Involvement of protein phosphatase-1-mediated MARCKS translocation in myogenic differentiation of embryonic muscle cells

A prominent event in the differentiation of skeletal muscle cells is the fusion of mononucleated myoblasts to form multinucleated myotubes(Bischoff and Holtzer, 1969;O'Neill and Stockdale, 1972). This process absolutely requires reorganization of cytoskeleton and accompanies various cellular events, including accumulation of muscle-specific proteins and redistribution of membrane components(Fulton et al., 1981;Nadal-Ginard, 1978;Pauw and David, 1979). Numerous studies using cultured myoblasts have suggested that phosphorylation and dephosphorylation of proteins play an important role in mediating differentiation of myoblasts (Toutant and Sobel, 1987; Rapuano et al.,1989; Adamo et al.,1989; Kim et al.,1992a; Kim et al.,1993; Baek et al.,1994). Okadaic acid, an inhibitor of protein phosphatases, has been shown to block membrane fusion of differentiating myoblast(Kim et al., 1991;Kim et al., 1992b). The activity of the protein phosphatases sensitive to okadaic acid has also been reported to increase progressively in differentiating myoblasts(Srinivasan and Begum,1994).

The myristoylated alanine-rich C kinase substrate (MARCKS) is an abundant,high affinity cellular substrate for PKC. It has been implicated in various cellular events, such as cell mobility, neurosecretion and cell transformation(Aderem, 1992a;Blackshear, 1993). One of the striking features of MARCKS is its phosphorylation-dependent translocation between the cytosol and the plasma membrane. The activation of PKC in various systems leads to phosphorylation of MARCKS and its reversible translocation from the plasma membrane to the cytosolic compartment of the cells(Thelen et al., 1991;Wang et al., 1989). MARCKS is able to crosslink filamentous actins in the plasma membrane, and the phosphorylation of MARCKS dissociates the filamentous actin crosslinking and translocates it to the cytosol (Hartwig et al., 1992). Therefore, the subcellular localization of MARCKS is undoubtedly essential to its function in the regulation of the cytoskeleton dynamics, particularly in the interaction of actin filaments at the plasma membrane.

Ser/Thr-specific protein phosphatases are divided into two groups, type-1 and type-2 (Ingebritsen and Cohen,1983a; Ingebritsen and Cohen,1983b). The type-1 protein phosphatase (PP-1) preferentially dephosphorylates the β-subunit of phosphorylase kinase and is inhibited by nanomolar concentrations of two heat-stable proteins, termed protein phosphatase inhibitor-1 and -2(Ingebritsen and Cohen, 1983a;Ingebritsen and Cohen, 1983b),as well as by micromolar concentrations of okadaic acid(Bialojan and Takai, 1988;Haystead et al., 1989;Cohen et al., 1990). By contrast, the type-2 protein phosphatases (PP-2), comprising PP-2A, PP-2B and PP-2C, specifically dephosphorylate the α-subunit of phosphorylase kinase (Ingebritsen and Cohen,1983a; Ingebritsen and Cohen,1983b). Their activities are not affected by the inhibitor proteins, but PP-2A is specifically inhibited by another heat stable protein(I₁^PP2A) and extremely sensitive to nanomolar concentrations of okadaic acid (Cohen,1989; Li et al.,1996). The bombesin- or vasopressin-stimulated phosphorylation of MARCKS is dynamically reduced upon removal of the PKC activator in Swiss 3T3,indicating that the phosphatase activity against MARCKS is present and could regulate the phosphorylation state of the protein(Rodriguez-Pena et al., 1986). Moreover, okadaic acid, which inhibits PP-1 and PP-2A, shows little or no effect on MARCKS phosphorylation on its own, but strongly inhibits the dephosphorylation of MARCKS, which were phosphorylated upon activation of PKC(Clarke et al., 1993).

