Main

Several experimental systems have been used to study early embryonic muscle development. Skeletal muscle development and gene expression have been extensively analyzed in vitro because appropriate differentiated cell lines are available(1–3). On the other hand, no permanent cell lines of beating cardiomyocytes are available, and studies of developing heart cells have made use of primary cultures of precardiac areas of blastoderm(4, 5), organ cultured hearts(6, 7), and dissociated heart cell cultures of embryonic or neonatal origin(8, 9). Likewise, the in vitro analysis of smooth muscle development has relied heavily on primary cultures of cells from vascular(10) or gastrointestinal(11) sources. Cultured embryonal carcinoma cells also have been useful for the analysis of smooth muscle and cardiac muscle cell differentiation and gene expression(12–14).

ES cell lines have been established from the inner cell mass of mouse blastocysts(15, 16). They can be maintained in the undifferentiated state by growing them on a feeder layer of primary embryonic fibroblasts(17, 18), or on the embryonic fibroblastic cell line (STO)(15, 16). Treating the cultures with differentiation inhibiting activity/leukemia inhibitory factor also maintains the cells in an undifferentiated state(19, 20). Upon removal of the inhibiting factor or feeder cells, ES cells undergo differentiation into EBs of increasing complexity in which early aspects of embryogenesis including angiogenesis and the development of the visceral yolk sac, blood islands, and myocardium can occur(17, 21).

We have used reverse transcriptase-PCR, transcript-specific probes for Northern analysis of mRNA levels, and selectively reactive antibodies to characterize the pattern of expression of the muscle isoactins in ES cells and EBs at various stages of differentiation. Although four distinct actins comprise the major isoforms found in cardiac, skeletal, enteric, and vascular muscle of adult vertebrates, most of these tissues contain more than one isoactin at some developmental stage(22–27). Each of these isoactins is expressed in a unique temporal and tissue-specific pattern that is conserved across species, strongly suggesting a functional significance for this isoform multiplicity(28). Here, we demonstrate that the differentiated EBs transcribe and translate all muscle isoactins in a manner similar to the early stages of developing mouse embryos. Our data extend previous observations in this system(17, 29–33) and support the notion that the formation of EBs from ES cells can be used to investigate gene function and regulation involved in both striated and smooth muscle development.

METHODS

ES Cells. The mouse blastocyst cell line, ES-D3 derived from the 129 Sv+ strain(17), was maintained morphologically undifferentiated by culture on a layer of mitomycin C-inactivated mouse embryonic fibroblasts. The cells were grown in high glucose Dulbecco's modified Eagle's medium (Whittaker Bioproducts), supplemented with 15% heat-inactivated FCS (Flow Laboratories Inc.) and 0.1 mM 2-mercaptoethanol (Sigma Chemical Co.). Cells were grown without antibiotics at 37°C in 10% CO2. In vitro differentiation was induced by removal of the feeder layer and growth in suspension culture in bacterial Petri dishes containing Dulbecco's modified Eagle's medium, 15% FCS for the first 4 d and 20% serum thereafter. Medium was changed or added every day. When the time course of expression of isoactin mRNAs in EBs was examined, RNA was prepared for Northern analysis from the contents of an entire dish. In some experiments, individual EBs were used for PCR analysis.

Oligonucleotides. Specific oligonucleotide probes for theα-cardiac, α-skeletal, γ-enteric, and α-vascular actins were synthesized (Applied Biosystems Inc.) by using unique sequences from the 3′-UT regions(43, 52–55). These probes are: α-cardiac actin, 5′-GGCTAAGAGAGAGACATCTC-3′;α-skeletal actin, 5′-CGAGTCAATCTATGTACACG-3′;α-vascular actin, 5′-CACAGTTGTGTGCTAGAGGCAG-3′; andγ-enteric actin, 5′-AAGGCTTGAAGGTTTTAATGATCTGTGGCTGGTGACCAAGTCTTGTGGGGATCCCAGAGAAGG-3′. For PCR analysis, an oligonucleotide comprising a sequence within the coding region was used as the 5′-actin primer: 5′-GCTCCCAGCACCATGAAGATCAAG-3′. This sequence encodes amino acids 320-327 and is shared by all four of the mouse muscle actins(43, 54) except for a single base pair difference(T → A) in amino acid 320 of the α-vascular isoform. To confirm that the expected products were generated, an internal oligonucleotide primer, located between the 5′-primer and the 3′-UT primer, corresponding to residues 336-346, was used to detect any DNA containing this sequence: 5′-TACAAGCTTGGATTGGCGGCTCCATCCTGGC-3′. This sequence is also conserved among the muscle actins.

