During the first four cleavage rounds of the Caenorhabdi- tis elegans embryo, five somatic founder cells AB, MS, E, C and D are born, which later form the tissues of the embryo. The classical criterion for a cell-autonomous specification of a tissue is the capability of primordial cells to produce this tissue in isolation from the remainder of the embryo. By this criterion, the somatic founder cells MS, C and D develop cell-autonomously. Laser ablation experiments, however, reveal that within the embryonic context these blastomeres form a network of duelling cellular interactions.

During normal development, the blastomere D inhibits muscle specification in the MS and the C lineage inhibits muscle specification in the D lineage. These inhibitory interactions are counteracted by two activating inductions. As described before the inhibition of body wall muscle in MS is counteracted by an activating signal from the ABa lineage. Body wall muscle in the D lineage is induced by MS descendants, which suppress an inhibitory activity of the C lineage. The interaction between the D and the MS lineage occurs through the C lineage. An interesting feature of these cell-cell interactions is that they do not serve to dis- criminate between equivalent cells but are permissive or nonpermissive inductions.

No evidence was found that the C-derived body wall muscle also depends on an induction, which suggests that possibly three different pathways coexist in the early embryo to specify body wall muscle, two of which are, in different ways, influenced by cell-cell interactions and a third that is autonomous.

This work supplies evidence that cells may acquire transient states during embryogenesis that influence the specification of other cells in the embryo. These states, however, may not be reflected in the developmental potentials of the cells themselves. They can only be scored indi- rectly by their action on the specification of other cells in the embryo.

Blastomeres that behave cell-autonomously in isolation are nevertheless subjected to cell-cell interactions in the embryonic context. Why this should be is an intriguing question. The classical notion has been that blastomeres are specified autonomously in nematodes. In recent years, it was established that at least five inductions are required to determine the AB descendants of C. elegans, whereas the P1 descendants have been typically viewed to develop more autonomously. It appears now that inductions also play a major role during the determination of P1-derived blastomeres.

Two basic modes of cell determination are proposed for the specification of cell fates during embryogenesis. In cell- autonomous development, cells are instructed by intrinsic cues, whereas in nonautonomous development cells depend on signals from other cells to choose a certain pathway of differ- entiation. The classical experiment to determine the mode by which a specific structure is specified is to isolate the pri- mordium of the structure in question from the embryo to prevent cell-cell interactions from occurring. If cells in isolation do not acquire the fate expected from the fate map, it is assumed that the normal fate of these cells is specified nonautonomously by means of cell-cell interactions with other cells. However, if the cells of the primordium execute the fate expected from the fate map in isolation from the embryo, cells are considered to be specified autonomously. Classical devel- opmental biology introduced another criterion to distinguish the state of specification of a primordium. When grafted to ectopic positions in the embryo, only primordia that are not reprogrammed to follow a new fate reflecting their new position in the embryo are considered to be terminally deter- mined. The biological meaning of this test, however, remains obscure when the fate of cells can be changed even though they are already restricted to their normal fate according to the fate map (for review see Slack, 1991). Recently I showed, using the early embryogenesis of C. elegans, that the experimentally observed reprogramming of a primordium could be taken as an indication that it is subjected to competing cell-cell interactions in the embryo (Schnabel, 1994).

In vertebrates, the early specification pathway of most fates requires interactions (for review see Slack, 1991). In invertebrates, especially in nematodes, some cell fates appear to be specified cell-autonomously (Laufer et al., 1980; Cowan and McIntosh, 1985; Edgar and McGhee, 1986; Schierenberg, 1988; Bowerman et al., 1992a; Mello et al., 1992). The body of the nematode C. elegans is assembled by a stereotypic lineage. During the first four cleavages of the fertilised zygote, the stem cell like P cells (P0-P3) undergo asymmetric divisions to produce the somatic founder cells AB, MS, E, C and D. The blastomere AB produces hypodermis, neurons and pharynx. MS produces pharynx and body wall muscle. E is the only precursor of intestine. C gives rise to body wall muscle and hypodermis. The blastomere D produces only muscle. The last P cell, P4, is the precursor of the germ line (Sulston et al., 1983) (Fig. 1).

Fig. 1.

The early embryonic lineage. During the first four cleavage divisions of the zygote P0 the stem-cell like P cells produce, in unequal divisions, five somatic founder cells and the germ line precursor P4. The major tissues produced by the somatic lineages are indicated below the lineages. The numbers of body wall muscle cells contributed by the different founder cells are indicated. The different lineages have different, characteristic cleavage rates. The time scale indicated on the left corresponds to development in minutes at 25°C.

Fig. 1.

The early embryonic lineage. During the first four cleavage divisions of the zygote P0 the stem-cell like P cells produce, in unequal divisions, five somatic founder cells and the germ line precursor P4. The major tissues produced by the somatic lineages are indicated below the lineages. The numbers of body wall muscle cells contributed by the different founder cells are indicated. The different lineages have different, characteristic cleavage rates. The time scale indicated on the left corresponds to development in minutes at 25°C.

The original notion was that nematodes develop strictly cell- autonomously using determinants for the specification of cells and tissues (zur Strassen, 1959). Recent evidence has shown that the specification of the AB lineage depends on at least five inductions (Priess and Thomson, 1987; Schnabel, 1991; Wood, 1991; Bowerman et al., 1992b; Hutter and Schnabel, 1994; Mango et al., 1994; Mello et al., 1994; Moskowitz et al., 1994; Hutter and Schnabel, 1995; Hutter and Schnabel, unpublished data). There is experimental evidence that is consistent with the notion that the other founder cells are specified cell autonomously (Laufer et al., 1980; Cowan and McIntosh, 1985; Edgar and McGhee, 1986; Schierenberg, 1988; Bowerman et al., 1992a; Mello et al., 1992). The potential to form hypodermis, muscle and intestine is segregated from the zygote P0 into the founder cells through the corresponding lineages. The specification of intestine in E nevertheless depends also on a cell-cell interaction required to polarise the EMS blastomere (Schierenberg, 1987; Goldstein, 1992, 1993). The MS, C and D blastomeres produce their normal tissues when all other blastomeres present in the embryo are ablated with a laser microbeam, which suggests, in accordance with the classical criterion for cell autonomous specification, that these tissues can be formed cell-autonomously (Laufer et al., 1980; Cowan and McIntosh, 1985; Edgar and McGhee, 1986; Bowerman et al., 1992a; Mello et al., 1992). However, it was shown recently that one of the blastomeres, MS, despite its cell autonomous potential to form muscle, is still subjected to coun- teracting interactions that are inhibiting and activating the pro- duction of body wall muscle from MS within the embryonic context (Schnabel, 1994). It thus appears that blastomeres execute their cell autonomous potentials when isolated but nev- ertheless are exposed, during normal development, to inhibitory cell-cell interactions which in turn are overcome by activating interactions. Here I describe a systematic search for other competing, or duelling, cell-cell interactions in the early embryo. By laser ablating different blastomeres or combina- tions of blastomeres, it is shown that the formation of body wall muscle from D is inhibited by an interaction from the C lineage which in turn is overcome by an activating signal from the MS lineage. I further show that the inhibitory signal sup- pressing muscle specification described earlier for MS (Schnabel, 1994) is derived from the D blastomere. This inhibitory signal is transferred by the two descendants of C.

Methods for culturing and handling of C. elegans have been described by Brenner (1974) and Wood (1988). The standard wild-type strain N2 is that of Brenner (1974). Laser ablation of blastomeres, immuno- staining of irradiated embryos and the lineage analyses were carried out as described previously (Schnabel, 1991, 1994; Hutter and Schnabel, 1994). Laser operations of embryos were carried out at 25°C. The operated embryos were incubated at either 15°C for approximately 9 hours, or at 20°C for approximately 6 hours or at 25°C for approximately 5 hours. The progress of embryogenesis was monitored by observing the untreated embryos on the slides. All treated embryos were scored for proper laser ablations just before fixation. Properly irradiated blastomeres either did not divide at all or divided only a few times. Embryos were processed for antibody stainings when they reached the developmental stage shown in Fig. 2A corresponding to about 430 minutes of development at 20°C (Sulston et al., 1983). Untreated embryos developed normally under these conditions. All embryos were stained with mAb NE8 4C6.3 (antibody collection of the LMB MRC Cambridge England; Goh and Bogaert, 1991; see also Schnabel, 1994).

Fig. 2.

