors Is Conserved in Human Cancer Cells (A) Expression of Sox3 and Snail1 in different human cancer cell lines. Epithelial MCF7 cells express high levels of Sox3 and low levels of Snail1. Conversely,

mesenchymal MDA231, MDA435, and A375P lines are almost devoid of Sox3 transcripts and show high Snail1 expression.

(B) Efficient Sox3 downregulation by specific siRNA in MCF7 cells is accompanied by an increase in Snail1 and associated mesenchymal markers Adam12 and

Fibronectin and a decrease in the epithelial marker Claudin1.

(C) TGF-b mediated Snail1 induction in MCF7 cells is accompanied by a decrease in Sox3 expression. This decrease is Snail1-dependent, as it is prevented by

transfection with Snail1 siRNA.

(D and E) Sox3 stable overexpression in MDA435 leads to morphological changes compatible with a partial mesenchymal to epithelial transition (D). These

morphological changes are associated with a decrease in Snail2 expression and an increase in E-cadherin expression (E).

(F) The migratory behavior of MDA435-Sox3 stable transfectants was tested in culture using a wound healing assay. Control MDA435 cells cover the wound in

12 hr, in clear contrast with Sox3-expressing cells.

(G) Invasive behavior of MDA435-Sox3 cells. The nuclei of cells that invaded the collagen matrix were stained with DAPI and quantified. MDA435 cells expressing

Sox3 significantly decreased their invasive properties.

(H) Diagram showing the relationship between Snail and Sox3 factors in gastrulating embryos and in cancer cells. In gastrulating embryos, the mutual repression

between Sox3 and Snail1 regulates the decision to ingress at the primitive streak. Sox3 expressing cells do not ingress, ensuring the development of ectodermal

derivatives in developing embryos. In cancer cells, Snail induces EMT and its expression correlates with invasive properties. Sox3 represses Snail, thereby

preventing EMT to maintain epithelial integrity.

Developmental Cell

Sox3 and Snail Define Embryonic Territories

Developmental Cell 21, 546–558, September 13, 2011 ª2011 Elsevier Inc. 553

Developmental Cell

Sox3 and Snail Define Embryonic Territories

this transcription factor. We observed morphological changes

compatible with a partial mesenchymal-to-epithelial transition

(Figure 8D), concomitant with the decrease in Snail2 expression

and the reactivation of E-cadherin transcription (Figure 8E). We

next examined cell behavior and found that Sox3 induced

a dramatic decrease in cell motility as assessed in a wound

healing assay in culture (Figure 8F) and a decrease cell invasion

in collagen gels (Figure 8G). In summary, the antagonistic rela-

tionship between Snail1 and Sox3 is maintained in cancer cells,

and their relative expression correlates with their morphological,

motility and invasive properties.

DISCUSSION

One of the earliest cellular decisions in metazoans is the subdi-

vision of the early embryo into the domains that will give rise to

the different embryonic layers. In amniote embryos, the first

subdivision occurs at the primitive streak, where ingressing cells

will later become mesoderm and endoderm and the noningress-

ing cells will become ectoderm. Here we show that the partition-

ing of these cellular domains at early primitive streak stages is

regulated by interactions between two transcription factors,

Snail and Sox3, which direct ingressing versus noningressing

behaviors, respectively. In the chick embryo, Snail2 and Sox3

act as mutual transcriptional repressors; cells that express

high levels of Snail2 are devoid of Sox3 transcripts and ingress

through the primitive streak. In contrast, cells expressing high

levels of Sox3 lack Snail2 expression and stay in the epiblast.

Both Sox3 ectopic expression and Snail2 inhibition close to the

primitive streak prevent cell ingression while still permitting the

movement toward the midline. In turn, Snail2 overexpression

or Sox3 inhibition increases cell ingression at the primitive streak.

This indicates that the interplay between Snail2 and Sox3

controls the delamination from the epiblast, thereby ensuring

the subdivision of the embryo into two populations, ingressing

and noningressing, that will later give rise to the mesendoderm

and to much of the ectoderm, respectively. At later stages, the

posterior neural tube arises from a population of bipotent axial

stem cells set aside within the region of the tail bud (Wilson

et al., 2009). These stem cells can become posterior neural tissue

or undergo EMT to become paraxial mesoderm, and their ingres-

sion depends on the repression of Sox2 by Tbx6, as recently

shown in the mouse (Takemoto et al., 2011).

Our data show that Sox3 and Snail behave as direct repres-

sors of each other. Snail is a well-known transcriptional

repressor that controls cell movements both in embryonic devel-

opment and tumor progression (Thiery et al., 2009). Interestingly,

although the genes of the SoxB1 subgroup (Sox1, Sox2, and

Sox3) are usually described as transcriptional activators (Uchi-

kawa et al., 1999), here we show that like the SoxB2 subgroup

genes (Sox14 and Sox21), Sox3 can also function as a transcrip-

tional repressor depending on the context, as also described for

Sox2 during the differentiation of ES cells (Navarro et al., 2008).