RNA isolation and analysis. Total cellular RNA was isolated from mouse embryonic fibroblasts, undifferentiated ES cells, ES cell-derived EBs, and CD-1 mouse tissues using RNazol (Cinna Scientific) as described previously(56). For Northern blot analysis, 20 μg of RNA were electrophoresed through a 1.2% agarose, 6.6% formaldehyde gel, and blotted onto a Nytran membrane (Schleicher & Schuell). After transfer, the filters were baked at 80°C for 2 h under vacuum. Each of the four oligonucleotide probes was hybridized at 45°C for 12 h in 6 × saline sodium citrate (SSC) (1 × SSC = 0.15 M sodium chloride, 0.015 M sodium citrate), 1% SDS, 100 mg/mL denatured sonicated salmon sperm DNA, and 5× Denhardt's solution. The filters were washed twice for 5 min at room temperature and then twice for 30 min at 45°C in 2 × SSC. The autoradiographs were exposed at -70°C with Kodak X-Omat film using Cronex Lighting Plus screens (Dupont) and developed after 5 d. In all cases, the probe was labeled by using [γ-32P]ATP with T4 polynucleotide kinase and the concentration of the probe was 3 ng/mL.

PCR analysis. Unless otherwise noted, 1 μg of total cellular RNA was used as template for each PCR amplification. When single EBs were used, tRNA was included during the RNazol extraction, and the RNA obtained was divided equally into four separate reactions to detect the four muscle actin mRNAs. Negative control reactions containing 1 μg of yeast tRNA were included in each experiment.

The first strand cDNA was synthesized using the specific 3′-UT region primer for each muscle isoactin. The RNA sample was added to the reaction buffer (100 mM Tris-Cl, pH 9, 40 mM ammonium sulfate, 3 mM MgCl2), 25 pmol of the appropriate primer, 1 mM each dATP, dCTP, dGTP, and dTTP, and 2.5 U of avain myeloblastosis virus reverse transcriptase (Life Technologies, Inc.) in a final volume of 20 μL. The primer was allowed to anneal to the nucleic acid at 21°C for 10 min, and then incubated at 42°C for 45 min. The reaction was stopped by incubation at 95°C for 10 min, after which the volume was brought up to 50 μL with reaction buffer containing 25 pmol of 5′-actin primer and 2.5 U of AmpliTaq polymerase (Perkin-Elmer Co.). The reaction was overlaid with mineral oil, and amplification was carried out in an Ericomp Inc. temperature cycler. The sequence of the cycles was: initial denaturation, 3 min at 94°C, followed by 32 cycles of amplification (denaturation, 1 min at 94°C; annealing, 30 s at 60°C; elongation, 1 min at 72°C), and a final extension, 10 min at 72°C. For the gene amplification experiment, 1 ng of mouse genomic DNA was added to the same reaction buffer, with 25 pmol of each pair of primers, 1 mM of each dNTP, and 2.5 U of AmpliTaq in a final volume of 50 μL. The cycle sequence differed only in the annealing temperature (55°C).

