C. Elegans Polarity and Gastrulation

Development 129, 387-397 (2002) 387 Printed in Great Britain © The Company of Biologists Limited 2002 DEV8905

Cell polarity and gastrulation in C. elegans

Jeremy Nance1,2 and James R. Priess1,2,* 1Division of Basic Sciences, Fred Hutchinson Cancer Research Center, Seattle, Washington 98109, USA 2Howard Hughes Medical Institute, Seattle, Washington 98109, USA *Author for correspondence (e-mail: [email protected])

Accepted 17 October 2001

SUMMARY

Gastrulation in C. elegans embryos involves formation of a apical ﬂattening associated with an apical concentration of blastocoel and the ingression of surface cells into the non-muscle myosin. We provide evidence that ingression blastocoel. Mutations in the par-3 gene cause abnormal times are determined by genes that control cell fate, though separations between embryonic cells, suggesting that the interactions with neighboring cells can prevent ingression. PAR-3 protein has a role in blastocoel formation. In normal development, PAR proteins localize to either the apical or basal surfaces of cells prior to blastocoel formation; we Key words: Blastocoel, Apical-basal polarity, Gastrulation, demonstrate that this localization is determined by cell Ingression, Morphogenesis, Cell fate, Cell shape, Non-muscle contacts. Cells that ingress into the blastocoel undergo an myosin, LIT-1, HMR-1, NMY-2, PAR-2, PAR-3, PAR-6, C. elegans