We have recently shown that MARCKS translocates from the cytosol to the plasma membrane while mononucleated myoblasts fuse to form multinucleated myotubes (Kim et al., 2000). This MARCKS translocation was in part due to a decrease in the expression of PKCθ, which is the major enzyme responsible for MARCKS phosphorylation,during the myogenic process. We have also demonstrated that treatment with PMA, an activator of the protein kinase, blocks both MARCKS translocation and myoblast fusion. In order to clarify further whether the MARCKS shuttling is indeed involved in myogenesis, we examined the effect of okadaic acid on translocation of MARCKS during the differentiation process. In the present studies, we show that okadaic acid prevents the translocation of MARCKS from the cytosol to the plasma membrane and that the dephosphorylating activity against MARCKS markedly increases during the course of myogenesis. In addition, PP-1 was found to be the major enzyme responsible for the MARCKS dephosphorylation. Moreover, overexpression of MARCKS carrying a mutation that prevents myristoylation and thus blocks its membrane translocation impaired the fusion of cultured myoblasts. Therefore, we suggest that the protein phosphatase-1-mediated MARCKS localization at the membrane is involved in the fusion of embryonic muscle cells.

The cDNA for chick MARCKS and its mutant forms (N/S and A₂G₂: see below) and polyclonal anti-MARCKS antibody were kindly provided by P. J. Blackshear and D. J. Stumpo (NIEHS). Culture media, horse serum, antibiotic/antimycotics solution, and other culture reagents were from Gibco-BRL. [γ-³²P]ATP was obtained from Amersham; FuGENE 6 and complete protease inhibitor cocktail were from Boeringher Mannheim; okadaic acid, tautomycin, Ro-31-8220, protein phosphatase inhibitor-2, and protein phosphatase-2A inhibitor were from Calbiochem;rhodamine-phalloidin was from Molecular Probes. pcDNA3.1/myc-His and Vectashield were obtained from Invitrogen and Vector Laboratories,respectively. Antibodies against PP-1, c-Myc, and phospho-MARCKS (Ser152/156)were from Transduction Laboratories, Santa Cruz Biotechnology, and Cell Signaling Technologies, respectively. All secondary antibodies were from Jackson Immunoresearch. All other materials were purchased from Sigma, unless otherwise indicated.

Myoblasts from breast muscle of 12-day-old chick embryos were prepared as described previously (Ha et al.,1979; Kim et al.,1992a). The cells were plated on collagen-coated culture dishes at a concentration of 5×10⁵ cells/ml in minimum essential medium(MEM) containing 10% (v/v) horse serum, 10% (v/v) chick embryo extract, and 1%(v/v) antibiotic/antimycotics solution. One day after the cell seeding, the culture medium was changed with the same medium but containing 2% embryo extract for induction of differentiation. The cells were subjected to Giemsa staining, and the degree of myoblast fusion was determined as the percentage of the number of nuclei in fused cells to the total number of nuclei in 10 randomly chosen the fields under a microscope. Cells containing more than three nuclei were regarded as fused cells.

Myoblasts cultured for appropriate periods were washed three times with ice-cold Earl's balanced salt solution (EBSS), harvested by centrifugation,and kept frozen at -70°C until use. The cells were resuspended in 50 mM Tris-HCl (pH 7.4), 5 mM MgCl₂, 1 mM dithiothreitol (DTT), and 1× complete protease inhibitor cocktail, and preparation of their subcellular fractions was performed as described previously(Kim et al., 2000). Protein concentration was determined by a previously described method(Bradford, 1976) or by using bicinchoninic acid when Triton X-100 was present(Smith et al., 1985). Bovine serum albumin was used as a standard.

Phospho-MARCKS was prepared using purified MARCKS and partially purified PKCθ preparation, which was obtained by the heparin-agarose chromatography of the muscle extracts as described previously(Kim et al., 2000). Reaction mixtures contained 20 mM Hepes (pH 7.4), 0.25 mM EDTA, 0.25 mM EGTA, 1 mM CaCl₂, 1 mM DTT, 5 mM MgCl₂ and 0.15 mM[γ-³²P]ATP (0.05 mCi/ml). After incubating them for 3 minutes at 25°C, the samples were heated for 3 minutes at 95°C and centrifuged at 15,000 g for 30 minutes. The resulting ³²P-labeled MARCKS was used as the substrate for assaying the phosphatase activity. The enzyme activity was determined by incubation of ³²P-labeled MARCKS with the soluble extracts obtained from the cultured myoblasts in 50 mM Hepes(pH 7.4), 1 mM EDTA, 1 mM EGTA, 5 mM MgCl₂, 1 mM DTT, 1×complete protease inhibitor cocktail, and 100 nM Ro-31-8220 or staurosporine. Incubations were performed for 30 minutes at 30°C and stopped by addition of 2% SDS solution containing 1% 2-mercaptoethanol. The samples were heated for 3 minutes at 95°C and subjected to polyacrylamide gel electrophoresis in 10% slab gels under denaturing conditions (SDS-PAGE)(Laemmli, 1970), followed by autoradiography. When inhibitors were included, the extracts were incubated with them for 15 minutes prior to the addition of ³²P-labeled MARCKS.

