After fertilization in C. elegans, activities encoded by the maternally expressed par genes appear to establish cellular and embryonic polarity. Loss-of-function mutations in the par genes disrupt anterior-posterior (a-p) asymmetries in early embryos and result in highly abnormal patterns of cell fate. Little is known about how the early asymmetry defects are related to the cell fate patterning defects in par mutant embryos, or about how the par gene products affect the localization and activities of developmental regulators known to specify the cell fate patterns made by individual blastomeres. Examples of such regulators of blastomere identity include the maternal proteins MEX-3 and GLP-1, expressed at high levels anteriorly, and SKN-1 and PAL-1, expressed at high levels posteriorly in early embryos. To better define par gene functions, we examined the expression patterns of MEX-3, PAL-1 and SKN-1, and we analyzed mex-3, pal-1, skn-1 and glp-1 activities in par mutant embryos. We have found that mutational inactivation of each par gene results in a unique phenotype, but in no case do we observe a complete loss of a-p asymmetry. We conclude that no one par gene is required for all a-p asymmetry and we suggest that, in some cases, the par genes act independently of each other to control cell fate patterning and polarity. Finally, we discuss the implications of our findings for understanding how the initial establishment of polarity in the zygote by the par gene products leads to the proper localization of more specifically acting regulators of blastomere identity.

The generation of asymmetry before mitosis is a general mechanism for producing cells with different fates during growth and development (for a review, see Gonczy and Hyman, 1996). For example, sperm entry provides an extrinsic cue to polarize the 1-cell zygote in C. elegans, defining the anterior-posterior (a-p) body axis (Hird and White, 1993; Goldstein and Hird, 1996). Polarization of the nematode zygote results in posterior displacement of the first mitotic spindle, producing a smaller P1 and a larger AB daughter, two cells with very different fates (Sulston et al., 1983). AB and P1 differ in their cell cycle times and in the orientation of their mitotic spindles: AB divides first with a transversely oriented spindle to make two daughters of equal size, while P1 divides after AB, with its spindle aligned parallel to the long axis and displaced posteriorly, dividing unequally (Hyman and White, 1987; Hyman, 1989; see Fig. 1). The different developmental potentials of P1 and AB also are apparent when each is cultured in isolation: only P1 and not AB can produce pharyngeal cells, intestinal cells and body wall muscle cells (Laufer et al., 1980; Priess and Thomson, 1987; Draper et al., 1996). Consistent with their different developmental potentials, genetically identified regulators of pattern formation are differentially localized to P1 and AB, and their descendants (see Fig, 1).

Fig. 1.

Schematic of polarization and patterning in the early C. elegans embryo. Blastomere names (P0, P1, AB, P2, EMS, ABa, ABp, P3, C, E and MS) are adjacent to the corresponding cells. (A) Early 1-cell-stage embryo, called P0, after fertilization. The maternal pronucleus is at the left end of the zygote, and the paternal pronucleus, with an associated centriole, is at the right. Sperm entry typically occurs opposite the maternal pronucleus. After fertilization, the maternal pronucleus completes meiosis, producing two polar bodies (small open circle). After completion of meiosis, an actindependent cytoplasmic flux occurs (curved and dashed arrows). (B) Dividing 1-cell-stage embryo, with the first mitotic spindle aligned on the long axis and displaced posteriorly. The polarized distributions of PAR-1 and PAR-2 (brown), and PAR-3 (tan), are inherited by the P0 descendants P1, P2 and P3, precursors to the final germline progenitor P4. For simplification, the PAR distributions are shown only at the 1-cell stage. (C) Dividing 2-cell-stage embryo, with AB dividing earlier and transversely and P1 dividing later and longitudinally. The GLP-1 transmembrane receptor, related to Drosophila Notch, is first detectable late in the 2-cell stage at the interface of P1 and AB (purple). The putative transcription factor SKN-1 accumulates to high levels in the nucleus of P1 by late in the 2-cell stage (blue). Cytoplasmic MEX-3 (green) is a putative RNAbinding protein with KH domains. Dark green in AB indicates higher levels relative to the light green in P1. (D) A 4-cell-stage embryo, when a dorsal-ventral axis is established. The transcription factors PAL-1 (red) and SKN-1 (blue) are present in the nuclei of P2 and EMS. PAL-1 specifies the identity of two P2 descendants, C and D. Low levels of MEX-3 persist in some posterior blastomeres at later stages (not shown). (E) A 12-cell-stage embryo. Two Aba granddaughters (shaded grey) respond to a signal from MS by adopting fates that include the production of pharyngeal cells. By the 12-cell-stage SKN-1 is no longer detectable, but PAL-1 expression persists. C produces epidermis and body wall muscle, MS produces pharynx and body wall muscle, and E produces all of the intestine. P3 divides to produce D, a body wall muscle precursor, and P4, the germline progenitor. See Sulston et al., 1983, for a complete description of the embryonic lineage. See text for additional references.

Fig. 1.

Schematic of polarization and patterning in the early C. elegans embryo. Blastomere names (P0, P1, AB, P2, EMS, ABa, ABp, P3, C, E and MS) are adjacent to the corresponding cells. (A) Early 1-cell-stage embryo, called P0, after fertilization. The maternal pronucleus is at the left end of the zygote, and the paternal pronucleus, with an associated centriole, is at the right. Sperm entry typically occurs opposite the maternal pronucleus. After fertilization, the maternal pronucleus completes meiosis, producing two polar bodies (small open circle). After completion of meiosis, an actindependent cytoplasmic flux occurs (curved and dashed arrows). (B) Dividing 1-cell-stage embryo, with the first mitotic spindle aligned on the long axis and displaced posteriorly. The polarized distributions of PAR-1 and PAR-2 (brown), and PAR-3 (tan), are inherited by the P0 descendants P1, P2 and P3, precursors to the final germline progenitor P4. For simplification, the PAR distributions are shown only at the 1-cell stage. (C) Dividing 2-cell-stage embryo, with AB dividing earlier and transversely and P1 dividing later and longitudinally. The GLP-1 transmembrane receptor, related to Drosophila Notch, is first detectable late in the 2-cell stage at the interface of P1 and AB (purple). The putative transcription factor SKN-1 accumulates to high levels in the nucleus of P1 by late in the 2-cell stage (blue). Cytoplasmic MEX-3 (green) is a putative RNAbinding protein with KH domains. Dark green in AB indicates higher levels relative to the light green in P1. (D) A 4-cell-stage embryo, when a dorsal-ventral axis is established. The transcription factors PAL-1 (red) and SKN-1 (blue) are present in the nuclei of P2 and EMS. PAL-1 specifies the identity of two P2 descendants, C and D. Low levels of MEX-3 persist in some posterior blastomeres at later stages (not shown). (E) A 12-cell-stage embryo. Two Aba granddaughters (shaded grey) respond to a signal from MS by adopting fates that include the production of pharyngeal cells. By the 12-cell-stage SKN-1 is no longer detectable, but PAL-1 expression persists. C produces epidermis and body wall muscle, MS produces pharynx and body wall muscle, and E produces all of the intestine. P3 divides to produce D, a body wall muscle precursor, and P4, the germline progenitor. See Sulston et al., 1983, for a complete description of the embryonic lineage. See text for additional references.

Of the maternal genes known to control pattern formation in C. elegans, mutations in six, par-1 through par-6, cause the earliest and most extensive polarity defects in the zygote, eliminating many early asymmetries (Kemphues et al., 1988; Kirby et al., 1990; Morton et al., 1992; Cheng et al., 1995; Guo and Kemphues, 1996; Kemphues and Strome, 1997). For example, in all par mutant embryos P1 and AB divide synchronously, and they are of equal size in all but par-4 mutant embryos. Furthermore, of the four PAR proteins examined thus far, all are present at low levels throughout the cytoplasm and at much higher levels in the cytoplasmic cortex, with the cortical enrichment being polarized along the a-p axis for PAR-1, PAR-2 and PAR-3 (Levitan et al., 1994; Etemad-Moghadam and Kemphues, 1995; Guo and Kemphues, 1995; Boyd et al., 1996); J. Watts, D. Morton and K. Kemphues, personal communication). Initially, PAR-1 and PAR-2 are present cortically through-out mature oocytes but, after fertilization, both are detected only in the posterior cortex of the 1-cell zygote, called P0 (Guo and Kemphues, 1995; Boyd et al., 1996; see Fig. 1). Complementary to PAR-1 and PAR-2, cortical PAR-3 is present only in the anterior of P0, and the posterior boundary of the cortical PAR-3 in P0 abutts the anterior boundary of PAR-1 and PAR-2 (Etemad-Moghadam and Kemphues, 1995; Kemphues and Strome, 1997; see Fig. 1). PAR-4 is unique in that although enriched cortically it is not polarized along the a-p axis (J. Watts, D. Morton, and K. Kemphues, personal communication).

While four par genes have been identified molecularly, how they establish or regulate polarity remains almost entirely unkown. PAR-1 contains a predicted N-terminal ser/thr kinase domain and a C-terminal domain that interacts with a non-muscle conventional myosin (Guo and Kemphues, 1996a). PAR-2 contains a putative ATP-binding site and a zinc-binding domain of the ‘RING finger’ class (Levitan et al., 1994). PAR-3 is a novel protein with three PDZ repeats, and PAR-4 contains a ser/thr kinase domain (Etemad-Moghadam and Kemphues, 1995); J. Watts and K. Kemphues, personal communication). Intriguingly, PAR-3 extends posteriorly in par-2 mutant embryos, and PAR-2 extends anteriorly in par-3 mutants, suggesting that par-2 and par-3 interact during the specification of a-p polarity (Etemad-Moghadam and Kemphues, 1995; Boyd et al., 1996). Indicating another inter-action, PAR-1 is not enriched cortically in par-2 mutant embryos (Guo and Kemphues, 1995). Because par mutant embryos exhibit very early defects in polarization of the zygote, and because the PAR proteins themselves become asymmetrically distributed shortly after fertilization, the par genes likely encode machinery that polarizes the zygote directly in response to sperm entry, initiating processes that ultimately localize more specifically acting regulatory factors to individual blastomeres (Guo and Kemphues, 1996b; Kemphues and Strome, 1997). Homologs of par-1 called MARK kinases have been identified in yeast and in mammals (Levin et al., 1987; Levin and Bishop, 1990; Drewes et al., 1997). The involvement of these closely related kinases in regulating polarity in yeast and microtubule stability in mammalian cells indicates that understanding par gene functions in C. elegans is of general importance.

