Requirement of cell division for muscle actin expression in the primary muscle cell lineage of ascidian embryos

An intriguing problem in developmental biology is how different genes are activated at specific times in restricted parts of the embryo. Cytoplasmic factors localized in the egg and segregated to different blastomeres have been postulated to control the spatial pattern of gene expression during embryogenesis (Morgan, 1934; Davidson, 1986). Muscle cell differentiation in ascidian embryos is a classic example of cell fate specification by cytoplasmic factors (reviewed by Jeffery, 1985). Early ascidian development is characterized by a series of invariant cleavages in which cytoplasmic regions of the egg are segregated to different embryonic cells. Eventually, a tadpole larva is formed containing about forty differentiated muscle cells in its locomotory tail. Most of the tail muscle cells are descendants of the B4.1 (vegetal-posterior) blastomeres of the 8-cell embryo (Conklin, 1905; Nishida & Satoh, 1983, 1985; Nishida, 1987) and are designated primary muscle cells. The primary muscle cells receive most of the myoplasm, a localized cytoplasmic region of the egg that is segregated differentially during cleavage.

Evidence that primary muscle cells are specified by cytoplasmic factors is derived from three kinds of experiments. First, when early embryos are arrested at various cleavages with cytochalasin and allowed to develop until controls reach the larval stage, muscle components appear only in cells that would normally form primary muscle cells (Whittaker, 1973; Satoh, 1979; Crowther & Whittaker, 1983; Nishikata et al.1988). Second, when presumptive muscle blastomeres are separated from the embryo and cultured in isolation, they express muscle cell components autonomously (Whittaker et al. 1977; Crowther & Whittaker, 1984; Deno et al. 1984, 1985; Nishikata et al. 1987). Third, when cytoplasm destined to enter B4.1 cells is introduced into other cells and these cells are subsequently arrested in cleavage, they sometimes express acetylcholinesterase (AChE) (Whittaker, 1982; Bates,1987), an enzyme normally made in muscle cells. These observations suggest that muscle-forming substances are segregated to the primary muscle cell lineage during cleavage and can alter cell fate when introduced into atypical regions of the embryo.

A few of the most caudal muscle cells in the larval tail do not arise from B4.1 cells; these are designated secondary muscle cells. Secondary muscle cells are descendants of the b4.2 (animal-posterior) and A4.1 (vegetal-anterior) blastomeres of the 8-cell embryo (Nishida & Satoh, 1983, 1985; Zalokar & Sardet, 1984; Nishida, 1987). Whether secondary muscle cell lineages are specified by egg cytoplasmic factors or inductive interactions is still unresolved (Deno et al. 1984, 1985; Meedel et al. 1987).

We have used muscle actin cDNA clones as probes to examine muscle cell development in embryos of the ascidian Styela (Tomlinson et al. 1987a). Styela embryos contain three major isoforms of actin, two cytoskeletal actins and one muscle-specific actin (Tomlinson et al. 1987b). The cytoskeletal actins and their mRNAs are stored in the egg, but muscle actin is synthesized during development, directed primarily by zygotic mRNA (Tomlinson et al. 1987a). Although low levels of muscle actin mRNA probably exist in the egg, these maternal transcripts disappear during the early cleavages and are replaced by zygotic mRNA beginning at gastrulation. From gastrulation to about mid-tailbud stage, zygotic muscle actin mRNA accumulates exclusively in primary muscle cells, but afterwards transcripts also appear in secondary muscle cells (Tomlinson et al. 1987a).

We have examined the expression of muscle actin in cleavage-arrested Styela embryos. Previous investigations have come to different conclusions concerning the role of cell division in the expression of muscle cell components in ascidian embryos. According to Whittaker and collaborators (Whittaker, 1973; Crowther & Whittaker, 1983, 1984, 1986), AChE and myofibril-like structures are produced in embryos arrested at the 1cell stage. In contrast, Nishikata et al. (1987) have shown that myosin heavy chain is not expressed when cleavage is arrested before the 8-cell stage. In the present investigation, evidence is presented that muscle actin expression requires at least three cell divisions, occurs exclusively in primary muscle lineage cells, and can be uncoupled from AChE development in cleavage-arrested (CA) Styela embryos.