Immunofluorescence micrographs showing body wall muscle cell differentiation in normal and laser-ablated embryos. All embryos are stained with mAb NE8 4C6.3. The two panels in each row except for A, D and E show two different focal planes of each embryo. (A1) Wild-type embryo at approximately 400 minutes of development (20°C; Sulston et al., 1983) shortly after elongation has started. Two of the four body wall muscle rows formed in the embryo are seen. At this stage, the cytoplasm of the body wall muscle cells is staining. The nuclei can be seen and counted. In normal embryos, 81 body wall muscle cells stain. (A2) Wild-type embryo at approximately 460 minutes of development. The muscle cells start to produce filaments; however the nuclei can still be identified. Soon after this stage, the general cytoplasmic staining disappears as the sarcomere develops and it is impossible to count the number of body wall muscle cells. Therefore manipulated embryos were evaluated only when the cytoplasmic staining was still present. Shortly before hatching the antibody stains faintly at least four more muscles which are not body wall muscles. These muscles are therefore not considered further in this work. (B-I) Examples of laser ablated embryos from the experimental series shown in Fig. 3. (B) Embryo in which all cells except for P3, the mother of D, were ablated. A total of 20 muscle cells were observed in this embryo. This is the equivalent normally produced by D. (C) Embryo in which all cells except for C were ablated. I counted 31 muscle cells in this embryo. C normally produces 32 muscle cells. (D) An embryo with an ablated EMS blastomere. I counted 31 muscle cells in this embryo. If C and D would produce their full complements of muscle this embryo should contain 53 (AB 1, C 32, D 20) muscle cells. (E) An embryo where EMS and C were ablated. The number of muscle cells observed in this embryo was 20, the number normally produced by D. (F) Embryo with ablated ABa and C blastomeres. A total of 34 muscle cells were observed in this embryo. Therefore MS should have produced 14 muscle cells in the embryo. (G, H) Embryos where ABa and D were ablated at the cleavage from two to four C descendants (G) or about 10 minutes later when the two E descendants cleave (H). The earlier D ablation still partially relieves the inhibition of MS-derived muscles. I counted 48 muscle cells in this embryo (G). When D is ablated later (H) the MS muscle is already fully suppressed. I counted 33 muscle cells in this embryo. (I) An embryo where first EMS was ablated and then later the two C descendants Ca and Cp were ablated just before the onset of their mitoses. Most of the inhibition of D already occurred in this embryo, only 6 muscle cells could be detected. Bar 10 μm.

Fig. 2.

Immunofluorescence micrographs showing body wall muscle cell differentiation in normal and laser-ablated embryos. All embryos are stained with mAb NE8 4C6.3. The two panels in each row except for A, D and E show two different focal planes of each embryo. (A1) Wild-type embryo at approximately 400 minutes of development (20°C; Sulston et al., 1983) shortly after elongation has started. Two of the four body wall muscle rows formed in the embryo are seen. At this stage, the cytoplasm of the body wall muscle cells is staining. The nuclei can be seen and counted. In normal embryos, 81 body wall muscle cells stain. (A2) Wild-type embryo at approximately 460 minutes of development. The muscle cells start to produce filaments; however the nuclei can still be identified. Soon after this stage, the general cytoplasmic staining disappears as the sarcomere develops and it is impossible to count the number of body wall muscle cells. Therefore manipulated embryos were evaluated only when the cytoplasmic staining was still present. Shortly before hatching the antibody stains faintly at least four more muscles which are not body wall muscles. These muscles are therefore not considered further in this work. (B-I) Examples of laser ablated embryos from the experimental series shown in Fig. 3. (B) Embryo in which all cells except for P3, the mother of D, were ablated. A total of 20 muscle cells were observed in this embryo. This is the equivalent normally produced by D. (C) Embryo in which all cells except for C were ablated. I counted 31 muscle cells in this embryo. C normally produces 32 muscle cells. (D) An embryo with an ablated EMS blastomere. I counted 31 muscle cells in this embryo. If C and D would produce their full complements of muscle this embryo should contain 53 (AB 1, C 32, D 20) muscle cells. (E) An embryo where EMS and C were ablated. The number of muscle cells observed in this embryo was 20, the number normally produced by D. (F) Embryo with ablated ABa and C blastomeres. A total of 34 muscle cells were observed in this embryo. Therefore MS should have produced 14 muscle cells in the embryo. (G, H) Embryos where ABa and D were ablated at the cleavage from two to four C descendants (G) or about 10 minutes later when the two E descendants cleave (H). The earlier D ablation still partially relieves the inhibition of MS-derived muscles. I counted 48 muscle cells in this embryo (G). When D is ablated later (H) the MS muscle is already fully suppressed. I counted 33 muscle cells in this embryo. (I) An embryo where first EMS was ablated and then later the two C descendants Ca and Cp were ablated just before the onset of their mitoses. Most of the inhibition of D already occurred in this embryo, only 6 muscle cells could be detected. Bar 10 μm.

The C. elegans embryo produces 81 body wall muscle cells derived from the AB (1), MS (28), C (32) and D (20) blas- tomeres (Fig. 1). The specification of the single body wall muscle cell derived from AB is not considered further in this work. By isolating the blastomeres C and P3, the mother of D, I confirmed earlier work suggesting that body wall muscle is produced cell autonomously from these blastomeres (Cowan and McIntosh, 1985; Mello et al., 1992; Figs 2B,C, 3A).

As described earlier, the production of body wall muscle in the MS lineage depends on an induction from the AB lineage since an inhibitory signal emanating from the posterior P2 lineage must be overcome (Schnabel, 1994). During this work, I often use the term lineage for convenience in a slightly unusual way to avoid an ongoing repetition of many blas- tomere names. For example, the term MS lineage is used to refer to the MS blastomere itself or its descendants or to several generations of descendants. To search for further cell-cell interactions required for the specification of tissues, embryos of the appropriate stage were mounted on microscope slides and blastomeres were ablated with a laser microbeam. Developed embryos were evaluated by staining with the mon- oclonal antibody NE8 4C6.3 specific for body wall muscle (see Schnabel, 1994) and by following the embryonic lineages of embryos whose development was recorded with a 4-Dimen- sional Microscope (Fig. 4; Hird and White, 1993; see also Schnabel, 1991; Hutter and Schnabel, 1994). It was shown before that laser ablations of individual blastomeres do not affect the remainder of the embryo unspecifically (Schnabel, 1994; Hutter and Schnabel, 1994).

Fig. 3.

Interactions modulating body wall muscle specification in the early embryo (key to figure in lower right panel). (A1) Muscle development in untreated embryos. (A2) An isolated C blastomere produces body wall muscle cell-autonomously. (A3) After isolation of P3, the D blastomere produces body wall muscle cell- autonomously. (B) Within the embryonic context, the EMS lineage is required to overcome an inhibition of muscle production conferred by the C lineage. After the ablation of EMS (B1), approximately 20 muscle cells are missing. (B2) The additional ablation of P3 does not affect muscle specification further. (B3) If C is ablated in addition to EMS, the inhibition of body wall muscle in D is relieved, which suggests that the C lineage is suppressing muscle specification in the D lineage. (B4) The two C descendants Ca and Cp are the inhibitors. After a late ablation of these blastomeres, most muscle derived from D is already suppressed. (C) Timing of the activation of muscle in the D lineage. (C2) Ablation of MS still fully inhibits muscle production in D. After ablation of the two descendants MSa and MSp (C3), the activation occurred partially; after the ablation of the four descendants of MS, at the onset of gastrulation the activation occurred fully (C4). (D) The blastomere D is the inhibitor of body wall muscle production in the MS lineage and the ABa lineage is required to overcome this inhibition. (D1-3) Again the initial set of experiments demonstrating that muscle formation in MS is exposed to duelling interactions (Schnabel, 1994) to facilitate the understanding of the following experiments. (D4-12) Experiments determining the timing of the inhibition and identifying the blastomere D derived from the P2 lineage (Fig. 1) as the blastomere inhibiting muscle production in MS. An early ablation (D4) of the P3 blastomere relieves the inhibition of the MS lineage completely. After a late ablation of this blastomere some muscles are already inhibited (D5). (D7,8) The ablation of the two P3 descendants D and P4 at the time when Ca and Cp divide relieves the inhibition of MS muscles partially, after the two E descendants have divided complete relief is observed. Ablation of only P4 does not (D9), ablation of D, however, does (D10) relieve the inhibition of MS-derived muscles, which indicates that the D lineage is the source of the inhibitory signal. (D11,12) The ablation of D at the time when the two C descendants Ca and Cp divide relieves the inhibition of MS muscles partially; after the two E descendants have divided complete relief is observed. (E) Some further ablations. (E1) The AB lineage does not produce supernumerary body wall muscle cells in isolation. (E2,3) Neither the ABa nor the ABp lineage are involved in the inhibition of muscle in the D lineage.

Fig. 3.