To better understand the par genes, we investigated how mutations in par-1, par-2, par-3 and par-4 affect the function of four other maternally expressed gene products encoded by skn-1, glp-1, pal-1 and mex-3. Mutations in these latter genes cause defects in the fates of individual blastomeres without causing more general defects in polarity characteristic of par mutants. The skn-1 gene encodes a putative transcription factor required to specify the fate of one P1 daughter, EMS (Bowerman et al., 1992, 1993; Blackwell et al., 1994). skn-1 also activates a signal that induces ABa descendants to produce pharyngeal cells (Shelton and Bowerman, 1996), and glp-1 encodes a Notch-like receptor required for ABa descendants to respond to this signal (Priess and Thomson, 1987; Austin and Kimble, 1989; Yochem and Greenwald, 1989; Evans et al., 1994; Fig. 1). PAL-1 is a homeodomain protein and specifies the production of body wall muscle and epidermis by the P2 daughter of P1 (Hunter and Kenyon, 1996). MEX-3 is a putative RNA-binding protein required to prevent translation of pal-1 mRNA in anterior blastomeres in the early embryo (Draper et al., 1996; Hunter and Kenyon, 1996).

By examining the expression and function of SKN-1, GLP-1, PAL-1 and MEX-3, we show that par-1, par-2 and par-3 mutant embryos, although superficially similar in terminal phenotype, each exhibit a unique set of cell fate patterning defects. par-4 mutant embryos show even less phenotypic similarity to the other par mutants with respect to the develop-mental pathways that we analyze. These results indicate that the par genes have substantially distinct roles in segregating different cell fate specification activities. Therefore we think it unlikely that the par genes operate in a single pathway. Instead, we suggest that the par genes encode part of a possibly complex network of interacting gene products that link polarization in the zygote to the patterning of specific blastomere identities.

Strains and alleles

Nematode cultures were maintained as described (Brenner, 1974). The genotypes of strains used for analysis of mutant phenotypes were as follows: +/DnT1 (IV;V) IV; par-1(e2012) rol-4(sc8)/DnT1(IV;V) V, skn-1(zu67)/DnT1(IV;V) IV; par-1(e2012) rol-4(sc8)/DnT1(IV;V), par-2(lw32) unc-45(e286ts)/sC1[dpy-1(e1) let] III, par-2(lw32) unc-45(e286ts)/sC1[dpy-1(e1) let] III; dpy-13(e184sd) skn-1(zu67)/nT1(IV;V) IV; +/nT1(IV;V) V, par-2(lw32) dpy-1(e1) unc-32(e189) glp-1(e2142ts)/ dpy-17(e224) unc-32(e189) glp-1(e2142ts) III, par-3(it71) lon-1(e185)/qC1[dpy-19(e1259ts) glp-1(q339)] III, par-3(it71) lon-1(e185) unc-32(e189) glp-1(e2142ts)/ dpy-17(e164) unc-32(e189) glp-1(e2142ts) III, lon-1(e185) par-3(it71)/qC1[dpy-19(e1259ts) glp-1(q339)] III; skn-1(zu67)/DnT1(IV;V) IV; +/DnT1(IV;V)V, par-4(it47ts) V, glp-1(e2142ts) unc-32(e189) III; par-4(it47ts) V, mDp1 (f,IV); unc-5(e53) skn-1(zu67) IV; par-4(it57ts) V. All glp-1 strains and par-4 strains were maintained at 15°C and adults shifted to 25°C 2 hours before collection of embryos. DpyUnc par-2 glp-1 mothers were identified because they were more severely Dpy than DpyUnc glp-1 siblings. In all ablation experiments, the presence of par-2 or par-3 were confirmed by using only those embryos in which P1 and AB were of equal size and by confirming the phenotype of unablated siblings using Nomarski optics to score characteristic cell types before transferring partial embryos for antibody staining. In ablations using par-4(it47ts), mutant embryos were identified by waiting to score the synchrony of the P1 and AB cleavages.

The skn-1 allele zu67 is the strongest skn-1 allele and corresponds to a premature stop codon in the second coding exon (C. Schubert, B. Bowerman and J. Priess, unpublished data). At 25°C, the glp-1 allele e2142ts specifically eliminates the ability of GLP-1 to respond to the 12-cell-stage MS induction of pharyngeal cell production by ABa descendants, without affecting the response to an earlier signal from P2 (Hutter and Schnabel, 1994; Mello et al., 1994). par-1(e2012) is the most penetrant par-1 allele identified with respect to the lack of intestinal cells (Kemphues et al., 1988). par-2(lw32) is a strong allele with a nonsense mutation predicted to encode 233 of 628 predicted amino acids (Levitan et al., 1994). par-3(it71) appears to be a protein null (Cheng et al., 1995; Etemad-Moghadam and Kemphues, 1995). The par-4 allele it57ts is a strong allele when grown at 25°C (Cheng et al., 1995).

Embryo manipulations and microscopy

1-cell and 2-cell wild-type and mutant embryos were collected by cutting open gravid adults in a watch glass filled with M9 buffer, mounted on 3% agarose pads under a coverslip as described (Sulston et al., 1983). Blastomeres were killed using 10-25 pulses from a Laser Science, Inc., VSL-337 laser microbeam directed into a Zeiss Axioskop (Avery and Horvitz, 1989). Partial embryos were allowed to develop overnight for 15-20 hours at room temperature for analysis of pharyngeal muscle cell production, or for 8-10 hours at room temperature for analysis of body wall muscle cell production. Antibody staining procedures for pharyngeal muscle and body wall muscle, and for SKN-1, PAL-1 and MEX-3 were as described (Bowerman et al., 1993; Draper et al., 1996; Hunter and Kenyon, 1996). Intestinal cell-specific gut granules were scored using polarizing optics (Bowerman et al., 1992). For analysis of unablated embryos, homozygous mutant mothers were collected and either allowed to lay embryos for several hours or cut open to obtain embryos; the embryos were then allowed to develop and analyzed for pharyngeal muscle and body wall muscle cells as for partial embryos. We used Kodak Technical Pan film and HC110 developer, a Polaroid Sprint Scan 35 and Adobe Photoshop Version 4.0 for making figures.

Specification of pharyngeal cell fates in the early embryo

To define requirements for the par genes in segregating cell fate specification activities, we first examined the patterning of pharyngeal cell fates in par mutant embryos. Wild-type embryos produce 37 pharyngeal muscle cells, 19 mostly anterior muscles derived from ABa and 18 mostly posterior muscles from MS (Sulston et al., 1983; see Fig. 2A). As illustrated in Fig. 1, two granddaughters of ABa require glp-1 function to produce pharyngeal cells in response to an inductive signal from MS at about the 12-cell stage (Priess et al., 1987; Mango et al., 1994; Hutter and Schnabel, 1995). Consequently, in glp-1 mutant embryos only MS produces pharyngeal muscle cells (Priess et al., 1987; see Fig. 2B). skn-1 function is required both for specifying pharyngeal cell production by MS and for activating the signal from MS that induces production of ABa-derived pharyngeal cells (Bowerman et al., 1992; Shelton and Bowerman, 1996). Consequently skn-1 mutant embryos fail to produce any pharyn-geal cells (Fig. 2C). The requirements for the par genes in restricting pharyngeal cell specification activities to individual blastomeres have been examined most extensively in par-1 mutant embryos (Bowerman et al., 1993; Draper et al., 1996; Hunter and Kenyon, 1996). Here, we report an analysis of glp-1 and skn-1 function in par-2, par-3 and par-4 mutant embryos.

Fig. 2.

glp-1 and skn-1 function and the production of pharyngeal muscle cells in par-1, par-2, par-3 and par-4 mutant embryos. Fluorescence micrographs of terminally differentiated wild-type (A) and mutant (B-O) embryos stained with monoclonal antibody 9.2.1 to visualize pharyngeal muscle cells (Miller et al., 1983). Embryos are oriented with anterior to the left. Genotypes are indicated by labels to left and at top. (A) Wild-type embryo near hatching with both P1- and AB-derived pharynx. (B) glp-1(e2142ts) embryo with only P1-derived pharynx. (C) skn-1(zu67) embryo with no pharynx (D, E) par-1(e2012) and glp-1(e2142ts); par-1(e2012) embryos made indistinguishable amounts of pharyngeal muscle. (F) skn1(zu67); par-1(e2012) embryo with no pharyngeal muscle [see Bowerman et al., 1993, for quantitation of par-1 and glp-1; par-1 and skn-1; par-1 pharyngeal muscle phenotypes]. (G) par-2(lw32) embryos made large clumps of pharyngeal muscle, but (H) par- 2(lw32) glp-1(e2142ts) embryos made much less pharyngeal muscle, often in two small clumps. (I) par-2(lw32); skn-1(zu67) embryo with no pharyngeal muscle: 111/128 par-2; skn-1 embryos made no pharyngeal muscle and 17/128 made only 2-8 pharyngeal muscle cells. (J,K) par-3(it71) and par-3(it71) glp-1(e2142ts) embryos made indistinguishable amounts of pharyngeal muscle. (L) par-3(it71); skn-1(zu67) embryo with no pharynx: 170/189 par-3; skn-1 embryos made no pharyngeal muscle and 19/189 made 2-16 pharyngeal muscle cells. Similar results were obtained using par-3(it62); skn-1(zu67) embryos (data not shown). (M,N) par-4(it47ts) and glp-1(e2142ts); par-4(it47ts) mutant embryos with large clumps of pharyngeal muscle cells located posteriorly. 234/424 par-4 embryos (55%) made pharyngeal muscle cells (10 to >50 cells/ embryo; 30 to 40 cells in most embryos). 190/424 (45%) made no pharyngeal muscle cells. (O) skn-1(zu67); par-4(it47) embryo with no pharyngeal muscle cells. 372/395 skn-1; par-4 embryos made no pharyngeal cells; 23/395 made from 2 to 9 pharyngeal muscle cells. See Materials and Methods for a description of the strains and alleles used and for fixation and antibody staining procedures.