The ascidians Styelaplicata and Styela clava were used in these experiments. S. plicata was purchased from Marinus Inc. (Long Beach, CA) and S. clava was collected in the vicinity of Woods Hole. MA. The animals were maintained as described previously (Tomlinson et al. WSla). Gametes were obtained from dissected gonads or by natural spawning (West & Lambert, 1975). Eggs were inseminated by mixing gametes from two or more individuals and embryos were cultured at 17°C as described previously (Jeffery et al. 1983). Under these conditions, first cleavage occurred at approximately 60 min. second cleavage at 80min. third cleavage at 105min. fourth cleavage at 120min. fifth cleavage at 150min and sixth cleavage at 180 min after insemination. Although Styela are usually self-sterile hermaphrodites, cultures derised from two or more individuals often contain low levels of self-fertilized embryos that are more advanced in development than artificially fertilized embryos.

Embryos were arrested at various cleavages by treatment with cytochalasin B or D (Sigma Chemical Corp.. St Louis. MO) dissolved in Millipore-filtered sea water (MFSW) at a final concentration of 2 – 6 μg ml^-1. Cytochalasin treatment blocks cytokinesis without affecting DNA synthesis or nuclear division in ascidian embryos (Whittaker, 1973). Cytochalasin-treated embryos were incubated at 17 °C in MFSW until controls reached the hatching stage (about 15 h).

To label proteins, controls and CA embryos were incubated with [³⁵S]methionine (1290 Ci mmole^-1; Amersham, Arlington Heights, IL) at a final concentration of 5 μ Ciml^-1. When controls reached the hatching stage, the embryos were washed three times in 15 ml of ice-cold extraction buffer (Tomlinson et al. 1987b), which consisted of 0 · 01 m-Tris-HCl, 0 · 0lm-NaCl, 0 · 001 M-MgCL (pH7 · 2) containing 10 μg ml^-’ each of phenylmethylsulphonyl fluoride and leupeptin. After washing, the samples were homogenized in 5 ml of ice-cold extraction buffer, centrifuged at 12000g, and stored at —70°C. Two-dimensional polyacrylamide gel electrophoresis (O’Farrell, 1975) was carried out as described by Tomlinson et al. (1987b). The gels were dried, exposed to Kodak XRP film for 4 – 8 weeks and autoradiographed.

In situ hybridization was conducted as described by Tomlinson et al. (1987a). Embryos were fixed in ice-cold, 3:1 95 % ethanol: acetic acid (or 5 % formalin-MFSW for in situ hybridization of AChE-stained embryos; see below) for 20 min, dehydrated by successive treatments with ice cold 70 %, 80 %, 90%, 95% and 100% ethanol (10 min each), treated successively with ice-cold 100% ethanol: 100% toluene (3:1, 1:1, 1:3; 10 min each) and cleared in 100% toluene for 10 min at 20°C. The cleared embryos were incubated successively with 100% toluene: 100% paraplast (3:1. 1:1, 1:3; 20min each) at 58°C and embedded in 100% paraplast. The embedded specimens were sectioned serially at 8 μn and sections were attached to glass slides subbed with gelatin. Before hybridization, the slides were treated with 1 μg ml^-1 proteinase K in 0 · 1 m-Tris-HCl. 0 · 0 · 5m-EDTA (pH 8 · 0) for 30min at 37°C. rinsed in 0 · 1 m-phosphate buffer (pH 7 · 2) for 5 min at 20°C, postfixed in 4% paraformaldehyde in phosphate buffer for 30min at 4 °C. treated with freshly prepared 0 · 25% acetic anhydride, 0 · 1 M-triethanolamine for 10min at 20°C, dehydrated through an ethanol series to 100% ethanol and airdried. About 1 × 10⁷cts min^-1 of tritiated muscle actin RNA probe was applied to each slide in 60 μl of hybridization buffer [2 × SSC, 50% formamide. 1 × Denhardt’s Solution (Denhardt, 1966). and 10% dextran sulphate]. The muscle actin RNA probe was transcribed in vitro from the Bluescribe subclone SpMA3C as described previously (Tomlinson et al. 1987a). SpMA3C is a 666-base HacIII fragment of the cDNA clone SpMA. which contains parts of the coding and 3’ noncoding regions of a S. plicata muscle actin mRNA (Tomlinson et al. 1987a). Hybridization was carried out for approximately 12 h at 45°C. The slides were washed in 250 ml of 4 × SSC for 15 min at 20°C. treated with 20 μg ml^-1 pancreatic RNase A in 0 · 01 m-Tris-HCl. 0 · 5.m-NaCl. 0 · 001 m-EDTA (pH 7 · 0) for 30min at 37°C and washed successively for 30 min each in 21 of 2 × SSC at 20°C. 1 × SSC at 45° C and 0 · 1 × SSC at 45°C. Previous studies have shown that SpMA3C detects muscle actin mRNA when sections are washed at this stringency (Tomlinson et al. 1987a). Controls for these experiments consisted of in situ hybridization of similar sections with a reverse orientation SpMA3C RNA probe prepared as described previously (Tomlinson et al. 1987a). These controls resulted in background levels of signal.