INTRODUCTION 1992; Lin et al., 1995; Rocheleau et al., 1997). These endodermal and mesodermal precursors must then move from Numerous studies on C. elegans embryos have revealed the the outer surface of the embryo into the interior for the basic strategies that establish the anterior-posterior and dorsal- formation of a functional intestine and muscular system. ventral axes. Axis specification in C. elegans embryos begins Animal embryos use several strategies to position with fertilization of the egg, where the point of sperm entry endodermal and mesodermal precursors into their interior, a defines the posterior pole (Goldstein and Hird, 1996). process called gastrulation. In many embryos, such as those of Fertilization induces the association of a group of proteins, sea urchins and amphibians, the early embryonic cleavages collectively called PAR proteins, to either the anterior or generate a cluster of cells (the blastula), and a central cavity posterior cortex. For example, PAR-3 and PAR-6 associate (the blastocoel) develops within this cluster. Certain cells with the anterior pole while PAR-2 and PAR-1 associate with detach from their neighbors on the cell surface and enter the the posterior pole (Boyd et al., 1996; Etemad-Moghadam et al., blastocoel (ingression), or large groups of cells fold into the 1995; Guo and Kemphues, 1995; Hung and Kemphues, 1999). blastocoel (invagination and involution) (Gilbert and Raunio, The localization of the PAR proteins is interdependent; 1997). In some animals, gastrulation occurs without formation mutations in par-3 cause the anterior mislocalization of PAR- of a blastocoel. In ctenophores, for example, ectodermal cells 2, and mutations in par-2 cause the posterior mislocalization spread over and internalize endodermal cells (epiboly) (Komai, of PAR-3 (Boyd et al., 1996; Etemad-Moghadam et al., 1995). 1968; Martindale and Henry, 1999). The dorsal-ventral axis is determined as the embryo divides Gastrulation in C. elegans begins at the 26-cell stage, when from two cells to four (reviewed by Schnabel and Priess, 1997). two endodermal precursors ingress from the surface of the The division of the anterior cell generates two initially embryo into the interior. Shortly thereafter, mesodermal equivalent daughters that express a receptor related to the precursors and germline precursors follow the endodermal Notch protein. The division of the posterior cell generates non- precursors into the interior of the embryo (Sulston et al., 1983). equivalent daughters, one of which expresses a ligand for the Very little is known about the cellular or molecular basis for receptor. In an apparently random manner, only one of the these events. Several mutations have been identified that receptor-expressing cells contacts the ligand-expressing cell; prevent or delay ingression of the endodermal precursors, most this interaction leads to the specification of dorsal cell types of which cause the endodermal precursors to divide such as hypodermis (skin). prematurely (Denich et al., 1984; Knight and Wood, 1998). In addition to anterior-posterior and dorsal-ventral axes, the Inhibiting embryonic transcription also causes the endodermal early embryo must establish an outer-inner polarity. Cells on precursors to divide precociously and prevents ingression the ventral surface of the embryo become committed to (Powell-Coffman et al., 1996). Ingression of the endodermal endodermal and mesodermal fates through a combination of precursors is likely to require embryonic transcription of one cell signaling events and asymmetrically localized factors or more genes in a chromosomal region called the endoderm (Bowerman et al., 1993; Bowerman et al., 1992; Goldstein, determining region, or EDR; a chromosomal deficiency that 388 J. Nance and J. R. Priess deletes the EDR delays or prevents ingression (Zhu et al., Bowerman, 1996). Embryo combination experiments were performed 1997). There are at least two genes in the EDR that are involved at 15°C. To inhibit transcription, embryonic culture medium was in specification of the endodermal cell fate, suggesting a supplemented with 50 µg/ml α-amanitin (Sigma). Laser ablations possible link between cell fate and ingression (Zhu et al., were performed as described elsewhere (Mello et al., 1992). 1997). 4D images were acquired as described (Thomas et al., 1996) using In this report we investigate the cellular basis for gastrulation 4D Grabber v1.32 software (C. Thomas, Integrated Microscopy Resource, University of Wisconsin-Madison, USA). Movies were in C. elegans. PAR proteins that show anterior-posterior analyzed using 4D Viewer v4.11 (C. Thomas, Integrated asymmetry in 1-cell and 2-cell embryos subsequently develop Microscopy Resource, University of Wisconsin-Madison, USA) outer-inner, or apical-basal, polarity as the blastula forms. We and Nematode Navigator software (kindly provided by J. Pitt, demonstrate that the apical-basal localization of the PAR http://www.fhcrc.org/labs/priess/nn.html). The volume of the proteins is dependent on cell-cell contacts by generating double blastocoel and embryo (± s.d.) were calculated using NIH Image1.62 embryos with abnormal patterns of cell contact. We show that (Wayne Rasband, National Institutes of Health, USA). ingression is associated with changes in the shape of the Ingressions were scored when a cell sank permanently from the ingressing cells and a redistribution of non-muscle myosin. surface of the embryo. Comparisons of ingression times were based Finally, we present evidence that cell fate, rather than cell on the analysis of 4D videorecordings of three embryos; all cells that position, is the predominant factor regulating ingression, ingressed during gastrulation were followed in each embryo. Ingression times were normalized to cell division timings presented although steric interactions can play an important role. by Sulston et al. (Sulston et al., 1983) using the interval between the MSa and MSaaa divisions as a measure of developmental time. Descriptions of wild-type gastrulation were based on observations of MATERIALS AND METHODS these and additional 4D videorecordings. Ectopic ingressions in mex- 1 mutants were scored in the ABpr lineage. lit-1 temperature shifts Nematode culture and strains were from 15°C to 25°C at the 24-cell stage. Because of the variable Nematodes were cultured and manipulated as described (Brenner, expressivity of the lit-1 mutant phenotype, analysis of MS ingression 1974). Unless otherwise indicated, experiments were performed on was restricted to the 7/11 embryos showing simultaneous ingression the wild-type N2 (var. Bristol). The following mutants were utilized: of all wishbone cells. In the E ablation experiments, analysis was chromosome II LG II: unc-4(e120) (Brenner, 1974), mex-1(zu120) restricted to the anterior MS central cells that do not contact the E (Mello et al., 1992); unc-32(e189) (Brenner, 1974), lit-1(t1512ts) corpse. (Kaletta et al., 1997), unc-45(e286ts) (Brenner, 1974), par-2(lw32) (Cheng et al., 1995), par-3(it71) (Cheng et al., 1995), lon-1(e185) Electron microscopy (Brenner, 1974); LGV: zuDf2 (Zhu et al., 1997). The following Mixed-stage embryos were fixed, embedded, sectioned and stained for integrated transgenes containing green fluorescent protein (GFP) transmission electron microscopy (TEM) as described (Priess and reporters were used: zuIs3 (end-1::GFP) (J. Nance, unpublished), Hirsh, 1986). Several thick sections were cut from each block to ruIs32 (pie-1::GFP::HIS-11) (Pratis et al., 2001), pxIs6 (pha- determine the precise stage and orientation of specific embryos. 4::GFP::HIS-11) (Portereiko and Mango, 2001), itIs153 (pie-1::PAR- Sections were analyzed from >50 embryos embedded in three 2::GFP) (Wallenfang and Seydoux, 2000). Mex-1 embryos were mex- different blocks. Embryos for scanning electron microscopy were 1 unc-4. Par-2 embryos were par-2 unc-45. Par-3 embryos were par-3 fixed as for TEM, omitting the tannic acid treatment. Embryos were lon-1; zuIs3. transferred by mouth pipet to a poly-L-lysine coated coverslip, then dehydrated and dried as described (Braet et al., 1997). dsRNA-mediated interference (RNAi) Standard techniques were used to synthesize double-stranded RNA Immunostaining and fluorescence microscopy (dsRNA) from T7 promoter-tagged, PCR-amplified cDNA or Embryos were fixed and processed for overnight immunostaining at genomic DNA. For analysis of ama-1(RNAi) embryos (2-4 µg/µl 4°C as described (Leung et al., 1999). Cultured blastomeres were dsRNA made from cDNA region 2876-3450), young adult fixed in –20°C methanol for 15 minutes, then immunostained at 37°C hermaphrodites carrying both a maternal (pie-1::GFP::HIS-11) and for 1 hour. A strain containing the itIs153 reporter was used for zygotic (end-1::GFP) GFP reporter were injected in the gonad or analysis of PAR-2 in cultured blastomeres. Unless indicated soaked overnight with dsRNA (Fire et al., 1998; Tabara et al., 1998). otherwise, immunofluorescence studies were based on the observation After allowing worms to recover for 30 hours, embryos were analyzed of 15-100 embryos at the indicated stage. The following dilutions of for end-1::GFP expression. In a representative experiment, the end- antibodies/antisera were used: 1:200 mouse anti-GFP (‘α’-GFP; 1::GFP reporter was detectable in only 5/224 ama-1(RNAi) embryos. Clontech); 1:10 rabbit α-HMR-1 (Costa et al., 1998); 1:50 rabbit α- For analysis of hmr-1(RNAi) embryos (2.5 µg/µl dsRNA made from NMY-2 (Guo and Kemphues, 1996); 1:10 rabbit α-PAR-2 (Boyd et genomic DNA region 13344-13989), approximately 20 hours after al., 1996); 1:10 chicken α-PAR-3 (Tabuse et al., 1998); 1:15 rabbit α- injection of dsRNA into wild-type adults, embryos were collected and PAR-3 (Etemad-Moghadam et al., 1995); 1:10 rabbit α-PAR-6 (Hung fixed for immunofluorescence. Embryos were immunostained to and Kemphues, 1999); 1:1000 rabbit α-PGL-1 (Kawasaki et al., confirm HMR-1 depletion. 1998). Secondary antibodies were conjugated with Alexa Fluor 488 (Molecular Probes) or Cy-3 (Jackson ImmunoResearch Laboratories, Embryo handling and analysis Inc.). In some experiments, DNA was stained with 60 ng/ml 4,6- To examine blastocoel formation, 1-cell or 2-cell embryos were diamidino-2-phenylindole (DAPI). pipetted onto a poly-L-lysine (Sigma)-coated coverslip containing Confocal images were acquired on a Leica TCS SP spectral clay spacers; the coverslip was inverted onto a slide for observation. confocal microscope. All other images were acquired using a standard In all other experiments 2-cell or 4-cell embryos were mounted for epifluorescence microscope equipped with a digital camera; some videomicroscopy on 4% agarose (see Sulston et al., 1983). Isolated epifluorescence images were numerically deconvolved using AB blastomeres were mounted in embryonic culture medium (Shelton DeltaVision algorithms (Applied Precision, Inc.). and Bowerman, 1996) on slides and covered with coverslips coated with 1.5% agarose; techniques used to isolate and culture embryos or Analysis of NMY-2 localization individual blastomeres were essentially as described (Shelton and For the analysis of NMY-2 localization in MS descendants, embryos C. elegans polarity and gastrulation 389 expressing pha-4::GFP::HIS-11 were stained with antibodies or after they separate, temporarily obscuring regions of the antisera against PGL-1, GFP and NMY-2; GFP and PGL-1 staining blastocoel. In contrast to the often large spaces separating basal were used as reference points to identify specific MS descendants surfaces (Fig. 1E), lateral surfaces either are juxtaposed or are after fixation. NMY-2 was scored as ‘enriched’ when the apical separated by only small spaces (Fig. 1D). concentration in an ingressing cell appeared greater than in other How might the various surfaces of embryonic cells become surface cells, exclusive of dividing cells. specified, such that the blastocoel forms at the basal surface? In the 4-cell embryo, the basal surface corresponds to the region where the two daughters of the AB blastomere (ABa RESULTS and ABp) contact the two daughters of the P1 blastomere (EMS and P2) (Fig. 1B). We asked whether blastocoel Formation of the blastocoel formation required interactions between specific AB and P1 Gastrulation in C. elegans involves the movement of a subset of cells from the ventral surface of the embryo into the blastocoel. These movements, described here as cell ingressions, occur over a period of about 3 hours and begin at the 26-cell stage of embryogenesis. Some, or all, of the descendants of each of the early embryonic blastomeres ingress (see Fig. 4A,B). The ingressing cells form tissues such as the endoderm (descendants of the E blastomere), germline (P4 descendants) and mesoderm (D descendants and a subset of AB, MS, and C descendants; see Fig. 4A). Prior to gastrulation, the embryo consists of a hull of cells one cell in thickness. These cells surround a small, central cavity called the blastocoel (arrowheads, Fig. 1A). For convenience, we distinguish three types of membrane surfaces for each cell. The apical membrane faces the perimeter of the embryo, the basal membrane faces the blastocoel, and the lateral membranes face adjacent cells within the hull (Fig. 1A). In three-dimensional timelapse (‘4D’) videorecordings of living embryos, the blastocoel forms as the basal surfaces of diametrically opposed cells gradually detach from each other. Small openings between basal surfaces are first visible at the 4-cell stage (arrowhead, Fig. 1B) and numerous openings of variable size are present by the 26-cell stage (arrowheads, Fig. 1C). The blastocoel achieves its maximum volume by the 26- cell stage when it measures 450±100 µm3 (n=5). This volume is only 2% of the volume of the embryo (25,800±980 µm3), and less than half the volume of a single cell (about 1000 µm3). Because of the small size of the blastocoel, the basal surfaces of diametrically opposed cells transiently come into contact