Eight microliters of each reaction were analyzed by electrophoresis in a 2% agarose gel. Subsequently, for Southern blot analysis the products were blotted onto a nylon membrane and baked at 80°C for 2 h. The internal oligonucleotide was hybridized in 6 × SSC, 1% SDS, 1 × BLOTTO, and 200 mg/mL denatured salmon sperm at 45°C for 12 h. The probe concentration was 3 ng/mL. The filter was washed in 2 × SSC three times for 5 min at room temperature and for 1 h at 45°C.

Immunofluorescence. EBs were grown on 1,3-aminopropyl-triethoxysilane (Sigma Chemical Co.)-coated glass slides in Dulbecco's modified Eagle's medium supplemented with 15 and 20% fetal bovine serum as described above. The EBs attach to the slides and undergo differentiation similar to that in suspension culture, including the appearance of areas that rhythmically contract. After 9 and 13 d in culture, slides were rinsed in PBS and treated with methanol at -10°C for 7 min. After air drying, the slides were stored at -70°C until they were rehydrated in PBS for use.

The specificity of the antibodies have been described previously. MAb HUC 1-1 has been shown to be specific for all muscle isoactins(27) and MAb B4 to have a preferential reactivity with the γ-enteric isoform(27, 57). MAbs 1A4 and 5C5 (Sigma Chemical Co.) were used to detect vascular and striated isoactins, respectively(49, 50). P3X63, the“nonimmune” MOPC21 IgG from the mouse P3 myeloma, was used as a primary antibody to control for nonspecific interactions. Goat anti-mouse IgG, rhodamine-conjugated (Boehringer Mannheim), was used at a dilution of 1:50. Dilutions of both the MAb and the secondary antibodies were made in PBS solution (Whittaker Bioproducts) containing 1 mg/mL BSA (Sigma Chemical Co.). Both the primary and secondary antibodies were incubated with the cells for 90 min in a humidified chamber at room temperature, followed by three washes in PBS. The slides were mounted in Bacto FA mounting fluid (Difco Laboratories). A Zeiss ICM 405 microscope (Carl Zeiss Inc.) equipped for epifluorescence was used to view and photograph the specimens. Photographs were taken using Kodak Technical Pan film 2415. The film was developed with HC 110.

RESULTS AND DISCUSSION

ES cell differentiation. ES cell derived EBs have been shown to differentiate into a variety of cell types, identified as cardiac, skeletal, and smooth muscle by light and electron microscopy(17). In our hands, about one third of the EBs growing in suspension or attached to glass slides for 8 d of culture in differentiation conditions exhibit areas which rhythmically contract at rates that vary between 13 and 78 beats/min. Multiple areas with contractions occasionally were found in the same EB, often with independent rates of contraction. Using the patch-clamp technique, Maltsev et al.(34) characterized these areas into different phenotypes corresponding to sinonodal, atrial, and ventricular cells. Interestingly, the beating areas also showed variable responses when exposed to a β-adrenergic drug such as isoproterenol. As shown by others(35), real time video display (corroborated by at least two investigators) showed that the rate of contraction in these regions generally increased. A positive chronotropic response of as much as 100% was observed in our study; however, certain beating areas were unaffected.

PCR analysis of EBs. Reverse transcriptase-PCRs were used to detect mRNA transcripts in the small number of muscle cells found in individual EBs. We synthesized an oligonucleotide within the coding region(5′-actin primer, see Fig. 1) that when paired with each specific muscle isoactin oligonucleotide will generate a unique size fragment when the RNA is amplified. Moreover, the gene for each muscle actin generates a larger fragment due to the inclusion of intronic sequences. Thus, the products resulting from amplification of contaminating DNA or primary RNA can be easily distinguished from the product derived from the mRNA. The results of the PCR amplification of α-cardiac, α-skeletal,γ-enteric, and α-vascular actin mRNA yields products of 196, 315, 242, and 205 bp, respectively, whereas the predicted gene products obtained with the same primers using DNA are ≈700, 389, 3404, and ≈1400 bp, respectively. The locations of the primers in each gene and the resulting amplification products are shown in Fig. 1. All of the predicted amplified products from the actin genes were obtained except that for the enteric actin gene, presumably because of the large size of intron 7 in this gene(36). As described under“Methods,” the PCR amplification products corresponding to the skeletal, cardiac, and vascular actin genes were shown to be derived from a muscle actin gene by Southern blotting (not shown) using an oligonucleotide corresponding to a sequence located between the PCR primers that were used for each actin.