Fig. 2.

glp-1 and skn-1 function and the production of pharyngeal muscle cells in par-1, par-2, par-3 and par-4 mutant embryos. Fluorescence micrographs of terminally differentiated wild-type (A) and mutant (B-O) embryos stained with monoclonal antibody 9.2.1 to visualize pharyngeal muscle cells (Miller et al., 1983). Embryos are oriented with anterior to the left. Genotypes are indicated by labels to left and at top. (A) Wild-type embryo near hatching with both P1- and AB-derived pharynx. (B) glp-1(e2142ts) embryo with only P1-derived pharynx. (C) skn-1(zu67) embryo with no pharynx (D, E) par-1(e2012) and glp-1(e2142ts); par-1(e2012) embryos made indistinguishable amounts of pharyngeal muscle. (F) skn1(zu67); par-1(e2012) embryo with no pharyngeal muscle [see Bowerman et al., 1993, for quantitation of par-1 and glp-1; par-1 and skn-1; par-1 pharyngeal muscle phenotypes]. (G) par-2(lw32) embryos made large clumps of pharyngeal muscle, but (H) par- 2(lw32) glp-1(e2142ts) embryos made much less pharyngeal muscle, often in two small clumps. (I) par-2(lw32); skn-1(zu67) embryo with no pharyngeal muscle: 111/128 par-2; skn-1 embryos made no pharyngeal muscle and 17/128 made only 2-8 pharyngeal muscle cells. (J,K) par-3(it71) and par-3(it71) glp-1(e2142ts) embryos made indistinguishable amounts of pharyngeal muscle. (L) par-3(it71); skn-1(zu67) embryo with no pharynx: 170/189 par-3; skn-1 embryos made no pharyngeal muscle and 19/189 made 2-16 pharyngeal muscle cells. Similar results were obtained using par-3(it62); skn-1(zu67) embryos (data not shown). (M,N) par-4(it47ts) and glp-1(e2142ts); par-4(it47ts) mutant embryos with large clumps of pharyngeal muscle cells located posteriorly. 234/424 par-4 embryos (55%) made pharyngeal muscle cells (10 to >50 cells/ embryo; 30 to 40 cells in most embryos). 190/424 (45%) made no pharyngeal muscle cells. (O) skn-1(zu67); par-4(it47) embryo with no pharyngeal muscle cells. 372/395 skn-1; par-4 embryos made no pharyngeal cells; 23/395 made from 2 to 9 pharyngeal muscle cells. See Materials and Methods for a description of the strains and alleles used and for fixation and antibody staining procedures.

Pharyngeal cell fate specification pathways in par-2, par-3 and par-4 mutant embryos

To analyze the activities of skn-1 and glp-1 in par mutant embryos, we examined the patterning of pharyngeal cell fates in par mutant embryos in which either glp-1 or skn-1 activity was absent (Fig. 2). As shown previously for par-1 mutants, we have found that most par-2 and par-3 mutants produce a large excess of pharyngeal muscle cells in both the anterior and posterior portions of the embryo (Fig. 2G,J). The pharyngeal phenotype of par-4 mutant embryos is more variable: large numbers of pharyngeal muscle cells are present posteriorly in about 50% of par-4 mutant embryos (Fig. 2M), but no pharyngeal cells are made in the remaining 50% (Fig. 2 legend). As expected, all par mutant embryos require skn-1 function to produce any pharyngeal cells (Fig. 2F,I,L,O). We next asked if the production of pharynx was dependent on glp-1 activity. We found that par-3; glp-1 and par-4; glp-1 double mutant embryos produce indistinguishable numbers and distributions of pharyngeal cells compared to par-3 and par-4 single mutants (compare Fig. 2J and K, M and N). In contrast, par-2; glp-1 double mutant embryos produce substantially fewer pharyngeal cells than do par-2 mutant embryos (compare Fig. 2G and H). Thus par-2 mutant embryos are unique in producing glp-1-dependent pharynx anteriorly. In par-1, par-3 and par-4 mutant embryos, all pharyngeal cell production is skn-1- dependent and none appears glp-1-dependent.

Pharyngeal cell production by P1 and AB in par mutant embryos

To determine conclusively the origins of the pharyngeal muscle made by par mutant embryos, we analyzed the abilities of mutant P1 and AB blastomeres, ‘isolated’ by laser ablation (see Materials and Methods), to produce pharynx. Based on the a-p distribution of pharyngeal cells in terminally differentiated mutant embryos (Fig. 2), it appears that in par-1, par-2 and par-3 mutants, both P1 and AB can produce pharyngeal cells. However, of those par-4 mutant embryos that make pharyngeal cells, they do so only in the posterior part of the embryo, presumably from P1. Using laser ablation experiments, we found that, as reported previously for par-1 mutant embryos (Bowerman et al., 1993), both P1 and AB from par-3 mutant embryos almost always produced large numbers of pharyngeal muscle cells, and they did so independent of glp-1 function (Table 1). In contrast to par-1 and par-3, isolated AB blastomeres from par-2 mutant embryos often produce no or few pharyngeal muscle cells (Table 1), consistent with the observation that par-2 embryos produce glp-1-dependent pharyngeal cells anteriorly (Fig. 2G,H). To test more conclusively if AB descendants in par-2 mutant embryos produce glp-1-dependent pharyngeal cells, we killed the P1 descendants in par-2 mutant embryos after the time at which pharyngeal induction occurs in wild-type embryos (Mango et al., 1994; Hutter and Schnabel, 1995). We found that, when P1 descendants were killed at these later stages, AB descendants always produced large numbers of pharyngeal muscles cells (Table 1). Further-more, when P1 descendants in par-2; glp-1 double mutant embryos were killed at these later stages, AB descendants did not make pharynx (Table 1). We conclude that production of pharynx by AB descendants in par-2 mutant embryos requires glp-1-dependent cell interactions. Finally, both in par-4 single mutant embryos and in glp-1; par-4 double mutant embryos, about 50% of isolated P1 blastomeres produced large numbers of pharyngeal cells, while AB descendants never produced pharynx (Table 1). This finding correlates well with the observation that 50% of intact par-4 mutant embryos produce pharyngeal cells posteriorly (Fig. 2G).

Table 1.

Production of pharyngeal muscle cells and intestinal cells by AB and P1

Production of pharyngeal muscle cells and intestinal cells by AB and P1
Production of pharyngeal muscle cells and intestinal cells by AB and P1

Pharyngeal cell production by P1 and AB daughters in par mutant embryos

We further analyzed the patterning of pharynx by examining the abilities of individual blastomeres in 4-cell-stage par mutant embryos to produce pharyngeal muscle. In wild-type embryos only one P1 daughter and one AB daughter produce pharynx (Sulston et al., 1983). However, in par-1 mutant embryos, both P1 and both AB daughters usually produce pharyngeal cells (Bowerman et al., 1993). Because the cleavage axes of P1 and AB in par-2 and par-4 are highly variable, it was not possible to reproducibly identify individual 4-cell-stage blastomeres. As the daughters cannot be distinguished, we report the number of randomly isolated P1 daughters that made pharyngeal cells/the number of embryos analyzed in par-2, par-3 and par-4 mutant embryos. In par-2 mutants, 8/9 isolated P1 daughters made pharyngeal cells and 6/6 AB daughters made pharyngeal cells, with the latter ablations done at the 12-cell stage. In par-3 mutants, 3/3 isolated anterior P1 daughters, 4/4 isolated posterior P1 daughters, 5/6 isolated anterior AB daughters and 3/3 isolated posterior AB daughters made pharygneal muscle. In par-4 mutants, 5/9 isolated P1 daughters and 0/6 isolated AB daughters made pharyngeal muscle. In all cases scored positive, 5-20 pharngeal muscle cells were present. We conclude that the mechanisms normally limiting pharyngeal cell production to only one P1 and one AB daughter are not functional in par mutant embryos.

SKN-1 distribution in par mutant embryos

As the final step in our analysis of pharyngeal cell fate patterning in par mutant embryos, we examined the spatial and temporal regulation of SKN-1 expression. In wild-type embryos, SKN-1 is present at high levels only in P1 and its descendants from late in the 2-cell stage until the 8-cell stage. Peak levels of SKN-1 are detected at the 4-cell stage in EMS and P2; little or no SKN-1 is detectable in ABa and ABp (Bowerman et al., 1993; see Fig. 3A). As shown previously, SKN-1 is present at roughly equal levels in all 4-cell-stage par-1 blastomeres (Bowerman et al., 1993; Guo and Kemphues, 1995; see Fig. 3B). To determine if mislocalization of SKN-1 is a general property of par mutants, we examined par-2, par-3 and par-4 mutant embryos. We found that, like par-1 mutants, most par-3 mutants have roughly equal SKN-1 levels in all 4-cell-stage blastomeres (Fig. 3B,D). However, SKN-1 localiza-tion appears normal in par-2 and par-4 mutant embryos (Fig. 3C,E). In contrast to the defects in spatial regulation, temporal regulation of SKN-1 expression was not affected by mutations in any of the par genes (see Fig. 3 legend). The ubiquitous expression of SKN-1 in par-1 and par-3 mutant embryos cor-relates with the ability of both P1 and AB to produce skn-1-dependent, glp-1-independent pharyngeal cells. Localization of SKN-1 expression to P1 descendants in par-2 and par-4 mutant embryos correlates with the ability of only P1 but not AB descendants to produce glp-1-independent pharynx in par-2 and par-4 mutants. The absence of ABa-derived pharyngeal cells in par-4 mutant embryos could be due to a failure in signaling by P1 descendants, or to a defective response by AB descendants.