We demonstrate that Sox3 directly binds to the promoter of its

target gene, Snail2. This is in agreement with recent findings in

the zebrafish embryo showing that Sox3 can repress the expres-

sion of Bozozok, a homeobox gene important for the formation of

the dorsal organizer and subsequent gastrulation movements

(Shih et al., 2010). We also show that Sox3 represses the activity

554 Developmental Cell 21, 546–558, September 13, 2011 ª2011 Els

of a Snail2 promoter construct in vivo, now confirming that, in

addition to a transcriptional activator, Sox3 can be considered

as a bona fide transcriptional repressor, depending on context.

Our description of the involvement of Sox3/Snail interactions

in defining embryonic territories in the chick is consistent with

the gastrulation defects observed in chimeras obtained after

injection of Sox3 null ES cells in mouse blastocysts. This gastru-

lation phenotype observed in the chimeras was difficult to

explain, considering that Sox3 is never expressed in the meso-

derm at gastrulation stages. However, our data on the regulation

of Snail1 and Sox3 expression in embryoid bodies obtained from

mouse Sox3 null ES cells and our analysis of gastrulating Sox3

null; Sox2 heterozygous embryos provide a simple explanation

for the gastrulation defects first observed in the chimeras con-

taining Sox3 null ES cells and confirm the conservation of the

interplay between Sox3 and Snail in defining ectodermal and

mesendodermal territories.

Our data further clarify recent data showing multiple defects,

including gastrulation defects, in zebrafish embryos after down-

regulation of the full complement of SoxB1 genes (Okuda et al.,

2010). The downregulation of individual SoxB1 genes does not

give rise to severe defects in the fish, reflecting the overlap in

the expression patterns of several family members, particularly

for Sox3 and the fish-specific Sox19a and 19b (Okuda et al.,

2006). In addition, in the fish embryo, the formation and migration

of mesendoderm does not involve a full EMT, but rather cells very

quickly re-express the homolog of E-cadherin (Cdh1) and

migrate as a cohesive group (Montero et al., 2005). In the fish,

Cdh1 re-expression is necessary for the proper migration of

the mesendodermal cells (Montero et al., 2005) and indeed it

occurs concomitantly with the loss of snail1a, which is only

transiently expressed in the involuting mesoderm (Blanco

et al., 2007). Furthermore, a dominant-negative form of Sox3 in

the zebrafish induces the formation of multiple organizers (Shih

et al., 2010), indicating that in the fish, as we show here in the

amniote embryo, Sox3 needs to be downregulated for gastrula-

tion to proceed normally. Interestingly, both in fish and Xenopus,

gastrulation starts concomitantly with the transient activation of

Snail expression (Blanco et al., 2007; Mayor et al., 2000). Thus,

the antagonism between Sox3 and Snail factors shown here in

the amniote embryo, although not directly examined in zebrafish

or Xenopus, may contribute to the initiation of the gastrulation

process and may therefore be conserved not only in amniotes

but also in anamniote embryos.

A crucial issue that also emerges at gastrulation is the neces-

sary coordination between cell fate and cell behavior. Snail has

been considered as a mesodermal determinant from studies in

Drosophila where it is a repressor of nonmesodermal genes.

However, we would like to argue that the role of Snail is indepen-

dent from cell fate determination, its main role being the regula-

tion of cell behavior. Our data indicate that Snail2 can trigger an

ectopic EMT and cell delamination from the epiblast but is

unable to induce the expression of mesodermal or endodermal

markers. This is in agreement with data from Drosophila indi-

cating that low levels of snail do not repress nonmesodermal

genes in the presumptive mesoderm while still able to promote

invagination (Ip et al., 1994) and it is also consistent with data

from the mouse showing that Snail1-deficient embryos can

form mesodermal tissue that expresses Brachyury and Tbx6

evier Inc.

Developmental Cell

Sox3 and Snail Define Embryonic Territories

but these cells are unable to migrate due to defects in EMT and

the continued expression of E-cadherin (Carver et al., 2001; Mur-

ray and Gridley, 2006). Indeed, fate determination and cell

delamination seem to be two independent processes, both

driven by FGF signaling through FGFR1 in the gastrulating

mouse, exerted on one hand by maintaining the expression of

Snail and on the other by controlling the expression of the meso-

dermal genes Brachyury and Tbx6 (Ciruna and Rossant, 2001).

Thus, our data show that the definition of the two main embry-

onic territories in the early gastrulating embryo, namely the

ectoderm and the mesendoderm, is governed by the control of

cell behavior driven by the antagonistic role between Snail and

Sox3 factors independently from cell specification, which is

concomitantly coordinated by FGF signaling. Recent data from

ascidian embryos indicate that the subsequent decision for

the mesendoderm to subdivide into mesoderm or endoderm is

determined by the asymmetric partitioning of the Not transcrip-

tion factor in the cells destined to become mesodermal, a mech-

anism that is likely to be conserved in vertebrates because Not

has been already described in Xenopus, fish, and chick embryos

(Takatori et al., 2010).

