the caenorhabditis elegans gene lin-26 can trigger ...worms.zoology.wisc.edu/pdfs/quintin.pdf ·...
TRANSCRIPT
Developmental Biology 235, 410–421 (2001)doi:10.1006/dbio.2001.0294, available online at http://www.idealibrary.com on
The Caenorhabditis elegans Gene lin-26Can Trigger Epithelial Differentiationwithout Conferring Tissue Specificity
Sophie Quintin, Gregoire Michaux, Laura McMahon,Anne Gansmuller, and Michel Labouesse1
Institut de Genetique et de Biologie Moleculaire et Cellulaire, CNRS/INSERM/ULP,B.P. 163, F-67404 Illkirch Cedex, C.U. de Strasbourg, France
How epithelial cell fates become specified is poorly understood. We have previously shown that the putative C2H2zinc-finger transcription factor LIN-26 is required for the differentiation of ectodermal and mesodermal epithelial cells inCaenorhabditis elegans. Here, we report that ectopic LIN-26 expression during early gastrulation transforms mostblastomeres into epithelial-like cells. Specifically, LIN-26 induced the expression of three epithelial markers: the adherensjunction protein JAM-1; DLG-1, which is essential for the assembly of JAM-1 at junctions; and CHE-14, which is involvedin apical trafficking. Furthermore, ultrastructural studies revealed that ectopic LIN-26 expression induced the formation ofadherens-like junctions. However, ectopic lin-26 expression did not confer any tissue-specific cell fate, such as theepidermal cell fate, as evidenced from the observation that several epidermal-specific genes were not induced. Conversely,we show that epidermal cells displayed some polarity defects in lin-26 mutants. We conclude that lin-26 can induceepithelial differentiation and that epitheliogenesis is not a default pathway in C. elegans. © 2001 Academic Press
Key Words: C. elegans; epithelial cell; cell fate specification; adherens junction; apical trafficking; cell polarity; epidermis;uterus; support cell; lin-26.
ettpc
tgcdlibee(
INTRODUCTION
Despite the importance and widespread occurrence ofepithelial cells, the mechanisms that establish epithelialcell identity remain largely unknown. This contrasts forinstance with the genetic programs that control myogenesisand neurogenesis, which have been described at the mo-lecular level (Arnold and Braun, 2000; Brunet and Ghysen,1999). Two key features make epithelial cells unique. First,they originate from all three embryonic germ layers,whereas muscles or neurons come from a single germ layer.Second, some cells only transiently acquire epithelial char-acteristics before they become mesenchymal (Birchmeierand Birchmeier, 1993), whereas myogenic and neuronalcharacteristics are stable. In addition, cells that perma-nently retain epithelial attributes (e.g., in the epidermis,kidney, and intestine) can become tumorigenic, in whichcase they generally lose their polarised and adhesive prop-
1 To whom correspondence should be addressed. Fax: (33) 3 88 65
32 01. E-mail: [email protected].410
rties (Thiery et al., 1988). The epithelial–mesenchymalransition and the loss of adhesive properties followingumorigenesis highlight the plasticity of the epithelialhenotype, and may reflect that epithelial cell fate specifi-ation involves unusual differentiation pathway(s).It is currently unclear whether the specification of epi-
helial cell fates requires a specific battery of regulatoryenes, whether genes that specify the identity of tissuesontaining epithelial cells are sufficient to induce epithelialifferentiation, or, as it has been suggested, whether epithe-ial differentiation is a default pathway (Frisch, 1997). Thedentification of genes that control epitheliogenesis haseen hampered in part by the lack of markers unique topithelial cells (Davies and Garrod, 1997). Although adher-ns junctions (AJs) are found predominantly in epitheliaGeiger et al., 1992; Gumbiner, 1996; Nose et al., 1988;Yeaman et al., 1999), E-cadherin and a- and b-catenins,which coassemble to form AJs, are not exclusively ex-pressed in epithelial cells. So far, genes whose functionwould be devoted to the regulation of epithelial-specific
factors have not been described. In support of the notion0012-1606/01 $35.00Copyright © 2001 by Academic Press
All rights of reproduction in any form reserved.
2ehe
ec1
imtetLecccw
pimt
411Epithelial Differentiation in C. elegans
that epithelial differentiation is integrated within the ge-netic program that controls organ specification is the ob-servation that the WT1 protein, which is essential forkidney development but also acts in a few nonepithelialtissues, can bind the E-cadherin promoter (Hosono et al.,000). Likewise, the search for genes that control thexpression of keratinocyte-specific intermediate filamentsas identified several positive factors which do not actxclusively in epithelia (Sinha et al., 2000). The epithelial
cell default-phenotype hypothesis, on the other hand, isalso supported by several observations. First, many genesthat control the expression of junction components orepithelial intermediate filament proteins in vertebratesappear to encode factors that act as negative regulators ofepithelial differentiation. For instance, the adenovirus E1aprotein indirectly turns on the expression of epithelialgenes by counteracting the activity of the general repressorsCtBP and p300/CBP (Frisch, 1994; Grooteclaes and Frisch,2000). Second, the zinc-finger protein Snail, which bindsthe E-cadherin promoter, is preferentially expressed in cellsthat undergo an epithelial–mesenchymal transition (EMT)and is up-regulated in carcinomas (Batlle et al., 2000; Canot al., 2000). The related protein Slug can facilitate EMT byausing the dissociation of desmosomes (Savagner et al.,997).C. elegans has only recently been used as a model to
nvestigate epithelial biology. The intestine and the epider-is (also called hypodermis) are the two epithelial tissues
hat have so far been the most extensively studied in thembryo. Genetic analysis has established that generation ofhese tissues involves two classes of genes (for a review, seeabouesse and Mango, 1999). In the epidermis, the genelt-1, which encodes a GATA factor, specifies epidermalell identity. (Note that there are two classes of epidermalells: cells from the so-called “major epidermis” whichover the body, and cells from the “minor epidermis”hich cover the tip of the head and tail; elt-1 acts only in
the former.) elt-1 is first expressed at the onset of gastrula-tion and its inactivation leads to a complete absence of themajor epidermis (Page et al., 1997). Two likely elt-1 targetsare elt-3, which encodes another GATA factor almostexclusively found in a subset of epidermal cells, and lin-26,which encodes a putative zinc-finger transcription factorexpressed in ectodermal and mesodermal epithelial cells(Chanal and Labouesse, 1997; Gilleard and McGhee, 2001;Gilleard et al., 1999; Labouesse et al., 1994, 1996; Page etal., 1997). Inactivation of elt-3 does not appear to impairepidermal development (Gilleard and McGhee, 2001; Gil-leard et al., 1999). In contrast, mutations in lin-26 lead toembryonic lethality due to the degeneration of ectodermalepithelial cells, and to sterility due to malformation of thesomatic gonad epithelium (den Boer et al., 1998; Labouesseet al., 1994, 1996). How lin-26 acts is not known. Forexample, in the epidermis, lin-26 could act by repressingthe expression of neuronal genes or by activating theexpression of epidermal genes.