Interactions modulating body wall muscle specification in the early embryo (key to figure in lower right panel). (A1) Muscle development in untreated embryos. (A2) An isolated C blastomere produces body wall muscle cell-autonomously. (A3) After isolation of P3, the D blastomere produces body wall muscle cell- autonomously. (B) Within the embryonic context, the EMS lineage is required to overcome an inhibition of muscle production conferred by the C lineage. After the ablation of EMS (B1), approximately 20 muscle cells are missing. (B2) The additional ablation of P3 does not affect muscle specification further. (B3) If C is ablated in addition to EMS, the inhibition of body wall muscle in D is relieved, which suggests that the C lineage is suppressing muscle specification in the D lineage. (B4) The two C descendants Ca and Cp are the inhibitors. After a late ablation of these blastomeres, most muscle derived from D is already suppressed. (C) Timing of the activation of muscle in the D lineage. (C2) Ablation of MS still fully inhibits muscle production in D. After ablation of the two descendants MSa and MSp (C3), the activation occurred partially; after the ablation of the four descendants of MS, at the onset of gastrulation the activation occurred fully (C4). (D) The blastomere D is the inhibitor of body wall muscle production in the MS lineage and the ABa lineage is required to overcome this inhibition. (D1-3) Again the initial set of experiments demonstrating that muscle formation in MS is exposed to duelling interactions (Schnabel, 1994) to facilitate the understanding of the following experiments. (D4-12) Experiments determining the timing of the inhibition and identifying the blastomere D derived from the P2 lineage (Fig. 1) as the blastomere inhibiting muscle production in MS. An early ablation (D4) of the P3 blastomere relieves the inhibition of the MS lineage completely. After a late ablation of this blastomere some muscles are already inhibited (D5). (D7,8) The ablation of the two P3 descendants D and P4 at the time when Ca and Cp divide relieves the inhibition of MS muscles partially, after the two E descendants have divided complete relief is observed. Ablation of only P4 does not (D9), ablation of D, however, does (D10) relieve the inhibition of MS-derived muscles, which indicates that the D lineage is the source of the inhibitory signal. (D11,12) The ablation of D at the time when the two C descendants Ca and Cp divide relieves the inhibition of MS muscles partially; after the two E descendants have divided complete relief is observed. (E) Some further ablations. (E1) The AB lineage does not produce supernumerary body wall muscle cells in isolation. (E2,3) Neither the ABa nor the ABp lineage are involved in the inhibition of muscle in the D lineage.

Fig. 4.

Lineage analyses of ablated embryos. The upper part of the figure depicts the C and D lineages, the lower part MS-derived lineages producing body wall muscle and again the D lineage. The C lineage produces mostly hypodermis and muscles, the MS lineage produces pharynx (not shown) and body wall muscles. The D lineage produces only muscles. The figure shows the results of lineage analyses using the 4-D microscope. After the ablation of either EMS (A) or MS (B), the C-derived muscles develop normally as far as this can be judged using the microscope. Muscle precursors show the typical migration pattern and assemble into rows. However, the D-derived muscle cells undergo additional mitoses, an indication that these cells are not specified properly. After initially migrating with the other muscle precursors, cells often leave the muscle rows later and then undergo the additional mitosis. The C-derived hypodermal cells mostly developed normally. For a few cells I observed an additional mitosis. The evaluation of the hypodermal lineages proved to be very difficult in these embryos, which may indicate that these cells were somehow not completely normal. The significance of these additional mitoses remains to be clarified. (C) The lineage of the C-derived hypodermis was determined in two embryos where MS and P3 were ablated. I easily scored the hypodermal cells to be normal in these embryos. A normal differentiation of the hypodermis should be expected in these embryos if the P3 lineage not only acts as an inhibitor of body wall muscle differentiation in the MS lineage but also influences hypodermal differentiation in the C lineage which transfers the signal inhibiting muscle in the MS lineage. (D) Lineage analysis of ABa ablated embryos. MS-derived muscles either undergo an additional mitosis or differentiate into large round cells which somehow resemble hypodermal cells in these embryos. This confirms my earlier conclusion that the expression of body wall muscle depends on an induction from the ABa lineage (Schnabel, 1994). The D lineage produced body wall muscles in these embryos. (E) The ablation of C partially relieves the inhibition of MS-derived body wall muscles in ABa ablated embryos. The immunochemical analysis suggests that the MS lineage produces approximately 14 muscle cells in doubly ablated embryos. This is reflected in the terminal differentiation of the analysed lineages normally producing muscle in the MS lineage. Approximately half of the cells differentiated aberrantly. (*) The cell MSapppp arrested one cleavage too early and developed into a large cell with hypodermal appearance. The D lineage produced body wall muscles in this embryo.

Fig. 4.

Lineage analyses of ablated embryos. The upper part of the figure depicts the C and D lineages, the lower part MS-derived lineages producing body wall muscle and again the D lineage. The C lineage produces mostly hypodermis and muscles, the MS lineage produces pharynx (not shown) and body wall muscles. The D lineage produces only muscles. The figure shows the results of lineage analyses using the 4-D microscope. After the ablation of either EMS (A) or MS (B), the C-derived muscles develop normally as far as this can be judged using the microscope. Muscle precursors show the typical migration pattern and assemble into rows. However, the D-derived muscle cells undergo additional mitoses, an indication that these cells are not specified properly. After initially migrating with the other muscle precursors, cells often leave the muscle rows later and then undergo the additional mitosis. The C-derived hypodermal cells mostly developed normally. For a few cells I observed an additional mitosis. The evaluation of the hypodermal lineages proved to be very difficult in these embryos, which may indicate that these cells were somehow not completely normal. The significance of these additional mitoses remains to be clarified. (C) The lineage of the C-derived hypodermis was determined in two embryos where MS and P3 were ablated. I easily scored the hypodermal cells to be normal in these embryos. A normal differentiation of the hypodermis should be expected in these embryos if the P3 lineage not only acts as an inhibitor of body wall muscle differentiation in the MS lineage but also influences hypodermal differentiation in the C lineage which transfers the signal inhibiting muscle in the MS lineage. (D) Lineage analysis of ABa ablated embryos. MS-derived muscles either undergo an additional mitosis or differentiate into large round cells which somehow resemble hypodermal cells in these embryos. This confirms my earlier conclusion that the expression of body wall muscle depends on an induction from the ABa lineage (Schnabel, 1994). The D lineage produced body wall muscles in these embryos. (E) The ablation of C partially relieves the inhibition of MS-derived body wall muscles in ABa ablated embryos. The immunochemical analysis suggests that the MS lineage produces approximately 14 muscle cells in doubly ablated embryos. This is reflected in the terminal differentiation of the analysed lineages normally producing muscle in the MS lineage. Approximately half of the cells differentiated aberrantly. (*) The cell MSapppp arrested one cleavage too early and developed into a large cell with hypodermal appearance. The D lineage produced body wall muscles in this embryo.

Specification of body wall muscle in D

If the specification of the 52 body wall muscle cells derived from the blastomeres C and D occurred only cell autonomously, then the ablation of any other blastomere in the early embryo should not interfere with muscle development in these lineages. The ablation of ABa or ABp in the 4-cell-stage embryo has indeed no effect on the specification of muscle in either C or D (Schnabel, 1994). However, after ablation of the EMS blastomere muscle development is affected. If muscles were specified cell autonomously in the C and D lineages, one would expect to find 53 (C 32, D 20, AB 1) instead of the normal 81 muscle cells since MS, the daughter of EMS, con- tributes 28 cells to the body wall muscle. However, after the ablation of EMS the number is reduced to 31±2 (± s. d.) muscle cells (Figs 2D, 3B). The observed reduction of body wall muscles indicates that either the C or D or even both lineages require an induction from the EMS lineage to produce body wall muscle. Since the number of missing muscle cells is very close to the number produced by D (20), I tested by a double ablation of EMS and P3, the mother of D, whether indeed the contribution of D is missing. The rationale of this experiment is that the additional ablation of a blastomere that is unable to produce muscles due to a missing interaction should not make any difference. If, however, the missing muscles were normally produced by a cell different from the ablated one the number of muscle cells counted after this ablation should be further decreased. The number of muscle cells counted after the double ablation is the same (30±2) as when only EMS is ablated (31±2) (Fig. 3B). I took this as evidence that it is indeed the body wall muscle derived from D that is missing in EMS ablated embryos. This was confirmed by directly observing the muscle differentiation using the 4-Dimensional Microscope. In embryos where either EMS or MS were ablated, prospective muscle cells in the D lineage underwent an additional round of mitoses, which suggests that these cells are not specified properly (Fig. 4). Muscle cells derived from C, which are not affected according to the immunochemical analyses, differen- tiated normally into body wall muscles when observed directly using the 4-Dimensional Microscope.

The D lineage is subjected to an inhibitory signal

The result concerning muscle specification in D raises the same dilemma observed earlier when studying the muscle specifica- tion in the MS lineage. The isolated founder cell can produce its tissue apparently autonomously; however, it is possible to demonstrate that the specification of this same tissue within the embryonic context requires an induction. This dilemma can only be resolved if an inhibitory interaction normally sup- presses the autonomous tissue specification (Schnabel, 1994). In this situation, the repression of muscle cell specification should be relieved if the activating and inhibiting signals are both removed. This is what also happens in an isolation exper- iment. The inductions here referred to as ‘activating’ are indeed, as will be apparent later, ‘counter-inhibitory’ interac- tions which serve to overcome other inhibitory interactions. For the matter of simplicity, the counter-inhibitory signals are nevertheless referred to as activating signals throughout the manuscript because, in the presence of the inhibition, the experimental effect observed is an activation.