The washed slides were dehydrated to 95% ethanol, air dried and autoradiographed as described by Jeffery et al. (1983). The autoradiographs were stained through the emulsion with Harris hematoxylin and mounted with Permount for microscopic viewing and photography. Harris hematoxylin stains nuclei (or aggregations of nuclei in CA embryos) purple and the cytoplasm blue. The myoplasm often can be distinguished in hematoxylin-stained sections because it stains a lighter blue than the general cytoplasm.

Prior to assaying for AChE activity, embryos were fixed with 5 % formalin in MFSW for 20min on ice. AChE activity was determined by the method of Karnovsky & Roots (1964) as applied by Whittaker (1973). Preincubation of the specimens in 0-05 M-eserine sulphate, an AChE inhibitor, completely abolished AChE staining.

Muscle actin transcripts and AChE activity were assayed simultaneously by subjecting sections of histochemically stained embryos to in situ hybridization with the SpMA3C probe. Following AChE histochemistry conducted as described above, embryos were chilled and postfixed for 1 – 2 h in ice-cold 3:1 ethanol: acetic acid. The postfixed embryos were dehydrated, cleared and embedded; and sections were subjected to in situ hybridization as described above. After autoradiography, the sections were stained for 10 s with Harris hematoxylin and mounted. The cells that contained AChE stained reddish-brown, AChE negative cells stained light blue and cells containing muscle actin mRNA were identified by their accumulation of grains. To test the reliability of this method, AChE-stained positive and negative embryos were separated manually after histochemistry and subjected to in situ hybridization procedures. All embryos that were initially selected as being AChE positive had some reddish-brown stained cells in sections, whereas AChE negative embryos contained only blue-stained cells.

Presumptive muscle cells were identified by their location in sectioned embryos and the presence of large amounts of myoplasm. Myoplasmic pigment granules were identified by their staining characteristics in hematoxylin- or Milligan’s trichrome-stained sections. Milligan’s trichrome staining (Milligan, 1946; Tomlinson et al. 1987a) was conducted as follows. After treatment with 100% toluene to remove the paraplast, sections were rinsed successively in 100 % and 95 % ethanol, allowed to mordant in potassium dichromate-HCl for 2 min, and stained with acid fuchsin for 1 min. The stain was fixed in phosphomolybdic acid for 3 min. Subsequently, acid fuchsin-stained slides were counterstained with orange G for 5 min, with fast green for 10 – 15 min and washed with 1 % acetic acid for 3 min. Finally, the stained sections were dehydrated in 95 % and 100 % ethanol, cleared and mounted in Permount. Milligan’s trichrome stains the myoplasm of primary muscle lineage cells dark red; other types of cells are stained light pink or blue-green.