Pioneer work in the chick embryo showed that when a subset

of ectodermal cells is specified to become the nervous system,

the expression of another transcription factor, Churchill, acts

as an important switch, preventing the ingression of prospective

neural plate cells through the anterior part of the primitive streak

from late stage HH4 onward (Sheng et al., 2003). The Sox3/Snail

switch that we describe here to control cell ingression occurs

before neural induction and before the onset of Churchill ex-

pression. The two mechanisms are sequentially implemented in

the embryo, with the Snail/Sox3 axis acting first to ensure that

a subpopulation of ectodermal cells stays in the epiblast. There-

fore, the interplay between Sox3 and Snail controls the first

subdivision of embryonic territories. Subsequently, upon neural

induction, Churchill ensures that the subpopulation of ecto-

dermal cells already specified as neural precursors do not ingress

through the primitive streak to become mesoderm or endoderm.

Finally, our data obtained in human cancer cell lines suggest

that the mutual repression between Sox3 and Snail is also in

place. Epithelial tumor cells express high levels of Sox3 and

low levels of Snail1, and the opposite is true for mesenchymal

tumor cells. Not only are Snail1 and Sox3 expression levels

associated with the morphological and invasive phenotype, but

also interference with Sox3 or Snail1 expression induces recip-

rocal changes in their expression, compatible with the existence

of a loop of mutual repression, as described in embryos. These

data may have important implications in tumor biology, as Snail

reactivation and EMT contributes to the first steps of the meta-

static cascade in carcinomas and it is considered a target of

anti-invasive drugs (Thiery et al., 2009; Nieto, 2011). Therefore,

it is important to identify not only how Snail is reactivated in

tumors but also to identify its negative regulators. We propose

Sox3 as a likely candidate.

EXPERIMENTAL PROCEDURES

Chick Embryos and Explant Cultures

Fertilized hen eggs were purchased from Granja Gilbert (Tarragona, Spain).

The eggs were incubated and opened, and the embryos were explanted for

Developmen

EC culture as described elsewhere (Flamme, 1987; Chapman et al., 2001).

Embryos were staged according to Eyal-Giladi and Kochav (1976) (EG) and

Hamburger and Hamilton (1951) (HH), selecting HH2 embryos for experiments.

Mouse Embryos, ES Cells, and Human Cell Lines

Mouse embryos were obtained by crossing C57 and CBA mice. Embryos

dissected at 7.5 dpc were fixed overnight in 4% paraformaldehyde. CCE

mouse ES cells (Robertson et al., 1986; Keller et al., 1993) were cultured in

cell culture dishes with DMEM (Invitrogen) supplemented with 10% serum

and LIF (1000 U/ml, Chemicon International). Embryoid bodies were grown

on bacteriological grade plastic dishes in the same medium in the absence

of LIF, and total RNA was extracted with Trizol (Life Technologies) after

different times in culture. MCF7, MDA231, MDA435, and A375P human tumor

cell lines were purchased from the ATCC (Virginia, USA) and were cultured in

DMEM supplemented with 10% heat inactivated serum and 0.1% penicillin-

streptomycine (Invitrogen). Cells were transfected with negative control for

siRNAs or with those directed against Snail1 (3 sequences tested) or Sox3

(5 sequences tested) using Lipofectamine RNAiMAX following the manufac-

turer’s instructions (Invitrogen). For RNA/Lipofectamine complex formation,

siRNAs were used at a working concentration of 100 nM. Because they

were able to downregulate expression with different efficiencies, we present

the data obtained with the most efficient oligonucleotide in each case, whose

sequences are shown in Table S1. RNA was isolated 2 days after transfection

for the Sox3 interference experiment. When indicated, 2 ng/ml TGFb was

added to the cells 24 hr after Snail1 siRNA transfection. Total RNA was ob-

tained at 1 hr, 24 hr, and 48 hr after TGFb administration, and in all cases

Snail1 induction was impaired (Figure S3). Total RNA was purified using the

illustra RNAspin Mini kit including DNaseI treatment (GE Healthcare). Stealth

siRNA (Invitrogen) sequences were as follows: 50-UCCCAGAUGAGCAUUGG CAGCGAGG-30 against human Snail1 (SNAI1) and 50-AGUUCCAGGGUU AUUCUGUUACAUU-30 against human SOX3.

Viral Production and Generation of Sox3 Stable Expressing Cells

Retroviral production and infection were carried out as previously described

(Mani et al., 2008). After infection, MDA435 cells expressing either pBabe-

PURO or pBabe-PURO-Sox3 were selected with 10 mg/ml puromycin for

2 weeks.