To address this issue, we analysed the consequences of
Copyright © 2001 by Academic Press. All right
ectopic lin-26 expression during neuronal differentiationand in noncommitted cells by monitoring the expression ofseveral epithelial-specific and epidermal-specific markers.In particular, we examined the expression of the genesjam-1, dlg-1, and che-14, which are exclusively expressed inepithelial cells in C. elegans and characterise two aspects ofthe epithelial polarised phenotype (junction and polarisedtrafficking). The gene jam-1 encodes a protein of unknownfunction which is associated with adherens junctions (Fran-cis and Waterston, 1991; Mohler et al., 1998). dlg-1 isrequired to aggregate JAM-1 at adherens junctions andencodes a Lethal Discs Large homologue (McMahon et al.,submitted). che-14 encodes an apical multipass transmem-brane protein that has recently been implicated in apicaltrafficking (Michaux et al., 2000). In contrast to jam-1 anddlg-1, which are found in all epithelial cells, che-14 expres-sion is restricted to the nonneuronal ectoderm (epidermalcells and support cells). We show that ectopic lin-26 expres-sion during a specific time window in embryogenesis canconfer a nonepidermal epithelial-like identity to all blas-tomeres. In parallel, we analyse the expression and subcel-lular localisation of the JAM-1, DLG-1, and CHE-14 pro-teins in lin-26-deficient embryos. We propose that lin-26functions to promote epithelial differentiation and acts,directly or indirectly, on some of the genes that play acritical role in organising structures or processes that con-tribute to the polarised epithelial phenotype.
MATERIALS AND METHODS
Constructs and Transgenic Animals
The lin-26B cDNA from pML500 (Dufourcq et al., 1999) wascloned between the KpnI and SacI sites of the vector pPD49.78carrying the hsp16 heat-shock promoter (gift from A. Fire). Thishs::lin-26 construct (pML501) was coinjected into wild-type ani-mals (N2) at a concentration of 50 ng/ml with the dominant markerRF4 (Mello and Fire, 1995). One extrachromosomal array wasntegrated by x-ray irradiation to generate the alleles mcIs22 and
cIs23, which were outcrossed four times against N2. Similarly,he lin-26B cDNA was cloned between the AccI and SacI sites of
the vector pPD57.56 carrying the mec-3 promoter (gift from A.Fire). Several control constructs were made. The lin-26C splicedvariant, which lacks the second LIN-26 Zn-finger (Dufourcq et al.,1999), was cloned into pPD49.78. Two stop codons were introducedafter the initiation codon of lin-26B in pML501 to generate aprotein with the predicted sequence MLSKFVVVEVSN-SNNT*TLV*. Last, cDNAs for lir-1 isoforms A, C, D, E (Dufourcqet al., 1999) were cloned into pPD49.78. Transgenic lines wereestablished in N2 animals with these control plasmids.
To examine the expression of markers in various backgrounds,we proceeded as follows. The dpy-7::gfp (gift from I. Johnstone),che-14::gfp (Michaux et al., 2000), and dlg-1::gfp (McMahon et al.,submitted) constructs were coinjected with pMH33 [dpy-20(1)plasmid; Clark et al., 1995] in a dpy-20(e1282);mcIs23 background.Alternatively, the integrated jam-1::gfp (jcIs1; Mohler et al., 1998)and elt-1::gfp (vpIs2; gift from J. Gilleard), and established arrayscarrying dpy-7::gfp or let-502::gfp (Wissmann et al., 1999) were
crossed with mcIs22 or mcIs23 animals.s of reproduction in any form reserved.
ctrfe(A21
w1
sm
lm
eteoa
lFsnhisdrt
L1bMwgBs5coahseTmNm2Istrb
agdsmAp
412 Quintin et al.
To examine whether elt-1 (Page et al., 1997) was required forjam-1 induction after ectopic LIN-26 expression, we generated amcIs23;elt-1(zu180) strain or performed RNA interference againstelt-1 in mcIs23 animals. A double-stranded elt-1 RNA preparationcorresponding to exons 1–3 (625 nucleotides) was injected intohermaphrodites. The phenotype of elt-1(RNAi) embryos was iden-tical to that of elt-1(zu180) mutants by Nomarski or by stainingwith antibodies against LIN-26 or JAM-1. To examine the distri-bution of epithelial markers in lin-26-deficient embryos, we per-formed RNA interference against lin-26 (exons 3–5) (Bosher et al.,1999) in animals carrying che-14::gfp, dlg-1::gfp, or jam-1::gfptransgenes, or stained embryos carrying the null mutation lin-26(mc15) (den Boer et al., 1998) with the antibody MH27 (verysimilar results were obtained with both methods).
Heat-Shock Experiments and AnimalManipulations
Embryos or larvae were placed at 33°C for 25 min, then returnedto 20°C for a minimum of 4 h. In many cases, embryos werecollected after dissection of transgenic mothers and their stage wasdetermined under the microscope before the heat-shock (HS). Toexamine the consequences of ectopic lin-26 expression, we alsorecorded the embryonic lineage of heat-shocked embryos using a4-D microscope (Horner et al., 1998), using a microscope equippedwith a temperature-control device (running water around theobjective and under the slide). Transgenic hs::lin-26 embryos wereonstantly recorded for 7 h starting at the 4-cell stage and throughhe heat shock (30 min at 33°C), which was applied when embryoseached the 28-cell stage. To test whether the progeny of the firstour blastomeres can respond to the effect of ectopic LIN-26, weliminated three cells at the 4-cell stage using a laser microbeamLaser Science) in the jcIs1;mcIs23 and in the jcIs1 backgrounds.fter the operation, embryos were allowed to recover for 1 h at0°C before HS, and jam-1::gfp expression was examined at least0 h after the HS.
Touch response of animals carrying the mec-3::lin-26 constructas tested after gentle touch with an eyelash (Way and Chalfie,989).
Antibody Staining and Electron Microscopy
Immunostaining with polyclonal antibodies against LIN-26(Labouesse et al., 1996), PHA-4 (pharynx and intestine; Horner etal., 1998), or with the monoclonal antibodies MH27 (adherensjunctions; Francis and Waterston, 1991), NE8–4C6 (body wallmuscles; Schnabel, 1994), 1CB4 (intestinal cells; Okamoto andThomson, 1985), and 3NB12 (subset of pharyngeal muscles; Priessand Thomson, 1987) were performed as described (Chanal andLabouesse, 1997). Electron microscopy was performed as described(Legouis et al., 2000) on mcIs23 embryos that had been heathocked 5–7 h prior to embryo processing or on lin-26(mc15)utants that had reached midembryogenesis.
RESULTS
Early Ectopic LIN-26 Expression Prevents OrganFormation
To determine whether lin-26 is a negative or a positive
regulator, we expressed the major lin-26 cDNA isoform pCopyright © 2001 by Academic Press. All right
in-26B under the control of the neuronal-specific promoterec-3 (mec-3::lin-26 construct), which is an early-acting
gene in touch neurons (Way and Chalfie, 1989), and underthe control of the heat-shock responsive promoter hsp16(hs::lin-26 construct) (Stringham et al., 1992). The effects ofctopic lin-26 expression were analysed by assaying theouch response of transgenic mec-3::lin-26 animals, or byxamining the development of animals in which expressionf the hs::lin-26 construct had been induced (see Materialsnd Methods).Two lines of evidence suggest that ectopic expression of
in-26 in neurons does not affect neuronal differentiation.irst, animals carrying the mec-3::lin-26 construct weretill touch-sensitive (data not shown), indicating that toucheuron differentiation was unaffected. Second, inducings::lin-26 expression at the time of neuronal differentiationn embryos (350-cell to 550-cell stages) or in larvae (L1tage) resulted in viable animals that could move normallyespite robust LIN-26 expression in neuroblasts and neu-ons (data not shown). We conclude that lin-26 is unlikelyo function as a repressor of neuronal differentiation.