Figure 1
figure 1

PCR analysis of muscle actin mRNA and genes. Pairs of oligonucleotides were used to amplify either RNA from heart, skeletal muscle, stomach, and aorta, respectively, or mouse genomic DNA. PCR yields a product of the predicted size for each of the four muscle actin mRNAs which can be detected by staining with ethidium bromide. Likewise, when genomic DNA is carried through this procedure, larger fragments are generated with the oligonucleotide pairs for the skeletal, cardiac, and vascular actins due the inclusion of intronic sequences. On the other hand, the enteric actin genomic fragment is predicted to be about 3.4 kb in length(36), and this size would not be expected to be synthesized or detected under these conditions. The marker lanes contain HaeIII-digested φX174 DNA.

Full size image

Pooled RNA samples from the feeder cells and EBs cultured for different lengths of time in the absence of feeder cells (d 0, 5, 10, 15, and 20) were examined by PCR analysis to look for each of the four muscle actin transcripts (Fig. 2). This very sensitive method detectedα-skeletal, γ-enteric, and α-vascular transcripts in the“embryonic fibroblasts” that were used as feeder cells. This most likely reflects the presence of a wide variety cell types in this preparation, because the entire torsos of d 16 mouse embryos (minus the heart and liver) were used to prepare these cells. Similarly, the lack of cardiac actin expression in the feeder cells is probably due to the fact that the heart is not included in the cultured cells. On the other hand, the α-vascular actin mRNA was clearly present in the ES cultures at d 0 by PCR analysis. This may reflect a low level of expression in undifferentiated ES cells, but our immunofluorescence results do not detect any staining for vascular smooth muscle actin. Thus, this signal may simply represent low levels of contamination by feeder cells. However, because only traces of theγ-enteric and α-skeletal actin PCR products were obtained at d 0, the ES cells and embryoid bodies clearly are not grossly contaminated by the embryonic fibroblasts containing vascular actin. By 5 d of culture in the absence of feeder cells, all four muscle actin mRNAs were easily detected by PCR analysis, although only very low levels of the PCR product corresponding to the enteric actin mRNA were generated.

Figure 2
figure 2

PCR analysis of muscle actin transcripts in ES cells and EBs. The same RNA prepared for the Northern analyses (Fig. 4). from ES cells (d 0) or embryoid bodies after 5, 10, 15, and 20 d in suspension culture was amplified by PCR for 32 cycles. Control (CT) RNA was prepared from from heart, leg skeletal muscle, stomach, or aorta to demonstrate the adequacy and specificity of the oligos used to detect the cardiac, skeletal, enteric, or vascular actin PCR products, respectively. The PCR products were visualized by staining with ethidium bromide.

Full size image

Individual EBs from d 1-6 were collected, and each one was subjected to RNA amplification by PCR to determine the early time course for the appearance of the four muscle isoactin-transcripts upon induction of differentiation (Fig. 3). In this experiment, the transcript forα-vascular actin is readily detected on the 1st d of differentiation, whereas the mRNAs for α-cardiac and α-skeletal actin show a much lower signal, and that of enteric actin cannot be detected. Although the relative amount of the PCR product for each of the four muscle actin mRNAs varies with time of culture, the significance of these variations is not clear, because no effort was made to generate quantitative PCR results. In fact, it should be noted that the variations that are observed in the PCR products for any of the actins do not reflect the signals that were obtained using Northern blot analysis (Fig. 3). Thus, the results of our PCR analysis can be regarded only as qualitative or semiquantitative, at best. Nevertheless, because equal amounts of RNA were used for each reaction at a given time point, a stronger signal within a given set of reactions for each individual actin suggests a higher accumulation of the transcripts for this isoform. The PCR product derived from enteric actin mRNA is detectable at d 3 and increases through d 5. The cardiac actin reaction product is present at low levels through d 4 of culture with increasing levels at d 4 and 6. The skeletal actin reaction product shows a gradual increase in intensity with increasing time in culture. Again, the vascular actin PCR product is present from d 1 through d 6. The general increase observed for all four PCR products representing the four muscle actins is probably due to a combination of an increase in the number of cells producing these mRNAs along with the accumulation of specific transcripts within those cells.