Myoblasts cultured for 48 hours were resuspended in 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5 mM MgCl₂ and 1× complete protease inhibitor cocktail. The cells were disrupted by sonication on ice and centrifuged at 15,000 g for 30 minutes. The supernatants were incubated with 10 μg of anti-PP-1 IgG or preimmune IgG at 4°C for 2 hours in a rotating shaker. They were then added with 40 μl of 50% (v/v)protein A-Sepharose in the same buffer and incubated for 1 hour. After centrifugation, the supernatants were subjected to phosphatase assay. They were also subjected to SDS-PAGE as above, followed by autoradiography.

Since a stable cell line cannot be established with the primary culture of myoblasts and since these cells have a relatively low efficiency of transfection, Myc-tagged MARCKS were prepared for identification of the transfected cells. The N/S mutant carries the Asn residue in the phosphorylation domain of MARCKS in place of Ser(Swierczynski and Blackshear,1996). In the A₂G₂ mutant, the Gly residue next to the N-terminus of MARCKS was substituted with Ala. PCR reactions were performed using the cDNAs for the wild-type MARCKS and its mutant forms as the templates, and oligonucleotide primers carrying the HindIII and XbaI restriction sites. The sequences of the primers are as follows:primer 1, TGTTAAGCTTGCCACCATGGGTGCC, which includes a HindIII site;primer 2, TGTTAAGCTTGCCACCATGGCTGCC, which has an altered sequence that can complement with the A₂G₂ mutation site as well as the HindIII site; primer 3, GCGGTCTAGACTCCGCCGGCTCGGC, which has a XbaI site. The PCR products were cut off by treatment with HindIII and XbaI and ligated into the pcDNA3.1/myc-His plasmid that had been cut with the same enzymes. Thus, in the resulting plasmids, the cDNAs for MARCKS and its mutant forms are under control of the cytomegalovirus (CMV) promoter/enhancer. These constructs also have a polyadenylation signal of the bovine growth hormone (BGH)terminator at their 3′-termini.

The cells were seeded on 100 mm plates at a density of 1.0×10⁵ cells/ml and cultured for 24 hours. To prepare DNA for transfection, 12 μl of FuGENE 6 was diluted with 0.4 ml of MEM and incubated for 5 minutes at room temperature. The diluted solution was slowly added to the pcDNA derivatives (4 μg), mixed gently, and incubated for 15 minutes. The cultured myoblasts were rinsed once with MEM and slowly added with the FuGENE 6-treated DNA solution. After incubation for 4 hours, the transfected cells were rinsed with and cultured in the differentiation medium for 48 hours.

Myoblasts that had been cultured for 24 hours were transfected with the pcDNA derivatives and incubated for 24 hours. The cells were washed twice in PBS and fixed in 2% paraformaldehyde for 15 minutes. They were again washed twice in PBS and incubated in PBS containing 0.5% saponin and 2% horse serum to permeabilize the cells and saturate nonspecific binding sites. The cells were incubated with the monoclonal antibody raised against c-Myc (9E10) and a polyclonal antibody raised against creatine kinase for 1 hour, and then with fluorescein isothiocyanate (FITC)-conjugated donkey anti-mouse IgG and phycoerythrin (PE)-conjugated goat anti-rabbit IgG, respectively, for the next 1 hour. They were then washed twice in PBS, separated by collagenase, and resuspended in PBS. The samples were analyzed with a FACStarplus (Becton Dickinson), and data were analyzed with CellQuest software.