Fig. 1. Blastocoel formation. All images represent sections through the center of embryos. (A) 26-cell embryo indicating the apical (a), lateral (l) and basal (b) surfaces of a cell lining the blastocoel (arrowheads). The E daughters are indicated by asterisks. (B) 4-cell embryo showing a small basal separation (arrowhead); cell names are listed. (C) Electron micrograph of a 28-cell embryo where the E daughters (dashed outline) have partially ingressed. Prominent spaces between cells are indicated by arrowheads. (D,E) Membrane contacts at the same, high magniﬁcation showing examples of (D) a small space (arrow) between lateral surfaces and (E) a large space associated with basal surfaces; arrows in E illustrate close apposition of membranes around the space. (F) 16-cell partial embryo derived from an isolated AB blastomere; note the large central cavity (arrowhead). (G) 16-cell par-3 mutant embryo showing abnormally large separations (arrowheads) between lateral membranes. (H) Double embryo after ﬁxation showing a cavity (arrowhead) in the center of AB descendants; a subset of the eight AB descendants present are visible and labeled 1 or 2 according to origin. The P1 descendants (MS, E, P3, and C) from each half of the double embryo are labeled. N, nucleus. Intact embryos (A-C,G) are 50 µm. Bar (for D,E),1 µm. 390 J. Nance and J. R. Priess

Fig. 3. PAR localization in double embryos. Double embryos were immunostained for the PAR protein listed above each panel. The subset of AB descendants visible are indicated by 1 or 2, according to origin, and names of some of the P1 descendants are indicated. Both germline blastomeres (P2) are indicated with asterisks in A. (C,D) Upper and lower focal planes of the same double embryo. Note the blastocoel visible at the center of the 1 and 2 cells in C. At Fig. 2. PAR localization. (A,B) PAR-2 in 4-cell (A) and 7-cell (B) the time of fixation, double embryos were equivalent to the following embryos. (C,D) PAR-3 in early 4-cell (C) and 8-cell (D) embryos. single embryo stages: (A) 4-cell plus 4-cell, (B) 8-cell plus 8-cell, (E) 28-cell embryo showing PAR-2 (green), PAR-3 (red) and DNA (C,D) 7-cell plus 8-cell. Bar, 10 µm. (blue); the ingressing E daughters are indicated with yellow asterisks. The embryo is oriented as in Fig. 4D. (F) Nomarski micrograph of 50-cell embryo indicating cell surfaces facing the blastocoel 1H). We draw several conclusions from these experiments. (arrowheads). (F′) PAR-3 localization in embryo shown in F. PAR-3 First, the surfaces of cells at the 2-cell and 4-cell stages of is localized to surfaces facing the blastocoel (arrowheads) in addition embryogenesis are adhesive; thus this adhesion must be to localization at the apical surface (arrow). (G) PAR-3 in an 8-cell overcome for cell surfaces to separate during blastocoel ama-1(RNAi) embryo. (H) PAR-2 in an 8-cell par-3(it71) embryo; formation. Second, interactions between AB descendants are PAR-2 is localized to all cell surfaces (arrows). In A-E and G the sufficient to form a blastocoel. Finally, we conclude that cell germline precursor is indicated with a cyan asterisk. contact can reorient the polarity of AB descendants; in double embryos a blastocoel can form at a surface that would normally descendants by separating AB from P1 at the 2-cell stage and have been an apical surface. allowing each blastomere to develop in isolation (n=8). The AB descendants generated by the first two cycles of cell PAR proteins and apical-basal polarity in early division formed a tightly adherent spherical cluster of cells. By embryos the third cycle of cell division, the clusters developed a central Analysis of the PAR proteins has focused primarily on their cavity of variable size that resembled the normal blastocoel anterior-posterior localization at the 1-cell stage and their roles (Fig. 1F). Because in these experiments P1 descendants form in subsequent anterior-posterior polarity (see Introduction). a linear array of cells rather than a spherical cluster (see Fig. After cell division begins, however, this anterior-posterior 1H) (Schierenberg, 1987), we did not examine the behaviour asymmetry is reiterated only in the lineage of cells that form of these cells further. We asked whether AB descendants from the germline (Boyd et al., 1996; Etemad-Moghadam et al., separate embryos could generate a blastocoel by combining 1995; Guo and Kemphues, 1995; Hung and Kemphues, 1999) two 2-cell embryos, or two early 4-cell embryos, head to head (the germline cells P2 and P3 are indicated by cyan asterisks with the anteriormost surfaces of the AB cells in contact in Fig. 2). In contrast, some somatic (non-germline) cells show (n=16). At the point of contact, the surfaces of the combined an apical-basal polarity in PAR localization (Boyd et al., 1996; blastomeres flattened and adhered tightly. All of these ‘double’ Etemad-Moghadam et al., 1995; Guo and Kemphues, 1995; embryos developed a central cavity of variable size surrounded Hung and Kemphues, 1999). By the 4-cell stage, the posterior by both sources of AB descendants (labeled 1 and 2 in Fig. PAR protein PAR-2 is localized to basolateral surfaces (arrows, C. elegans polarity and gastrulation 391