Using the hs::lin-26 construct, we observed that ectopicIN-26 affects development when it is induced prior to the00-cell stage. C. elegans embryogenesis can schematicallye divided into four periods (for a review, see Labouesse andango, 1999). From fertilisation until the 28-cell stage,hich corresponds to the onset of gastrulation, maternal
enes control patterning and blastomere fate specification.etween the 28- and 100-cell stages, a few zygotic genespecify organ and tissue identities. Between the 100- and50-cell stages, cell differentiation takes place under theontrol of zygotic genes. Lastly, morphogenesis takes placence the embryo has reached the 550-cell stage (Figs. 1And 1B). We observed three phenotypic classes amongeat-shocked embryos (Fig. 1 and Table 1). The mosteverely affected embryos were observed when LIN-26xpression was induced between the 20- and 50-cell stages.hese embryos (denoted class II) arrested with approxi-ately 200 cells and lacked most tissues, as judged byomarski microscopy (Fig. 1D; class II embryos will be theain focus of this work). Embryos heat shocked prior to the
0-cell stage (Fig. 1C) were less severely affected than classI embryos, arresting with about 500 cells. Embryos heathocked between the 50- and 100-cell stages (class III) failedo elongate beyond the comma stage and showed a stronglyeduced pharynx (Fig. 1E). Beyond the 100-cell stage, em-ryos were not affected (Fig. 1F).As shown in Table 1, none of the phenotypes described
bove were detected in control experiments, and, moreenerally, our heat-shock conditions did not affect theevelopment of most control embryos. Specifically, weubmitted wild-type embryos to the same heat-shock treat-ent, we introduced two stop codons close to the lin-26UG, and finally we expressed under the same heat-shockromoter the minor isoform lin-26C that lacks the second
redicted zinc-finger of the LIN-26 protein. In this lasts of reproduction in any form reserved.
oeg
eo
egMHehwwcs
jmToceoigrc
aer
413Epithelial Differentiation in C. elegans
control experiment, we could detect protein expressionusing polyclonal antibodies recognising all isoforms (datanot shown). In parallel, we also showed that ectopic expres-sion of lir-1, which encodes proteins with two zinc fingersclosely related to those of LIN-26 (Dufourcq et al., 1999),does not affect embryonic development (Table 1). Thesecontrols demonstrate that the effects observed in hs::lin-26embryos are entirely specific to LIN-26 activity, and fur-thermore that they require the presence of both predictedLIN-26 zinc-fingers.
To confirm the absence of certain tissues in class II andclass III embryos, we stained them with several tissue-specific antibodies. We found that class II embryos had nobody wall muscles (Fig. 2D), no pharynx, no rectum (Fig.2F), and no intestine (Fig. 2H). Likewise, we observed that
FIG. 1. Ectopic lin-26 expression has stronger effects when per-formed during early gastrulation. mcIs23 embryos were stagedunder the microscope, submitted to a heat-shock (HS), and photo-graphed 6–8 h later. Wild-type embryos at (A) midembryogenesisand (B) end of embryogenesis. (C) Representative class I embryo:embryos heat-shocked prior to the 20-cell stage arrested with nomorphogenesis but a normal cell number (see Table 1 for pen-etrance). (D) Representative class II embryo: embryos heat-shockedbetween the 20- and 50-cell stages arrested with half as many cells(see Table 2; note that cells are bigger than in A). (E) Representativeclass III embryo: embryos heat-shocked between the 50- and100-cell stages arrested during elongation with pharyngeal defects(arrows point at where the pharynx basement membrane should be;compare to A). (F) Representative embryo resulting from HS afterthe 100-cell stage; these embryos developed normally to the pretzelstage. In this and Figs. 2, 3, 5, and 6, anterior is to the left, dorsal isup, and the scale bar represents 10 mm.
class III embryos had fewer pharyngeal cells (Fig. 2J). These s
Copyright © 2001 by Academic Press. All right
bservations indicate that cells that should becomendodermal or mesodermal in class II embryos and pharyn-eal in class III embryos have adopted another fate.In summary, LIN-26 induced the strongest effects if
ctopically expressed during early gastrulation, when organr tissue precursors become assigned to their identities.
LIN-26 Expression Promotes an EpithelialDifferentiation Program
To determine the fates adopted by cells in class II em-bryos, we examined whether they could express markersnormally present in tissues that express lin-26 (e.g., supportcells, epidermis, somatic gonad; Table 2). Strikingly, wecould detect the adherens junction protein JAM-1 aroundmost LIN-26-positive cells (Figs. 3E–3H) in a pattern thatwas generally more irregular and punctate than in wild-typeepithelial cells (Figs. 3A–3D). Similarly, we detected ectopicDLG-1::GFP expression in class II embryos in a pattern thatwas quite similar to that observed for JAM-1 (Fig. 3N), andectopic CHE-14::GFP expression at the membrane in manycells in class II embryos (Fig. 3P, Table 2). Ectopic JAM-1expression was also present in class I embryos but in fewercells (data not shown), suggesting that heat-shock effectsare weaker at an early stage. In class III embryos, the JAM-1pattern in the pharynx was discontinuous and abnormal(data not shown), often forming a small clump. It was oftendifficult to determine with certainty whether this stainingresulted from partial suppression of pharyngeal develop-ment or from an ectopic jam-1 induction. We note, how-ver, that, in pha-4 mutant embryos or in embryos homozy-ous for a complete deletion of the pha-4 locus, a similarH27 pattern is observed (Chanal and Labouesse, 1997;orner et al., 1998). Consistent with the observation that
mbryonic patterning and viability are affected only ifs::lin-26 expression is induced prior to the 100-cell stage,e did not observe ectopic expression of epithelial markershen the heat-shock treatment was applied after the 100-
ell stage, despite strong LIN-26 induction (data nothown).The simultaneous overexpression of che-14, dlg-1, and
am-1 after ectopic lin-26 expression strongly suggests thatost cells in class II embryos acquired epithelial characters.o confirm this hypothesis, we examined the ultrastructuref individual cells in class II embryos by electron micros-opy. Indeed, we could recognise distinct junctions both inxternal and internal cells (Figs. 4B and 4C), which wereften present at multiple positions along the membranenstead of a unique subapical position. This feature, to-ether with the fact that CHE-14::GFP was detected in aather uniform pattern around cells, indicates that cells inlass II embryos might have an enlarged apical domain.The previous observations suggest that most cells have
dopted an epithelial-like cell fate in class II hs::lin-26mbryos. To test whether all blastomeres can equallyespond to LIN-26, we asked whether this reprogramming
till occurs when each of the first four blastomeres (ABa,s of reproduction in any form reserved.