The cells plated on collagen-coated glass coverslips were fixed with 3.7%paraformaldehyde in PBS for 10 minutes, followed by permeabilization with 0.5%Triton X-100 in PBS. All subsequent dilutions and washes were carried out with PBS containing 0.1% Triton X-100 (PBST). Nonspecific binding sites were saturated by incubation with 3% horse serum and 10% gelatin in PBST for 1 hour. The cells were incubated with the antibody raised against c-Myc for 1 hour and then with FITC-conjugated donkey anti-mouse IgG for the next 1 hour. For detection of filamentous actin, the cells were incubated with rhodamine-phalloidin. The coverslips were mounted in Vectashield and viewed in a laser scanning confocal microscope (Carl Zeiss LSM 510). FITC and rhodamine were excised by 488 nm argon and 543 nm HeNe laser, and the images were filtered by bandpass 505-530 nm and longpass 585 nm filters, respectively. The acquired images were processed with Adobe PhotoShop and printed with Tektronix phaser 450. No fluorescence was detected if primary antibody was omitted.

In order to determine whether dephosphorylation of MARCKS is involved in translocation of the protein from the cytosol to the plasma membrane, the primary culture of chick myoblasts was treated with okadaic acid or tautomycin, potent inhibitors of protein phosphatases, at the time of medium change. Also, the cells that had been treated with okadaic acid or tautomycin for 24 hours were freed of the drug by changing the medium. These cells were further cultured for the next 48 and 24 hours, respectively, and subjected to analysis for subcellular distribution of MARCKS. Total cell lysates and membrane fractions were prepared from the cells and subjected to SDS-PAGE,followed by immunoblot analysis using an anti-MARCKS antibody.Fig. 1A shows that treatment with okadaic acid (lane O) or tautomycin (lane T) and subsequent removal of the drug from the medium (lanes F_O and F_T) show little or no effect on the total level of MARCKS. However, the drug treatment strongly inhibited the translocation of MARCKS to the membrane and this inhibitory effect could be reversed upon removal of the drug from the medium(Fig. 1B). We also examined the effect of the drug on membrane fusion of the cells that had been cultured under the same conditions. As shown in Fig. 1C, both okadaic acid and tautomycin reversibly blocked the cell fusion. These results suggest that dephosphorylation of MARCKS leads to translocation of the protein to the plasma membrane in cultured myoblasts and that the cell fusion is closely correlated with the membrane translocation of MARCKS.

To determine whether the dephosphorylating activity against MARCKS might change during myogenesis, soluble extracts were obtained from myoblasts that had been cultured for various periods and assayed for their dephosphorylation activity by incubation with ³²P-labeled MARCKS. To prevent the endogenous activity of PKCs, the extracts were pretreated with Ro-31-8220, a potent inhibitor of the enzymes. The samples were then subjected to SDS-PAGE followed by autoradiography. The ability of the soluble extracts to dephosphorylate ³²P-labeled MARCKS markedly increases(Fig. 2A) as the cells fuse to form myotubes (Fig. 2C). Moreover, this increase in the dephosphorylating activity was particularly evident in the extracts obtained from cells that are competent or committed for membrane fusion (i.e. cultured for 36-48 hours). However, treatment with okadaic acid or tautomycin strongly inhibited both dephosphorylating activity and membrane fusion (Fig. 2A and C, respectively). To determine whether the increase in the dephosphorylating activity indeed correlates with the phosphorylation state of MARCKS in the cells, the same extracts were subjected to immunoblot analysis using an antibody that reacts only with the phosphorylated form of MARCKS. As shown in Fig. 2B, the level of the phospho-MARCKS proteins decreased in parallel with the increase in the dephosphorylating activity. Moreover, treatment with okadaic acid or tautomycin blocked the decrease in the level of phospho-MARCKS. These results suggest that the MARCKS-specific phosphatase activity is upregulated just before or during the onset of myoblast fusion and may be required for myoblast fusion.