Fig. 3.

SKN-1 distribution in par-1, par-2, par-3 and par-4 mutant embryos. Fluorescence micrographs showing fixed 4-cell-stage wild-type and par mutant embryos stained for SKN-1 protein (left column) and with DAPI to visualize DNA in nuclei (right column). Embryos are oriented with anterior to the left as determined by the position of DAPI-stained polar bodies (visible in I and J; out of the focal plane in the other DAPI images). (A,F) Wild-type embryo with high levels of SKN-1 in P2 and EMS. Lower levels are just detectable in ABa and ABp. Note that the chromosomes are more highly condensed in ABa and ABp than in P2 and EMS. In DAPI-stained par mutant embryos (G-I), chromosome condensation appears equivalent as all four cells divide synchronously. (B,G) par-1(2012) mutant embryo. In 22/24 4-cell-stage par-1 embryos, equal levels of staining were detected in all blastomeres; in 2/24 embryos, slightly lower levels of SKN-1 were detected in the two more anterior blastomeres. Similar results were obtained with par-1(b274) embryos (data not shown). (C,H) par-2(lw32) mutant embryo. In 10/13 4-cell-stage par-2 embryos, SKN-1 was detectable in the two most posterior blastomeres and undetectable or barely detectable in the two most anterior blastomeres; in 3/13 slightly lower levels of SKN-1 were detected in the two more anterior blastomeres. Similar results were obtained using par-2(it5ts) embryos (data not shown). (D,I) par-3(it71) mutant embryo. In 8/14 4-cell-stage par-3 mutant embryos, SKN-1 was evenly distributed; in 3/14, one 4-cell-stage blastomere stained less brightly than the other three; in 2/14, the two anterior blastomeres stained less brightly; in 1/14, the two posterior blastomeres stained less brightly. In all cases, SKN-1 was distributed more evenly than in wild-type. Similar results were obtained using par-3(it62) embryos (data not shown). (E,J) par-4(it57) mutant embryo. In 9/11 4-cell-stage par-4 mutant embryos, high levels of SKN-1 were present in the two smaller posterior blastomeres with no or little SKN-1 detectable in the two larger anterior blastomeres; in 2/11 nearly equal levels of SKN-1 were detected in all four blastomeres. As in wild-type embryos, SKN-1 was detectable at low levels in all 2-cell-stage and 4-cell-stage par mutant embryos examined [n=18, 8, 16, and 14 for par-1, par-2, par-3 and par-4 mutant embryos, respectively]. SKN-1 was detectable in some but not all 8-cell-stage mutant embryos: 4/18 par-1, 2/9 par-2, 3/14 par-3 and 2/7 par-4 embryos had detectable staining, frequencies similar to wild-type (Bowerman et al., 1993). In all embryos, SKN-1 was not detectable by the 16-cell stage (data not shown).

Fig. 3.

SKN-1 distribution in par-1, par-2, par-3 and par-4 mutant embryos. Fluorescence micrographs showing fixed 4-cell-stage wild-type and par mutant embryos stained for SKN-1 protein (left column) and with DAPI to visualize DNA in nuclei (right column). Embryos are oriented with anterior to the left as determined by the position of DAPI-stained polar bodies (visible in I and J; out of the focal plane in the other DAPI images). (A,F) Wild-type embryo with high levels of SKN-1 in P2 and EMS. Lower levels are just detectable in ABa and ABp. Note that the chromosomes are more highly condensed in ABa and ABp than in P2 and EMS. In DAPI-stained par mutant embryos (G-I), chromosome condensation appears equivalent as all four cells divide synchronously. (B,G) par-1(2012) mutant embryo. In 22/24 4-cell-stage par-1 embryos, equal levels of staining were detected in all blastomeres; in 2/24 embryos, slightly lower levels of SKN-1 were detected in the two more anterior blastomeres. Similar results were obtained with par-1(b274) embryos (data not shown). (C,H) par-2(lw32) mutant embryo. In 10/13 4-cell-stage par-2 embryos, SKN-1 was detectable in the two most posterior blastomeres and undetectable or barely detectable in the two most anterior blastomeres; in 3/13 slightly lower levels of SKN-1 were detected in the two more anterior blastomeres. Similar results were obtained using par-2(it5ts) embryos (data not shown). (D,I) par-3(it71) mutant embryo. In 8/14 4-cell-stage par-3 mutant embryos, SKN-1 was evenly distributed; in 3/14, one 4-cell-stage blastomere stained less brightly than the other three; in 2/14, the two anterior blastomeres stained less brightly; in 1/14, the two posterior blastomeres stained less brightly. In all cases, SKN-1 was distributed more evenly than in wild-type. Similar results were obtained using par-3(it62) embryos (data not shown). (E,J) par-4(it57) mutant embryo. In 9/11 4-cell-stage par-4 mutant embryos, high levels of SKN-1 were present in the two smaller posterior blastomeres with no or little SKN-1 detectable in the two larger anterior blastomeres; in 2/11 nearly equal levels of SKN-1 were detected in all four blastomeres. As in wild-type embryos, SKN-1 was detectable at low levels in all 2-cell-stage and 4-cell-stage par mutant embryos examined [n=18, 8, 16, and 14 for par-1, par-2, par-3 and par-4 mutant embryos, respectively]. SKN-1 was detectable in some but not all 8-cell-stage mutant embryos: 4/18 par-1, 2/9 par-2, 3/14 par-3 and 2/7 par-4 embryos had detectable staining, frequencies similar to wild-type (Bowerman et al., 1993). In all embryos, SKN-1 was not detectable by the 16-cell stage (data not shown).

Intestinal cell production by P1 and AB in par mutant embryos

We next examined par mutant embryos with respect to another a-p asymmetry, the production of intestinal cells by only P1 and not AB (Sulston et al., 1983). Like the production of pharynx, the production of intestine by P1 requires skn-1 function (Bowerman et al., 1992). Although skn-1 is required in par mutant embryos to specify the excess pharynx that they make, most par mutant embryos fail to produce any intestinal cells (Kemphues et al., 1988). Indeed, par-1 and par-4 mutant embryos almost never make intestine. However, the fraction of par-2 and par-3 mutant embryos that make intestinal cells is sufficiently high to to determine the abilities of P1 and AB to produce intestine (Table 1). We found that, in par-2 mutants, only P1 and not AB is capable of producing intestinal cells. By contrast, in par-3 mutant embryos, P1 and AB both produced intestinal cells at a similar frequency. The production of intestinal cells by only P1 in par-2 mutant embryos, and by both P1 and AB in par-3 mutant embryos is consistent with the normal and abnormal expression pattern of SKN-1 in, respectively, par-2 and par-3 mutant embryos.

Patterning of body wall muscle in par mutant embryos

The final asymmetry that we analyzed is in the different abilities of P1 and AB to produce body wall muscle. In wild-type embryos, P1 produces 80 of the 81 body wall muscle cells made during embryogenesis in C. elegans (Sulston et al., 1983). Although the production of body wall muscle by some P1 descendants is influenced by cell signals (Schnabel, 1994, 1995), P1 nevertheless produces many body wall muscle cells after removing or killing AB, while AB fails to produce body wall muscle after removal of P1 (Priess and Thomson, 1987; Draper et al., 1996). To further define the consequences of mutationally inactivating the par genes, we examined the abilities of P1 and AB to produce body wall muscle in par mutant embryos.

As with pharyngeal muscle production, the different abilities of P1 and AB in par mutant embryos to produce body wall muscle is apparent when differentiated intact mutant embryos are fixed and stained with antibodies specific for a body wall muscle myosin. par-1 (Fig. 4C) and par-3 (Fig. 4G) mutants both produce body wall muscle throughout the embryo, suggesting that both P1 and AB make muscle, an observation confirmed by laser ablation experiments (Table 2). In par-2 mutant embryos, body wall muscle cells are made only in the posterior part of the embryo (Fig. 4E). Laser ablation experiments confirmed the impression from intact embryos that only P1 and not AB produces body wall muscle in par-2 embryos (Table 2). The body wall muscle phenotype of par-4 mutant embryos is, like the pharyngeal phenotype, more variable. par-4 mutant embryos produce posteriorly localized body wall muscle cells in about 50% of the embryos; the remaining embryos do not produce body wall muscle (Table 2, Fig. 4 legend). To determine if the production of body wall muscle and pharynx are coupled in par-4 mutant embryos, we double-stained differentiated par-4 mutants with antibodies both to body wall muscle and to pharyngeal muscle. We found that about 43% of par-4 mutant embryos produce both pharyngeal and body wall muscles posteriorly (48/112). Surprisingly, though, some embryos produce only pharyngeal muscle cells (3/112) and some produce only body wall muscle (17/112). Thus, in some par-4 mutant embryos, the specification of pharyngeal muscle and body wall muscle appear unlinked. Uncoupling of skn-1-dependent specification of pharynx from muscle has been noted previously: Schnabel (1994) showed that MS requires glp-1 function to produce body wall muscle but not to produce pharyngeal muscle.

Table 2.