Figure 3
figure 3

PCR detection of the muscle actin mRNAs in individual EBs. Amplification was performed using the RNA present in individual EBs that were harvested after 1-6 d in suspension culture. The total RNA derived from each EB was divided into four aliquots, and each portion was amplified using an oligo pair for one of the muscle isoactins. One-sixth of each reaction was electrophoresed on a 2% agarose gel. The PCR products were visualized by staining with ethidium bromide. CT = controls as in Fig. 2.

Full size image

Preparation of specific probes for the isoactin mRNAs and Northern blot analysis. To assess the level of transcripts encoding each of the muscle actins in the ES cell system, we selected unique domains within the 3′-UT region of each isoform to specifically detect α-cardiac,α-skeletal, γ-enteric, and α-vascular actin mRNAs. The degree of conservation of the coding sequences among the muscle isoactins is very high, whereas the 3′-UT regions have been shown to differ significantly among these mRNAs(37–39).

To demonstrate transcript distribution and specificity, each oligonucleotide was hybridized to total cellular RNA isolated from mouse tissues that had been blotted onto a nylon membrane (Fig 4, upper). The data show that each oligonucleotide has the predicted tissue specificity and reacts with mRNA of the appropriate size(40). That is, each probe detected the expected mRNA in the corresponding mature muscle; in addition, lower levels of coexpression in other tissues were noted as observed previously using tissues from the adult rat(40). As expected, α-cardiac actin mRNA was the prevalent transcript in the heart, with low levels of expression seen in skeletal muscle and aorta. The α-skeletal actin mRNA was the predominant transcript found in skeletal muscle, and was also detected in the heart, aorta, and stomach. The γ-enteric actin mRNA was the most abundant transcript present in the stomach with low levels in the aorta. Theα-vascular actin mRNA is abundant in stomach and readily detectable in the aorta. On the other hand, only trace amounts of α-vascular actin mRNA appear to be present in the heart and no mRNA for this actin is detected in the skeletal muscle mRNA.

Figure 4
figure 4

Northern blot analysis using transcript-specific probes. RNA (20 μg) from mouse embryonic fibroblasts (MEF), ES cells (d 0), EBs after 5, 10, 15, and 20 d in suspension culture or RNA from aorta (AO), heart (HT), leg skeletal muscle(SK), and stomach (ST) was fractionated on a 1.2% agarose/formaldehyde gel, blotted onto a Nytran membrane, and hybridized to the labeled oligonucleotide probe as indicated. The gels were stained with ethidium bromide to confirm uniform loading of the RNA (not shown). The autoradiograms were developed after 3-5-d exposure.

Full size image

The presence of the muscle actin transcripts was assessed in ES cells and in EBs at various stages of differentiation by Northern analysis. This method is clearly much less sensitive than reverse transcriptase-PCR but provides a more accurate indication of the relative abundance of each muscle actin mRNA present in these samples. For example, in the embryonic fibroblasts, which serve as feeder cells for the ES cell cultures, a significant amount ofα-vascular actin mRNA and a trace amount of enteric actin mRNA were detected by Northern analysis. In contrast, similar levels of skeletal, vascular, and enteric actin PCR products were observed (Fig. 2). However, in the undifferentiated ES cells, no muscle actin transcripts were detected by Northern analysis (not shown). Again, this is probably due to the fact that the ES cells comprise at least 90% of the total cells in the culture dish and is consistent with the interpretation that the vascular actin PCR product probably reflects the presence of a small number of“embryonic fibroblasts” in the ES cells. After 5 d the EBs display very low to undetectable levels of the muscle actin mRNAs on Northern blots. By d 10, there were very significant amounts of α-vascular transcripts, a low level of γ-enteric, and a very low level of both theα-cardiac and α-skeletal actin mRNAs. Only the α-vascular and γ-enteric actin mRNAs were detected with this method at d 15-20. Overall, these data suggest that α-vascular actin is the predominant muscle isoactin transcribed during EB differentiation under these conditions and that α-cardiac, α-skeletal, and γ-enteric actin mRNAs are transiently expressed.