Fig. 2A,B), and is not detected on the apical surfaces of somatic cells. By the middle of the 4-cell stage, the anterior PAR proteins PAR-3 and PAR-6 are present over the entire cortex of each somatic blastomere (arrows, Fig. 2C). However by the end of the 4-cell stage and at later stages PAR-3 and PAR-6 are concentrated in a broad ‘cap’ centered on the apical surface (arrow, Fig. 2D). The apical-basal polarity of the PAR proteins persists through early gastrulation, although the level of PAR-2 diminishes (Fig. 2E). After gastrulation begins, cells in the interior of the embryo accumulate PAR-3 and PAR-6 on their blastocoel-facing surfaces (arrowheads, Fig. 2F,F′); these and later changes in PAR distribution were not analyzed further (see also Leung et al., 1999). We used 4D videomicroscopy to examine formation of the blastocoel in par-2 and par-3 mutant embryos. par-2 mutant embryos (n=6) developed a central cavity resembling the normal blastocoel (data not shown). par-3 mutant embryos (n=6) also developed a central cavity, but in addition showed frequent separations between the lateral membranes of cells (arrowheads, Fig. 1G); these lateral separations persisted through several cell divisions. The lateral separations in par-3 mutant embryos were comparable to the spaces observed between the basal surfaces of cells in wild-type embryos, but were much larger than the spaces normally present between lateral surfaces. These results suggest that PAR-3 is required, directly or indirectly, for the normal adhesiveness of lateral membranes. Apical-basal polarity of PAR proteins requires cell interactions How do the PAR proteins switch from anterior-posterior polarity to apical-basal polarity? Germline blastomeres, which maintain anterior-posterior PAR polarity during the early embryonic stages, are transcriptionally quiescent, while Fig. 4. Cell ingression. (A) Schematic lineage diagram of early somatic blastomeres that switch to apical-basal asymmetry are blastomeres indicating AB and P1 descendants. (B) Lateral view of a transcriptionally active (Seydoux et al., 1996). However, we three-dimensional model of nuclei at the 26-cell stage; anterior is found that PAR-3 and PAR-2 switch from anterior-posterior to left. The subset of descendants of the early blastomeres that ingress, apical-basal polarity after embryonic transcription is prevented or give rise to ingressing cells, are indicated by the color scheme in by double-stranded RNA inhibition (RNAi) of the ama-1 gene A; non-ingressing cells (a subset of AB descendants) are shown in [Fig. 2G and data not shown; the ama-1 gene encodes the large grey. (C) Two diagrams of the ventral surfaces of embryos outlining subunit of RNA polymerase II (Rogalski and Riddle, 1988)]. the regions where various cells ingress and listing the periods of cell In addition, a blastocoel appeared to form normally in ama- ingressions in minutes from the 2-cell stage. The left diagram 1(RNAi) embryos (n=6) and in wild-type embryos exposed to indicates ingression of the E daughters and the MS ‘wishbone’ α-amanitin (n=4). Thus apical-basal PAR polarity, as well as descendants (red), and the right diagram indicates later ingressions including the MS ‘central’ descendants (red oval). (D-G) Nomarski blastocoel formation, appears to be determined by maternally light micrographs of representative stages of gastrulation; times as provided gene products. indicated. (D) 28-cell embryo; lateral view as in B. An MS We asked whether cell contacts determine apical-basal PAR descendant (red arrowhead; MSap) and P4 (cyan arrowhead) are polarity by constructing and analyzing double embryos with shown spreading across the apical surfaces of the E daughters ectopic cell contacts, as described above. 2-cell embryos were (yellow asterisks). (E) Ventral view showing MS ‘central’ joined head to head, then allowed to develop between 15 and descendants (red asterisks) and a subset of C descendants (green 45 minutes (1-3 additional cell cycles) before ﬁxation. We asterisks) prior to ingression. The arrow indicates the cleft created by found that PAR-2 was excluded from the apical regions of ingression of the D descendants. (F) Ventral view showing part of the somatic blastomeres in the double embryos, but was localized cleft created by ingression of the MS descendants; neighboring cells to all basal and lateral membranes including ectopic sites of (arrows) will eventually move to cover the cleft. (G) Ventral view showing epiboly of the skin cells (arrows) to cover remaining cells cell contact (n=6; arrow, Fig. 3A). In contrast, PAR-3 and PAR- on the ventral surface. 6 were localized predominately to the contact-free, apical surfaces of somatic blastomeres (n=6 and 8, respectively; Fig. 3B-D). As in wild-type embryos, there appeared to be a cap of localized exclusively to sites of cell contact in a wild-type 4- PAR-3 and PAR-6 toward the center of the apical cortex. cell embryo (Costa et al., 1998), a pattern very similar to that We asked if PAR localization was dependent on HMR-1, a of PAR-2. However hmr-1(RNAi) embryos, with markedly C. elegans homologue of vertebrate cadherin; HMR-1 is depleted HMR-1 levels, had the wild-type pattern of both PAR- 392 J. Nance and J. R. Priess

2 and PAR-3 localization (data not shown). We next asked if the apical-basal polarity in PAR localization was dependent on par gene function. As reported previously by others (Boyd et al., 1996; Etemad-Moghadam et al., 1995), we observed that PAR-2 was not restricted to the posterior pole in 1-cell stage par-3 mutant embryos, and that PAR-3 was not restricted to the anterior pole in par-2 mutant embryos. During the 2-cell to 16-cell stages, PAR-2 was not restricted to basal and lateral surfaces in par-3 mutant embryos, and was instead associated with the entire cortex of each blastomere (Fig. 2H) (see also Boyd et al., 1996). PAR-3 showed the wild-type pattern of Fig. 5. Cell extensions over apical localization in par-2 mutant embryos (n=40), but failed ingressing cells. to accumulate at the cortex in par-6 mutants (n=15). We (A) Scanning electron conclude that PAR-3 is required to exclude PAR-2 from the micrograph of the ventral surface of a 28-cell embryo, apical surface. PAR-3 localization to the apical surface appears anterior is to the left. to be independent of both HMR-1 and PAR-2, but requires (B) Tracing of A to indicate PAR-6. cell names. The exposed apical surfaces of the Ingression of cells during gastrulation ingressing E daughters are Ingression occurs on the ventral surface of the embryo. Many tinted. (C) High of the cells that ingress are born on the ventral surface (E, MS magnification of the region and P4 descendants; Fig. 4B). Other cells, such as the C corresponding to the boxed descendants (green cells in Fig. 4B), are born in dorsal or area in B. The arrow indicates lateral positions and move to the ventral surface prior to the thin, leading process from an MS descendant. ingression. We analyzed 4D videorecordings of gastrulating embryos to determine when and where cells ingress. Cells do not ingress from a single location, but rather enter the Discussion). After the cycles of ingression are completed, the blastocoel from several positions over the ventral surface of the ventral surface of the embryo is occupied primarily by embryo; these patterns of ingression are summarized in Fig. neuronal precursors. These precursors do not appear to enter 4C, and some of the major groups of ingressing cells are the blastocoel by ingression, but are internalized by epiboly of described here briefly. skin cells [Fig. 4G; reviewed by Simske and Hardin (Simske The daughters of the E blastomere are the first to ingress, and Hardin, 2001); see also Sulston et al. (Sulston et al., beginning at the 26-cell stage (90 minutes after the first 1983)]. embryonic division). Ingression of the E daughters (yellow asterisks in Fig. 4D) takes about 16 minutes, and creates a Control of cell ingression transient opening on the surface of the embryo. This opening Cells in the left half of the ‘wishbone’ group of MS is sealed as neighboring cells, primarily MS descendants and descendants ingress at the same time and have the same fate P4 (red and cyan arrowheads, respectively, Fig. 4D), spread as cells in the right half of the wishbone [see Fig. 4C; see also across the surface of the E daughters. In scanning electron Sulston et al. (Sulston et al., 1983)]. The observation that cells micrographs of the ventral surface, the MS descendants show with identical fates, but with different positions, can ingress at wedge-shaped processes that are oriented toward the ingressing the same time suggests that cell fate could determine the time E daughters (Fig. 5A,B). The leading edges of these processes of ingression. In their analysis of the C. elegans cell lineage, taper into thin sheets over the surfaces of the E daughters (Fig. Sulston et al. (Sulston et al., 1983) documented several 5C). Although the P4 blastomere moves across the surfaces of examples of pairs of cells that were located in different parts the E daughters, it does not extend processes resembling those of the embryo but that had identical, or nearly identical, from the MS descendants (Fig. 5). patterns of division and differentiation. We selected nine such By 150 minutes most of the ventral surface of the embryo examples of AB descendants (cell pairs numbered a and b is occupied by the 16 descendants of the MS blastomere. 12 in Table 1), and determined their ingression times from of the MS descendants ingress at about 150 minutes (red videorecordings of live embryos. Cells with the same fate ‘wishbone’ shape in Fig. 4C). The 4 remaining MS usually had remarkably similar ingression times (8/9 cases). descendants stay on the surface of the embryo for an additional For example, the ingression times of cells 4a and 4b differed cell cycle before ingressing at about 190 minutes (red oval by only 1 minute (Table 1). In contrast, cells that were born in shape in Fig. 4C, and red asterisks in Fig. 4E). The ingression equivalent cell cycles, but that had different fates could differ of the ‘wishbone’ group of 12 MS descendants, as well as the significantly in ingression times. For example, the ingression ingression of the P4 descendants, each creates a transient times of cells 4a and 9b differed by 77 minutes (Table 1). surface gap that is quickly enclosed by neighboring cells. In contrast, the surface gaps created by ingression of the D Patterning ingressions in MS descendants descendants (arrow, Fig. 4E) and by ingression of the If cell fate, rather than position, is the sole determinant of remaining MS descendants (arrows, Fig. 4F) persist for 20-30 ingression, mutations that affect cell fate should alter the minutes. In all cases examined, sister cells ingressed either pattern of ingression. Mutations in the mex-1 gene cause AB simultaneously or within a few minutes of each other (see descendants to express a transcription factor that normally C. elegans polarity and gastrulation 393