stsbeGpsatnentpippt
ac
etcaMe
414 Quintin et al.
ABp, EMS, P2) is isolated. Using a laser microbeam, weeliminated three blastomeres at the 4-cell stage in a straincarrying integrated jam-1::gfp and hs::lin-26 constructs, andubsequently induced lin-26 expression by a heat-shock at aime corresponding to the normal 20- to 50-cell stage. Ashown in Fig. 5, we observed that each of the first fourlastomeres was competent to generate ectopic jam-1-xpressing cells upon ectopic lin-26 expression. The extra-FP seen in these partial embryos was scattered andresent in a large proportion of cells, whereas it was onlyeen in a restricted area in corresponding controls. Inddition, it was always more similar to the JAM-1 distribu-ion observed in normal epidermal cells (see Fig. 3A) than inormal intestinal cells (see Fig. 3C), even in isolated EMSmbryos which normally produce pharynx and gut. Weoticed that the P2 blastomere did not respond as stronglyo ectopic LIN-26, since only 53% of isolated P2 embryosresented ectopic JAM-1::GFP, compared to 92% concern-ng EMS embryos (Fig. 5). This could be due to specificroperties of P2, which generates the germline and isrotected against somatic transcription by the transcrip-ional repressor pie-1 (Seydoux et al., 1996). Zhu and
coworkers had noticed a similar phenomenon for the P2
blastomere when attempting to ectopically express end-1,which specifies intestine identity (Zhu et al., 1998). Weconclude that all blastomeres at the 4-cell stage can bereprogrammed to express epithelial markers after induction
TABLE 1Expression of the Full-Length LIN-26B Protein Is Essential to Obs
Construct Heat-shock conditi
no (N2 strain) nono (N2 strain) Within 1 h after egg layinhs::lin-26B(stop)a Within 1 h after egg layinhs::lin-26Ca Within 1 h after egg layinhs::lin-26Ba Within 1 h after egg layinhs::lin-26Bf Within 1 h after egg layinhs::lin-26Bg Embryos with ,20 cellshs::lin-26Bg Embryos with .20 cells, ,hs::lin-26Bg Embryos with .50 cells, ,hs::lin-26Bg Embryos with .100 cellshs::lir-1Aa,h Within 1 h after egg layin
a At least two independent transgenic lines carrying extrachromobehaved similarly.
b Animals were allowed to lay eggs for 1 h, then transferred to aafter removal of mothers. Most embryos had on average between 2
c Experiments in which heat-shock was performed on precisely sare the most sensitive to heat-shock treatment. Among all embry2-fold stage, and the others hatched and developed to normal ferti
d Among the 91 embryos examined, 1 arrested prior to the comme Among the 149 embryos examined, 2 arrested prior to the comf These animals carried the integrated array mcIs23. Animals cag Experiments in which heat-shock was performed on preciselyh None of the other lir-1 isoforms gave a phenotype.
of LIN-26 at the gastrulation stage. i
Copyright © 2001 by Academic Press. All right
In summary, ectopic lin-26 expression between the 20-nd 50-cell stages transforms most cells into epithelial-likeells.
LIN-26 Does Not Confer Epidermal Identity
The apparent epithelial transformation that was observedin class II embryos and the observation that the JAM-1pattern in isolated EMS embryos was epidermal-like raisethe possibility that early blastomeres have adopted the fatesof epidermal precursors. To assay this possibility, we firstexamined whether the class II phenotype depends on theactivity of elt-1, which specifies epidermis identity (Page etal., 1997). When elt-1 is inactive, embryos fail to expressjam-1 and lin-26 in their “major epidermis” (Fig. 3I). Westill observed ectopic JAM-1 expression after ectopicLIN-26 expression, both in internal and external cells ofelt-1-deficient embryos (Fig. 3J). Thus, ectopic LIN-26 ex-pression can bypass the requirement for elt-1 to inducejam-1 expression.
To test more directly whether ectopic lin-26 conferspidermal identity, we examined the expression of markershat are exclusively expressed in the epidermis, namely theuticle collagen dpy-7 (Gilleard et al., 1997), and the musclettachment antigens recognised by the antibodies MH4 andH5 (Francis and Waterston, 1991). We also examined the
xpression of let-502, which controls actin reorganisation
a Reprogramming of Early Blastomeres
% embryonic lethality n
,1 5006c 397
15d 918e 149
72 11488 13087 95
cells 81 91cells 55 85
20 609 223
l arrays were examined. Result for only one line is given; the other
plate. Eggs were submitted to heat-shock treatment immediatelyd 50 cells.
embryos indicate that wild-type embryos with less than 10 cellsamined, 1 arrested at the comma stage, 30 developed beyond theults.age, 13 arrested after the 2-fold stage, and the others hatched.tage, 10 arrested at the pretzel stage, and the others hatched.g the integrated array mcIs22 behaved similarly in all our tests.d embryos in the integrated mcIs23 background (see Fig. 1).
erve
ons
gb
gb
gb
gb
gb
50100
gb
soma
new0 antagedos exle ad
a stma srryinstage
n the epidermis and the somatic gonad (Wissmann et al.,
s of reproduction in any form reserved.
ioere
DJCDLE
w
415Epithelial Differentiation in C. elegans
1999), and of elt-1. We found that these markers wereexpressed in approximately the same number of cells rela-tive to controls and in much fewer cells than jam-1 (Table2, Fig. 6, and data not shown).
Together, our results demonstrate that lin-26 does notnduce epidermal differentiation. Furthermore, they ruleut an alternative explanation for the effects observed afterctopic lin-26 expression, which would be that lin-26 up-egulates the major regulator of epidermal development,
FIG. 2. Ectopic LIN-26 expression prevents the formation of mosttissues. Immunostaining of wild-type (A, C, E, G, I), class II (B, D,F, H), or class III embryos (J) with antibodies specific for LIN-26 (A,B), muscles (C, D; mAb NE8-4C6), the pharynx and rectum (E, F;PHA-4 antiserum), the intestine (G, H; mAb 1CB4), or a subset ofpharyngeal muscles (I, J; mAb 3NB12). Besides LIN-26, none ofthese markers are expressed in class II embryos (D, F, H) at any timepoint after induction except in a few rare cells. We counted 8 6 53NB12-positive cells in class III (n 5 34; J), compared to 18 6 1 in
ild-type embryos (n 5 21; I).
lt-1.
Copyright © 2001 by Academic Press. All right
Partial Polarity Defects in lin-26 Mutant Embryos
If ectopic lin-26 expression can induce ectopic expressionof che-14, dlg-1, and jam-1, an expectation would be thattheir expression is impaired in lin-26 mutants. We reportedearlier that jam-1 expression is strongly down-regulatedand that the uterus does not form a lumen in animalsspecifically lacking lin-26 in the somatic gonad but not inthe ectoderm (den Boer et al., 1998). We thus examinedwhether this is also the case in lin-26-deficient embryos.Surprisingly, che-14, dlg-1, and jam-1 were still expressedat apparently normal levels in these embryos (Figs. 3L and3R, and data not shown). However, we noticed more subtledefects in the subcellular distribution of JAM-1 and CHE-14. For instance, the JAM-1 pattern showed internal clumpsin 45% of the embryos (n 5 59; arrowheads in Fig. 3L) orless frequently breaks (arrows in Fig. 3L), and CHE-14::GFPwas detected along the lateral membrane as well as theapical membrane in 42% of the embryos (n 5 79; arrows inFig. 3R), raising the possibility of epithelial polarity defects.Consistent with this possibility, analysis of the ultrastruc-ture of epidermal cells in lin-26-null embryos revealed thatadherens junctions were positioned at a more basal positionthan normal and were more extended than normal in 40%of the cells in lin-26(mc15) embryos (Fig. 4D). Therefore,using three different criteria, we observed cell polaritydefects in lin-26-deficient embryos.