Migration and Invasion Assays

For migration assays, cells were seeded in six-well culture dishes at a density

of 1 3 106 cells/well. A wound was made in the center of the culture 24 hr later,

and phase-contrast pictures were taken at different time intervals. Invasion

assays on collagen type-IV gels were performed as previously described

(Cano et al., 2000). Briefly, 6 3 104 cells of each type were seeded onto the

upper surface of the filters. After 12 hr of incubation, cells attached in the

lower part of the filters were fixed in methanol, stained with 4,6-diamidinophe-

nylindole (DAPI) and counted.

Electroporation of Chicken Embryos

Explanted embryos at HH2 were placed, vitelline membrane and filter paper

down, over an electroporation chamber (NEPAGEN) containing a platinum

electrode connected to the negative pole. A solution containing expression

plasmids (2 mg/ml in PBS with 0.1% Fastgreen and 6% sucrose) was injected

between the vitelline membrane and the epiblast. An anodal electrode was

placed over the hypoblast to cover the injected area. A train of electric pulses

(5 pulses, 4 Volts, 50 ms, and 0.5 Hz) was applied using an Intracept TSS10

pulse stimulator (Intracell). In all experiments, the nonelectroporated right

side of the embryo was used as a control. The embryos were then cultured

at 38�C (Chapman et al., 2001) to the desired stage. Embryos were photo- graphed with a Leica MZFLIII dissecting microscope to record GFP expression

and fixed overnight in 4% paraformaldehyde (PFA) in PBS at 4�C to be pro- cessed for in situ hybridization or immunohistochemistry. Cell death that might

have resulted from the electroporation procedure was excluded as a factor to

influence cell behavior (Figure S2).

Time-Lapse Confocal Imaging

Six hours after electroporation, chicken embryos at stage HH3+ were washed

in PBS and placed into a glass-bottom culture 35 mm Petri dish (MatTek)

tal Cell 21, 546–558, September 13, 2011 ª2011 Elsevier Inc. 555

Developmental Cell

Sox3 and Snail Define Embryonic Territories

containing egg albumin. The dish was then located into an incubation chamber

at 38�C surrounding a Leica inverted confocal microscope for image acquisi- tion. One image was captured each 10 min for a total of 8 hr. Movies were

assembled using the ImageJ software (http://rsbweb.nih.gov/ij/). Individual

cells were tracked using the ‘‘Manual Tracking’’ plug-in by F. Cordelières

(http://rsbweb.nih.gov/ij/plugins/track/track.html). Ingression was quantified

as the percentage of ingressing and noningressing cells after tracking 20 cells

per field in three fields per movie.

DNA Constructs

pCX-Snail2, pCX-DN-Snail2, and pCX-GFP expression vectors were previ-

ously described (Morales et al., 2007). Full-length Sox3 or a truncated domi-

nant negative form of Sox3 similar to that previously described for the Sox2

Xenopus gene (Kishi et al., 2000) were cloned in pCX at the EcoRI restriction

site. Snail2 and Sox3 promoters were PCR amplified from chick genomic

DNA using Phusion High Fidelity DNA polymerase (Finnzyme) (see Table S1

for primers sequences), sequenced, and inserted in pGL2 basic using the

KpnI and MluI restriction sites. For viral production, the coding sequence of

human Sox3 was amplified by PCR (see Table S1 for primers sequences)

and inserted in the pBabe-PURO vector using the EcoRI restriction site.

Whole-Mount In Situ Hybridization

Whole-mount in situ hybridization was carried out as described previously

(Nieto et al., 1996) omitting the proteinase K treatment. Digoxigenin-labeled

probes were synthesized from the full-length chicken cDNAs of Brachyury,

Otx2, RhoB (Liu and Jessell, 1998), and Snail2 (Nieto et al., 1994) and from

Expressed Sequence Tags (EST; Boardman et al., 2002) for Dlx5

(ChEST808h7), Tbx6L (ChEST90h8), and Sox17 (pgr1n.pk001.g24; Chapman

et al., 2007). Chicken and mouse Sox3 sequences were PCR amplified from

chicken and mouse genomic DNA (Table S1) and cloned in pGEMT-easy.

Mouse Snail1 probe was previously described (Sefton et al., 1998). Hybridized

probes were detected using an alkaline phosphatase-conjugated anti-digoxi-

genin antibody (Roche, 1:1000) in the presence of NBT/BCIP substrates

(Roche). For whole-mount fluorescent in situ hybridization, embryos were

processed as previously described (Acloque et al., 2008). Briefly, probes

were labeled using digoxigenin- or fluorescein-coupled nucleotides (Roche,

1:1000) and were sequentially developed with POD-conjugated anti-fluores-

cein or anti-digoxigenin antibodies (Roche). Peroxidase activity was succes-

sively detected with the TSA-plus Cy3 and Fluorescein kits (Perkin Elmer). In

some cases, the embryos were subjected to immunostaining with anti-GFP

antibody (Invitrogen, 1:1000). After hybridization and/or immunohistochem-

istry, embryos were fixed in 4% paraformaldehyde in PBS, washed, and pho-

tographed under a Leica M10 dissecting scope. Some embryos were subse-

quently embedded in paraffin (Fibrowax) or gelatin, sectioned at 10 mm or

40 mm, respectively, and photographed using a Leica DMR microscope under

Nomarski optics and equipped with an Olympus DP70 digital camera.