Production of body wall muscle cells by P1 and AB

Production of body wall muscle cells by P1 and AB
Production of body wall muscle cells by P1 and AB
Fig. 4.

skn-1 function and body wall muscle cell production in par-1, par-2, par-3 and par-4 mutant embryos. Fluorescence micrographs of differentiated wild-type (A) and mutant (B-J) embryos stained with monoclonal antibody 5.6 to detect body wall muscle cells (Miller et al., 1983). (A) Wild-type embryo showing two of four longitudinal quadrants of body wall muscle. (B) skn-1(zu67) embryo with body wall muscle in posterior part of embryo. (C) par-1(e2012) embryo with body wall muscle throughout embryo. (D) par-1(e2012); skn-1(zu67) with no body wall muscle [see Bowerman et al., 1993, for quantitation]. (E) par-2(lw32) embryo with body wall muscle in posterior part of embryo. (F) par-2(lw32); skn-1(zu67) embryo with body wall muscle in posterior part of embryo. (G) par-3(it71) embryo with body wall muscle througout embryo. (H) par-3(it71); skn-1(zu67) embryo with body wall muscle in posterior part of embryo [see Table 2 for quantitation of body wall muscle production by par-2, par-3, par-2; skn-1 and par-3; skn-1 mutant embryos]. (I) par-4(it47ts) embryo with body wall muscle in posterior part of embryo. 144/283 par-4 embryos (51%) made body wall muscle cells (20-50 cells/embryo). 139/283 (49%) made no body wall muscle cells. (J) skn-1(zu67); par-4(it47ts) embryo with body wall muscle in posterior part of embryo. 228/423 skn-1; par-4 embryos (54%) made body wall muscle cells (15-48 cells/embryo); 195/423 made no body wall muscle.

Fig. 4.

skn-1 function and body wall muscle cell production in par-1, par-2, par-3 and par-4 mutant embryos. Fluorescence micrographs of differentiated wild-type (A) and mutant (B-J) embryos stained with monoclonal antibody 5.6 to detect body wall muscle cells (Miller et al., 1983). (A) Wild-type embryo showing two of four longitudinal quadrants of body wall muscle. (B) skn-1(zu67) embryo with body wall muscle in posterior part of embryo. (C) par-1(e2012) embryo with body wall muscle throughout embryo. (D) par-1(e2012); skn-1(zu67) with no body wall muscle [see Bowerman et al., 1993, for quantitation]. (E) par-2(lw32) embryo with body wall muscle in posterior part of embryo. (F) par-2(lw32); skn-1(zu67) embryo with body wall muscle in posterior part of embryo. (G) par-3(it71) embryo with body wall muscle througout embryo. (H) par-3(it71); skn-1(zu67) embryo with body wall muscle in posterior part of embryo [see Table 2 for quantitation of body wall muscle production by par-2, par-3, par-2; skn-1 and par-3; skn-1 mutant embryos]. (I) par-4(it47ts) embryo with body wall muscle in posterior part of embryo. 144/283 par-4 embryos (51%) made body wall muscle cells (20-50 cells/embryo). 139/283 (49%) made no body wall muscle cells. (J) skn-1(zu67); par-4(it47ts) embryo with body wall muscle in posterior part of embryo. 228/423 skn-1; par-4 embryos (54%) made body wall muscle cells (15-48 cells/embryo); 195/423 made no body wall muscle.

PAL-1 distribution in par mutant embryos

To further investigate the body wall muscle phenotype of par mutant embryos, we analyzed the distribution of the homeodomain protein PAL-1. In wild-type embryos, body wall muscle production is specified by both PAL-1 and SKN-1 (Mello et al., 1992; Bowerman et al., 1993; Hunter and Kenyon, 1996), with PAL-1 detectable at the 4-cell stage in only EMS and P2 (Hunter and Kenyon, 1996; see Fig. 5). PAL-1 continues to be expressed in the descendants of both P1 daughters through the 28-cell stage (Table 3), although expression appears stronger in P2 than in EMS descendants (Hunter and Kenyon, 1996). Previous studies have shown that PAL-1 is not expressed in par-1 mutant embryos (Hunter and Kenyon, 1996). To determine if loss of PAL-1 expression is a general property of par mutants, we examined par-2, par-3 and par-4 embryos (Fig. 5; Table 3). In par-2 mutants, PAL-1 is expressed in a wild-type pattern, although the expression level appears to be reduced compared to wild-type (Fig. 5H). The distribution of PAL-1 in par-3 mutant embryos was variable. In about half the 4-cell to 32-cell par-3 mutant embryos, PAL-1 was detected in a normal posterior-localized pattern (Fig. 5I). Embryos that did not express PAL-1 in a wild-type pattern either failed to produce any detectable PAL-1 or expressed PAL-1 in all cells. The levels of PAL-1 observed in anterior blastomeres of par-3 mutant embryos were often lower than observed in posterior blastomeres (Fig. 5J), but in some cases the levels were equal (Fig. 5K). Finally, we found that similar to par-1, early stage par-4 mutant embryos do not make detectable levels of PAL-1 (Fig. 5L).

Table 3.

PAL-1 staining in par mutant embryos

PAL-1 staining in par mutant embryos
PAL-1 staining in par mutant embryos
Fig. 5.

PAL-1 distribution in par-2, par-3 and par-4 mutant embryos. Fluorescence micrographs showing fixed 4-cell-stage wild-type and par mutant embryos stained with antibodies specific for PAL-1 (right column) and with DAPI (left column). See Table 3 for quantitation. (A,G) Wild-type embryo with high levels of PAL-1 in P2 and EMS. (B,H) par-2(lw32) 4-cell-stage embryos usually have PAL-1 present only in two posterior blastomeres (H); one anterior nucleus is out of the focal plane (B). (C-E, I-K) par-3(it71) mutant embryos. PAL-1 is present only in the posterior blastomeres of some 4-cell-stage embryos (I), at higher levels posteriorly than anteriorly (J), and sometimes at equal levels in all blastomeres, as in an 8-cell-stage embryo (K). PAL-1 is not detectable in many par-3 mutant embryos (see Table 3). (F,L) par-4(it57ts) 4-cell-stage mutant embryo with no detectable PAL-1 (L).

Fig. 5.

PAL-1 distribution in par-2, par-3 and par-4 mutant embryos. Fluorescence micrographs showing fixed 4-cell-stage wild-type and par mutant embryos stained with antibodies specific for PAL-1 (right column) and with DAPI (left column). See Table 3 for quantitation. (A,G) Wild-type embryo with high levels of PAL-1 in P2 and EMS. (B,H) par-2(lw32) 4-cell-stage embryos usually have PAL-1 present only in two posterior blastomeres (H); one anterior nucleus is out of the focal plane (B). (C-E, I-K) par-3(it71) mutant embryos. PAL-1 is present only in the posterior blastomeres of some 4-cell-stage embryos (I), at higher levels posteriorly than anteriorly (J), and sometimes at equal levels in all blastomeres, as in an 8-cell-stage embryo (K). PAL-1 is not detectable in many par-3 mutant embryos (see Table 3). (F,L) par-4(it57ts) 4-cell-stage mutant embryo with no detectable PAL-1 (L).

skn-1 and pal-1 activities and the specification of body wall muscle in par mutant embryos

In wild-type embryos, pal-1 is required for somatic P2 descendants to make body wall muscle, and skn-1 is required for the normal specification of body wall muscle in EMS descendants (Mello et al., 1992; Bowerman et al., 1993; Draper et al., 1996; Hunter and Kenyon, 1996). Since null mutations in pal-1 are zygotic lethals (L. Edgar and W. Wood, personal communication), we inhibited maternal pal-1 function by microinjecting pal-1 antisense RNA into the syncitial ovary of pal-1(+) her-maphrodites (Hunter and Kenyon, 1996). Depleting pal-1 function by injection of pal-1 antisense RNA into skn-1 mutant hermaphrodites produces embryos that entirely lack body wall muscle cells, suggesting that skn-1 and pal-1 define a minimum set of body wall muscle specification activities (Hunter and Kenyon, 1996; but see below).

To determine the contributions of skn-1 and pal-1 to the specification of body wall muscle in par mutant embryos, we removed or inhibited their activities in each par mutant. In par-1 embryos both P1 and AB produce body wall muscle, SKN-1 is detected in all early blastomeres, and PAL-1 is not detectable (Bowerman et al., 1993; Hunter and Kenyon, 1996). In par-1; skn-1 double mutant embryos, body wall muscle cells are not made (Bowerman et al., 1993; Draper et al., 1996; see Fig. 4D). Thus skn-1 activity is necessary for all body wall muscle pro-duction by both P1 and AB in par-1 mutant embryos.

In par-2 mutant embryos, only P1 produces body wall muscle (Table 2), and both SKN-1 and PAL-1 are localized normally to posterior blastomeres (Figs 3, 5; Table 3). In par-2; skn-1 double mutants, the posterior half of the embryo still produces body wall muscle (Fig. 4F), and laser ablation exper-iments confirm that P1 makes muscle (Table 3). Similarly, depleting pal-1 function in par-2 mutants by injection of pal-1 antisense RNA results in embryos that still produce posteri-orly localized body wall muscle (data not shown). These results suggest that, in par-2 mutant embryos, either or both SKN-1 and PAL-1 can specify body wall muscle.

In par-3 mutant embryos, both P1 and AB produce body wall muscle cells, and both SKN-1 and PAL-1 can be mis-localized to anterior blastomeres (Figs 3D, 5D, 5E; Table 3). However, in par-3; skn-1 double mutants, body wall muscle is always made by P1 descendants even though PAL-1 is not detectable in 30% of par-3 mutant embryos. These data suggest that either undetectable amounts of PAL-1 can function to specify body wall muscle production, or that additional body wall muscle factors are active in par-3 mutant embryos. Injection of pal-1 antisense RNA into par-3; skn-1 double mutant hermaphro-dites did significantly reduce or eliminate body wall muscle production (18/31 embryo produced four or fewer muscle staining cells), indicating that the posteriorly localized PAL-1 can specify body wall muscle in par-3 mutant embryos.