In terms of smooth muscle differentiation, the PCR and Northern data show a slow rate of accumulation for the γ-enteric actin transcripts, whereasα-vascular actin mRNA is present at all time points with a maximum level in EBs at about 10 d of culture. The slow accumulation of the enteric actin message relative to the vascular smooth muscle actin message in individual EBs is consistent with earlier studies(40) showing that the vascular isoform appears before the enteric form in the smooth muscles of the developing fetus(29). Moreover, it is important to note that, although α-vascular actin is expressed in a variety of cell types(myofibroblasts, pericytes, and embryonic fibroblasts)(41, 42) in addition to smooth muscle, the expression of γ-enteric actin appears to be restricted to only smooth muscle(27, 40) and postmeiotic spermatocytes(43). Because the latter cells are clearly not present, the detection of this mRNA strongly argues that smooth muscle cells are indeed present in EBs as previously suggested from light and electron microscopic examination(17).

It is interesting to note that the sequential accumulation of muscle actin mRNAs in embryoid bodies parallels the normal events of early striated muscle development(25). Also, these findings are in accordance with previous work demonstrating that the first muscle cells to appear in the embryo are in the myocardium and the somitic myotome(23, 44). In the mouse embryo the cardiac compartment is formed between 7.5 and 8 d post coitum(45) and these early cardiomyocytes express both striated actin isoforms at high levels(44) withα-cardiac actin mRNA being the main form expressed throughout development. In the 18-somite mouse embryo (9.25 d post coitum) bothα-cardiac actin transcripts and the corresponding protein are abundant in cardiac myocytes. However, in the myotomes, the levels of the striated actin mRNAs are very low but detectable by in situ hybridization techniques(46) at this time. According to Lyons et al.(46), the accumulation of α-actin mRNA and protein appears to occur before that of myosin heavy chain mRNA and protein. Other muscle structural proteins such as tropomyosin, α-actinin(47), titin, and nebulin(48) also have been detected in embryonic mouse skeletal muscle. These proteins appear to be expressed in a specific order(9, 48). In the EBs, α- and β-cardiac myosin heavy chain and embryonic skeletal myocin heavy chain(30), phospholamban(31), cardiac alkali and regulatory myosin light chains(32), and tropomyosin(33) have been reported to be present.

Immunofluorescence analysis. Indirect immunofluorescence was carried out to demonstrate the expression of the muscle actins and determine their cellular distribution. Unfortunately, the current repertoire of MAbs does not include reagents specific for each of the muscle isoactins. In this study, four MAbs that show a selective reactivity toward one or more of the muscle actins were used to examine the cellular distribution of certain isoactins in differentiating EBs. All four muscle actins share the epitope for MAb HUC 1-1(26). Likewise, MAb B4 shows a selective reactivity toward enteric actin(26, 27) under the fixation conditions, although it will bind to the other three muscle actins in methanol-fixed cells(9). Only vascular smooth muscle actin reacts with MAb 1A4(49), whereas MAb 5C5 binds specifically to the two striated actins, cardiac and skeletal(50). None of these four MAbs binds to the two cytoplasmic actins.