Table 1. Difference in ingression times of AB descendants Pair* 1b 2b 3b 4b 5b 6b 7b 8b 9b Time (minutes)‡ 193 221 217 212 269 232 245 296 288 1a193028241976395210395 2a212199505720338476 3a 210 17 11 7 2 59 22 35 86 78 4a 211 18 10 6 1 58 21 34 85 77 5a 238 45 17 21 26 31§ 6 7 58 50 6a 241 48 20 24 29 28 9 4 55 47 7a 243 50 22 26 31 26 11 2 53 45 8a 287 94 66 70 75 18 55 42 9 1 9a 288 95 67 71 76 19 56 43 8 0

*Each number/letter combination represents an AB descendant; pairs of descendants with the same number prefix have similar lineages. Pairs 1a,b and 2a,b ingress in the 8th AB cell cycle while all other pairs ingress in the 9th AB cell cycle. 1a,b: ABaraapaa, ABaraappa; 2a,b: ABalpaaap, ABarapaap; 3a,b: ABalpapppp, ABarapappp; 4a,b: ABalpapppa, ABarapappa; 5a,b: ABalpappaa, ABarapapaa; 6a,b: ABalpaapap, ABaraaapap; 7a,b: ABalpaapaa, ABaraaapaa; 8a,b: ABalpaaaap, ABarapaaap; 9a,b: ABalpaaaaa, ABarapaaaa. ‡Time when ingression initiates (in minutes from the 2-cell stage). Values listed in the remainder of the matrix are differences between these times. Differences among pairs 1 and 2, which ingress one cell cycle earlier than all other pairs, are boxed. §Although the ancestry and subsequent differentiation of cell 5b is most similar to cell 5a, several AB descendants share the identical ‘sublineage’ (Sulston et al., 1983). specifies the MS fate (Bowerman et al., 1993; Mello et al., Rocheleau et al., 1999). We therefore shifted embryos 1992). We found that several AB descendants that do not homozygous for the temperature-sensitive mutation lit- ingress in wild-type embryos did ingress in mex-1 mutant 1(t1512ts) to the restrictive temperature before the birth of the embryos analyzed over the period corresponding to normal MS central cells; this temperature shift should have caused the ingression (n=2). However, many of the descendants of the central cells to have the same pattern of POP-1 expression as transformed AB cells did not ingress, in contrast to wild-type the wishbone cells. However, the central cells did not ingress MS descendants, and the pattern of ingression varied between at the same time as the wishbone cells in any of these each of the mex-1 embryos. experiments (n=7; Fig. 6C). We next asked whether the To understand how ingression times are determined in the presence of the E descendants in the blastocoel might block wild-type MS lineage, we focused on the difference in ingression of the central cells by killing the E blastomere in ingression times between the MS ‘wishbone’ descendants and temperature-shifted lit-1 mutant embryos. In 4 of 6 embryos their closely related relatives in the center of the wishbone (hereafter called the central cells; see Fig. 4C). As described above, the MS central cells (asterisks, Fig. 6A) remain on the surface for one cell cycle after the wishbone cells ingress. The central cells that remain on the surface overlie E descendants in the center of the embryo, while the wishbone cells ingress along the perimeter of the E descendants. We therefore asked whether preventing E ingression would allow the central cells to ingress with the wishbone cells. Although killing the E blastomere with a laser microbeam prevented ingression of the E daughters, the central cells did not ingress along with the wishbone cells (12/12 embryos, Fig. 6B). We obtained identical results in experiments where the fate of the E blastomere was altered by a chromosomal deficiency of the endoderm determining region (Zhu et al., 1997); although the E daughters and many of the subsequent E descendants remained on the surface of the mutant embryos, the wishbone Fig. 6. Control of MS ingression. All panels are ventral views of and central cells showed the wild-type difference in ingression embryos after ingression by the MS wishbone descendants; the MS times (6/6 embryos). central descendants are indicated by asterisks. The central By lineage, the central cells are posterior relatives of the descendants remain on the surface in (A) a normal embryo, (B) an anterior wishbone cells. In several cases studied, anterior- embryo after ablation of the E blastomere (dotted outline) and (C) a temperature-shifted lit-1 mutant embryo. (D) The central posterior differences between closely related cells in C. elegans descendants ingress (arrow) along with the wishbone descendants in are determined by anterior-posterior asymmetry in the a temperature-shifted lit-1 mutant embryo after ablation of the E expression of the transcription factor POP-1 (Lin et al., 1998; blastomere (dotted outline). The lit-1 chromosome was marked with Lin et al., 1995), and POP-1 asymmetry can be controlled by an unc-32 mutation; the embryo shown in B is a temperature-shifted the kinase LIT-1 (Kaletta et al., 1997; Meneghini et al., 1999; unc-32 control. 394 J. Nance and J. R. Priess

Schierenberg (Junkersdorf and Schierenberg, 1992)], suggesting apical ﬂattening is a cell-autonomous process. Cell shape changes often are driven by actin and non-muscle myosin. We therefore examined the distribution of NMY-2, a C. elegans non-muscle myosin (Guo and Kemphues, 1996). NMY-2 is present at low levels at the cell cortex during all stages of the cell cycle (Fig. 7C,D), and is strongly enriched in the contractile ring or midbody of mitotic cells (arrows, Fig. 7C,D). NMY-2 began to show enrichment on the apical surfaces of the E daughters just prior to ingression (24-cell stage, 6/36 embryos). This enrichment was evident in most E daughters at the onset of ingression (26-cell stage, 36/43 embryos; Fig. 7D), and in all E daughters near the end of ingression (28-cell stage, 20/20 embryos. The level of NMY-2 at the apical surface decreased once the E daughters entered the blastocoel (data not shown). We observed a similar apical enrichment of NMY-2 on ingressing MS descendants (arrowhead, Fig. 7F).