Immunohistochemistry

For immunohistochemistry, electroporated embryos were fixed in PFA 4% in

PBS. For laminin detection, 10 mm cryostat sections were treated with 0.1%

Triton X-100 (Sigma) in PBS, blocked with 10% FBS in PBS and incubated

overnight at 4�C with anti-laminin (primary antibody (Sigma) at 1:1000 dilution. For E-cadherin detection, chick embryos were embedded in paraffin and

sectioned at 8 mm. Immunostaining was performed by standard procedures

using anti-GFP antibody (rabbit polyclonal, Invitrogen; 1:500) and anti-E-cad-

herin (mouse monoclonal, BD Bioscience; 1:250). After washing, sections

were incubated for 1 hr with Alexa488 (Invitrogen, 1:1000) and Cy3 conjugated

(Jackson; 1:1000) secondary antibodies and photographed using a Leica DMR

microscope.

Embryoid bodies were fixed with ice-cold methanol, rehydrated, and immu-

nostained by standard procedures using anti-E-cadherin (mouse monoclonal

ECCD-2, Takara; 1:250) or anti-Snail1 (Abcam). Images were acquired using

a Leica inverted confocal microscope.

Chromatin Immunoprecipitation

Chicken embryos were electroporated either with GFP and control myc-Tag,

GFP and myc-Snail2, or GFP and myc-Sox3 expression plasmids. Eight hours

after electroporation, GFP-positive tissues were dissected from HH5 embryos.

556 Developmental Cell 21, 546–558, September 13, 2011 ª2011 Els

Tissues were crosslinked with 1% formaldehyde in PBS for 10 min at room

temperature and quenched with 0.125 M glycine for 5 min at room tempera-

ture. Tissues were then washed three times in PBS and resuspended in SDS

lysis buffer (1% SDS, 10 mM EDTA, and 50 mM Tris [pH 8.1]) using 100 ml of

buffer for a pool of 15 embryos (corresponding to approximately 1 3 105 cells).

Lysates were sonicated in an ultrasonic cell disrupter (Bioruptor, Diagenode

SA, Belgium) for 8 min, with alternating 30 s off and on, frozen in liquid nitrogen

and stored at �80�C. Each sonicated lysate (about 1 3 105 cells) was diluted in 900 ml of buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM

Tris [pH 8.1], and 167 mM NaCl) in the presence of protease inhibitors. Ten

microliters was recovered as the input fraction, and the rest was divided and

incubated overnight at 4�C with anti-myc ChIP grade (ab9132, Abcam, UK), anti-H3 ChIP grade (ab1791, Abcam, UK), or rabbit IgG control (Diagenode,

Belgium) using 1 mg of antibody for tissue lysate. Immunoprecipitation of

crosslinked Protein/DNA was performed adding 60 ml of a slurry of Protein

A Agarose beads (Roche) previously saturated with salmon sperm DNA

(1 mg/ml) and BSA (1 mg/ml). Complexes were washed using Low Salt

Washing Solution (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris

[pH 8.1], and 150 mM NaCl), High Salt Washing Solution (0.1% SDS, 1% Triton

X-100, 2 mM EDTA, 20 mM Tris [pH 8.1], and 500 mM NaCl), LiCl Washing

Solution (0.25 M LiCl, 1% IGEPAL, 1% deoxycholic acid, 1 mM EDTA, and

10 mM Tris [pH 8.1]), and two times TE Solution (10 mM Tris [pH 8.1] and

1 mM EDTA). Protein/DNA complexes were then eluted in 200 ml of elution

buffer (1% SDS and 100 mM NaHCO3) and dissociated by incubating the

samples at 65�C for 5 hr in the presence of 200 mM NaCl (added to the elution buffer). DNA was then purified using affinity columns and amplified by PCR and

real-time PCR using H3 samples as a reference.

PCR and Real-Time PCR

DNA obtained from the ChIP experiments was amplified using QPCR

PromSox3 SnailRE and QPCR PromSnail2 SoxRE primers (see Table S1 for

sequences). Reverse transcription was performed using random priming and

Superscript Reverse Transcriptase (Life Technologies), according to the

manufacturer’s guidelines. Real-time PCRs were performed using Absolute

SYBR Green mix (Thermo Scientific) in a Step One Plus machine or ABI PRISM

7500 thermocycler (Applied Biosystems). MCF7, MDA231, and MDA435

breast cancer and A375P melanoma cell lines cDNAs were amplified to

examine human Sox3 and Snail1 expression, using 36B4 as internal control

(Côme et al., 2006) and applying relative quantification using the 2-DDCt

method. For embryoid bodies variations in input RNA were corrected by sub-

tracting the number of PCR cycles obtained for b-actin. All primers sequences

are described in Table S1.