To examine the role of cell division in muscle actin expression, embryos were treated with cytochalasin at various times during early development. Cytochalasin treatment blocks cytokinesis, but nuclear division continues, resulting in embryos with multinucleate cells (Whittaker, 1973; Fig. 2A – G). The mean number of nuclei counted in sectioned 1-cell CA embryos was 136 ±18 (n = 10), suggesting that these embryos undergo at least seven nuclear divisions and accompanying rounds of DNA replication.

In an initial series of experiments [³⁵S]methionine was added to controls and CA embryos, the embryos were incubated Until controls reached the hatching stage, and labelled proteins were identified by twodimensional gel electrophoresis and autoradiography. As shown previously (Tomlinson et al. 1987b), three isoforms of actin are synthesized during early development; two of these are cytoskeletal actins and the third is a muscle actin (Fig. 1F). The cytoskeletal actins were synthesized in 1-, 2-, 4-, 8- and 16-cell CA embryos (Fig. 1A-E), indicating that there is no general effect of blocking cell division on actin synthesis. In contrast, muscle actin synthesis was not seen in 1- and 2-cell CA embryos (Fig. 1A,B), and occurred only at very low levels in 4-cell CA embryos (Fig. 1C). The first substantial labelling of muscle actin occurred in 8-cell CA embryos (Fig. 1D), and this increased to about control levels in 16-cell CA embryos (Fig. 1E). The results suggest that early cleavages are required for muscle actin synthesis.

Further information on the number of cleavages required for muscle actin expression was obtained by investigating mRNA accumulation in CA embryos. In this series of experiments, clutches of embryos from eight different animals were examined by in situ hybridization with a muscle actin cDNA probe. As shown previously (Tomlinson et al. 1987a), this probe detected transcripts confined to bands of tail muscle cells in sections of tailbud-stage control embryos (Fig. 2H). Each clutch of embryos arrested at the 1-, 2- and 4-cell stage behaved consistently with regard to accumulation of muscle actin mRNA. Of more than 4000 of these early CA embryos examined, none exhibited grains above background (Fig. 2A – C; Table 1). Since weak muscle actin synthesis was observed in 4-cell CA embryos by gel electrophoresis (Fig. 1C), either detection of mRNA by in situ hybridization was less sensitive than gel electrophoresis or the mass cultures of CA embryos were asynchronous. The latter explanation is likely because mass cultures of embryos are usually contaminated by low levels of more advanced embryos, the products of self-fertilization (see Materials and methods). In five clutches of embryos, the first muscle actin mRNA accumulated in 16-cell CA embryos, while in three clutches mRNA accumulation was detected in 8-cell CA embryos (Table 1). Thus, the number of cleavages required to begin muscle actin mRNA ac-cumulation varies from three to four in CA embryos from different individuals. To be certain that the transcripts observed in CA embryos represent zygotic muscle actin mRNA, in situ hybridization was conducted on CA embryos processed at various times before and during cytochalasin treatment. These embryos initially showed detectable levels of maternal muscle actin mRNA, but this signal disappeared during the drug incubation period and was replaced (in embryos arrested after the third cleavage or later) beginning when controls gastrulated (data not shown), indicating that zygotic muscle actin transcripts were being detected in CA embryos. In summary, the results suggest that at least three cell divisions are required for the expression of muscle actin.