DISCUSSION

Embryo architecture and strategies of gastrulation In comparison to other animal systems (see Introduction), Fig. 7. Cell-shape changes in ingressing cells. Images are of living gastrulation in C. elegans involves both cell ingression and embryos (A,B,E) or are confocal sections of fixed embryos after epiboly. Embryogenesis occurs within a fixed space limited by immunostaining for NMY-2 (C,D,F). (C,D,F) Confocal sections. All the eggshell, and the volume of the blastocoel is very small, embryos are oriented as in Fig. 4D, and arrowheads indicate apical less than the volume of a single cell at the onset of gastrulation. surfaces. (A-D) The E daughters are indicated (yellow asterisks) Thus the first cell ingressions must be coupled to simultaneous before ingression (A,C) and during ingression (B,D). redistributions in the masses of other embryonic cells. During (E-F) Ingression of MS descendants (red asterisks). MSapaa is ingression of the E daughters, this redistribution appears to be shown in (E) and two of the MS ‘wishbone’ descendants are shown in (F), in which NMY-2 is green and DNA blue. A,B,E and F are accomplished primarily by the MS descendants that shown at twice the magnification of C and D. immediately spread across the site of ingression. At later embryonic stages the ingressing cells are relatively small and there is less need for the surrounding cells to redistribute their tested, the central cells ingressed at the same time as the mass; at these stages surface gaps persist over the sites of wishbone cells (arrow, Fig. 6D). Because the lit-1(t1512ts) ingression for long periods of time. allele is incompletely penetrant, we consider it likely that POP- Gastrulation in embryos like Drosophila involves sheets of 1 expression was not altered in the two embryos with normal cells that invaginate to form furrows and pockets. While ingression times (see Kaletta et al., 1997). We conclude that adherens junctions link the surface cells of Drosophila the ingression pattern of the MS central descendants is embryos at the time of gastrulation (Oda et al., 1998; Tepass controlled by redundant mechanisms involving cell fate and Hartenstein, 1994), adherens junctions are not visible in C. specification and interactions with E descendants. elegans embryos until most of the cell movements associated with gastrulation are complete (Costa et al., 1998) (our Cell shape changes in ingressing cells unpublished observations). The absence of adherens junctions In living embryos, the apical surfaces of the E daughters in early C. elegans embryos may thus facilitate ingression by initially appear rounded (arrowhead, Fig. 7A), similar to the small groups of cells. In sea urchins, primary mesenchyme apical surfaces of neighboring cells. Shortly before ingression, cells appear to lose adherens junction connections with their apical surfaces flatten (Fig. 7B), and remain flattened neighboring surface cells during ingression (Miller and during ingression; small, transient protrusions of the apical McClay, 1997a; Miller and McClay, 1997b). In our survey of membranes appear during flattening (data not shown). We ingression times in C. elegans embryos, we noted that sister observed a similar apical flattening during ingression of the MS cells ingressed at approximately the same time; solitary cells descendants (Fig. 7E), suggesting that apical flattening is a did not ingress. A likely possibility is that the midbody linkage general characteristic of ingressing cells. between sister cells prevents either sister from ingressing Apical flattening during ingression could be autonomous or alone. Both electron microscopic and immunocytochemical a response to tension generated by neighboring cells. In normal studies have shown that the embryonic cells in C. elegans development, the MS descendants extend processes over the E remain coupled by midbodies to their sisters through most of daughters during ingression (see Fig. 5). However, we found the cell cycle (Krieg et al., 1978) (J. N. and J. R. P., unpublished that the E daughters underwent apical flattening and ingressed observations). even after the MS blastomere was killed with a laser Ingressing cells in C. elegans embryos show an apical microbeam [10/10 embryos; see also Junkersdorf and flattening and an apical accumulation of non-muscle myosin. C. elegans polarity and gastrulation 395