Luciferase Assays

Chicken embryos were electroporated with pRL-CMV (Promega) as an internal

control and either pGL2b (Promega), pGL2-promSox3, pGL2 delpromSox3,

pGL2-promSnail2, or pGL2 delpromSnail2, in the presence or absence of

pCX-Snail2 (for Sox3 promoter experiments) or pCX-Sox3 (for Snail2 promoter

experiments). For each assay, three electroporated embryos were pooled to

get one measurement and experiments were made in triplicate (9 embryos

per experiment). Tissues were lysed using Passive lysis buffer (Promega)

and activity measured with Dual luciferase assays (Promega) using a Berthold

luminometer.

Statistical Analysis

In figures including statistical analyses, the values represent mean + SD of

three independent experiments (ANOVA analysis, * p < 0.1, ** p < 0.01,

*** p < 0.001).

SUPPLEMENTAL INFORMATION

Supplemental Information includes three figures, one table, and twelve movies

and can be found with this article online at doi:10.1016/j.devcel.2011.07.005.

ACKNOWLEDGMENTS

We are very grateful to other laboratory members for helpful discussions and to

Susan Chapman for providing the Sox17 clone. Hervé Acloque has performed

evier Inc.

Developmental Cell

Sox3 and Snail Define Embryonic Territories

the majority of the experiments in this work. Oscar Ocaña has contributed to

the analysis of embryos in Figure 7 and performed the experiments shown in

Figure 8. Ander Matheu, Karine Rizzoti, and Clare Wise have contributed to

Figure 7. Work in the two laboratories is supported by grants from the Spanish

Ministry of Science and Innovation (BFU2008-01042, CONSOLIDER-INGENIO

2010 CSD2007-00017, and CSD2007-00023 to M.A.N.) and the Generalitat

Valenciana (Prometeo 2008/049 to M.A.N.), and by the UK Medical Research

Council (U117512772 to R.L.-B.).

Received: November 29, 2010

Revised: May 19, 2011

Accepted: July 10, 2011

Published online: September 12, 2011

REFERENCES

Acloque, H., Wilkinson, D.G., and Nieto, M.A. (2008). In situ hybridization anal-

ysis of chick embryos in whole-mount and tissue sections. Methods Cell Biol.

87, 169–185.

Acloque, H., Adams, M.S., Fishwick, K., Bronner-Fraser, M., and Nieto, M.A.

(2009). Epithelial-mesenchymal transitions: the importance of changing cell

state in development and disease. J. Clin. Invest. 119, 1438–1449.

Avilion, A.A., Nicolis, S.K., Pevny, L.H., Perez, L., Vivian, N., and Lovell-Badge,

R. (2003). Multipotent cell lineages in early mouse development depend on

SOX2 function. Genes Dev. 17, 126–140.

Barrallo-Gimeno, A., and Nieto, M.A. (2005). The Snail genes as inducers of

cell movement and survival: implications in development and cancer.

Development 132, 3151–3161.

Bertocchini, F., and Stern, C.D. (2002). The hypoblast of the chick embryo

positions the primitive streak by antagonizing nodal signaling. Dev. Cell 3,

735–744.

Blanco, M.J., Barrallo-Gimeno, A., Acloque, H., Reyes, A.E., Tada, M., Allende,

M.L., Mayor, R., and Nieto, M.A. (2007). Snail1a and Snail1b cooperate in the

anterior migration of the axial mesendoderm in the zebrafish embryo.

Development 134, 4073–4081.

Boardman, P.E., Sanz-Ezquerro, J., Overton, I.M., Burt, D.W., Bosch, E., Fong,

W.T., Tickle, C., Brown, W.R., Wilson, S.A., and Hubbard, S.J. (2002). A

comprehensive collection of chicken cDNAs. Curr. Biol. 12, 1965–1969.

Brennan, J., Lu, C.C., Norris, D.P., Rodriguez, T.A., Beddington, R.S., and

Robertson, E.J. (2001). Nodal signalling in the epiblast patterns the early

mouse embryo. Nature 411, 965–969.

Cano, A., Pérez-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 expres-

sion. Nat. Cell Biol. 2, 76–83.

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.

Chapman, S.C., Collignon, J., Schoenwolf, G.C., and Lumsden, A. (2001).

Improved method for chick whole-embryo culture using a filter paper carrier.

Dev. Dyn. 220, 284–289.