Since muscle actin is expressed differentially in primary and secondary muscle lineages of normal embryos (Tomlinson et al. 1987a), it was of interest to identify the cells that accumulate muscle actin mRNA in CA embryos. In previous cleavage-arrest experiments, the maximal number of blastomeres that expressed muscle components was equal to the number of cells belonging to the primary muscle lineage at each stage of development, i.e. two cells in the 2-, 4-, and 8-cell embryo, four cells in the 16-cell embryo, six cells in the 32-cell embryo, and eight cells in the 64-cell embryo (Whittaker, 1973; Crowther & Whittaker, 1983, 1984; Nishikata et al. 1987). More recently, AChE activity has also been detected in a third cell (presumably a precursor of a secondary muscle cell) in Ascidia embryos arrested at the 8-cell stage (Meedel et al. 1987). Since in situ hybridization was used to determine the spatial expression of muscle actin in the present investigation, it was necessary to determine the total number of labelled cells by examining serial sections of CA embryos. Fig. 3 shows alternate serial sections of an embryo that was arrested at the 32-cell stage. By following labelled cells through these sections, it can be seen that a total of four cells have accumulated muscle actin mRNA in this embryo. Table 2 summarizes the results of similar in situ hybridization experiments conducted with serial sections of 8-, 16-, 32- and 64-cell CA embryos. In each case, the number of cells that accumulated muscle actin mRNA was no greater than the maximal number of primary muscle lineage cells. Further experiments were conducted to identify the labelled cells in CA embryos. In serially sectioned embryos, labelled cells sometimes could be identified as primary muscle cell precursors by their position within the embryo. These cells could also be identified by their enrichment in myoplasm, either in the same hematoxylin-stained sections or in alternate serial sections stained with Milligan’s trichrome. The results obtained by hematoxylin and Milligan’s trichrome staining are shown in Table 3. Although all primary muscle cells did not express muscle actin mRNA, each labelled cell was a member of the primary muscle lineage. The results imply that secondary muscle lineage cells do not express muscle actin in embryos arrested up to the 64-cell stage.

The myoplasm is a unique cytoplasmic domain containing pigment granules, cytoskeletal filaments (Jeffery & Meier, 1983) and maternal actin mRNA (Jeffery, 1984). After being segregated to the B4.1 cells during cleavage, myopl asmic pigment granules subsequently leave the periphery and enter the perinuclear region of presumptive muscle cells (Conklin, 1905). Fig. 4 shows the intracellular distribution of muscle actin mRNA in primary muscle cells of 16-cell CA embryos. Muscle actin mRNA did not accumulate uniformly in primary muscle cells; it was localized in the perinuclear region (Fig. 4; also see Fig. 2D), the same area in which myoplasmic pigment granules were detected by Milligan’s trichrome staining. Occasionally, grains accumulated on one side of the nucleus where there was also a concentration of myoplasmic pigment granules (Fig. 4B). The localization of muscle actin mRNA in the perinuclear region was less evident in embryos arrested at later cleavages when myoplasm fills the cytoplasm of primary muscle cells (Fig. 2F,G). These results suggest that zygotic muscle actin transcripts tend to accumulate in the myoplasm of CA embryos and suggest that this region may be involved in the organization of muscle cells.

When CA embryos were assayed for AChE activity, the results were different from those obtained for muscle actin expression (Fig. 2; Table 1). Although less than 10% of 1- and 2-cell CA embryos expressed AChE, about 50 % of 4-cell CA embryos developed AChE, and this value increased to about 80 % in 8- and 16-cell CA embryos. These results are similar to those obtained for cleavage-arrested Ciona embryos (Whittaker, 1973). To determine whether AChE can develop in the absence of muscle actin expression, embryos arrested at the 16-cell stage were subjected to AChE histochemistry; the AChE-positive and -negative embryos were separated manually, fixed and processed for in situ hybridization to determine whether they accumulated muscle actin mRNA. The results show that none of the AChE-negative embryos accumulated muscle actin mRNA, but mRNA was detected in some of the AChE-positive embryos (Table 4). Thus, AChE can be expressed in embryos that do not express muscle actin; however, this experiment does not indicate whether muscle actin and AChE are always expressed in the same cells of a single embryo. To determine whether AChE and muscle actin mRNA are expressed in the same cells, CA embryos were subjected to AChE histochemistry; subsequently, these embryos were postfixed, sectioned, subjected to in situ hybridization and stained with hematoxylin as described in Materials and methods. AChE-positive cells stained reddish-brown, AChE negative cells stained blue, and cells containing muscle actin mRNA were identified by labelling. As indicated in Table 5, many cells that express muscle actin also express AChE. And, as expected from earlier experiments (Table 4), some AChE-positive cells did not accumulate muscle actin mRNA. However, some cells that accumulated muscle actin mRNA did not show AChE activity (Table 5). The results suggest that muscle actin and AChE expression can be uncoupled in some cells of CA embryos.