In par-4 mutants, only P1 produces body wall muscle, and it does so in only 50% of the embryos (Fig. 4; Table 2). PAL-1 is not expressed in par-4 mutants and skn-1 is posteriorly localized (Figs 3, 5). Surprisingly, 50% of par-4; skn-1 double mutants still make posteriorly localized body wall muscle (Table 2). Since PAL-1 is not expressed in par-4 mutant embryos, there may be a SKN-1/PAL-1-independent activity that can specify body wall muscle production in par-4 mutant embryos (see Discussion).

MEX-3 distribution in par mutant embryos

MEX-3 is a putative RNA-binding protein with KH domains that is required to repress pal-1 translation in oocytes and early embryos (Draper et al., 1996). MEX-3 is distributed throughout oocytes but at the 4-cell stage is present cytoplasmically at higher levels in anterior than in posterior blastomeres (Draper et al., 1996; see Fig. 6A). Previous studies have shown that, in mex-3 mutant embryos, PAL-1 is strongly expressed in all blastomeres, and all blastomeres produce body wall muscle (Draper et al., 1996; Hunter and Kenyon, 1996). Furthermore, it has been proposed that the higher levels of MEX-3 in anterior blastomeres at the 4-cell stage might limit pal-1 translation to posterior blastomeres. Consistent with this hypothesis are the observations that MEX-3 is present at high levels in all blastomeres and PAL-1 is not detectable in par-1 mutant embryos. Moreover, removing mex-3 activity from par-1 embryos relieves the translational repression and results in PAL-1 being produced at high levels in all blastomeres (C. P. H., unpublished data). Similarly a mex-3; skn-1; par-1 triple mutant embryo produces large numbers of body wall muscle cells from both P1 and AB (Draper et al., 1996). Thus, low levels of MEX-3 correlate with PAL-1 expression in wild-type embryos and in mex-3 and par-1 mutants.

Fig. 6.

MEX-3 distribution in par-2, par-3 and par-4 mutant embryos. Fluorescence micrographs showing fixed 4-cell-stage wild-type and par mutant embryos stained with antibodies specific for MEX-3 protein (left column) and with DAPI (right column). For quantitation of results, see Table 4. (A,E) Wild-type 4-cell-stage embryo with high levels of MEX-3 in ABa and ABp (A). MEX-3 is also a P-granule component (Draper et al., 1996) as indicated by the punctate staining in P2 (A). (B,F) par-2(lw32) 4-cell-stage embryo with MEX-3 at higher levels in anterior than in posterior blastomeres (B). (C,G) par-3(it71) 4-cell-stage embryo with MEX-3 at high levels in all blastomeres (C). MEX-3 is also detectable in P-granules in P2 and EMS; some of these are visible in the par-2 embryo (B). (D,H) par-4(it57ts) 4-cell-stage embryo with MEX-3 at high levels in all blastomeres (D).

Fig. 6.

MEX-3 distribution in par-2, par-3 and par-4 mutant embryos. Fluorescence micrographs showing fixed 4-cell-stage wild-type and par mutant embryos stained with antibodies specific for MEX-3 protein (left column) and with DAPI (right column). For quantitation of results, see Table 4. (A,E) Wild-type 4-cell-stage embryo with high levels of MEX-3 in ABa and ABp (A). MEX-3 is also a P-granule component (Draper et al., 1996) as indicated by the punctate staining in P2 (A). (B,F) par-2(lw32) 4-cell-stage embryo with MEX-3 at higher levels in anterior than in posterior blastomeres (B). (C,G) par-3(it71) 4-cell-stage embryo with MEX-3 at high levels in all blastomeres (C). MEX-3 is also detectable in P-granules in P2 and EMS; some of these are visible in the par-2 embryo (B). (D,H) par-4(it57ts) 4-cell-stage embryo with MEX-3 at high levels in all blastomeres (D).

To further test the relationship between MEX-3 and PAL-1, we stained par-2, par-3 and par-4 mutant embryos with a monoclonal antibody specific for MEX-3 (Fig. 6; Table 4). Similar to wild-type embryos, in par-2 mutants PAL-1 is detected in both P1 daughters. Consistent with the translational repression model, we found that the MEX-3 distribution also appeared wild-type, with higher levels at the 4-cell stage in AB daughters than in P1 daughters. In par-3 mutants, PAL-1 is present at detectable levels in all blastomeres in some 4-cell and 8-cell-stage embryos, and only posteriorly in others (Fig. 5; Table 3). Surprisingly, we found that MEX-3 was present at high levels in the cytoplasm of all 4-cell-stage blastomeres in all par-3 mutant embryos examined (Fig. 6C). Thus, in par-3 mutant embryos, MEX-3 levels do not correlate with the pattern of PAL-1 expression. Finally, par-4 mutant embryos resemble par-1 mutants: no PAL-1 expression was detected in par-4 mutant embryos (Fig. 5L) and MEX-3 was present cytoplasmically at high levels in all 4-cell-stage par-4 mutant embryos (Fig. 6D). Like mex-3; par-1 double mutant embryos, mex-3; par-4 double mutant embryos express PAL-1 in all blastomeres (C. P. H., unpublished data). Thus the ubiquitously distributed MEX-3 appears to repress all PAL-1 expression in both par-1 and par-4 mutants. In summary, although the MEX-3 and PAL-1 expression patterns are complementary in wild-type, par-1, par-2 and par-4 mutant embryos, they are not so correlated in par-3 mutant embryos.

Table 4.

MEX-3 staining in par mutant embryos

MEX-3 staining in par mutant embryos
MEX-3 staining in par mutant embryos

The par genes sometimes act independently of each other to polarize and pattern early C. elegans embryos

We have shown that the phenotypes of par-1, par-2, par-3 and par-4 mutant embryos, with respect to the function of skn-1, glp-1, pal-1 and mex-3, each are unique. We briefly summarize each mutant and then discuss the implications of this diversity in phenotype. (i) par-1 mutations, as shown previously, result in ubiquitous expression of GLP-1, MEX-3 and SKN-1, and cause a loss of PAL-1 expression. par-1 mutant embryos produce skn-1-dependent, glp-1-independent pharynx and body wall muscle from all 4-cell-stage blastomeres. (ii) par-2 mutant embryos show normal localization of SKN-1, PAL-1 and MEX-3 at the 4-cell stage. Both posterior blastomeres in 4-cell-stage par-2 mutant embryos produce skn-1- and pal-1-dependent body wall muscle and skn-1-dependent, glp-1-independent pharynx. Both anterior blastomeres produce glp-1-dependent pharynx. (iii) par-3: As in par-1 mutants, SKN-1 and MEX-3 are expressed ubiquitously in par-3 mutant embryos. But PAL-1, instead of being absent, is expressed and sometimes mis-localized. All 4-cell-stage blastomeres produce skn-1-dependent, glp-1-independent pharynx and body wall muscle. (iv) par-4 mutant embryos are unusual in that about 50% of them fail to produce any pharynx or body wall muscle, even though SKN-1 is expressed normally. As in par-1 mutants, MEX-3 is expressed ubiquitously and PAL-1 is absent.

Based on their lack of phenotypic similarity, we suggest that the par genes do not regulate pattern formation by operating in a single genetic pathway to control the polarity of early embryonic cells. At the same time, some par phenotypes show simple epistatic relationships: most notably, par-3 is epistatic to par-2 with respect to mitotic spindle axis orientation in P1 and AB, with respect to SKN-1 distribution in P1 and AB descendants, and with respect to the abilities of P1 and AB to produce pharynx and body wall muscle (Cheng et al., 1995; Tables 1 and 2). Thus some par gene functions probably involve steps arranged in a linear pathway. However, for the par genes as a group, we think it most likely that they function as part of a complex network of interconnected pathways in which some par gene products regulate aspects of blastomere polarity and cell fate patterning independently of others. Epistasis analysis of cell fate specification pathways in other par double mutants may shed more light on this issue.

In addition to each par mutant exhibiting a unique phentoype, we note that none of the four par genes that we analyzed are required for all a-p asymmetries. par-1 mutants show the most extensive loss of a-p asymmetry: (1) the first cleavage is equal; (2) P-granules, SKN-1, GLP-1 and MEX-3 are present at equal levels in anterior and posterior blastomeres, and (3) PAL-1 is absent from all blastomeres. However, as in wild-type, the mitotic spindle is oriented longitudinally in the posterior 2-cell-stage blastomere and transversely in the anterior blastomere, indicating that some a-p asymmetry remains in par-1 mutants. par-3 mutant embryos, like par-1 mutants, show extensive losses in a-p asymmetry: P1 and AB both divide longitudinally, and SKN-1, MEX-3 and, to a lesser extent, GLP-1 are all mis-localized. However, PAL-1 protein and function appear to be localized to posterior blastomeres in many par-3 mutant embryos. par-4 mutants show a loss of asymmetry in MEX-3 and GLP-1 distribution and fail to express PAL-1 but, as in wild type, the first cleavage is unequal and SKN-1 is present only in posterior blastomeres. Surprisingly, par-2 mutant embryos appear to show substantially less loss of a-p asymmetry. Although the first cleavage in par-2 mutant embryos always produces two equal-sized daughters, SKN-1, GLP-1, PAL-1 and MEX-3 are localized normally in most par-2 mutant embryos (our data; Crittenden et al., 1996), and P-granules are present only in the posterior blastomere at the 2-cell stage (Kemphues and Strome, 1997). The observation that no single par gene is required for all a-p asymmetries is consistent with our conclusion that the par genes in some cases act independently of each other to generate polarity in the early embryo.