Concerning the inhibition of muscle in the D lineage, I could show that, after ablation of EMS and C, the D lineage produces its normal equivalent of 20 (20±1) muscle cells (Figs 2E, 3B), which suggests that the C lineage normally suppresses muscle development in D. The role of the C lineage in suppressing body wall muscle in D is discussed later in more detail. EMS is required to overcome this inhibition. Double ablations of either ABa or ABp and EMS do not relieve the suppression of muscle production from D (Fig. 3E).

These results indicate that the MS lineage is required for the proper specification of the D lineage, which is derived from the P2 lineage. I earlier reported that the production of body wall muscle in the MS lineage itself is inhibited by the P2 lineage (Schnabel, 1994). Therefore the interactions between these two lineages are reciprocal. The P2 lineage suppresses muscle specification in the MS lineage whereas the MS lineage is required to permit muscle specification in the blastomere D which is derived from the P2 lineage.

The timing of the interactions

It appears that a complex network of duelling interactions exists in the early C. elegans embryo. ABa is required for muscle production in MS, which is repressed by the P2 lineage. MS itself is required to overcome an inhibitory signal within the P2 lineage. Two formal scenarios could explain the complex interactions. (i) There exist discrete independent interactions among the blastomeres which could be required at different times of development. (ii) Alternatively, the complex interaction pattern could be generated by a network of recip- rocal interactions occurring simultaneously. To distinguish between these two possibilities I determined the timing of the inhibitory and activating signals, which are discussed in the following three sections.

Activation of muscle in the D lineage

To determine the timing of the induction of muscle in the D lineage, I ablated EMS or the blastomeres of the MS lineage at different stages of early development. In these experiments, the inhibiting signal derived from the C lineage remains intact, but the source of the activating signal is removed at different times of development. If the activation did not yet occur, the D-derived muscle should still be suppressed. If the ablation of the source is carried out after the activation occurred D should express its normal equivalent of body wall muscles. Ablation of EMS or MS completely suppresses muscle formation in the D lineage. After the ablation of the two MS descendants 40% of the D-derived muscles are formed, an indication that a partial activation already occurred. If the 4 MS descendants are ablated soon after their birth, the full equivalent of D-derived muscles is formed, which shows that the activation is already completed. I therefore conclude that the activation of body wall muscle in the D lineage occurs late at the 2 MS cell stage around the time when the D blastomere is born (Figs 3C, 5).

Inhibition of muscle in the MS lineage

In my previous work I showed, using immunochemical methods, that the specification of body wall muscle in the MS lineage requires a signal from the ABa lineage to counteract an inhibitory signal derived from the P2 lineage (Schnabel, 1994). The conclusion that it is the MS-derived body wall muscle that requires the signal from the ABa lineage was now confirmed by following MS lineages producing body wall muscle in the 4-Dimensional Microscope (Fig. 4D). To identify the P2-derived blastomere which produces the inhibitory signal and to determine the timing of the inhibitory signal, I first ablated ABa in embryos to remove the signal activating muscle in MS. To identify the source of the inhibitory signal which now can be scored in absence of the activating signal, I ablated additionally P3 very early or late in its cell cycle or both of its daughters, either around the time when the two C blastomeres are cleaving, or after the cleavage of the two E cells (Figs 1, 3D). An early ablation of P3 still fully relieves the inhibition of body wall muscles in the MS lineage. The consequences of the ablation of C, the other daughter of P2, will be discussed below in a different context. If both daughters of P3, P4 and D are ablated at the time when the two C descendants cleave, most of the body wall muscles are already suppressed (approx- imately 70%). Ablation of both blastomeres right at or after the cleavage of the two E descendants about 10 minutes later shows that the inhibition is already completed at that time of development (Fig. 3D). To identify which of the daughters of P3, D or P4 is producing the inhibitory signal, I ablated ABa and P4 or D. Ablation of P4 does not relieve the inhibition sug- gesting that the D lineage is the source of the inhibitory signal. Indeed, the ablation of the D blastomere right after its birth relieves most of the inhibition, approximately 75% of the MS- derived muscles are produced. If D is ablated 10 minutes later at the cleavage of the two C descendants most of the inhibi- tion of MS-derived muscles is observed and only approxi- mately 35% of the muscles are produced in the MS lineage. The MS-derived muscle is fully suppressed when D is ablated another 10 minutes later at or after the cleavage of the two E descendants. These results suggest that the D blastomere itself is the source of the inhibitory signal suppressing muscle in the MS lineage (Figs 2G,H, 3D).

I reported earlier that the ABp lineage also functions as an inhibitor if ABa and P2 are ablated (Schnabel, 1994). The results presented here show that this is very probably due to the P2 ablation itself. When ABa and additionally the daughters of P2, P3 and C are either ablated on their own or together ABp never elucidates an inhibitory activity (Figs 3D4 and 6A3-5).

Simultaneous exchange of signals among blastomeres

A graph showing the time courses of the interactions is depicted in Fig. 5. The curves corresponding to the activation of the muscle specification in the D lineage and to the inhibi- tion of muscle specification in the MS lineage cross at the late 24-cell / early 26-cell stage of the embryo. This is the time when the blastomere D is born. The time window in which the interactions occur corresponds to approximately 30 minutes at 25°C. Within the inherent limit of the resolution of the laser ablation experiments (it may take a few minutes after the ablation is finished until a blastomere is inactivated), it appears that the relation between the MS and D lineage is reciprocal. The activating and inhibiting signals are exchanged simulta- neously at about the time when the D blastomere is born.

Fig. 5.

Timing of the interactions modifying muscle development in the MS and D lineages. The upper part of the figure shows the inhibition or activation of body wall muscle cells when different blastomeres are ablated at the different embryonic stages indicated in the lower part of the figure. The horizontal bars indicate the times of existence of the different blastomeres. The triangles correspond to the activation of body wall muscle in the MS lineage by the ABa lineage (data from Schnabel, 1994). The activation occurs around the 12-cell stage of the embryo. The circles indicate the activation of the D-derived body wall muscle by the MS lineage (data from Fig. 3C). The activation occurs around the late 24- / early 26-cell stage and thus one cleavage later than the activation of MS. The squares indicate the inhibition of muscle production in the MS lineage by the P2 lineage and its descendants after the activating ABa lineage was ablated (data from Fig. 3D). The inhibition occurs exactly at the same time as the reciprocal activation of the D lineage by the MS lineage. The two crosses indicate the timing of the inhibition of the D-derived muscle by the C lineage. After an ablation of the activating MS lineage the ablation of the blastomere C still fully relieves the inhibition of D. When the two C descendants Ca and Cp are ablated shortly before they undergo mitoses most of the muscle derived from D is already inhibited. Therefore the inhibition of D appears to follow the same time course as the other interactions in the posterior of the embryo.

Fig. 5.

Timing of the interactions modifying muscle development in the MS and D lineages. The upper part of the figure shows the inhibition or activation of body wall muscle cells when different blastomeres are ablated at the different embryonic stages indicated in the lower part of the figure. The horizontal bars indicate the times of existence of the different blastomeres. The triangles correspond to the activation of body wall muscle in the MS lineage by the ABa lineage (data from Schnabel, 1994). The activation occurs around the 12-cell stage of the embryo. The circles indicate the activation of the D-derived body wall muscle by the MS lineage (data from Fig. 3C). The activation occurs around the late 24- / early 26-cell stage and thus one cleavage later than the activation of MS. The squares indicate the inhibition of muscle production in the MS lineage by the P2 lineage and its descendants after the activating ABa lineage was ablated (data from Fig. 3D). The inhibition occurs exactly at the same time as the reciprocal activation of the D lineage by the MS lineage. The two crosses indicate the timing of the inhibition of the D-derived muscle by the C lineage. After an ablation of the activating MS lineage the ablation of the blastomere C still fully relieves the inhibition of D. When the two C descendants Ca and Cp are ablated shortly before they undergo mitoses most of the muscle derived from D is already inhibited. Therefore the inhibition of D appears to follow the same time course as the other interactions in the posterior of the embryo.

To address the question whether the inhibition of body wall muscle production in the D lineage by the C lineage also occurs at the same time as the interactions between the MS and D lineages, I ablated EMS to remove the activation of muscle in D and then ablated the two C descendants Ca and Cp very shortly before they cleave again. As already described, the ablation of C itself completely relieves the inhibition of muscle in the D lineage. If the inhibition of body wall muscle in the D lineage by the C lineage would follow the same time course as the inhibition of body wall muscle in the MS lineage, one would expect that only about 30% of the D-derived body wall muscles should be formed when Ca and Cp are ablated shortly before their division. After ablation of these cells indeed approximately 7 instead of the normal 20 body wall muscles derived from D are produced. This corresponds to approxi- mately 30% of the D-derived muscles. Therefore, it is very probable that the interaction between the C and D lineage occurs at the same time as the other interactions (Figs 3B, 5).