TABLE 2Ectopic lin-26 Expression Does Not Turn on Epidermal Markers
Marker Wild-type n mcIs22 n
LIN-26 129 6 3a 38 152 6 24 40API 530a 5 216 6 36b 40
AM-1 (MH27) c 153 6 26d 25HE-14::GFP 7 6 3 30 43 6 29 31PY-7::GFP 56 6 4 13 31 6 13 48ET-502::GFP 39 6 8 13 39 6 15 50LT-1::GFP 61 6 9 15 58 6 7 15
a Data from Chanal and Labouesse, 1997 and Horner et al., 1998.b The strong reduction in cell number is not due to cell death, as
recording the lineage of two hs::lin-26 embryos during a heat-shockexperiment showed that cells from all early embryonic lineagesdivided much more slowly and stopped dividing earlier thannormal, but did not die (see Materials and Methods).
c Due to the nature of the JAM-1 pattern, it is difficult to countstaining cells in the head and the digestive tract of wild-typeembryos. The C. elegans lineage predicts that 85 epidermal cells,40 support cells, 9 arcade and 9 marginal cells in the pharynx, 20intestinal cells, 9 rectal cells, and 9 valve cells (total 181 out of 558cells; compare to 153 out of 161) would be expected to expressJAM-1 in wild-type embryos.
d The number of LIN-26-staining cells in the same set of embryos
was 161 6 26.s of reproduction in any form reserved.
FIG
.3.
Ect
opic
LIN
-26
expr
essi
onin
duce
sth
eex
pres
sion
ofep
ith
elia
lm
ark
ers.
Wil
d-ty
peco
mm
ast
age
(A–D
),cl
ass
IIh
s::l
in-2
6(E
–H),
elt-
1-de
fici
ent
(I),
elt-
1-de
fici
ent
hs:
:lin
-26
clas
sII
(J),
dlg
-1::g
fptr
ansg
enic
200-
cell
stag
e(K
),or
lin
-26-
defi
cien
t(L
)em
bryo
sw
ere
stai
ned
wit
hth
ean
tibo
dies
reco
gnis
ing
the
prot
ein
sin
dica
ted
abov
eor
belo
wth
epa
nel
s.(A
–H)P
ictu
res
are
show
nin
pair
ssh
owin
gth
eJA
M-1
patt
ern
alon
e(A
,C,
E,G
)an
daf
ter
mer
gin
gw
ith
LIN
-26
(B,D
,F,H
).G
FPau
tofl
uor
esce
nce
ofw
ild-
type
(M,O
,Q),
clas
sII
hs:
:lin
-26
(N,P
),or
lin
-26-
defi
cien
t(R
)em
bryo
sca
rryi
ng
eith
era
dlg
-1::g
fp(M
,N)o
ra
che-
14::g
fp(O
–R)c
onst
ruct
.(A
,B)E
xter
nal
foca
lpl
anes
and
(C,D
)in
tern
alfo
cal
plan
esof
aco
ntr
olem
bryo
atm
idem
bryo
gen
esis
.(E
,F)E
xter
nal
foca
lpl
anes
and
(G,H
)in
tern
alfo
cal
plan
esof
acl
ass
IIh
s::l
in-2
6em
bryo
show
ing
JAM
-1ex
pres
sion
inm
ost
cell
s.(I
,J)E
xter
nal
foca
lpla
ne
ofem
bryo
saf
ter
RN
Aii
nte
rfer
ence
agai
nst
elt-
1.T
her
ew
asli
ttle
JAM
-1le
ftin
the
elt-
1-de
fici
ent
embr
yo(I
,com
pare
toA
;th
ebr
ack
eted
haz
yst
ain
ing
corr
espo
nds
toou
tof
focu
sin
tern
alor
gan
s,w
hil
eJA
M-1
clu
mps
mar
ked
wit
har
row
sco
rres
pon
dsto
min
orep
ider
mal
cell
sw
her
eel
t-1
isn
orm
ally
inac
tive
),bu
tec
topi
cli
n-2
6ex
pres
sion
cou
ldov
erco
me
this
effe
ct(J
).(K
)Con
trol
dlg
-1::g
fpem
bryo
atth
e20
0-ce
llco
stai
ned
wit
han
tibo
dies
agai
nst
GFP
,L
IN-2
6.an
dJA
M-1
.In
con
tras
tto
clas
sII
embr
yos
wh
ich
arre
stw
ith
appr
oxim
atel
y20
0ce
lls,
wil
d-ty
peem
bryo
sw
ith
the
sam
ece
lln
um
ber
expr
ess
LIN
-26
inab
out
40ep
ider
mal
prec
urs
ors
but
not
JAM
-1n
orD
LG
-1(c
ompa
reto
H,N
);n
ote
that
clas
sII
hs:
:lin
-26
embr
yos
are
com
para
ble
wit
hth
isco
ntr
olem
bryo
for
cell
nu
mbe
rbu
tn
otfo
rde
velo
pmen
tala
geas
we
had
tow
ait
for
atle
ast
3h
afte
rth
eh
eat-
shoc
kto
see
expr
essi
onof
jam
-1,d
lg-1
::gfp
,or
che-
14::g
fp.(
L)E
xter
nal
foca
lpl
ane
ofan
embr
yoaf
ter
RN
Ain
terf
eren
ceag
ain
stli
n-2
6.T
her
ew
ere
inte
rnal
aggr
egat
es(a
rrow
hea
ds)
and
less
freq
uen
tly
brea
ks
inth
ead
her
ens
belt
(arr
ows)
.(M
)E
xter
nal
foca
lpl
ane
ofa
con
trol
dlg
-1::g
fpem
bryo
(th
epa
tter
nis
iden
tica
lto
that
ofJA
M-1
both
exte
rnal
lyan
din
tern
ally
).(N
)In
tern
alfo
cal
plan
eof
acl
ass
IId
lg-1
::gfp
;hs:
:lin
-26
embr
yo.
DL
G-1
::GFP
flu
ores
cen
cew
asde
tect
edin
mos
tce
lls.
(O)
Mos
tex
tern
al(a
pica
l)fo
cal
plan
ean
d(Q
)fo
cal
plan
eat
the
leve
lof
epid
erm
aln
ucl
eiof
the
sam
eco
ntr
olch
e-14
::gfp
embr
yo.A
tth
isst
age,
CH
E-1
4::G
FPis
dete
cted
mai
nly
atth
eap
ical
surf
ace
(O)a
nd
ina
pun
ctat
epe
rin
ucl
ear
patt
ern
(Q).
(P)I
nte
rnal
foca
lpl
ane
ofa
clas
sII
che-
14::g
fp;h
s::l
in-2
6em
bryo
show
ing
CH
E-1
4::G
FPfl
uor
esce
nce
inm
ost
cell
s(c
he-
14is
nor
mal
lyn
otex
pres
sed
inin
tern
alti
ssu
es).
(R)S
uba
pica
lfo
cal
plan
eaf
ter
RN
Ain
terf
eren
ceag
ain
stli
n-2
6.C
HE
-14:
:GFP
was
dete
cted
inla
tera
lm
embr
anes
(arr
ows)
,u
nli
ke
inco
ntr
olem
bryo
s(c
ompa
reto
Q).
416 Quintin et al.
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.