Mislocalization of regulatory proteins and their sufficiency to ectopically specify developmental programs in par mutant embryos

The status of different cell fate determination pathways in par mutant embryos also yields insight into the sufficiencies of different regulators to promote or repress specific developmental programs. We emphasize four such findings. (i) In par-3 mutant embryos, MEX-3 levels do not correlate with PAL-1 expression patterns (see below for discussion). (ii) In par-4 mutant embryos, there may be a body wall muscle specification pathway inde-pendent of either skn-1 or pal-1 function: even though no PAL-1 is detectable in par-4 mutant embryos, 50% of skn-1; par-4 double mutant embryos produce large numbers of body wall muscle cells. Thus it may be significant that 50% of skn-1 mutant embryos in which pal-1 function has been depleted by pal-1 RNA injection produce body wall muscle (Hunter and Kenyon, 1996). This latter result could be explained by incomplete depletion of pal-1 activity, or it might also indicate that a SKN-1/PAL-1-inde-pendent activity can specify body wall muscle. (iii) Also in par-4 mutants, we find that high levels of SKN-1 are not always sufficient to specify the production of pharyngeal cells and body wall muscle cells, and we note that the induction of anterior pharyngeal cells never occurs. (iv) Because PAL-1 is expressed normally in par-2 mutants, and PAR-1 is not cortically enriched in par-2 mutants (Guo and Kemphues, 1995), cortically localized PAR-1 is not required for PAL-1 expression in P1 descendants. Similarly, P-granules are localized to the posterior blastomere in 2-cell-stage par-2 mutant embryos, indicating the cortical PAR-1 is not required for P-granule segregation during the first embryonic cleavage (Kemphues and Strome, 1997).

PAL-1 translational control: repression, derepression and localization

Previous studies have shown that spatial regulation of PAL-1 expression occurs at least in part translationally (Hunter and Kenyon, 1996). One key regulator is the putative RNA-binding protein MEX-3, which is expressed at high levels in oocytes but becomes partially localized to anterior blastomeres in 4-cell-stage embryos (Draper et al., 1996). The complementarity of the MEX-3 and PAL-1 expression patterns suggest that the levels of MEX-3 might regulate the pattern of pal-1 translation: higher anterior MEX-3 levels may limit PAL-1 production to posterior blastomeres. This hypothesis is supported by the observations in par-1 mutant embryos that MEX-3 is present at high levels in all 4-cell-stage blastomeres and that PAL-1 is not detectable (Draper et al., 1996; Hunter and Kenyon, 1996). Thus it is conceivable that, in wild-type embryos, par-1 negatively regulates mex-3 in posterior blastomeres, resulting in lower levels of MEX-3 which permit translation of pal-1 mRNA posteriorly. Our analysis of body wall muscle specification in par-2 and par-4 mutant embryos support the original correlation between high levels of MEX-3 and translational repression of pal-1 mRNA. In 4-cell-stage par-2 mutant embryos, as in wild-type embryos, MEX-3 is present at higher levels in anterior blastomeres and PAL-1 is detectable only in posterior blastomeres. In par-4 mutant embryos, as in par-1 mutant embryos, MEX-3 is present at high levels both anteriorly and posteriorly and PAL-1 is not detectable. Alternatively, though, par-1 could act independently of mex-3 to relieve repression of pal-1 translation in posterior blastomeres, and the levels of MEX-3 might be an indirect result of such de-repression.

The correlation between MEX-3 levels and PAL-1 expression does not hold true in par-3 mutant embryos: MEX-3 is present at high levels both anteriorly and posteriorly, and yet PAL-1 is detectable often in posterior and occasionally in anterior blastomeres. Thus the levels of MEX-3 cannot account for the pattern of pal-1 translation in par-3 mutant embryos. One interpretation of these results is that par-1 and par-4 are required for an activity that relieves MEX-3-mediated repression of pal-1 translation without directly affecting MEX-3. This hypothesis is supported by the observations that PAL-1 is not detectable in par-1 or par-4 mutant embryos but is expressed in all blastomeres in mex-3; par-1 and mex-3; par-4 double mutant embryos. Furthermore, although PAL-1 can be mislocalized in par-3 mutant embryos, it is not detected at significant levels until the 4-cell stage. This separation of temporal and spatial regulation indicates that par-3 function may be required to localize a par-1- and par-4-dependent pal-1 derepressing activity to posterior blastomeres. Finally, it is important to bear in mind that abnormal interactions might occur in par mutant embryos due to the mislocalization of multiple regulatory factors. If so, the loss of correlation between MEX-3 and PAL-1 expression in par-3 mutant embryos could be misleading.

The mechanisms that localize different developmental regulators are not tightly coupled in early C. elegans embryos

Our results, summarized in Fig. 7, indicate that largely independent mechanisms control the localization of different regulatory proteins in the early C. elegans embryo. Even proteins with similar expression patterns respond differently to loss of par functions. For example, par-2 mutations uncouple the anterior localization of GLP-1 and MEX-3, while par-1 and par-4 mutations differently uncouple the posterior localization of SKN-1 and PAL-1. The apparent lack of coupling of the mechanisms that generate different a-p asymmetries in the early C. elegans embryo is in substantial contrast to the largely heirarchical localization of maternal gene products during the generation of a-p asymmetry in the Drosophila melanogaster oocyte (St. Johnston, 1993).

Fig. 7.

Summary of MEX-3, GLP-1, SKN-1 and PAL-1 expression patterns in par-1, par-2, par-3 and par-4 mutant embryos. The data for MEX-3, SKN-1 and PAL-1 are from data presented here; the data for GLP-1 are from Crittenden et al., 1996. GLP-1 is mislocalized in 37% of par-2 embryos and in 77% of par-3 embros; PAL-1 is mislocalized in 24% of par-3 mutant embryos. In all other cases, nearly all embryos show the pattern illustrated.

Fig. 7.

Summary of MEX-3, GLP-1, SKN-1 and PAL-1 expression patterns in par-1, par-2, par-3 and par-4 mutant embryos. The data for MEX-3, SKN-1 and PAL-1 are from data presented here; the data for GLP-1 are from Crittenden et al., 1996. GLP-1 is mislocalized in 37% of par-2 embryos and in 77% of par-3 embros; PAL-1 is mislocalized in 24% of par-3 mutant embryos. In all other cases, nearly all embryos show the pattern illustrated.

How directly or indirectly the par genes affect localization of SKN-1, GLP-1, PAL-1 and MEX-3 expression remains to be determined. The maternal RNAs for skn-1, glp-1 and pal-1 are distributed evenly throughout early embryos (Evans et al., 1994; Seydoux and Fire, 1994; Hunter and Kenyon, 1996). Presumably, polarization of the early embryo by PAR proteins in response to sperm entry localizes translational regulators, and perhaps regulators of protein or RNA stability, that in turn control the expression of the regulatory proteins analyzed here. Indeed, recent studies have shown that translational regulation requiring 3′UTR sequences from the corresponding mRNA appear sufficient to account for the localized expression of GLP-1 and PAL-1 (Evans et al., 1994; Hunter and Kenyon, 1996). However, the apparent uncoupling of the spatial regulation of different regulatory proteins in par mutant embryos suggest that the events linking par gene functions to translational regulation may be complex and indirect. While several genes that control specific blastomere identities, and six par genes, have been identified inC. elegans (Kemphues and Strome, 1997; Schnabel and Priess, 1997), the functional links between the events controlled by these two different groups of maternal genes remain to be elucidated.

We are grateful to Bruce Draper and Jim Priess for providing MEX-3 antibodies prior to publication. We thank the C. elegans stock center, funded by the NIH, and Jim Priess and Ken Kemphues for some of the strains used in these experiments; Sarah Crittenden and Judith Kimble, Bruce Draper and Jim Priess, and Jennifer Watts, Diane Morton and Ken Kemphues for sharing unpublished information; Michele Champagne, Danielle Hamill, Bruce Howard and Elena Kouzminova for construct-ing double mutants; David Miller and Susan Strome for antibodies; and Bruce Draper, Pierre Gönczy and Chris Shelton for helpful comments and discussions. C. P. H. would like to thank and acknowledge Cynthia Kenyon for her generosity in providing space and support (NIH GM37053) during some of the initial experiments of this investigation. C. P. H. was supported by a grant from the American Cancer Society – CA division. This worked was also supported by grants from the American Cancer Society (DB-61) and the NIH (GM49869) to B. B.