The D-derived signal suppressing muscle in the MS lineage is transmitted by the C lineage

The signals could be either exchanged by long-range diffusible signals or they could be transferred through cell-cell contacts among blastomeres. Since very short-range diffusible signals cannot be distinguished with the methods used here from signals transferred by direct cell-cell contacts, these two mech- anisms are not distinguished further in this work. Cell-cell contacts confer the specificity of the inductions specifying the AB lineage (Hutter and Schnabel, 1994; 1995; Moskowitz et al., 1995). To discriminate between the two possibilities for the inhibition of the MS lineage by D, I carried out double ablations ablating ABa to remove the signal activating muscle in MS, and C or E whose descendants could themselves be part of a signal chain between D and the MS descendants (the topography and the cell-cell contacts in 24-cell embryos are discussed below; Fig. 7). Ablation of E does not interfere with the inhibition of body wall muscle in MS by the D lineage. The ablation of C, however, partially relieves the inhibition of muscle production in MS. The immunochemical analysis of body wall muscle specification in these embryos suggests that approximately 14 body wall muscles are produced by the MS lineage (Figs 2F, 6A3-5). A lineage analysis shows that the MS lineage produces indeed some body wall muscles when C is ablated in addition to ABa. The D lineage develops normally after the ablation of ABa and C (Fig. 4E). The C lineage thus acts formally as an inhibitor of muscle production in the MS lineage. However, the C lineage depends completely on the P3 (D) lineage to acquire its inhibitory function. Ablation of P3 alone, which leaves the C lineage intact, completely relieves the inhibition of the MS lineage (Fig. 3D4). The C lineage on its own thus does not inhibit the MS lineage. Therefore I propose that the inhibitory signal from D is transferred through the C lineage. To find further evidence that the inhibitory signal is transferred through C descendants I ablated the MS activat- ing lineage ABa and the blastomere Cp immediately after its birth at a time when the inhibitory blastomere D is not yet born. During this experiment, Ca, which touches the MS descendants (Fig. 6), stays intact and could thus inhibit the MS lineage on its own. The ablated blastomere Cp, however, ‘isolates’ Ca from P3 and D, respectively. After this manipulation, the inhi- bition of body wall muscle production from MS is relieved to the same extent (approximately 50%) as when the mother of these cells, C, is ablated (Fig. 6). This indicates that Ca does not inhibit the MS lineage on its own, but that the D lineage is required for the inhibitory activity. It appears that Ca depends on Cp to execute an inhibitory function or to transmit an inhibitory signal.

Fig. 6.

The reciprocal interactions between the MS and D lineages are mediated by the C lineage. (A) The inhibitory signal suppressing body wall muscle formation in the MS lineage derived from the blastomere D is transferred by the descendants of the C blastomere. In all embryos, the ABa blastomere was ablated to remove the signal activating muscle in the MS lineage in order to permit the detection of the inhibitory signal derived from the blastomere D. (A1,2) As shown before in Fig. 3D, the ablation of the blastomere P3 relieves the inhibition of muscle in the MS lineage completely. Neither the ABp nor the C lineage act as inhibitors in this situation. (A3) Nevertheless the suppression of body wall muscle is relieved when the C lineage is ablated, which suggests that it is required to transfer the inhibitory signal. (A4,5) The ABp lineage does not act as an inhibitor when only C or C and P3 are ablated. This experiment excludes the possibility that ablation of C fully relieves the inhibition of the MS lineage but in turn body wall muscle is suppressed by the ABp lineage as with the ablation of P2. (A5) Late ablation of C and P3 removes the inhibition of muscle in the MS lineage to the same extent as ablating only P3 at the same time (Fig. 3D5). This suggests that the P3 lineage, that is the blastomere D, is the initial source of the signal whose transfer is hampered by the ablation of C. The experiment also provides evidence that it is the ablation of P2 itself that causes the ABp lineage to inhibit muscle in the MS lineage. (A6) The ablation of only the blastomere Cp has the same effect as the ablation of the whole C lineage. This corroborates the argument that the C lineage on its own does not inhibit muscle in the MS lineage since Ca is normal in this experiment and could still function as an inhibitor. Cp is located completely sideways from the direct line between the D blastomere and the MS descendants. This makes it rather probable that the inhibitory signal is really transferred either directly through or over the surface of the C descendants. It is rather improbable that the ablation of Cp should interfere with a free diffusion of a soluble factor between D and MS descendants. (A7) The ablation of the E lineage whose descendants span, like the descendants of C, the gap between the descendants of MS and the blastomere D has no effect, which suggests that this lineage is not involved in the transfer of a signal. (B) The MS-derived signal activating body wall muscle in the D lineage acts indirectly by inhibiting the inhibitory function of the C lineage. (B1,2) The ablation of Cp relieves the inhibition of muscle formation from D observed after the ablation of only EMS. (B3) The ablation of only Cp has, as expected for the removal of an inhibitor, no effect on D-derived muscles. (B4) The ablation of only Ca, however, blocks the differentiation of body wall muscles in the D lineage. This indicates that the MS lineage does not activate the D lineage directly and that the C lineage is required to transfer the signal (B5). In contrast the ablation of Ea has no effect on muscle specification.

Fig. 6.

The reciprocal interactions between the MS and D lineages are mediated by the C lineage. (A) The inhibitory signal suppressing body wall muscle formation in the MS lineage derived from the blastomere D is transferred by the descendants of the C blastomere. In all embryos, the ABa blastomere was ablated to remove the signal activating muscle in the MS lineage in order to permit the detection of the inhibitory signal derived from the blastomere D. (A1,2) As shown before in Fig. 3D, the ablation of the blastomere P3 relieves the inhibition of muscle in the MS lineage completely. Neither the ABp nor the C lineage act as inhibitors in this situation. (A3) Nevertheless the suppression of body wall muscle is relieved when the C lineage is ablated, which suggests that it is required to transfer the inhibitory signal. (A4,5) The ABp lineage does not act as an inhibitor when only C or C and P3 are ablated. This experiment excludes the possibility that ablation of C fully relieves the inhibition of the MS lineage but in turn body wall muscle is suppressed by the ABp lineage as with the ablation of P2. (A5) Late ablation of C and P3 removes the inhibition of muscle in the MS lineage to the same extent as ablating only P3 at the same time (Fig. 3D5). This suggests that the P3 lineage, that is the blastomere D, is the initial source of the signal whose transfer is hampered by the ablation of C. The experiment also provides evidence that it is the ablation of P2 itself that causes the ABp lineage to inhibit muscle in the MS lineage. (A6) The ablation of only the blastomere Cp has the same effect as the ablation of the whole C lineage. This corroborates the argument that the C lineage on its own does not inhibit muscle in the MS lineage since Ca is normal in this experiment and could still function as an inhibitor. Cp is located completely sideways from the direct line between the D blastomere and the MS descendants. This makes it rather probable that the inhibitory signal is really transferred either directly through or over the surface of the C descendants. It is rather improbable that the ablation of Cp should interfere with a free diffusion of a soluble factor between D and MS descendants. (A7) The ablation of the E lineage whose descendants span, like the descendants of C, the gap between the descendants of MS and the blastomere D has no effect, which suggests that this lineage is not involved in the transfer of a signal. (B) The MS-derived signal activating body wall muscle in the D lineage acts indirectly by inhibiting the inhibitory function of the C lineage. (B1,2) The ablation of Cp relieves the inhibition of muscle formation from D observed after the ablation of only EMS. (B3) The ablation of only Cp has, as expected for the removal of an inhibitor, no effect on D-derived muscles. (B4) The ablation of only Ca, however, blocks the differentiation of body wall muscles in the D lineage. This indicates that the MS lineage does not activate the D lineage directly and that the C lineage is required to transfer the signal (B5). In contrast the ablation of Ea has no effect on muscle specification.

Fig. 7.

Cell-cell contacts in a late 24-cell embryo. (A,B) Nomarski micrographs. Left ventral views. (A) Top optical section. (B) Bottom section. The D blastomere always touches the blastomere Cp. The C-derived and MS-derived blastomeres lie on the dorsal and ventral side of the embryo, respectively. They only touch in the inner of the embryo. (C) The contact between MSp and Ca can be seen in an electron micrograph kindly provided by T. Cole. Bars 10 μm.

Fig. 7.

Cell-cell contacts in a late 24-cell embryo. (A,B) Nomarski micrographs. Left ventral views. (A) Top optical section. (B) Bottom section. The D blastomere always touches the blastomere Cp. The C-derived and MS-derived blastomeres lie on the dorsal and ventral side of the embryo, respectively. They only touch in the inner of the embryo. (C) The contact between MSp and Ca can be seen in an electron micrograph kindly provided by T. Cole. Bars 10 μm.