Chapman, S.C., Matsumoto, K., Cai, Q., and Schoenwolf, G.C. (2007).

Specification of germ layer identity in the chick gastrula. BMC Dev. Biol. 7, 91.

Ciruna, B., and Rossant, J. (2001). FGF signaling regulates mesoderm cell fate

specification and morphogenetic movement at the primitive streak. Dev. Cell

1, 37–49.

Côme, C., Magnino, F., Bibeau, F., De Santa Barbara, P., Becker, K.F., Theillet,

C., and Savagner, P. (2006). Snail and slug play distinct roles during breast

carcinoma progression. Clin. Cancer Res. 12, 5395–5402.

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.

Eyal-Giladi, H., and Kochav, S. (1976). From cleavage to primitive streak

formation: a complementary normal table and a new look at the first stages

Developmen

of the development of the chick. I. General morphology. Dev. Biol. 49,

321–337.

Flamme, I. (1987). Prolonged and simplified in vitro culture of explanted chick

embryos. Anat. Embryol. (Berl.) 176, 45–52.

Fraser, S.E., and Stern, C.D. (2004). Early rostrocaudal patterning of the meso-

derm and neural plate. In Gastrulation: From Cells to Embryo, C.D. Stern, ed.

(Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press), pp. 419–432.

Hamburger, V., and Hamilton, H. (1951). A series of normal stages in the devel-

opment of the chick embryo. J. Morphol. 88, 49–92.

Harrisson, F., Callebaut, M., and Vakaet, L. (1991). Features of polyingression

and primitive streak ingression through the basal lamina in the chicken blasto-

derm. Anat. Rec. 229, 369–383.

Ip, Y.T., Maggert, K., and Levine, M. (1994). Uncoupling gastrulation and

mesoderm differentiation in the Drosophila embryo. EMBO J. 13, 5826–5834.

Keller, G., Kennedy, M., Papayannopoulou, T., and Wiles, M.V. (1993).

Hematopoietic commitment during embryonic stem cell differentiation in

culture. Mol. Cell. Biol. 13, 473–486.

Kishi, M., Mizuseki, K., Sasai, N., Yamazaki, H., Shiota, K., Nakanishi, S., and

Sasai, Y. (2000). Requirement of Sox2-mediated signaling for differentiation of

early Xenopus neuroectoderm. Development 127, 791–800.

Linker, C., and Stern, C.D. (2004). Neural induction requires BMP inhibition

only as a late step, and involves signals other than FGF and Wnt antagonists.

Development 131, 5671–5681.

Liu, J.P., and Jessell, T.M. (1998). A role for rhoB in the delamination of neural

crest cells from the dorsal neural tube. Development 125, 5055–5067.

Locascio, A., Manzanares, M., Blanco, M.J., and Nieto, M.A. (2002).

Modularity and reshuffling of Snail and Slug expression during vertebrate

evolution. Proc. Natl. Acad. Sci. USA 99, 16841–16846.

Mani, S.A., Guo, W., Liao, M.J., Eaton, E.N., Ayyanan, A., Zhou, A.Y., Brooks,

M., Reinhard, F., Zhang, C.C., Shipitsin, M., et al. (2008). The epithelial-mesen-

chymal transition generates cells with properties of stem cells. Cell 133,

704–715.

Mayor, R., Guerrero, N., Young, R.M., Gomez-Skarmeta, J.L., and Cuellar, C.

(2000). A novel function for the Xslug gene: control of dorsal mesendoderm

development by repressing BMP-4. Mech. Dev. 97, 47–56.

Montero, J.A., Carvalho, L., Wilsch-Bräuninger, M., Kilian, B., Mustafa, C., and

Heisenberg, C.P. (2005). Shield formation at the onset of zebrafish gastrula-

tion. Development 132, 1187–1198.

Morales, A.V., Acloque, H., Ocaña, O.H., de Frutos, C.A., Gold, V., and Nieto,

M.A. (2007). Snail genes at the crossroads of symmetric and asymmetric

processes in the developing mesoderm. EMBO Rep. 8, 104–109.

Murray, S.A., and Gridley, T. (2006). Snail1 gene function during early embryo

patterning in mice. Cell Cycle 5, 2566–2570.

Navarro, P., Chambers, I., Karwacki-Neisius, V., Chureau, C., Morey, C.,

Rougeulle, C., and Avner, P. (2008). Molecular coupling of Xist regulation

and pluripotency. Science 321, 1693–1695.

Nieto, M.A. (2002). The snail superfamily of zinc-finger transcription factors.

Nat. Rev. Mol. Cell Biol. 3, 155–166.

Nieto, M.A. (2011). The ins and outs of the epithelial to mesenchymal transition

in health and disease. Annu. Rev. Cell Dev. Biol. 27, in press. Published online

July 6, 2011. 10.1146/annurev-cellbio-092910-154036.