MARCKS that was phosphorylated can be almost completely dephosphorylated by the purified catalytic subunit of PP-2A(Clarke et al., 1993). The purified catalytic subunit of PP-1 (PP-1C) can also dephosphorylate MARCKS,although somewhat less efficiently than the catalytic subunit of PP-2A(Clarke et al., 1993). To identify the protein phosphatase(s) that is responsible for dephosphorylation of MARCKS during myogenesis, the soluble extracts obtained from the cells that had been cultured for 48 hours were incubated with ³²P-labeled MARCKS in the presence of increasing concentrations of tautomycin or okadaic acid. As shown in Fig. 3A, both tautomycin (open circle) and okadaic acid (closed circle) inhibited the phosphatase activity in a dose-dependent fashion. The concentrations required for a half-maximum inhibition (IC₅₀) by tautomycin and okadaic acid were about 1 nM and 60 nM, respectively. Since PP-2C is insensitive to both inhibitors and since the inhibition of PP-2B by tautomycin needs a more than 5000-fold higher concentration of the agents(Mackintosh and Klumpp, 1990),it appears unlikely that PP-2B and PP-2C act on phosphorylated MARCKS in differentiating myoblasts. In addition, it has been reported that PP-2A requires levels of okadaic acid at least 100-fold lower than those of PP-1 for complete inhibition, although the drug can inhibit both PP-1 and PP-2A(Bialojan and Takai, 1988). These results suggest that PP-1 is responsible for dephosphorylation of MARCKS.

To clarify further the involvement of PP-1 in MARCKS dephosphorylation, the same cell extracts were subjected to immunoprecipitation by treatment with an antibody raised against PP-1C. The resulting supernatants were then incubated with ³²P-labeled MARCKS, followed by SDS-PAGE and autoradiography. As shown in Fig. 3B, treatment with the anti-PP-1C IgG completely prevented the dephosphorylating activity(lane c), unlike treatment with preimmune IgG (lane b). We also examined the effects of phosphatase inhibitors on dephosphorylation of ³²P-labeled MARCKS. Fig. 3C shows that the dephosphorylating activity is strongly inhibited by protein phosphatase inhibitor-2 (lane b), which is a specific inhibitor of PP-1, but not by protein phosphatase-2A inhibitor(I₁^PP2A) (lane c). Moreover, the protein phosphatase activity was inhibited by treatment with both inhibitor proteins to an extent similar to that seen with protein phosphatase inhibitor-2 alone (lane d). Thus, it is likely that PP-1 is the major enzyme responsible for dephosphorylation of MARCKS during the myogenic differentiation.

To determine whether the increase in the dephosphorylating activity against ³²P-labeled MARCKS during myogenesis(Fig. 2A) might be due to a change in the expression of PP-1C, the same extracts were subjected to immunoblot analysis using the anti-PP-1C IgG. However, the protein level of PP-1C remained nearly constant during the entire culture periods(Fig. 4). PP-1 is known to consist of multiple subunits, including catalytic, targeting, and regulatory subunits. Thus, it is possible that the expression of either targeting or regulatory subunit or both of them may change during myogenesis and result in an increase in the capacity of dephosphorylating MARCKS.

To clarify whether MARCKS translocation is involved in myoblast fusion, we first examined the effects of the mutations that could influence membrane translocation of MARCKS. The cDNAs for two different mutant forms of MARCKS,N/S and A₂G₂, were used for cell transfection(Swierczynski and Blackshear,1996). In N/S, the Ser phosphorylation site is replaced by Asn and therefore cannot be phosphorylated by protein kinases. In A₂G₂, the Gly residue, which is located next to the N-terminus of MARCKS and serves as the myristoylation site, is substituted with Ala. Thus, N/S should be targeted to the plasma membrane, while A₂G₂ should remain in the cytosol. The two forms of MARCKS were also Myc-tagged for their identification upon immunoblot analysis using anti-Myc antibodies.