I have shown previously that the ABp descendants, which are also located in the posterior of the embryo, do not have a role in propagating the D-derived signal. After ablation of the entire AB lineage, which removes the activating signal from ABa but leaves EMS, C and P3 (D) intact, the MS-derived muscle is fully suppressed and the C- and D-derived muscles are expressed normally (Schnabel, 1994).

The MS-derived signal activating muscle in the D lineage is also transmitted by the C lineage

As discussed above, the C lineage acts as an inhibitor of muscle development in the D lineage. This inhibition is overridden by the MS lineage. Again, I wanted to examine by which path the activating signal is transmitted. ABp-derived blastomeres which are located in the posterior of the embryo do not appear to have a role in the transfer of the signal as the ablation of ABp alone has no influence on the transfer of the activation to D (Schnabel, 1994). I directly wanted to test the question of whether the signal is also transferred by the C lineage. This question, however, leads to the experimental dilemma that the ablation of the C lineage, which possibly transmits the signal, also removes the source of the inhibition, which is derived from the same lineage. It was shown earlier that the ablation of C fully relieves the inhibition of D (Fig. 3B3). However, if the two descendants of C, Ca and Cp are ablated very late shortly before their division, most of the D-derived muscle is already inhibited (Fig. 3B4).This suggests that both or one of the descendants function as an inhibitor. The fact that the inhibitory signal is only sent after the two C descendants are born permits to design two experiments to test whether the acti- vating signal is also transmitted through the C descendants. The first experiment is to ablate EMS and Cp. Ablation of EMS removes the activation of the D-derived muscle, which allows testing of whether Cp, which touches D (Fig. 7), can act as an inhibitor by itself. Indeed, an additional early ablation of Cp relieves the inhibition of the D-derived muscle (Fig. 6B2), which suggests that the Cp lineage inhibits the muscle in D. As expected for an inhibitor, ablation of only Cp has no effect on D (Fig. 6B3). If Cp on its own acts as an inhibitor, it is possible to test whether the activating signal from the MS lineage is transferred by the C descendants or by the remainder of the embryo by ablating only Ca. Ablation of only Ca leaves both the activating lineage MS and the inhibiting lineage Cp intact. If the activating signal is transmitted by a path different from the C descendants to D, the ablation of only Ca should have no effect. If, however, the activating signal passes through the C descendants, the ablation of Ca alone should block the transmission of the signal which in turn should permit Cp to suppress muscle in D. As shown in Fig. 6B4, the ablation of only Ca suppresses the D-derived muscle. As a control, I also ablated only Ea which had no effect (Fig. 6B5). These results suggest that both the activation and the inhibi- tion of the D-derived muscle occur via the C lineage. As will be discussed later, MS could possibly directly influence the C lineage and thus relieve the inhibition of the D lineage.

Cell-cell contacts among the interacting blastomeres

The presented experiments supply evidence that the MS lineage interacts with the D lineage, the D lineage with the MS lineage and the C with the D lineage. The reciprocal interac- tions between MS and D depend on the C descendants, which suggests that the signals do not diffuse freely in the posterior of the embryo but are transferred directly through the C descendants. This raises the possibility that the signals are passed through direct contacts among blastomeres. To see whether indeed direct cell-cell contacts exist between the inter- acting blastomeres, I analysed the cell-cell contacts in 24-cell embryos, the stage when the interactions occur. Cell-cell contacts between the descendants of D and C are obvious in embryos inspected under the light microscope (Fig. 7). The two MS and two C descendants lie, however, on opposite sides of the embryo; therefore, MSp and Ca could only have contact in the centre of the embryo. This contact can be seen in the light microscope but is very hard to depict in pictures. In an electron microscopic section of a 24-cell-stage embryo, published already by Krieg et al. (1978), the contact between MSp and Ca can clearly be seen (also shown in Fig. 7C; the picture was kindly provided by T. Cole). Thus interactions among the lineages MS and C and C and D, respectively, could occur through direct cell-cell contacts.

Is the C lineage itself influenced by the interactions?

None of the described ablation experiments had an effect on the body wall muscle specification in the C lineage. Thus, body wall muscle specification in this lineage appears to be different from that in the other blastomeres. The blastomere C also produces hypodermis. To test the possibility that the formation of hypodermis may depend on the MS lineage, which influ- ences the inhibitory activity of the C lineage, I evaluated the differentiation of the C-derived hypodermis in ablated embryos. Since the major part of the hypodermis is derived from the AB lineage and since I am not aware of a reliable staining method to score the C-derived hypodermis directly, I analysed the differentiation of hypodermal cells derived from the C lineage by using the 4-D Microscope. After the ablation of either EMS or MS, most prospective hypodermal cells did differentiate into hypodermis. A few hypodermal precursors, however, underwent additional mitoses (Fig. 4), which could be taken as an indication that these cells are not specified normally. The significance of this effect remains to be clarified.

Muscle specification appears autonomous but it is not

To study the problem of cell determination during embryoge- nesis, it is essential to know the basic mechanism(s) by which a certain structure is specified. The interpretation of mutant phenotypes of developmental genes depends on a detailed knowledge of the general pathway of cell determination. All mutant phenotypes of genes affecting descendants of the blas- tomere P1 have been interpreted in terms of a cell-autonomous specification of the tissues derived from this blastomere (Kemphues et al., 1985; Schnabel and Schnabel, 1990; Bowerman et al., 1992a; Mello et al., 1992). The main aim of this work is to show that the somatic founder cells MS and D, and possibly C, participate in cell-cell interactions during their early development.

The three lineages producing body wall muscle can still do so when the remainder of the embryo is laser ablated. By this criterion, body wall muscle is specified autonomously (Fig. 8A). If, however, only single blastomeres like ABa or EMS are ablated, muscle production is suppressed in the MS and D lineages, respectively. This result clearly contradicts the notion that body wall muscle is specified autonomously in the embryonic context. It indicates that in the embryo muscle must be induced. This contradiction can only be resolved by assuming that other interactions normally suppress body wall muscle in the embryonic context. I indeed identified two lineages producing inhibitory signals. Muscle formation in the MS lineage is suppressed by the D blastomere. The formation of body wall muscle in the D lineage itself is suppressed by the C lineage.

Fig. 8.

A network of duelling interactions in the early C. elegans embryo. (A) Upon isolation from the remainder of the embryo, the MS, C and D lineages produce body wall muscles cell- autonomously. Nevertheless body wall muscle specification in the embryo depends on cell-cell interactions. A formal description of the interactions modulating muscle specification is depicted in (B). (C) The interactions among the lineages are reciprocal. The ABa lineage activates body wall muscle in the MS lineage whereas the MS lineage specifies the left-right asymmetry of the ABa lineage. The interactions of the MS and D lineage occur through the C lineage and are very probably transmitted through cell-cell contacts among the descendants of these lineages. (D) Spaghetti model of the duelling interactions. The cell-cell interactions activating body wall muscle do so by inducing states in the MS and C lineages with altered competence (squares). At the 12-cell stage, the ABa lineage alters the state of the MS lineage to one that is no longer susceptible to the inhibitory signal that it is exposed to one cleavage later. The MS lineage suppresses the competence of the C lineage to inhibit the muscle production in the D blastomere at the late 24-/early 26-cell stage of the embryo. Different shapes indicate different states of competence of blastomeres. The squares are states that do not participate any more in the interactions. The fonts indicate different fates of the blastomeres. (ABa) Ground state, left-right symmetric. (ABa) Induced, left-right asymmetric. (MS) Ground state, normal fate, susceptible to inhibition. (MS) Induced state, resistant to inhibition. (𝕄𝕊) Abnormal fate induced, no body wall muscle produced. (C) Muscle specification in the C lineage is not influenced by the interactions. (D) Ground state, normal fate, susceptible to inhibition. (𝔻) Abnormal fate induced, no body wall muscle produced. The stippled circles enclose the states that participate in the interactions. Since embryos from mothers homozygous for one allele (e2141) of glp-1 are left-right symmetric but have normal body wall muscles an uninduced ABa lineage can still activate muscle in the MS lineage (Schnabel, 1994). All states of the MS lineage prevent the inhibitory function of the C lineage. It is currently not clear whether an inhibited D blastomere is able to suppress body wall muscle in the MS lineage.

Fig. 8.