In ascidian embryos, differential gene activity is manifested by the appearance of specific mRNAs, proteins and supramolecular structures in different cell lineages. The components that appear in the muscle cell lineage include AChE (Whittaker, 1973), myosin (Meedel, 1983; Nishikata et al. 1987), muscle actin (Tomlinson et al. 1987a,b), and myofibril-like structures (Crowther & Whittaker, 1983; 1986). Since there is evidence for segregation of factors that specify some of these components during cleavage (Whittaker, 1973), it was of interest to determine whether cell division is required for differential gene activity in the muscle cell lineage. In the present investigation, we have used two-dimensional gel electrophoresis, in situ hybridization with a cDNA probe, and enzyme histochemistry to examine the role of cell division in muscle actin expression and its relationship to AChE development in cleavage-arrested Styela embryos. As reported previously (Whittaker, 1973; Crowther & Whittaker, 1986), AChE development does not appear to require early cleavages. In contrast, the present studies suggest that at least three cleavages are required for muscle actin expression.

During normal development, muscle actin expression begins between the 64-cell stage and the beginning of gastrulation (Tomlinson et al. 1987a), but AChE is first detected between the late gastrula and neurula stages (Perry & Melton, 1983; Meedel & Whittaker, 1983; Meedel et al. 1987). Assuming AChE and muscle actin genes are activated concomitantly at the early gastrula stage, when zygotic transcription is thought to begin in ascidian embryos (Meedel & Whittaker, 1978), the methods used to detect muscle actin appear to be more sensitive than those used to assay AChE. Therefore, if muscle actin is expressed in early CA embryos, it would be expected to appear at least in those embryos that develop AChE. When more than 4000 different 1-, 2- and 4-cell CA embryos were examined by in situ hybridization, however, none were found to accumulate muscle actin mRNA. Furthermore, substantial muscle actin synthesis was not observed by 2D-gel electrophoresis before third cleavage. These results suggest that early and late CA embryos differ qualitatively in their ability to express muscle actin.

The inability of early CA embryos to express muscle actin is consistent with recent studies showing that several cleavages are necessary for myosin heavy chain expression in the muscle cell lineage (Nishikata et al. 1987), but conflicts with reports that myofibril-like structures develop in 1- and 4-cell CA embryos (Crowther & Whittaker, 1983, 1986). Despite the similarity in appearance of the reported structures to myofibrils in electron micrographs, there is no evidence that they actually contain muscle actin. Further studies involving more specific biochemical analysis of these assemblages are necessary before the actin species they contain can be identified with certainty. Consistent with our results, early cleavages are also required for the development of melanized inclusions in the brain sensory cell lineage of ascidian embryos (Whittaker, 1973), cilia in the trophoblast cell lineage of the mollusc Patella (Janssen-Dommerholt et al. 1983) and rhabditin granules and esterase activity in the endodermal cell lineage and paramyosin in the muscle cell lineage of Caenor-habditis elegans (Cowan & Macintosh, 1984; Edgar & McGhee, 1986). Thus, many cell lineage-specific components require cell division for differential gene expression during embryogenesis.