Apical flattening could promote ingression by redistributing even when space in the blastocoel is made available by killing cytoplasm toward the basal region of an ingressing cell, and the E blastomere. We propose that in normal development this create space at the surface of the embryo for neighboring cells mechanism prevents the central cells from attempting to to redistribute their cytoplasm laterally. The invagination of ingress when there is not sufficient space to do so. epithelial sheets in Drosophila gastrulation is also associated with an apical flattening and accumulation of non-muscle Establishing apical-basal polarity and the PAR myosin (Leptin and Grunewald, 1990; Sweeton et al., 1991; proteins Young et al., 1991). Thus the cellular basis for invagination by The events of gastrulation indicate that early embryonic cells epithelial sheets and for ingression by small groups of cells have apical-basal polarity; for example, NMY-2 accumulates may be conserved, with the principal difference being whether at the apical surface of ingressing cells, and the blastocoel the cells are linked by junctional complexes. forms at the basal surfaces. Apical-basal asymmetry is evident in embryonic cells of C. elegans as early as the 4-cell stage: Cell fate and ingression PAR-3 and PAR-6 are localized to the apical surfaces, and In principle, gastrulation could position a subset of unspecified PAR-2 and HMR-1(cadherin) are localized to the basal and cells into the interior of the embryo, where subsequent lateral surfaces of cells. Drosophila and vertebrate homologues events would specify the endodermal and mesodermal fates of PAR-3 and PAR-6, together with the atypical protein kinase appropriate for internal organs. Instead, C. elegans and many PKC-3, have been shown to function in establishing apical- other animals appear to specify the fates of endodermal and basal polarity in epithelial cells and neuroblasts (reviewed by mesodermal progenitors when these progenitors are on the Ohno, 2001). We have shown here that large cell separations embryo surface. Thus there must be direct or indirect are observed between both basal and lateral surfaces of cells mechanisms that couple gastrulation movements with cell fate. in par-3 mutants. This result suggests that par-3(+) may have Our present study provides evidence that cell fate is tightly a role in distinguishing the basal from lateral surfaces during linked to cell ingression in C. elegans. For example, mex-1 blastocoel formation in wild-type embryos, where large mutants with ectopic MS-like cells show ectopic ingressions separations only occur between basal surfaces. during the period MS descendants normally ingress, and a We have shown that cell contacts restrict PAR-3 and PAR-6 mutation in lit-1 that changes the fate of MS descendants can to the contact-free, apical surfaces. This mechanism differs, at change the ingression time of those descendants. Similarly, in least in part, from the mechanism that localizes these same Drosophila the cell fate regulators snail and twist appear to proteins to the anterior surface of the 1-cell embryo. While control invagination (reviewed by Leptin, 1995). PAR-2 has a role in determining PAR-3 localization at the 1- The molecular connections between cell fate and ingression cell stage, PAR-2 is not required for the apical localization of or invagination are not known. Embryonic transcription is PAR-3 at the 4-cell stage, nor is it required for blastocoel required for gastrulation in C. elegans (Powell-Coffman et al., formation. HMR-1(cadherin) also is localized to cell contacts, 1996), though we have shown here that transcription is not but does not appear to have a role in PAR localization or required for the apical-basal localization of the PAR proteins blastocoel formation (our present results) (Costa et al., 1998). or for blastocoel formation. Instead, embryonic transcription is Interestingly, genetic or immunological inhibition of E- required for the apical concentration of non-muscle myosin in cadherin function in early mouse embryos does not prevent ingressing cells and for the apical flattening of these cells. A individual cells from becoming polarized, but rather causes a simple model is that the cellular machinery required for randomization in the axis of polarity (Johnson et al., 1986; ingression or invagination is regulated by one master control Larue et al., 1994; Riethmacher et al., 1995). HMR-1 appears gene, and that this gene is in turn regulated by any one of to be the only ‘classical’ cadherin with a β-catenin binding site, several embryonically expressed transcription factors that similar to mouse E-cadherin (Hill et al., 2001), although the C. specify cell fate. By analogy, the Drosophila cell cycle elegans genome sequence predicts several cadherin-related regulator string has a large, complex promoter controlled by proteins whose functions and localization have not been numerous transcription factors that also control region-specific determined (Hill et al., 2001). differentiation, creating regions of mitosis that match regions Localization of PAR-3, or associated proteins, to the apical of common cell fate (Edgar et al., 1994; Johnston and Edgar, surface could in principle differentiate the basal surface from 1998; Lehman et al., 1999). the lateral surface. For example, the localization of ion The control of MS ingression in C. elegans appears to be an channels to the apical surface could create a gradient that example of fine-tuning in a developmental system. After affects the opposite, basal, surface differently than the lateral ingression of the E daughters, the largest available spaces in surfaces. Vectorial ion transport is essential for formation of the blastocoel are anterior to, and on both sides of, the E the blastocoel in mouse embryos, and channel proteins appear daughters. The MS wishbone cells overlie these spaces, and to be localized with apical-basal polarity in trophectodermal ingress into them. In contrast, the MS central cells overlie the cells lining the blastocoel [reviewed by Watson and Barcroft E daughters, and do not ingress at the same time as the (Watson and Barcroft, 2001)]. It will be interesting in future wishbone cells. Our analysis of lit-1 mutant embryos studies to determine how apical-basal polarity of the PAR demonstrates that the presence of the E daughters in the proteins directs subsequent asymmetries. In the 1-cell embryo, blastocoel is sufficient to prevent ingression of the central cells the anterior-posterior polarity of the PAR proteins establishes during ingression of the wishbone cells; we consider it likely a parallel gradient of MEX-5 in the cytoplasm. MEX-5 in turn that this is a steric interaction, but cannot exclude more specific functions to prevent the anterior expression of posterior cell signaling. However, the MS central cells are subject to an proteins (Schubert et al., 2000). However, MEX-5 is uniformly additional, lit-1-dependent, control that prevents ingression distributed in somatic blastomeres that have an apical-basal 396 J. Nance and J. R. Priess polarity in PAR protein distribution, suggesting that MEX-5 in the synchronization and orientation of polarization in 8-cell mouse does not mediate apical-basal polarity. blastomeres. J. Embryol. Exp. Morphol. 93, 239-255. Johnston, L. A. and Edgar, B. A. (1998). Wingless and Notch regulate cell- We thank Judith Austin, Ken Kemphues, Susan Mango and cycle arrest in the developing Drosophila wing. Nature 394, 82-84. Junkersdorf, B. and Schierenberg, E. (1992). Embryogenesis in C. elegans Geraldine Seydoux for providing transgenic strains; Ichiro Kawasaki after elimination of individual blastomeres or induced alteration of the cell and Susan Strome for PGL-1 antibodies; and Kathryn Baker and Ken division order. Roux’s Arch. Dev. Biol. 202, 17-22. Kemphues for the extremely generous gifts of PAR and NMY-2 Kaletta, T., Schnabel, H. and Schnabel, R. (1997). Binary specification of antibodies. We thank Judy Groombridge and Franque Remington for the embryonic lineage in Caenorhabditis elegans. Nature 390, 294-298. assistance with electron microscopy, Jason Pitt for providing advice Kawasaki, I., Shim, Y. H., Kirchner, J., Kaminker, J., Wood, W. B. and and Nematode Navigator software, and members of the Priess Strome, S. (1998). PGL-1, a predicted RNA-binding component of germ laboratory for useful discussions and suggestions. Some of the granules, is essential for fertility in C. elegans. Cell 94, 635-645. nematode strains used in this study were provided by the Knight, J. K. and Wood, W. B. (1998). Gastrulation initiation in Caenorhabditis Genetics Center, which is funded by the NIH National Caenorhabditis elegans requires the function of gad-1, which encodes a protein with WD repeats. Dev. Biol. 198, 253-265. Center for Research Resources (NCRR). J.N. and J.R.P. are supported Komai, T. (1968). Ctenophora. In Invertebrate Embryology (ed. M. Kume and by the Howard Hughes Medical Institute. K. Dan), pp. 117-124. Belgrade: Prosveta. Krieg, C., Cole, T., Deppe, U., Schierenberg, E., Schmitt, D., Yoder, B. and von Ehrenstein, G. (1978). The cellular anatomy of embryos of the REFERENCES nematode Caenorhabditis elegans. Dev. Biol. 65, 193-215. Larue, L., Ohsugi, M., Hirchenhain, J. and Kemler, R. (1994). E-cadherin null mutant embryos fail to form a trophectoderm epithelium. Proc. Natl. Bowerman, B., Draper, B. W., Mello, C. C. and Priess, J. R. (1993). The Acad. Sci. USA 91, 8263-8267. maternal gene skn-1 encodes a protein that is distributed unequally in early Lehman, D. A., Patterson, B., Johnston, L. A., Balzer, T., Britton, J. S., C. elegans embryos. Cell 74, 443-452. Saint, R. and Edgar, B. A. (1999). Cis-regulatory elements of the mitotic Bowerman, B., Eaton, B. A. and Priess, J. R. (1992). skn-1, a maternally regulator, string/Cdc25. Development 126, 1793-1803. expressed gene required to specify the fate of ventral blastomeres in the early Leptin, M. (1995). Drosophila gastrulation: from pattern formation to C. elegans embryo. Cell 68, 1061-1075. morphogenesis. Ann. Rev. Cell Dev. Biol. 11, 189-212. Boyd, L., Guo, S., Levitan, D., Stinchcomb, D. T. and Kemphues, K. J. Leptin, M. and Grunewald, B. (1990). Cell shape changes during gastrulation (1996). PAR-2 is asymmetrically distributed and promotes association of P in Drosophila. Development 110, 73-84. granules and PAR-1 with the cortex in C. elegans embryos. Development Leung, B., Hermann, G. J. and Priess, J. R. (1999). Organogenesis of the 122, 3075-3084. Caenorhabditis elegans intestine. Dev. Biol. 216, 114-134. Braet, F., de Zanger, R. and Wisse, E. (1997). Drying cells for SEM, AFM Lin, R., Hill, R. J. and Priess, J. R. (1998). POP-1 and anterior-posterior fate and TEM by hexamethyldisilazane: a study on hepatic endothelial cells. J. decisions in C. elegans embryos. Cell 92, 229-239. Microsc. 186, 84-87. Lin, R., Thompson, S. and Priess, J. R. (1995). pop-1 encodes an HMG box Brenner, S. (1974). The genetics of Caenorhabditis elegans. Genetics 77, 71- protein required for the specification of a mesoderm precursor in early C. 94. elegans embryos. Cell 83, 599-609. Cheng, N. N., Kirby, C. M. and Kemphues, K. J. (1995). Control of cleavage Martindale, M. Q. and Henry, J. Q. (1999). Intracellular fate mapping in a basal spindle orientation in Caenorhabditis elegans: the role of the genes par-2 metazoan, the ctenophore Mnemiopsis leidyi, reveals the origins of mesoderm and par-3. Genetics 139, 549-559. and the existence of indeterminate cell lineages. Dev. Biol. 214, 243-257. Costa, M., Raich, W., Agbunag, C., Leung, B., Hardin, J. and Priess, J. R. Mello, C. C., Draper, B. W., Krause, M., Weintraub, H. and Priess, J. R. (1998). A putative catenin-cadherin system mediates morphogenesis of the (1992). The pie-1 and mex-1 genes and maternal control of blastomere Caenorhabditis elegans embryo. J. Cell Biol. 141, 297-308. identity in early C. elegans embryos. Cell 70, 163-176. Denich, K. T. R., Schierenberg, E., Ishenghi, E. and Cassada, R. (1984). Meneghini, M. D., Ishitani, T., Carter, J. C., Hisamoto, N., Ninomiya- Cell-lineage and developmental defects of temperature-sensitive embryonic Tsuji, J., Thorpe, C. J., Hamill, D. R., Matsumoto, K. and Bowerman, arrest mutants of the nematode Caenorhabditis elegans. Roux’s Arch. Dev. B. (1999). MAP kinase and Wnt pathways converge to downregulate an Biol. 193, 164-179. HMG-domain repressor in Caenorhabditis elegans. Nature 399, 793-797. Edgar, B. A., Lehman, D. A. and O’Farrell, P. H. (1994). Transcriptional Miller, J. R. and McClay, D. R. (1997a). Changes in the pattern of adherens regulation of string (cdc25): a link between developmental programming junction-associated β-catenin accompany morphogenesis in the sea urchin and the cell cycle. Development 120, 3131-3143. embryo. Dev. Biol. 192, 310-322. Etemad-Moghadam, B., Guo, S. and Kemphues, K. J. (1995). Miller, J. R. and McClay, D. R. (1997b). Characterization of the role of Asymmetrically distributed PAR-3 protein contributes to cell polarity and cadherin in regulating cell adhesion during sea urchin development. Dev. spindle alignment in early C. elegans embryos. Cell 83, 743-752. Biol. 192, 323-339. Fire, A., Xu, S., Montgomery, M. K., Kostas, S. A., Driver, S. E. and Mello, Oda, H., Tsukita, S. and Takeichi, M. (1998). Dynamic behavior of the C. C. (1998). Potent and specific genetic interference by double-stranded cadherin-based cell-cell adhesion system during Drosophila gastrulation. RNA in Caenorhabditis elegans. Nature 391, 806-811. Dev. Biol. 203, 435-450. Gilbert, S. F. and Raunio, A. M. (1997). Embryology: Constructing the Ohno, S. (2001). Intercellular junctions and cellular polarity: the PAR-aPKC Organism. Sunderland, Massachusetts: Sinauer Associates, Inc. complex, a conserved core cassette playing fundamental roles in cell Goldstein, B. (1992). Induction of gut in Caenorhabditis elegans embryos. polarity. Curr. Opin. Cell Biol. 13, 641-648. Nature 357, 255-257. Portereiko, M. F. and Mango, S. E. (2001). Early morphogenesis of the Goldstein, B. and Hird, S. N. (1996). Specification of the anteroposterior axis Caenorhabditis elegans pharynx. Dev. Biol. 233, 482-494. in Caenorhabditis elegans. Development 122, 1467-1474. Powell-Coffman, J. A., Knight, J. and Wood, W. B. (1996). Onset of C. Guo, S. and Kemphues, K. J. (1995). par-1, a gene required for establishing elegans gastrulation is blocked by inhibition of embryonic transcription with polarity in C. elegans embryos, encodes a putative Ser/Thr kinase that is an RNA polymerase antisense RNA. Dev. Biol. 178, 472-483. asymmetrically distributed. Cell 81, 611-620. Pratis, V., Casey, E., Collar, D. and Austin, J. (2001). Creation of low-copy Guo, S. and Kemphues, K. J. (1996). A non-muscle myosin required for integrated transgenic lines in Caenorhabditis elegans. Genetics 157, 1217- embryonic polarity in Caenorhabditis elegans. Nature 382, 455-458. 1226. Hill, E., Broadbent, I. D., Chothia, C. and Pettitt, J. (2001). Cadherin Priess, J. R. and Hirsh, D. I. (1986). Caenorhabditis elegans morphogenesis: superfamily proteins in Caenorhabditis elegans and Drosophila The role of the cytoskeleton in elongation of the embryo. Dev. Biol. 117, melanogaster. J. Mol. Biol. 305, 1011-1024. 156-173. Hung, T. J. and Kemphues, K. J. (1999). PAR-6 is a conserved PDZ domain- Riethmacher, D., Brinkmann, V. and Birchmeier, C. (1995). A targeted containing protein that colocalizes with PAR-3 in Caenorhabditis elegans mutation in the mouse E-cadherin gene results in defective preimplantation embryos. Development 126, 127-135. development. Proc. Natl. Acad. Sci. USA 92, 855-859. Johnson, M. H., Maro, B. and Takeichi, M. (1986). The role of cell adhesion Rocheleau, C. E., Downs, W. D., Lin, R., Wittmann, C., Bei, Y., Cha, Y. H., C. elegans polarity and gastrulation 397