A network of duelling interactions in the early C. elegans embryo. (A) Upon isolation from the remainder of the embryo, the MS, C and D lineages produce body wall muscles cell- autonomously. Nevertheless body wall muscle specification in the embryo depends on cell-cell interactions. A formal description of the interactions modulating muscle specification is depicted in (B). (C) The interactions among the lineages are reciprocal. The ABa lineage activates body wall muscle in the MS lineage whereas the MS lineage specifies the left-right asymmetry of the ABa lineage. The interactions of the MS and D lineage occur through the C lineage and are very probably transmitted through cell-cell contacts among the descendants of these lineages. (D) Spaghetti model of the duelling interactions. The cell-cell interactions activating body wall muscle do so by inducing states in the MS and C lineages with altered competence (squares). At the 12-cell stage, the ABa lineage alters the state of the MS lineage to one that is no longer susceptible to the inhibitory signal that it is exposed to one cleavage later. The MS lineage suppresses the competence of the C lineage to inhibit the muscle production in the D blastomere at the late 24-/early 26-cell stage of the embryo. Different shapes indicate different states of competence of blastomeres. The squares are states that do not participate any more in the interactions. The fonts indicate different fates of the blastomeres. (ABa) Ground state, left-right symmetric. (ABa) Induced, left-right asymmetric. (MS) Ground state, normal fate, susceptible to inhibition. (MS) Induced state, resistant to inhibition. (𝕄𝕊) Abnormal fate induced, no body wall muscle produced. (C) Muscle specification in the C lineage is not influenced by the interactions. (D) Ground state, normal fate, susceptible to inhibition. (𝔻) Abnormal fate induced, no body wall muscle produced. The stippled circles enclose the states that participate in the interactions. Since embryos from mothers homozygous for one allele (e2141) of glp-1 are left-right symmetric but have normal body wall muscles an uninduced ABa lineage can still activate muscle in the MS lineage (Schnabel, 1994). All states of the MS lineage prevent the inhibitory function of the C lineage. It is currently not clear whether an inhibited D blastomere is able to suppress body wall muscle in the MS lineage.

Thus formally four lineages interact during body wall muscle specification (Fig. 8B). There is strong evidence that Analysis of the timing of the interactions showed that the inter- actions occur at two stages of development. At the 12-cell stage, the ABa lineage and the MS blastomere interact recip- rocally. Both inductions depend on the receptor encoded by the glp-1 gene (Austin and Kimble, 1987; Priess et al., 1987; Yochem and Greenwald, 1989; Hutter and Schnabel, 1994; Schnabel, 1994). This induction specifies features in both lineages. The MS blastomere induces the left-right asymmetry within the ABa lineage (Hutter & Schnabel, 1994) and the ABa lineage induces body wall muscles in the MS lineage (Schnabel, 1994). This induction, as will be discussed below, anticipates a counteracting inhibitory signal being part of the other interactions among the MS, C and D lineage, which occur about one cleavage later at the late 24- and the early 26-cell stage. The interaction between the MS and D lineage is formally also reciprocal. This reciprocity is not created by a direct cellular contact but in a more complicated way involving cells lying in between.

These latter interactions occur through or involve the two C descendants Ca and Cp (Fig. 8C) but are qualitatively very different. In the case of inhibition of MS-derived muscle, the data are consistent with the notion that the C descendants act three of the four lineages involved in this network of interac- tions are affected by the interactions. The question whether the C lineage is also affected by the interactions remains open.

Activation Inhibition as transmitters of the signal which is derived from the D blas- tomere. The C lineage on its own does not inhibit muscle pro- duction in the MS lineage. In this case, the affected partners of the interactions are interacting through intercalated blas- tomeres. It was not possible to block the inhibitory signal com- pletely. A possible explanation for this could be that the signal is transferred along the surface of the blastomeres, which may not be affected as severely by the ablations as the cytoplasm and the nuclei of the irradiated blastomeres.

It appears surprising that the disturbance of cell-cell inter- actions (Figs 3C3, D7, 4E and 6A4) causes a blastomere to execute a lineage pattern that in its terminal branches is partially normal and partially abnormal. It is thus possible to induce fates partially (see also Hutter and Schnabel, 1994, Fig. 2 for another example). This indicates that the cell determina- tion may be labile throughout the whole lineage specification. This somehow contradicts the general notion that develop- mental decisions always lead to the establishment of exactly one fate.

The situation concerning the modulation of muscle production in the D lineage appears to be different from that in MS. In this case, the most simple assumption is that the interactions occur directly between the activator and inhibitor and not directly with the affected D lineage. The MS lineage sup- presses the inhibitory function of the inhibiting C lineage. The effect is thus indirect. It appears that the C lineage displays a ground state that inhibits muscle production in the D lineage. This state is altered by the MS lineage to another state which is no longer inhibiting the D lineage and D can now act autonomously (Fig. 8D).

The interacting cells all have direct cell-cell contacts, which suggests that the signals are mediated by these contacts. This can be proven by showing that an experimental alteration of the contacts between blastomeres also alters the induction of fates, as it was demonstrated for the induction of the ABp fate by P2 and the induction of left-right asymmetrical fates in the ABa lineage by MS (Hutter and Schnabel, 1994; 1995; Moskowitz et al., 1994). This experiment depends, however, on the existence of other competent, i. e. equivalent blas- tomeres that are possibly induced by altered contacts. An inter- esting feature of the cell-cell interactions described here is that they do not serve to discriminate between equivalent cells, which is normally the purpose of ‘instructive’ inductions. In the classical terms, they qualify rather as ‘permissive’ or in the case of the inhibitory signals as ‘nonpermissive’ inductions (for review see Slack, 1991). Because in ‘permissive’ induc- tions equivalent cells do not exist, the appropriate experiments to test whether the interactions indeed are mediated through contacts cannot be executed.

A further interesting feature of the interactions is that they all appear to be independent from each other. They occur at different embryonic stages and by interfering with one of the interactions one does not necessarily influence any of the other interactions. The ablation of ABa suppresses the production of muscle in the MS lineage but it does not affect the general com- petence of the MS lineage to interact with the C lineage. The same is true for the interactions in the posterior of the embryo. Ablation of C removes the inhibition of D, but D still produces the signal inhibitory to the MS lineage, since the ablation of C only partially hinders the signal suppressing muscle in the MS lineage.

Classically nematodes were considered to develop cell- autonomously (for review see zur Strassen, 1959). However, this notion was already disputed at the beginning of this century (Boveri, 1910). In recent years, it was well established that the AB lineage of the C. elegans embryo is specified by at least five different inductions following the embryonic axes (Priess and Thomson, 1987; Schnabel, 1991; Wood, 1991; Bowerman et al., 1992b; Hutter and Schnabel, 1994; Mango et al., 1994; Mello et al., 1994; Moskowitz et al., 1994; Hutter and Schnabel, 1995; Hutter and Schnabel, unpublished data). The P1 lineage, however, has been typically viewed as devel- oping autonomously (Laufer et al., 1980; Cowan and McIntosh, 1985; Edgar and McGhee, 1986; Kemphues et al., 1988; Schierenberg, 1988; Schnabel and Schnabel, 1990; Bowerman et al., 1992a; Mello et al., 1992). This view was first challenged by Schierenberg (1987) and later by Goldstein (1992), who suggested that the P2 blastomere interacts with EMS blastomere to permit the expression of its potential to express intestine. My previous work (Schnabel, 1994) and this work now suggest that all P1-derived somatic founder cells participate in inductions.

Different pathways for body wall muscle specification

The work presented here shows that body wall muscle speci- fication in the MS and D lineages is susceptible to cell-cell interactions, though in each case differently, whereas the muscle specification in C appears to be cell-autonomous. This indicates that there exist possibly three different pathways for body wall muscle specification in the early C. elegans embryo. Genetic evidence for two different pathways of body wall muscle specification also emerged earlier during the course of the analysis of the gene skn-1 (Bowerman et al., 1992a, 1993; Mello et al. 1992). Since this gene is required for the specifi- cation of the mother of the MS blastomere, the EMS blas- tomere, it probably acts upstream of the MS specific pathway for muscle specification.

How are the functions of blastomeres integrated?

The observations presented here all address the question by which means the early functions of the blastomeres, for example the competence to emit and/or to receive signals, are related to the general identities of blastomeres which are reflected by the lineage pattern and tissues produced by these blastomeres. Two extreme scenarios may help to clarify the problem. One scenario is that all functions of a blastomere are strictly coupled. The blastomeres have one identity that integrates all the functions at a certain stage. For example, the MS blastomere has an integral identity that controls the functions inducing the AB and the C lineage but also activates the programs governing the control of the complex lineage pattern producing the tissues corresponding to the lineage. According to this scenario, it should not be possible to separate the different functions of the program of a blastomere easily. This could only be achieved by interfering with downstream functions for example by genetic means. This notion has been described earlier for P1-derived blastomeres using the term blastomere identity (Schnabel and Schnabel, 1990; Mello et al., 1992).

The other scenario is that blastomeres have roles during the early development that are not directly related to their function as precursors for tissues formed later in development. In this case, it should be possible to separate these functions. Consid- ering the results presented here I favour the latter scenario. It is possible to interfere e.g. with one part of the function of the MS lineage without affecting others. One can suppress the pro- duction of body wall muscle right after the blastomere is born without affecting the other features of the lineage, for example the potential to alter the state of the C lineage.