1bt
lt
ctm1t
ale b
417Epithelial Differentiation in C. elegans
DISCUSSION
We previously established that lin-26 is normally ex-pressed in three polarised cell types, epidermal cells, sup-port cells, and uterine cells, where it is essential for normaldifferentiation (den Boer et al., 1998; Labouesse et al., 1994,996). However, its precise function in these cells had noteen characterised. In the present study, our goal was first
FIG. 4. Ultrastructure of hs::lin-26 and lin-26(null) embryos. Wilprocessed for electron microscopy. A9–D9 are two-fold enlargemenelectron dense adherens-like junction is compact and located closhs::lin-26 embryos. There were multiple, generally quite extendeembryos that we examined. (D) lin-26(mc15) epidermal cell. The adfrom the apical surface in 43% of the junctions observed (n 5 42)cells, arrowhead), compared to 1% in control embryos (n 5 96). Sc
o determine for which specific aspect of cell differentiation d
Copyright © 2001 by Academic Press. All right
in-26 is required and second to identify its potentialargets.
This new analysis shows that ectopic LIN-26 expressionan reprogram the normal fates of all blastomeres to inducehe expression of three markers that play different roles inaintaining epithelial cell polarity (jam-1, dlg-1, and che-
4). Although the adherens junction pattern in embryoshat overexpress lin-26 was looking epidermal-like, we have
e (A), class II hs::lin-26 (B, C), and lin-26(mc15) (D) embryos werethe boxed areas in A–D. (A) Wild-type epidermal cell. The uniquethe apical surface. (B) External and (C) internal cell from class IIectron dense adherens-like junctions (arrows) in each of the fives-like junction was present but extended and located further away
in four mutant embryos (notice also the overlap between adjacentar, 200 nm.
d-typts ofe tod, elherenwith
emonstrated that lin-26 does not turn on epidermal mark-
s of reproduction in any form reserved.
s).
418 Quintin et al.
ers (five markers tested), strongly suggesting that lin-26does not confer epidermal identity per se. We could notdirectly test whether lin-26 confers support cell identitybecause there are no support cell markers available. How-ever, we do not think that this is the case, since cellsoverexpressing lin-26 did not send a projection as seen inthe che-14::gfp background (support cells normally send aprojection), which is a hallmark of support cells. Similarly,we do not think that lin-26 specifically turns on uterinedifferentiation, because it turned on che-14 which is nor-mally not expressed in uterine cells. Furthermore, ectopiclin-26 expression did not abolish elt-1 and dpy-7 expression(Table 2) in the same way as it repressed the expression ofintestinal and muscle markers (Fig. 2); we would haveexpected no dpy-7 or elt-1 expression if the embryo hadbecome uterine tissue.
An attractive interpretation of these data is that lin-26promotes epithelial differentiation within the nonneuronalectoderm and the somatic gonad. Genetic analysis in miceand flies has led to the notion of “master regulatory genes,”which are both necessary and sufficient to induce theformation of a particular tissue. Typical examples are the
FIG. 5. Each blastomere can respond to ectopic lin-26 expressioncolumn) embryos were operated to leave intact only one blastomereMethods. GFP autofluorescence micrographs of representative embrembryos with clear evidence of ectopic GFP fluorescence comparedEMS, and 6 isolated ABp in the jcIs1 background (wt control), whicontrol EMS embryo resembles the wild-type intestine JAM-1 pattis more similar to the normal JAM-1 epidermal pattern (large ring
myogenic genes MyoD and Myf5 (Arnold and Braun, 2000)
Copyright © 2001 by Academic Press. All right
or the proneural genes achaete/scute (Brunet and Ghysen,1999). We do not think that lin-26 qualifies as a “masterepithelial regulatory gene.” Although lin-26 is sufficient toinduce the expression of jam-1, dlg-1, and che-14, expres-sion of these genes is not deficient in lin-26-null embryos.Nonetheless, several observations are consistent withlin-26 playing an essential role in maintaining the expres-sion of jam-1, dlg-1, che-14, and other genes required forachieving epithelial cell polarity. First, we document theexistence of polarity defects in the epidermis of lin-26-deficient embryos (Fig. 4), which can easily account for thedegeneration of epidermal cells in lin-26-null embryos.Second, animals that are engineered to prevent lin-26 ex-pression in the somatic gonad generally fail to form auterine lumen and strongly down-regulate jam-1 in theuterus (den Boer et al., 1998). Finally, che-14 mutantsdisplay the same specific ultrastructural defects as the weakallele lin-26(n156) (Labouesse et al., 1994, 1996; Michaux etal., 2000), raising the possibility that che-14 could be adownstream target of lin-26.
In addition to its maintenance function, lin-26 could actin a redundant manner to induce jam-1, dlg-1, and che-14
nsgenic jcIs1 (middle left column) or jcIs1;hs::lin-26 (middle rightcolumn), and then heat-shock treated as detailed in Materials andre shown. The right column gives the proportion of jcIs1;hs::lin-26ntrol embryos. We tested 10 isolated ABa, 10 isolated P2, 7 isolatedl showed a reproducible pattern. Note that the GFP pattern in thenarrow zigzag line), whereas in the jcIs1;hs::lin-26 EMS embryo it
. Tra(left
yos ato coch alern (
expression during epithelial differentiation. Interestingly, it
s of reproduction in any form reserved.
afsMoeweciwettgedtfiidZr
iegieM
re5ta(1piMt1cec(dJ(oi(
419Epithelial Differentiation in C. elegans
has recently been reported that forced expression of theessential GATA factor elt-1 and of the dispensable GATAfactor elt-3 can turn on jam-1 expression in most cells, in amanner very similar to that observed with lin-26 (Gilleardand McGhee, 2001). A noteworthy difference between elt-1nd elt-3 on the one hand, and lin-26 on the other, is thatorced expression of elt-1 or elt-3 also induces the expres-ion of the epidermal collagen gene dpy-7 (Gilleard and
cGhee, 2001). Thus, a plausible model for the formationf the epidermis in C. elegans is the following (Fig. 7). First,lt-1 specifies tissue identity and turns on lin-26, elt-3, asell possibly as the other epidermal-specific GATA factors
lt-5 and elt-6 (K. Koh and J. Rothman, personal communi-ation). In turn, elt-1, lin-26, and elt-3 act redundantly tonduce epithelial-specific genes (jam-1, dlg-1, and che-14),hile elt-1 and elt-3 act redundantly to switch on
pidermal-specific genes (collagens). By extension, in otherissues where it is expressed, lin-26 could act redundantlyo promote epithelial differentiation, while other regulatoryenes would confer tissue identity and dictate the type ofpithelium being made. A comparable scheme has beenescribed for the genetic hierarchy controlling formation ofhe intestinal tube-shaped epithelium. First, the GATAactor end-1 together with at least one other gene specifyntestine identity and turn on the GATA factor elt-2, whichs essential for intestine differentiation, and for its functionuring postembryonic development (Fukushige et al., 1998;hu et al., 1997, 1998). Both end-1 and elt-2 then act in a
FIG. 6. Ectopic lin-26 expression does not confer epidermal iden-tity. GFP autofluorescence of elt-1::gfp (A, B, D, E) or dpy-7::gfp (C,F) in control embryos (A–C) and class II hs::lin-26 embryos (D–F).The control embryo in (A) and (B) corresponds to an embryo inwhich epidermal cell were recently born; elt-1 is expressed in all“major epidermal” cells at that stage, but its expression is main-tained only in the lateral seam cells as development proceeds (Pageet al., 1997; our unpublished observations). The control embryo in(C) is at the comma stage. (A, C, D, F) External focal planes, (B, E)internal focal planes. Note that in comparison with the imagesshown in Figs. 2B, 3G, 3N, and 3P, elt-1::gfp is not up-regulatedinternally (the same is true for dpy-7::gfp; not shown). Note alsothat several cells maintain expression of these transgenes, indicat-ing that ectopic lin-26 expression did not suppress epidermaldifferentiation in cells that presumably come from normal epider-mal lineages.