For convenience, the EBs were allowed to attach to glass slides rather than grow in suspension as described under “Methods.” The resulting EBs, although adherent to the slides, appeared to undergo differentiation at a similar rate to floating EBs, including the appearance of regions that beat rhythmically. Using the muscle actin-selective MAbs, two distinct populations of cells were identified in the EBs that contained one or more of the muscle actins. One population closely resembles cultured cardiomyocytes (Fig. 5,A, C, and E). These cells typically have one or two nuclei and, as seen with MAb HUC 1-1 (Fig. 5A), they display highly organized myofibrils that clearly contain muscle actin within the thin filaments of sarcomeres. Moreover, just as is the case in early cardiomyocytes in the developing embryo, these cells contain both vascular smooth muscle actin based on their staining with MAb 1A4(Fig. 5E) as well as at least one of the striated muscle actins as evidenced by their reactivity with MAb 5C5 (Fig. 5C). The overall pattern of staining is similar to that seen in embryonic chicken cardiomyocytes(9). These cardiomyocyte-like cells are generally found in small clusters within a field of nonmuscle cells that are not reactive with these muscle actin selective antibodies, which presumably correspond with the regions of beating cells. The second cell type expressing muscle actin in the EBs has a fibroblast-like morphology (Fig. 5,B,D,and F), and cells of this morphology show staining with MAbs HUC 1-1, B4, and 1A4, respectively. These cells are found in large numbers at the periphery of the attached EBs where they appear to have migrated after attachment of the EBs to the glass slides. The cells stain intensely for vascular smooth muscle actin using MAb 1A4(Fig. 5F) and could represent myofibroblasts(51), embryonic fibroblasts, or smooth muscle cells. Studies with MAb B4 (Fig. 5D), an antibody that shows a preferential reactivity toward the enteric smooth muscle actin, suggest that a somewhat smaller number of fibroblast-like cells in the EBs stain with this antibody compared with either MAb HUC 1-1 or MAb 1A4, although no attempt was made to quantitate the differences. In general, the fibroblast-like cells in EBs that contain these three muscle actin epitopes closely resemble those found in embryonic carcinoma cells(13) stained with these reagents. Importantly, these fibroblast-like cells clearly do not contain either of the striated actins because they do not react with MAb 5C5(not shown).

Figure 5
figure 5

Immunofluorescent staining of cultured EBs with antibodies against muscle actin. Embryoid bodies present in 9 or 13 d cultures were fixed as described under “Methods” and stained with MAbs selective for various muscle actins. A and B, MAb HUC 1-1 staining for any muscle actin; C, MAb 5C5 staining for cardiac and/or skeletal actin; D, MAb B4 staining for enteric smooth muscle actin; E and F, MAb 1A4 staining for vascular actin. Panels A, C, and E show staining with presumptive cardiomyocytes, whereas panels B, D, and F show the staining patterns observed with fibroblast-like or smooth muscle-like cells.

Full size image

Overall, the immunofluorescence data confirm and extend our Northern and PCR results. For example, vascular actin mRNA is the most abundant muscle actin message at any time in the EBs, and this isoactin is clearly detected by immunofluorescence in a large fraction of cells within the EBs. Likewise, the low level of cardiac and skeletal transcripts found by Northern analysis of the EBs is consistent with the relatively small number of cells that could be identified by immunoreactivity toward the striated muscle-specific actin MAb(5C5) in individual EBs compared with the number of cells containing vascular actin as shown by staining with MAb 1A4.

Previous studies have presented molecular and morphologic evidence for the presence of muscle elements in the EBs. In this report, we have characterized the expression of the four muscle isoactin mRNAs during the differentiation of ES cells to embryoid bodies, and we have demonstrated the presence of muscle actin(s) in two distinct cell populations by immunofluorescence. Overall, our results suggest that this system will provide a useful tool for studying the mechanisms involved in the regulation of actin gene expression during both striated muscle and smooth muscle differentiation. Moreover, this system may be useful in examining the functions of the muscle actins in these differentiating muscle cells after targeting each of the four genes encoding these proteins.