Ali, M., Priess, J. R. and Mello, C. C. (1997). Wnt signaling and an APC- The embryonic cell lineage of the nematode Caenorhabditis elegans. Dev. related gene specify endoderm in early C. elegans embryos. Cell 90, 707- Biol. 100, 64-119. 716. Sweeton, D., Parks, S., Costa, M. and Wieschaus, E. (1991). Gastrulation Rocheleau, C. E., Yasuda, J., Shin, T. H., Lin, R., Sawa, H., Okano, H., in Drosophila: formation of the ventral furrow and posterior midgut Priess, J. R., Davis, R. J. and Mello, C. C. (1999). WRM-1 activates the invaginations. Development 112, 775-789. LIT-1 protein kinase to transduce anterior/posterior polarity signals in C. Tabara, H., Grishok, A. and Mello, C. C. (1998). RNAi in C. elegans: elegans. Cell 97, 717-726. soaking in the genome sequence. Science 282, 430-431. Rogalski, T. M. and Riddle, D. L. (1988). A Caenorhabditis elegans RNA Tabuse, Y., Izumi, Y., Piano, F., Kemphues, K. J., Miwa, J. and Ohno, S. polymerase II gene, ama-1 IV, and nearby essential genes. Genetics 118, (1998). Atypical protein kinase C cooperates with PAR-3 to establish 61-74. embryonic polarity in Caenorhabditis elegans. Development 125, 3607- Schierenberg, E. (1987). Reversal of polarity and early cell-cell interaction in 3614. the embryo of C. elegans. Dev. Biol. 122, 452-463. Tepass, U. and Hartenstein, V. (1994). Development of intercellular junctions Schnabel, R. and Priess, J. R. (1997). Speciﬁcation of cell fates in the early in the Drosophila embryo. Dev. Biol. 161, 563-596. embryo. In C. elegans II (ed. D. L. Riddle, T. Blumenthal, B. J. Meyer and Thomas, C., DeVries, P., Hardin, J. and White, J. (1996). Four-Dimensional J. R. Priess), pp. 361-382. Plainview, New York: Cold Spring Harbor Imaging: Computer Visualization of 3D Movements in Living Specimens. Laboratory Press. Science 273, 603-607. Schubert, C. M., Lin, R., de Vries, C. J., Plasterk, R. H. A. and Priess, J. Wallenfang, M. R. and Seydoux, G. (2000). Polarization of the anterior- R. (2000). MEX-5 and MEX-6 function to establish soma/germline posterior axis of C. elegans is a microtubule-directed process. Nature 408, asymmetry in early C. elegans embryos. Mol. Cell 5, 671-682. 89-92. Seydoux, G., Mello, C. C., Pettitt, J., Wood, W. B., Priess, J. R. and Fire, Watson, A. J. and Barcroft, L. C. (2001). Regulation of blastocyst formation. A. (1996). Repression of gene expression in the embryonic germ lineage of Front. Biosci. 6, D708-730. C. elegans. Nature 382, 713-716. Young, P. E., Pesacreata, T. C. and Kiehart, D. P. (1991). Dynamic changes Shelton, C. A. and Bowerman, B. (1996). Time-dependent responses to glp- in the distribution of cytoplasmic myosin during Drosophila embryogenesis. 1-mediated inductions in early C. elegans embryos. Development 122, 2043- Development 111, 1-14. 2050. Zhu, J., Hill, R. J., Heid, P. J., Fukuyama, M., Sugimoto, A., Priess, J. R. Simske, J. S. and Hardin, J. (2001). Getting into shape: epidermal and Rothman, J. H. (1997). end-1 encodes an apparent GATA factor that morphogenesis in Caenorhabditis elegans embryos. BioEssays 22, 12-23. speciﬁes the endoderm precursor in Caenorhabditis elegans embryos. Genes Sulston, J. E., Schierenberg, E., White, J. G. and Thomson, J. N. (1983). Dev. 11, 2883-2896.