edundant manner to turn on several intestinal genes,
Copyright © 2001 by Academic Press. All right
ncluding jam-1 and the gut esterase gene ges-1 (Fukushiget al., 1998; Zhu et al., 1998). Although ELT-2 can bind thees-1 promoter in vitro and induce ges-1 expression in vivof expressed ectopically, ges-1 is normally expressed inlt-2-null mutants (Fukushige et al., 1998; Hawkins andcGhee, 1995).Our results have also shown that blastomeres can be
eprogrammed during a narrow time window of C. elegansmbryogenesis, at the beginning of gastrulation (28- to0-cell stages). In this respect, lin-26 behaves like otherissue-specific regulatory genes that have been tested using
similar approach, such as pha-4, end-1, elt-1, or elt-3Gilleard and McGhee, 2001; Horner et al., 1998; Zhu et al.,998). This time window corresponds to a period when theregastrula embryo has been patterned but organ/tissuedentities have not yet been assigned (Labouesse and
ango, 1999). Interestingly, the pharynx remained suscep-ible to ectopic lin-26 expression between the 50- and00-cell stages (class III embryos), and the phenotype oflass III embryos was similar to that of pha-4 mutantmbryos. pha-4 is first required between the 50- and 100-ell stages to specify the fates of pharyngeal precursorsHorner et al., 1998). In pha-4 mutants, the pharynx primor-ium is missing and there is an excess of LIN-26- andAM-1-expressing cells where pharyngeal cells should beChanal and Labouesse, 1997; Horner et al., 1998). Thesebservations are consistent with the suggestion that pha-4s required to repress lin-26 expression in the pharynxHorner et al., 1998).
FIG. 7. A model for the generation of the major epidermis. In thismodel (see text), elt-1 specifies epidermis identity; then lin-26,elt-1, and elt-3 act redundantly to induce differentiation as epithe-lial cells, while elt-1 and elt-3 are uniquely required to induceterminal differentiation as epidermal cells (which can be charac-terised as cuticle collagen expressing cells). Arrows correspond togenetic interactions. To simplify only well characterised genes are
indicated.s of reproduction in any form reserved.
420 Quintin et al.
So far, we have suggested that lin-26 promotes epithelialdifferentiation and might do so by inducing the expressionof epithelial-specific genes. In theory, the LIN-26 proteincould do so by directly binding to the promoters ofepithelial-specific genes, by binding to the promoter of anintermediate transcription activator or by repressing a re-pressor. Our experiments do not address this issue; how-ever, we do not favour at least one version of a doublerepressor model. We previously raised the possibility thatlin-26 could act in the nonneuronal ectoderm by repressingthe expression of neural-specific genes (Labouesse et al.,1996). The experiments presented in this study show thatlin-26 expression in neuroblasts or neurons does not affectneuronal differentiation, suggesting that lin-26 does notrepress a neuronal regulatory gene that would preventepithelial differentiation in neurons. Future experimentswill be aimed at determining whether LIN-26 can bind thepromoters of jam-1, dlg-1, and che-14.
It has been speculated that epithelial differentiationcould be a default pathway in vertebrate development(Frisch, 1997). In contrast, our experiments in C. eleganssuggest that a specific gene can promote epithelial differen-tiation without conferring any tissue specificity. One pos-sibility to account for these seemingly opposing conclu-sions could be that vertebrate studies were based on theanalysis of tumours and may not reflect the situationencountered in all epithelial tissues. In addition, we exam-ined the expression of genes that are required to organisejunctions or control apical trafficking, whereas previousstudies in vertebrates systems generally examined the ex-pression of structural proteins.
ACKNOWLEDGMENTS
We thank Elise Camut who made the initial hs::lin-26 construct.We are grateful to M.-A. Felix for use of her laser microscope. Wethank John Gilleard for sharing unpublished results, A. Fire, J.Gilleard, J. Hardin, I. Jonhstone, S. Mango, J. McGhee, B. Page, A.Wissmann, and Bob Waterston for reagents. We thank Julia Bosherfor critical reading of the manuscript, J. L. Vonesch for confocalimages, and J. M. Lafontaine and B. Boulay for photographs. Thiswork was supported by funds from the CNRS, INSERM, andHopital Universitaire de Strasbourg and by grants from the EEC-TMR program and the Association pour la Recherche contre leCancer (to M.L.).
REFERENCES
Arnold, H. H., and Braun, T. (2000). Genetics of muscle determi-nation and development. Curr. Top. Dev. Biol. 48, 129–164.
Batlle, E., Sancho, E., Franci, C., Dominguez, D., Monfar, M.,Baulida, J., and Garcia De Herreros, A. (2000). The transcriptionfactor snail is a repressor of E-cadherin gene expression inepithelial tumour cells. Nat. Cell Biol. 2, 84–89.
Birchmeier, C., and Birchmeier, W. (1993). Molecular aspects ofmesenchymal–epithelial interactions. Annu. Rev. Cell Biol. 9,
511–540.Copyright © 2001 by Academic Press. All right
Brunet, J. F., and Ghysen, A. (1999). Deconstructing cell determi-nation: Proneural genes and neuronal identity. BioEssays 21,313–318.
Cano, A., Perez-Moreno, M. A., Rodrigo, I., Locascio, A., Blanco,M. J., del Barrio, M. G., Portillo, F., and Nieto, M. A. (2000). Thetranscription factor snail controls epithelial–mesenchymal tran-sitions by repressing E-cadherin expression. Nat. Cell Biol. 2,76–83.
Chanal, P., and Labouesse, M. (1997). A screen for genetic locirequired for hypodermal cell and glial-like cell developmentduring Caenorhabditis elegans embryogenesis. Genetics 146,207–226.
Davies, J. A., and Garrod, D. R. (1997). Molecular aspects of theepithelial phenotype. BioEssays 19, 699–704.
den Boer, B. G., Sookhareea, S., Dufourcq, P., and Labouesse, M.(1998). A tissue-specific knock-out strategy reveals that lin-26 isrequired for the formation of the somatic gonad epithelium inCaenorhabditis elegans. Development 125, 3213–3224.
Dufourcq, P., Chanal, P., Vicaire, S., Camut, E., Quintin, S., denBoer, B. G., Bosher, J. M., and Labouesse, M. (1999). lir-2, lir-1,and lin-26 encode a new class of zinc-finger proteins and areorganized in two overlapping operons both in Caenorhabditiselegans and in Caenorhabditis briggsae. Genetics 152, 221–235.
Francis, R., and Waterston, R. H. (1991). Muscle cell attachment inCaenorhabditis elegans. J. Cell Biol. 114, 465–479.
Frisch, S. M. (1994). E1a induces the expression of epithelialcharacteristics. J. Cell Biol. 127, 1085–1096.
Frisch, S. M. (1997). The epithelial cell default-phenotype hypoth-esis and its implications for cancer. BioEssays 19, 705–709.