Myoblasts that had been cultured for 24 hours were transfected with the cDNAs and further cultured for 48 hours in the absence or presence of okadaic acid (Fig. 5B,C, panels a and b, respectively). Total cell lysates were prepared from the cells and subjected to immunoblot analysis. WhileFig. 5A shows the endogenous MARCKS proteins that were detected with an anti-MARCKS antibody,Fig. 5B represents the Myc-tagged, wild-type MARCKS (Wt) and its mutant forms (N/S and A₂G₂) that were expressed from the transfected cDNAs and detected with an anti-Myc antibody. Of note is that A₂G₂, which should lack the myristoyl group, migrated faster than the Myc-tagged, wild-type MARCKS or N/S in the gel. These results show that the transfected cells could overexpress the Myc-tagged MARCKS proteins. We then examined the effects of the mutations in the phosphorylation and myristoylation sites on MARCKS translocation in vivo. The soluble and membrane fractions were obtained from the cultured myoblasts that had been transfected with the cDNAs for the Myc-tagged wild-type MARCKS, N/S and A₂G₂. These fractions were then subjected to immunoblot analysis using an anti-Myc antibody as above. When the transfected cells were cultured in the absence of okadaic acid, the majority of the wild-type MARCKS(Wt) and N/S were recovered in the membrane fraction(Fig. 5Ca). Upon treatment with the drug, the wild-type MARCKS was recovered in the soluble fraction while the membrane translocation of N/S was not affected(Fig. 5Cb). In contrast,A₂G₂ remained exclusively in the soluble fraction,whether the transfected cells were cultured in the absence or presence of okadaic acid. These results indicate that dephosphorylation and myristoylation of MARCKS are essential for its translocation in vivo from the cytosol to the plasma membrane of the cultured myoblasts.

To determine whether translocation of MARCKS is indeed involved in membrane fusion of cultured myoblasts, the cells were transfected with the cDNAs for the Myc-tagged, wild-type MARCKS, N/S and A₂G₂ as above,cultured for 48 hours, and subjected to immunostaining using an anti-Myc antibody. They were then observed under a confocal microscope. The cells expressing the wild-type MARCKS (Wt) (Fig. 6A) and N/S (Fig. 6D) readily fused to form multinucleated myotubes with intense staining of the protein along the plasma membrane. In contrast, the cells expressing A₂G₂ had an unfused, bipolar shape with diffuse staining of MARCKS in the cytosol(Fig. 6G). Thus, it appears that translocation of MARKS from the cytosol to the plasma membrane is associated with or a requisite step for membrane fusion of cultured myoblasts.

To determine whether the MARCKS translocation during myogenesis is correlated with its capability of crosslinking filamentous actin, the same cells were also stained using pallotoxin, a selective ligand of filamentous actin but not of the globular form. The cytoskeletal networks of filamentous actin in cells expressing the wild-type MARCKS and N/S were co-localized with the membrane-translocated MARCKS (Fig. 6A-C and D-F, respectively). However, construction of the filamentous actin network was impaired in the cells expressing A₂G₂(Fig. 6G-I), although unfused myoblasts showed longitudinally penetrated actin filaments (data not shown). These results suggest that MARCKS translocation to the plasma membrane is tightly associated with its ability to crosslink filamentous actin, which may positively influence the membrane fusion of cultured myoblasts.

During myogenesis, the membrane fusion of myoblasts accompanies the accumulation of muscle-specific proteins, such as creatine kinase and myosin heavy chain. To examine whether the prevention of MARCKS translocation and hence blocking the myoblast fusion might influence the expression of muscle-specific proteins, the cells were transfected with the cDNAs for the Myc-tagged wild-type MARCKS, N/S and A₂G₂, as above. After culturing the cells for 48 hours, they were stained with both anti-Myc and anti-creatine kinase antibodies, followed by FACS analysis to measure the expression levels of creatine kinase in the transfected cells only.Fig. 7 shows that accumulation of creatine kinase could be detected not only in cells expressing the wild-type MARCKS and N/S but also in cells expressing A₂G₂ (although its level in cells expressing A₂G₂ is significantly less than that seen in those expressing the wild-type MARCKS and N/S), despite the fact that the latter cannot fuse to form myotubes, unlike the others. Similar results were obtained for the expression of myosin heavy chain (data not shown). These results strongly suggest that the expression of muscle-specific proteins is independent of the membrane translocation of MARCKS as well as myoblast fusion.