Blastomeres may acquire transient states not reflected in their own specification

The MS lineage is subjected to two duelling interactions, which occur at different times of embryogenesis. The activat- ing induction from the ABa lineage that sustains the specifica- tion of body wall muscle occurs before the interaction sup- pressing body wall muscle, which has an effect only if the activation did not occur. The activating signal therefore alters the state of the MS lineage to render it insensitive to the second signal. The phenomenon that two different states exist is also observed in the C lineage (Fig. 8D). The MS lineage switches the C lineage from a default state that is inhibitory for the D lineage into another state with no inhibitory capability. The two states of the C lineage are, in contrast to the MS lineage, not reflected in at least profound changes of the differentiation potential of the C lineage itself (Fig. 4). It thus appears that blastomeres acquire states in the embryonic context that cannot be diagnosed directly by evaluating the differentiation potential of the blastomeres in isolation or even in the embryo proper. The states are only reflected in their effects on other cells. This insight has implications for the general interpreta- tion of isolation experiments where primordia or combinations of primordia are analysed in cell culture. The behaviour of isolated cells may be very different from that of cells within the embryonic context. The behaviour of the C blastomere may serve as an example of such a phenomenon. An isolated C blas- tomere suppresses muscle formation in an opposed D blas- tomere but in an intact embryo this will never occur. It is quite conceivable that similar phenomena may occur in systems more complex than C. elegans.

Why do these interactions occur?

An intriguing question is why this network of duelling interactions should exist. I could not find any obvious purpose for these interactions using the given methods. I cannot exclude, however, that there exists a purpose that escapes my attention, possibly because the current concepts about embryogenesis are insufficient for such a situation. There is the possibility that these interactions have no essential purpose in the specifica- tion of cells at present. The interactions may simply reflect the evolution of the cell specification machinery. It is possible that the C. elegans embryo is just evolving from one specification system involving many cell-cell interactions to another system involving an autonomous specification or vice versa. It may be possible to test this by analysing early embryogenesis in other nematode species. Recently Sommer and Sternberg (1994) suggested that vulva formation in two nematode species Mesorhabditis and Teratorhabditis occurs autonomously, which contrasts the nonautonomous specification of this structure in C. elegans. This suggests that identical structures may be specified by completely different strategies within a single phylum. Therefore, at some time organisms must have existed that were just evolving from a nonautonomous to an autonomous strategy or vice versa. An interesting question is how such a profound change in the basic mechanism may be achieved under the very strong constraint that a functional animal must be maintained during this evolutionary process. It is feasible that different pathways are maintained or partially maintained until the fidelity of the new pathway is sufficiently optimised.

I thank Lewis Wolpert for suggesting that the observed phenomena may reflect the ongoing evolution of cell specification in the embryo and Tom Cole for supplying the electron micrograph, Heinke Schnabel, Richard Feichtinger and Harald Hutter for helpful discus- sions and for critical comments on the manuscript, Thierry Bogaert, Titus Kaletta, Don Moerman and Thomas Wilm for critically reading the manuscript.

Austin
,
J.
and
Kimble
,
J.
(
1987
).
glp-1 is required in the germ line for regulation of the decision between mitosis and meiosis
.
Cell
51
,
589
599
.
Boveri
,
T.
(
1910
).
Die Potenzen der Ascaris-Blastomeren bei abgeänderter Furchung
.
Festschrift zum sechzigsten Geburtstag Richard Hertwigs
III
,
133
213
Bowerman
,
B.
,
Eaton
,
B. A.
and
Priess
,
J. R.
(
1992a
).
skn-1, a maternally expressed gene required to specify the fate of ventral blastomeres in the early C. elegans embryo
.
Cell
68
,
1061
1075
.
Bowerman
,
B.
,
Tax
,
F. E.
,
Thomas
,
J. H.
and
Priess
,
J. R.
(
1992b
).
Cell interactions involved in the development of the bilaterally symmetrical intestinal valve cells during embryogenesis in Caenorhabditis elegans
.
Development
116
,
1113
1122
.
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
.
Brenner
,
S.
(
1974
).
The genetics of Caenorhabditis elegans
.
Genetics
77
,
71
94
.
Cowan
,
A. E.
and
McIntosh
,
J. R.
(
1985
).
Mapping the distribution of differentiation potential for intestine, muscle, and hypodermis during early development in Caenorhabditis elegans
.
Cell
41
,
923
932
.
Edgar
,
L. G.
and
McGhee
,
J. D.
(
1986
).
Embryonic expression of a gut-specific esterase in C. elegans
.
Dev. Biol.
114
,
109
118
.
Goh
,
P.
and
Bogaert
,
T.
(
1991
).
Positioning and maintenance of embryonic body wall muscle attachment in C. elegans requires the mup-1 gene
.
Development
111
,
667
681
.
Goldstein
,
B.
(
1992
).
Induction of gut in Caenorhabditis elegans embryos
.
Nature
357
,
255
257
.
Goldstein
,
B.
(
1993
)
Establishment of gut fate in the E lineage of
Caenorhabditis elegans: the roles of lineage-dependent mechanisms and cell interactions. Development
118
,
1267
1277
.
Hird
,
S.
and
White
,
J. G.
(
1993
).
Cortical and cytoplasmic flow polarity in early embryonic cells of Caenorhabditis elegans
.
J. Cell Biol.
121
(
6
),
1343
1355
.
Hutter
,
H.
and
Schnabel
,
R.
(
1994
).
glp-1 and inductions establishing embryonic axes in Caenorhabditis elegans
.
Development
120
,
2051
2064
.
Hutter
,
H.
and
Schnabel
,
R.
(
1995
).
Specification of anterior-posterior differences within the AB lineage in the C. elegans embryo: a polarising induction
.
Development
121
,
1559
1568
Kemphues
,
K. J.
,
Priess
,
J. R.
,
Morton
,
D. G.
and
Cheng
,
N.
(
1988
)
Identification of genes required for cytoplasmic localization in early C. elegans embryos
.
Cell
52
,
311
320
Krieg
,
C.
,
Cole,
T.
Deppe,
U.
,
Schierenberg
,
E.
,
Schmitt
,
D.
,
Yoder
,
B.
and
von Ehrenstein
,
G.
(
1978
)
The cellular anatomy of embryos of the nematode Caenorhabditis elegans
.
Dev. Biol.
65
,
193
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
.
Mango
,
S. E.
,
Thorpe
,
C. J.
,
Martin
,
P. R.
,
Chamberlain
,
S. H.
and
Bowerman
,
B.
(
1994
).
Two maternal genes, apx-1 and pie-1, are required to distinguish the fates of equivalent blastomeres in the early Caenorhabditis elegans embryo
.
Development
120
,
2305
2315
.
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
176
.
Mello
,
C. C.
,
Draper
,
B. W.
and
Priess
,
J. R.
(
1994
).
The maternal genes apx-1 and glp-1 and establishment of dorsal-ventral polarity in the early C. elegans embryos
.
Cell
77
,
95
106
.
Moskowitz
,
I. P. G.
,
Gendreau
,
S. B.
and
Rothman
,
J. H.
(
1994
).
Combinatorial specification of blastomere identity by glp-1 dependent cellular interactions in the nematode Caenorhabditis elegans
.
Development
120
,
3325
3338
.
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
48
,
241
250
.
Schierenberg
,
E.
(
1987
).
Reversal of cellular polarity and early cell-cell interactions in the embryo of Caenorhabditis elegans
.
Dev. Biol.
122
,
452
463
.
Schierenberg
,
E.
(
1988
).
Localisation and segregation of lineage-specific cleavage potential in embryos of Caenorhabditis elegans
.
Roux’s Arch. Dev. Biol.
197
,
282
293
.
Schnabel
,
H.
and
Schnabel
,
R.
(
1990
).
Early determination in the C. elegans embryo: a gene, cib-1, required to specify a set of stem-cell-like blastomeres
.
Development
108
,
107
119
.
Schnabel
,
R.
(
1991
).
Cellular interactions involved in the determination of the early C. elegans embryo
.
Mech. Dev.
34
,
85
100
.
Schnabel
,
R.
(
1994
).
Autonomy and nonautonomy in cell fate specification of muscle in the Caenorhabditis elegans embryo: a reciprocal induction
.
Science
263
,
1449
1452
.
Slack
,
J.
M. W. (1991) From egg to embryo, regional specification in early development. In
Developmental and Cell Biologies 26
. (ed.
P.W.
Barlow
,
D.
Bray
,
P.B.
Green
and
J.M.W.
Slack
). pp.
1
328
.
Cambridge
:
Cambridge University Press
.
Sommer
,
R. J.
and
Sternberg
,
W.
(
1994
).
Changes of induction and competence during the evolution of vulva development in nematodes
.
Science
265
,
114
118
.
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
.
Wood
,
W. B.
(
1988
).
The Nematode Caenorhabditis elegans
.
New York
:
Cold Spring Harbor Laboratory Press
.
Wood
,
W. B.
(
1991
).
Evidence from reversal of handedness in C. elegans embryos for early cell interactions determining cell fates
.
Nature
349
,
536
538
.
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
563
.
Zur Strassen,
O.
(
1959
).
Neue Beiträge zur Entwicklungsmechanik der Nematoden
.
Zoologica
107
,
1
142
.