Fukushige, T., Hawkins, M. G., and McGhee, J. D. (1998). TheGATA-factor elt-2 is essential for formation of the Caenorhab-ditis elegans intestine. Dev. Biol. 198, 286–302.
Geiger, B., Ayalon, O., Ginsberg, D., Volberg, T., Rodriguez Fernan-dez, J. L., Yarden, Y., and Ben-Ze’ev, A. (1992). Cytoplasmiccontrol of cell adhesion. Cold Spring Harbor Symp. Quant. Biol.57, 631–642.
Gilleard, J. S., Barry, J. D., and Johnstone, I. L. (1997). cis-regulatoryrequirements for hypodermal cell-specific expression of the Cae-norhabditis elegans cuticle collagen gene dpy-7. Mol. Cell. Biol.17, 2301–2311.
Gilleard, J. S., and McGhee, J. D. (2001). Activation of hypodermaldifferentiation in the Caenorhabditis elegans embryo by theGATA transcription factors elt-1 and elt-3. Mol. Cell. Biol. 21,2533–2544.
Gilleard, J. S., Shafi, Y., Barry, J. D., and McGhee, J. D. (1999).ELT-3: A Caenorhabditis elegans GATA factor expressed in theembryonic epidermis during morphogenesis. Dev. Biol. 208,265–280.
Grooteclaes, M. L., and Frisch, S. M. (2000). Evidence for a functionof CtBP in epithelial gene regulation and anoikis. Oncogene 19,3823–3828.
Gumbiner, B. M. (1996). Cell adhesion: The molecular basis oftissue architecture and morphogenesis. Cell 84, 345–357.
Hawkins, M. G., and McGhee, J. D. (1995). elt-2, a second GATAfactor from the nematode Caenorhabditis elegans. J. Biol. Chem.270, 14666–14671.
Horner, M. A., Quintin, S., Domeier M., Kimble, J., Labouesse, M.,and Mango, S. E. (1998). pha-4, an HNF-3 homolog, specifiespharyngeal organ identity in Caenorhabditis elegans. Genes
Dev. 12, 1947–1952.s of reproduction in any form reserved.
M
M
M
M
N
P
S
S
S
T
W
W
Y
Z
Z
421Epithelial Differentiation in C. elegans
Hosono, S., Gross, I., English, M. A., Hajra, K. M., Fearon, E. R., andLicht, J. D. (2000). E-cadherin is a WT1 target gene. J. Biol. Chem.275, 10943–10953.
Labouesse, M., Hartwieg, E., and Horvitz, H. R. (1996). TheCaenorhabditis elegans LIN-26 protein is required to specifyand/or maintain all non-neuronal ectodermal cell fates. Devel-opment 122, 2579–2588.
Labouesse, M., and Mango, S. E. (1999). Patterning the C. elegansembryo: Moving beyond the cell lineage. Trends Genet. 15,307–313.
Labouesse, M., Sookhareea, S., and Horvitz, H. R. (1994). TheCaenorhabditis elegans gene lin-26 is required to specify thefates of hypodermal cells and encodes a presumptive zinc-fingertranscription factor. Development 120, 2359–2368.
Legouis, R., Gansmuller, A., Sookhareea, S., Bosher, J. M., Baillie,D. L., and Labouesse, M. (2000). LET-413 is a basolateral proteinrequired for the assembly of adherens junctions in Caenorhab-ditis elegans. Nat. Cell Biol. 2, 415–422.cMahon, L., Legouis, R., Vonesch, J. L., Labouesse, M. (2001).Assembly of C. elegans apical junctions involves positioning andcompaction by LET-413 and protein aggregation by the MAGUKprotein DLG-1 J. Cell Sci., in press.ello, C., and Fire, A. (1995). DNA transformation. Methods Cell.Biol. 48, 451–482.ichaux, G., Gansmuller, A., Hindelang, C., and Labouesse, M.(2000). CHE-14, a protein with a sterol-sensing domain, isrequired for apical sorting in C. elegans ectodermal epithelialcells. Curr. Biol. 10, 1098–1107.ohler, W. A., Simske, J. S., Williams-Masson, E. M., Hardin, J. D.,and White, J. G. (1998). Dynamics and ultrastructure of develop-mental cell fusions in the Caenorhabditis elegans hypodermis.Curr. Biol. 8, 1087–1090.ose, A., Nagafuchi, A., and Takeichi, M. (1988). Expressed recom-binant cadherins mediate cell sorting in model systems. Cell 54,993–1001.
age, B. D., Zhang, W., Steward, K., Blumenthal, T., and Priess, J. R.(1997). ELT-1, a GATA-like transcription factor, is required forepidermal cell fates in Caenorhabditis elegans embryos. GenesDev. 11, 1651–1661.
avagner, P., Yamada, K. M., and Thiery, J. P. (1997). The zinc-
finger protein slug causes desmosome dissociation, an initial andCopyright © 2001 by Academic Press. All right
necessary step for growth factor-induced epithelial–mesenchymal transition. J. Cell Biol. 137, 1403–1419.
eydoux, G., Mello, C. C., Pettitt, J., Wood, W. B., Priess, J. R., andFire, A. (1996). Repression of gene expression in the embryonicgerm lineage of C. elegans. Nature 382, 713–716.
Sinha, S., Degenstein, L., Copenhaver, C., and Fuchs, E. (2000).Defining the regulatory factors required for epidermal geneexpression. Mol. Cell. Biol. 20, 2543–2555.
tringham, E. G., Dixon, D. K., Jones, D., and Candido, E. P. (1992).Temporal and spatial expression patterns of the small heat shock(hsp16) genes in transgenic Caenorhabditis elegans. Mol. Biol.Cell 3, 221–233.
hiery, J. P., Boyer, B., Tucker, G., Gavrilovic, J., and Valles, A. M.(1988). Adhesion mechanisms in embryogenesis and in cancerinvasion and metastasis. Ciba Found. Symp. 141, 48–74.ay, J. C., and Chalfie, M. (1989). The mec-3 gene of Caenorhab-ditis elegans requires its own product for maintained expressionand is expressed in three neuronal cell types. Genes Dev. 3,1823–1833.issmann, A., Ingles, J., and Mains, P. E. (1999). The Caenorhab-ditis elegans mel-11 myosin phosphatase regulatory subunitaffects tissue contraction in the somatic gonad and the embry-onic epidermis and genetically interacts with the Rac signalingpathway. Dev. Biol. 209, 111–127.eaman, C., Grindstaff, K. K., and Nelson, W. J. (1999). Newperspectives on mechanisms involved in generating epithelialcell polarity. Physiol. Rev. 79, 73–98.hu, J., Fukushige, T., McGhee, J. D., and Rothman, J. H. (1998).Reprogramming of early embryonic blastomeres into endoder-mal progenitors by a C. elegans GATA factor. Genes Dev. 15,3809–3814.hu, J., Hill, R. J., Heid, P. J., Fukuyama, M., Sugimoto, A., Priess,J. R., and Rothman, J. H. (1997). end-1 encodes an apparentGATA factor that specifies the endoderm precursors in Caeno-rhabditis elegans embryos. Genes Dev. 11, 2883–2896.
Submitted for publication March 5, 2001Revised April 10, 2001
Accepted April 10, 2001
Published online June 13, 2001s of reproduction in any form reserved.