We have previously shown that MARCKS translocates from the cytosol to the plasma membrane during the course of myogenic differentiation of cultured myoblasts and this translocation is mediated in part by the decrease in the expression of PKCθ, which is the major enzyme responsible for the phosphorylaion of MARCKS (Kim et al.,2000). In the studies described here, we demonstrated that the shuttling of MARCKS between the cytosol and the plasma membrane of cultured myoblasts could also be controlled by the protein phosphatase-mediated dephosphorylation of MARCKS. Treatment of the cultured myoblasts with okadaic acid or tautomycin was found to reversibly block MARCKS translocation, similar to its effect on the membrane fusion of the cells(Kim et al., 1991;Kim et al., 1992b). Moreover,the dephosphorylating activity against MARCKS was found to markedly increase in the cells that are competent or committed for membrane fusion. Thus, it is clear that dephosphorylation as well as phosphorylation of MARCKS is involved in the regulation of its subcellular localization in the embryonic muscle cells.

In a cell-free system, PP-1, PP-2A and PP-2C could dephosphorylate recombinant MARCKS or a synthetic peptide containing the phosphorylation domain of MARCKS. In intact Swiss 3T3 cells, okadaic acid, added at concentrations for inhibiting PP-1 and PP-2A but not PP-2C, exerts little effect on MARCKS phosphorylation on its own, but largely prevents the dephosphorylation of MARCKS that occurred following the activation of PKC by bombesin with subsequent receptor blockade(Clarke et al., 1993). Therefore, it has been suggested that, although dephosphorylation of MARCKS can be mediated by PP-2C in vitro, MARCKS might be dephosphorylated by okadaic-acid-sensitive PP-1 and/or PP-2A in Swiss 3T3.

A number of results obtained in the present studies suggest that PP-1 is the major protein phosphatase that is responsible for the dephosphorylating activity against MARCKS in cultured myoblasts. First, both tautomycin and okadaic acid blocked the dephosphorylating activity against MARCKS at concentrations that can inhibit PP-1 but not PP-2A. Second,immunoprecipitation by an antibody raised against the catalytic subunit of PP-1 (PP-1C), but not by preimmune IgG, prevented the MARCKS dephosphorylating activity. Third, the phosphatase activity was strongly inhibited by the protein phosphatase inhibitor-2, a specific protein inhibitor of PP-1, but not by the inhibitor protein specific to the protein phosphatase-2A. Thus, it appears clear that PP-1 participates in dephosphorylation of MARCKS during myogenesis.

However, it is noteworthy that the concentration of okadaic acid used to inhibit dephosphorylation effectively in vivo(Fig. 1, 25 nM) was much lower than the micromolar concentrations generally accepted to inhibit PP-1 in other cells (Bialojan and Takai,1988; Haystead et al.,1989; Cohen et al.,1990), which suggests that PP-2A may also participate in MARCKS dephosphorylation in cultured myoblasts. Furthermore, the concentration of tautomycin required for prevention of the MARCKS dephosphorylation in vivo was much higher than that needed in vitro (Figs1,3). Favre et al. have systematically demonstrated using MCF7 cells that the phosphatase inhibitors,including okadaic acid and tautomycin, have distinct cell permeability properties (Favre et al.,1997). Thus, it appears possible that okadaic acid shows much higher permeability to myoblasts than to other cells, whereas tautomycin has the opposite property.

Also noteworthy is the observation that the protein level of PP-1C remained nearly constant during the entire time course of myogenesis, despite the fact that the dephosphorylating activity against MARCKS markedly increased during the same period. In vivo, PP-1C exists as a complex with other proteins that target it to particular subcellular localizations, modify its substrate specificity, and regulates its enzyme activity(Lavoinne et al., 1991). In the rabbit skeletal muscle, PP-1C has been shown to interact with glycogen particle (PP-1G), sarcoplasmic reticulum (PP-1SR), myofibrils (PP-1M), and the cytoplasmic inhibitor-2 protein (PP-1I)(Dent et al., 1990). In insulin-stimulated L6 rat myoblasts, the protein level of PP-1G dramatically increases in response to the agonist, while that of PP-1C remains constant(Srinivasan and Begum, 1994). Thus, it appears possible that the increase in the dephosphorylation activity against MARCKS in cultured chick myoblasts may be regulated by the other subunits of PP-1, such as the targeting or regulatory subunits. Further studies are required to identify the subunit(s), whose expression may change and hence alter the activity of PP-1C during differentiation of chick myoblasts.