Austin
,
J.
and
Kimble
,
J.
(
1989
).
Transcript analysis of glp-1 and lin-12, homologous genes required for cell interactions during development of C. elegans
.
Cell
58
,
565
71
.
Avery
,
L.
and
Horvitz
,
H. R.
(
1989
).
Pharyngeal pumping continues after laser killing of the pharyngeal nervous system of C. elegans
.
Neuron
3
,
473
85
.
Blackwell
,
T. K.
,
Bowerman
,
B.
,
Priess
,
J. R.
and
Weintraub
,
H.
(
1994
).
Incorporation of homeodomain and bZIP elements into a DNA binding domain by C. elegans SKN-1 protein
.
Science
266
,
621
628
.
Bowerman
,
B.
,
Draper
,
B. W.
Mello
,
C. C.
and
Priess
,
J. R.
(
1993
).
The maternal gene skn-1 encodes a protein that is distributed unequally in early C. elegans embryos
.
Cell
74
,
443
452
.
Bowerman
,
B.
,
Eaton
,
B. A.
and
Priess
,
J. R.
(
1992
).
skn-1, a maternally expressed gene required to specify the fate of ventral blastomeres in the early C. elegans embryo
.
Cell
68
,
1061
75
.
Boyd
,
L.
,
Guo
,
S.
,
Levitan
,
D.
,
Stinchcomb
,
D. T.
and
Kemphues
,
K. J.
(
1996
).
PAR-2 is asymmetrically distributed and promotes association of P granules and PAR-1 with the cortex in C. elegans embryos
.
Development
122
,
3075
3084
.
Brenner
,
S.
(
1974
).
The genetics of Caenorhabditis elegans
.
Genetics
77
,
71
94
.
Cheng
,
N. N.
,
Kirby
,
C. M.
and
Kemphues
,
K. J.
(
1995
).
Control of cleavage spindle orientation in Caenorhabditis elegans: the role of the genes par-2 and par-3
.
Genetics
139
,
549
559
.
Crittenden
,
S. L.
,
Rudel
,
D.
,
Binder
,
J.
,
Evans
,
T. C.
and
Kimble
,
J.
(
1996
).
Genes required for GLP-1 asymmetry in the early C. elegans embryo
.
Dev. Biol
.
181
,
36
46
.
Draper
,
B. W.
,
Mello
,
C. C.
,
Bowerman
,
B.
,
Hardin
,
J.
and
Priess
,
J. R.
(
1996
).
The maternal gene mex-3 encodes a KH domain protein and regulates blastomere identity in early C. elegans embryos
.
Cell
87
,
205
216
.
Drewes
,
G.
,
Ebneth
,
A.
,
Preuss
,
E.
,
Mandelkow
,
E.-M.
and
Mandelkow
,
E.
(
1997
).
tMARK, a novel family of protein kinases that phosphorylate microtubule-associated proteins and trigger microtubule disruption
.
Cell
89
,
297
308
.
Etemad-Moghadam
,
S. G.
and
Kemphues
,
K. J.
(
1995
).
Asymmetrically distributed PAR-3 protein contributes to cell polarity and spindle alignment in early C. elegans embryos
.
Cell
83
,
743
752
.
Evans
,
T. C.
,
Crittenden
,
S. L.
,
Kodoyianni
,
V.
and,
Kimble
,
J.
(
1994
).
Translational control of maternal glp-1 mRNA establishes an asymmetry in the C. elegans embryo
.
Cell
77
,
183
194
.
Goldstein
,
B.
and
Hird
,
S. N.
(
1996
).
Specification of the anteroposterior axis in Caenorhabditis elegans
.
Development
122
,
1467
1474
.
Gönczy
,
P.
and
Hyman
,
A. A.
(
1996
).
Cortical domains and the mechanisms of asymmetric cell division
.
Trends Cell Biol
.
6
,
382
387
.
Guo
,
S.
and
Kemphues
,
K. J.
(
1995
).
par-1, a gene required for establishing polarity in C. elegans embryos, encodes a putative ser/thr kinase that is asymmetrically distributed
.
Cell
81
,
611
620
.
Guo
,
S.
and
Kemphues
,
K. J.
(
1996a
).
A non-muscle myosin required for embryonic polarity in Caenorhabidits elegans
.
Nature
382
,
455
458
.
Guo
,
S.
and
Kemphues
,
K. J.
(
1996b
).
Molecular genetics of asymmetric cleavage in the early C. elegans embryo
.
Curr. Op. in Genetics Dev
.
6
,
408
415
.
Hird
,
S. N.
and
White
,
J. G.
(
1993
).
Cortical and cytoplasmic flow polarity in early embryonic cells of Caenorhabditis elegans
.
J. Cell Biol
.
121
,
1343
1355
.
Hunter
,
C. P.
and
Kenyon
,
C.
(
1996
).
Spatial and temporal controls target pal-1 blastomere-specification activity to a single blastomere lineage in C. elegans embryos
.
Cell
87
,
217
226
.
Hutter
,
H.
and
Schnabel
,
R.
(
1994
).
glp-1 and inductions establishing embryonic axes in C. elegans
.
Development
120
,
2051
2064
.
Hutter
,
H.
and
Schnabel
,
R.
(
1995
).
Establishment of left-right asymmetry in the Caenorhabditis elegans embryo: a multistep process involving a series of inductive events
.
Development
121
,
3417
3424
.
Hyman
,
A. A.
(
1989
).
Centrosome movement in the early divisions of Caenorhabditis elegans: a cortical site determining centrosome position
.
J. Cell Biol
.
109
,
1185
1194
.
Hyman
,
A. A.
and
White
,
J. G.
(
1987
).
Determination of cell division axes in the early embryogenesis of Caenorhabiditis elegans
.
J. Cell Biol
.
105
,
2123
2135
.
Kemphues
,
K. J.
,
Priess
,
J. R.
,
Morton
,
D. G.
and
Cheng
,
N. S.
(
1988
).
Identification of genes required for cytoplasmic localization in early C. elegans embryos
.
Cell
52
,
311
20
.
Kemphues
,
K. J.
and
Strome
,
S.
(
1997
).
Fertilization and establishment of polarity in the embryo
.
C. elegans II
.
Plainview, NY
,
Cold Spring Harbor Laboratory Press
.
335
360
.
Kirby
,
C.
,
Kusch
,
M.
and
Kemphues
,
K. J.
(
1990
).
Mutations in the par genes of C. elegans affect cytoplasmic reorganization during the first cell cycle
.
Dev. Biol
.
142
,
203
215
.
Laufer
,
J. S.
,
Bazzicalupo
,
P.
and
Wood
,
W. B.
(
1980
).
Segregation of developmental potential in early embryos of Caenorhabditis elegans
.
Cell
19
,
569
577
.
Levin
,
D. E.
,
Hammond
,
C. I.
,
Ralston
,
R. O.
and
Bishop
,
M. J.
(
1987
).
Two yeast genes that encode unusual protein kinases
.
Proc. Natl. Acad. Sci. USA
84
,
6035
6039
.
Levin
,
D. E.
and
Bishop
,
M. J.
(
1990
).
A putative protein kinase gene (kin1+) is important for growth polarity in Schizosaccharomyces pombe
.
Proc. Natl. Acad. Sci. USA
87
,
8272
8276
.
Levitan
,
D. J.
,
Boyd
,
L.
,
Mello
,
C. C.
,
Kemphues
,
K. J.
and
Stinchcomb
,
D. T.
(
1994
).
par-2, a gene required for blastomere asymmetry in Caenorhabditis elegans, encodes zinc-finger and ATP-binding motifs
.
Proc. Natl. Acad. Sci. USA
91
,
6108
6112
.
Drewes
,
G.
,
Ebneth
,
A.
,
Preuss
,
U.
,
Mandelkow
,
E.-M.
and
Mandelkow
,
E.
(
1997
).
tMARK, a novel family of protein kinases that phosphorylate microtubule-associated proteins and trigger microtubule disruption
.
Cell
89
,
297
308
.
Mango
,
S. E.
,
Lambie
,
E. J.
and
Kimble
,
J.
(
1994
).
The pha-4 gene is required to generate the pharyngeal primordium of Caenorhabditis elegans
.
Development
120
,
3019
3031
.
Mello
,
C. C.
,
Draper
,
B. W.
,
Krause
,
M.
,
Weintraub
,
H.
and
Priess
,
J. R.
(
1992
).
The pie-1 and mex-1 genes and maternal control of blastomere identity in early C. elegans embryos
.
Cell
70
,
163
76
.
Miller
,
D. M.
Ortiz
,
I. I.
,
Berliner
,
G. C.
and
Epstein
,
H. F.
(
1983
).
Differential localization of two myosins within nematode thick filaments
.
Cell
34
,
477
790
.
Morton
,
D. G.
,
Roos
,
J. M.
and
Kemphues
,
K. J.
(
1992
).
par-4, a gene required for cytoplasmic localization and determination of specific cell types in Caenorhabditis elegans embryogenesis
.
Genetics
130
,
771
790
.
Priess
,
J. R.
,
Schnabel
,
H.
and
Schnabel
,
R.
(
1987
).
The glp-1 locus and cellular interactions in early C. elegans embryos
.
Cell
51
,
601
611
.
Priess
,
J. R.
and
Thomson
,
J. N.
(
1987
).
Cellular interactions in early C. elegans embryos
.
Cell
34
,
85
100
.
Schnabel
,
R.
(
1994
).
Autonomoy and nonautonomy in cell fate specification of muscle in the C. elegans embryo: a reciprocal induction
.
Science
263
,
1449
1452
.
Schnabel
,
R.
(
1995
).
Duels with obvious sense: counteracting inductions involved in body wall muscle development in the Caenorhabditis elegans embryo
.
Development
121
,
2219
2232
.
Schnabel
,
R.
and
Priess
,
J. R.
(
1997
).
Specification of cell fates in early embryos
.
C. elegans II
. pp.
361
382
.
Plainview, NY
:
Cold Spring Harbor Laboratory Press
.
Seydoux
,
G.
and
Fire
,
A.
(
1994
).
Soma-germline asymmetry in the distribution of embryonic RNAs in C. elegans
.
Development
120
,
2823
2834
.
Shelton
,
C. A.
and
Bowerman
,
B.
(
1996
).
Time-dependent responses to glp-1-mediated inductions in early C. elegans embryos
.
Development
122
,
2043
2050
.
St.
Johnston
, D
. (
1993
).
Pole plasm and the posterior group genes
.
In The Development of Drosophila melanogaster
. pp.
325
364
.
Plainview, NY
:
Cold Spring Harbor Laboratory Press
.
Sulston
,
J. E.
,
Schierenberg
,
E.
,
White
,
J. G.
and
Thomson
,
J. N.
(
1983
).
The embryonic cell lineage of the nematode Caenorhabditis elegans
.
Dev. Biol
.
100
,
64
119
.
Yochem
,
J.
and
Greenwald
,
I.
(
1989
).
glp-1 and lin-12, genes implicated in distinct cell-cell interactions in C. elegans, encode similar transmembrane proteins
.
Cell
58
,
553
63
.