Nieto, M.A., Sargent, M.G., Wilkinson, D.G., and Cooke, J. (1994). Control of

cell behavior during vertebrate development by Slug, a zinc finger gene.

Science 264, 835–839.

Nieto, M.A., Patel, K., and Wilkinson, D.G. (1996). In situ hybridization analysis

of chick embryos in whole mount and tissue sections. Methods Cell Biol. 51,

219–235.

Okuda, Y., Yoda, H., Uchikawa, M., Furutani-Seiki, M., Takeda, H., Kondoh,

H., and Kamachi, Y. (2006). Comparative genomic and expression analysis

of group B1 sox genes in zebrafish indicates their diversification during

vertebrate evolution. Dev. Dyn. 235, 811–825.

tal Cell 21, 546–558, September 13, 2011 ª2011 Elsevier Inc. 557

Developmental Cell

Sox3 and Snail Define Embryonic Territories

Okuda, Y., Ogura, E., Kondoh, H., and Kamachi, Y. (2010). B1 SOX coordinate

cell specification with patterning and morphogenesis in the early zebrafish

embryo. PLoS Genet. 6, e1000936.

Rex, M., Orme, A., Uwanogho, D., Tointon, K., Wigmore, P.M., Sharpe, P.T.,

and Scotting, P.J. (1997). Dynamic expression of chicken Sox2 and Sox3

genes in ectoderm induced to form neural tissue. Dev. Dyn. 209, 323–332.

Rizzoti, K., Brunelli, S., Carmignac, D., Thomas, P.Q., Robinson, I.C., and

Lovell-Badge, R. (2004). SOX3 is required during the formation of the hypothal-

amo-pituitary axis. Nat. Genet. 36, 247–255.

Robertson, E., Bradley, A., Kuehn, M., and Evans, M. (1986). Germ-line trans-

mission of genes introduced into cultured pluripotential cells by retroviral

vector. Nature 323, 445–448.

Sefton, M., Sánchez, S., and Nieto, M.A. (1998). Conserved and divergent

roles for members of the Snail family of transcription factors in the chick and

mouse embryo. Development 125, 3111–3121.

Sheng, G., dos Reis, M., and Stern, C.D. (2003). Churchill, a zinc finger

transcriptional activator, regulates the transition between gastrulation and

neurulation. Cell 115, 603–613.

Shih, Y.H., Kuo, C.L., Hirst, C.S., Dee, C.T., Liu, Y.R., Laghari, Z.A., and

Scotting, P.J. (2010). SoxB1 transcription factors restrict organizer gene

expression by repressing multiple events downstream of Wnt signalling.

Development 137, 2671–2681.

Stern, C.D. (2004). Gastrulation in the chick. In Gastrulation: From Cells to

Embryo, C.D. Stern, ed. (Cold Spring Harbor, NY: Cold Spring Harbor

Laboratory Press), pp. 419–432.

558 Developmental Cell 21, 546–558, September 13, 2011 ª2011 Els

Stern, C.D., and Canning, D.R. (1990). Origin of cells giving rise to mesoderm

and endoderm in chick embryo. Nature 343, 273–275.

Takatori, N., Kumano, G., Saiga, H., and Nishida, H. (2010). Segregation of

germ layer fates by nuclear migration-dependent localization of Not mRNA.

Dev. Cell 19, 589–598.

Takemoto, T., Uchikawa, M., Yoshida, M., Bell, D.M., Lovell-Badge, R.,

Papaioannou, V.E., and Kondoh, H. (2011). Tbx6-dependent Sox2 regulation

determines neural or mesodermal fate in axial stem cells. Nature 470, 394–398.

Thiery, J.P., Acloque, H., Huang, R.Y., and Nieto, M.A. (2009). Epithelial-

mesenchymal transitions in development and disease. Cell 139, 871–890.

Uchikawa, M., Kamachi, Y., and Kondoh, H. (1999). Two distinct subgroups of

Group B Sox genes for transcriptional activators and repressors: their expres-

sion during embryonic organogenesis of the chicken. Mech. Dev. 84, 103–120.

Voiculescu, O., Bertocchini, F., Wolpert, L., Keller, R.E., and Stern, C.D. (2007).

The amniote primitive streak is defined by epithelial cell intercalation before

gastrulation. Nature 449, 1049–1052.

Wakamatsu, Y., Endo, Y., Osumi, N., and Weston, J.A. (2004). Multiple roles of

Sox2, an HMG-box transcription factor in avian neural crest development.

Dev. Dyn. 229, 74–86.

Wilson, V., Olivera-Martinez, I., and Storey, K.G. (2009). Stem cells, signals and

vertebrate body axis extension. Development 136, 1591–1604.

Wood, H.B., and Episkopou, V. (1999). Comparative expression of the mouse

Sox1, Sox2 and Sox3 genes from pre-gastrulation to early somite stages.

Mech. Dev. 86, 197–201.

evier Inc.