The membrane translocation of MARCKS appears to be essential or a requisite event in the membrane fusion of cultured myoblasts. Distribution of MARCKS changes from the cytosol to the plasma membrane of the cells and this translocation appears to be regulated by both the increase in the PP-1-mediated MARCKS dephosphorylation and the decrease in the PKCθ-mediated MARCKS phosphorylation during the course of myogenic differentiation (Kim et al.,2000). Furthermore, overexpression of A₂G₂,which cannot be targeted to the plasma membrane because it lacks a myristoylation site, impaired the cell fusion. By contrast, the cells expressing N/S, which lacks a phosphorylation site and thus readily translocates to the membrane, underwent fusion similarly to those expressing the wild-type MARCKS. In addition, formation of filamentous actin cytoskeleton is largely dependent on translocation of MARCKS to the plasma membrane and its distribution coincides with that of MARCKS in cells overexpressing the wild-type MARCKS or N/S. It has been reported that alterations in MARCKS distribution by PKC-mediated phosphorylation result in disintegration of the membrane skeleton, and cytochalasin D, an inhibitor of actin polymerization,enhances PKC-induced translocation of MARCKS to the cytosol(Douglas et al., 1997;Vaaraniemi et al., 1999). Together with the fact that MARCKS regulates the cytoskeleton dynamics by crosslinking the actin filaments in the plasma membrane and that myoblast fusion accompanies massive cytoskeleton reorganization, we suggest that the PP-1-mediated MARCKS localization at the membrane is required for the fusion of embryonic muscle cells.

Stumpo et al. have reported that MARCKS-deficient mice exhibit abnormal brain development but do not have any obvious defects in other major organs including muscle, questioning the involvement of MARCKS in myoblast differentiation (Stumpo et al.,1995). However, they also reported that MARCKS is highly expressed in many developing neuronal tissues in which phenotypic alterations were not seen in the MARCKS-deficient mouse, as well as in many extracranial tissues in the mouse. Therefore, they suggested that the MARCKS homologue MRP(MARCKS-related protein) (Aderem,1992b; Blackshear,1993) might also be highly expressed in these tissues during development and prevent them from exhibiting an abnormal phenotype. Similarly,MRP in embryonic muscle tissues may serve as a functional equivalent of MARCKS and prevent any defect in muscle development in MARCKS-deficient mice,although the most sensitive processes would be affected by the reduction in total amount of `MARCKS equivalents'(Stumpo et al., 1995).

Of interest was the finding that expression of muscle-specific proteins,such as creatine kinase and myosin heavy chain, could occur in the fusion-arrested myoblasts upon overexpression of A₂G₂. This result is consistent with our previous finding that okadaic acid blocks the membrane fusion of chick myoblasts with little effect on induction of muscle-specific proteins (Kim et al.,1991). It has also been shown that prevention of cell fusion, such as by prolonged culture in a low-calcium medium or treatment with cytochalasin B, does not interfere with the synthesis of muscle-specific proteins and the other differentiative processes, including transport and release of calcium(Holtzer et al., 1975;Constantin et al., 1995). Thus, the synthesis of muscle-specific proteins occurring independently of the blockade of membrane fusion by A₂G₂ expression could be an additional example of the uncoupling of biochemical differentiation from morphological differentiation, despite the fact that both processes normally occur simultaneously during the myogenic differentiation of embryonic skeletal muscle cells.

We are grateful to P. J. Blackshear and D. J. Stumpo, (NIH) for supplying the chicken MARCKS antibody and cDNAs for N/S and A₂G₂. We also thank J. W. Lee (Korea Basic Science Institute) and Y. I. Choi (Seoul National University, Korea) for their technical assistance in confocal microscopy and FACS analysis, respectively. J.H.K. was a recipient of a fellowship from the BK21 program supported by the Korea Ministry of Education. This work was supported by grants from the Korea Science and Engineering Foundation (through the Research Center for Proteinaceous Materials).