Why does blocking cell division affect muscle actin expression? Although a specific number of DNA replication cycles is required for expression of cell lineagespecific components in ascidian and other embryos (Satoh & Ikegami, 1981; Satoh, 1982; Edgar & McGhee, 1988), DNA synthesis does not appear to be limiting in CA embryos. It is clear that 1-cell CA embryos contain enough nuclei to have accomplished at least seven rounds of DNA synthesis. Also excluded is the possibility that inhibitors of specific gene expression that could be associated with other cell lineages are segregated away from presumptive muscle cells after a particular cleavage. When CA embryos activate muscle actin transcription at the 8- or 16-cell stage, the primary muscle cell lineage still includes other derivatives; for instance, some descendants of B4.1, B5.1 and B5.2 cells produce mesenchyme, endodermal and notochord cells (Nishida & Satoh, 1985; Nishida, 1987). A possible explanation for these results is that muscle actin expression requires the attainment of a critical ratio of muscle-determining factors to general cytoplasmic volume in the presumptive muscle cells. This ratio would increase progressively as these factors are passed exclusively to these cells during cleavage. The difference in onset of muscle actin expression in embryos obtained from different individuals may be related to variations in the initial titer of muscle-determining factors present in the egg and segregated to the muscle cell lineage during cleavage. Thus, CA embryos derived from eggs containing high titers of these factors could reach a threshold level after only three cleavages, while embryos derived from eggs with low titers may not attain this level until after fourth cleavage.

Although secondary muscle cells express muscle actin during normal development (Tomlinson et al. 1987a), they do not accumulate detectable amounts of muscle actin mRNA in CA embryos arrested up to the 64-cell stage. If the idea that muscle actin expression is triggered by the attainment of a critical ratio of muscle-determining factors to general cytoplasmic volume is correct and secondary muscle cells receive a lesser number of these factors from the egg, more cell divisions would be required for expression in presumptive secondary muscle cells than in primary muscle cells. Secondary muscle lineage cells could also require more cleavages to promote muscle actin expression because they are specified by inductive processes rather than by the inheritance of cytoplasmic factors from the egg (Deno et al. 1984, 1985; Meedel et al. 1987). Thus, cell division could be required to develop inductive activity in other embryonic cells or the competence to respond to induction in presumptive secondary muscle cells.

All cells that can express a given tissue-specific component according to their cell lineage do not always do so in CA embryos (Whittaker, 1973). Although the basis for this phenomenon is not understood, it permits us to determine whether two markers are always expressed coordinately in the same cell lineage. Two lines of evidence suggest that muscle actin and AChE can be uncoupled in the primary muscle cell lineage of CA embryos. First, since muscle actin but not AChE expression appears to require cell division, many early CA embryos express AChE, although they do not accumulate muscle actin mRNA. Second, muscle actin and AChE sometimes are expressed in separate cells of a single CA embryo. There are several possible explanations for these results, but the simplest is that muscle actin and AChE expression are controlled by different regulatory factors. In light of the fact that AChE activity develops in muscle, mesenchyme (Meedel & Whittaker, 1979) and possibly neural (Durante, 1959) and epithelial (Minganti & Falugi, 1980) cells, whereas muscle actin expression is limited to muscle cells (Tomlinson et al. 1987a), the existence of separate regulatory pathways for AChE and muscle actin may ensure that the latter is expressed only in muscle cell lineages.

An intriguing result of this study was the observation that muscle actin transcripts accumulate primarily in the myoplasm of cleavage-arrested embryos. Previous studies have shown that the myoplasm is a specific cytoskeletal domain, containing mitochondria, pigment granules and at least two different classes of cytoskeletal filaments (Jeffery & Meier, 1983). Maternal mRNA is localized in the myoplasm, possibly by its interaction with the cytoskeleton (Jeffery, 1984). This is the first suggestion that zygotic mRNA is associated with the myoplasm, and it will be interesting to determine whether this phenomenon is also based on an association of muscle actin mRNA with cytoskeletal elements. Although it cannot be ruled out that zygotic muscle actin localization occurs only in CA embryos, it is possible that this association promotes spatial coordination of translation and assembly of myofibrillar elements in differentiating muscle cells. These results suggest that the myoplasm may have an important organizational role in muscle cell differentiation.

This research was supported by grants from the NIH (HD-13970) and NSF (DCB-84116763 and DCB-8812110). I thank P. Kemp for technical assistance and Dr B. J. Swalla for critical reading of the manuscript.