Aggrecan domains expected to traffic through the exocytic pathway are misdirected to the nucleus.

In this article, we report the misdirected targeting of expressed aggrecan domains. Aggrecan, the chondroitin sulfate (CS) proteoglycan of cartilage, normally progresses through the exocytic pathway. Proteins expressed from constructs containing the putative aggrecan signal sequence (i.e., the first 23 N-terminal amino acids), specified globular (G) domains G1 and/or G3, and a segment of the CS domain were detected in the endoplasmic reticulum (ER) and Golgi complex. Although proteins expressed from constructs containing the putative signal and G3, but lacking G1, were detected to a limited extent in the secretory pathway, they primarily accumulated in nuclei. Discrete nuclear inclusions were seen when G3 was expressed. Immunoelectron microscopic characterization of the inclusions suggested the association of nuclear G3 with other proteins. When signal-free G3 constructs and those with G3 immediately following the N-terminal signal were expressed, abundant dispersed accumulations filled the nucleoplasm. The data suggest first, that signal-free and signal-containing G3 proteins enter the nucleus from the cytosol, and second, that the entry of signal-containing G3 proteins into the ER lumen is inefficient. Hsp25, Hsp70, and ubiquitin were colocalized with nuclear G3, indicating the involvement of chaperones and the degradative machinery in the formation and/or attempted disposal of the abnormal nuclear inclusions. Overall, the results focus attention on (1) intracellular protein trafficking at the ER membrane and the nuclear envelope and (2) chaperone interactions and mechanisms leading to abnormal protein deposition in the nucleus.

The nanomelic mutation in the aggrecan gene is expressed in chick chondrocytes and neurons.

We have established the presence of at least two large chondroitin sulfate proteoglycans in the developing chick brain, one that reacts exclusively with HNK-1, a carbohydrate epitope found on several neural specific molecules, and one that reacts with S103L, a defined peptide epitope in the CS-2 domain of the cartilage-specific chondroitin sulfate proteoglycan (CSPG), aggrecan. In order to determine the relationships between the two distinct S103L-reactive CSPGs from cartilage (chondrocytes) and brain (neurons), as well as among the three large CSPGs expressed in brain, S103L, HNK-1 and versican, we studied the expression of these multiple proteoglycan species in the brain of nanomelic chicks. We have previously shown that homozygous embryos expressing the nanomelic phenotype exhibit a single point mutation in the aggrecan gene. In the present study, the S103L CSPG is not accumulated or synthesized by embryonic chick CNS tissue or E8CH neuronal cultures derived from nanomelic chick embryo cerebral hemispheres. In contrast, expression of both versican and the HNK-1 CSPG was normal in the mutant embryo CNS. Pulse chase experiments demonstrated the presence of the 380 kDa precursor in normal neurons and the 300 kDa truncated precursor in nanomelic neurons. Northern blot analysis revealed normal-sized mRNA but reduced levels of expression of the S103L CSPG message in nanomelic neurons, while expression of the versican message was comparable in normal and nanomelic neurons. Most conclusively, the point mutation previously identified in nanomelic cartilage mRNA was also identified in nanomelic brain mRNA. Together these results provide evidence that a single aggrecan gene is expressed in both cartilage and CNS tissue leading to the production of identical core proteins which then undergo differential and tissue-specific post-translation processing, resulting in the characteristic tissue-specific proteoglycans. Furthermore, versican and the HNK-1 CSPG, although structurally and chemically similar to the S103L CSPG, are the products of separate genes.

The ins and outs of aggrecan.

Aggrecan is a large and highly complex macromolecule, uniquely structured to fill space in the extracellular matrix (ECM) of cartilage. Lethal chondrodystrophies resulting from mutations in the structural gene for aggrecan demonstrate the serious consequences of the absence of aggrecan. Other chondrodystrophies are testimony to the importance of post-translational modifications. Here, Barbara Vertel reviews the role of aggrecan in the ECM of cartilage, discusses genetic mutations affecting aggrecan and highlights intracellular features of its synthesis and processing.

The biochemically and immunologically distinct CSPG of notochord is a product of the aggrecan gene.

Using the monoclonal antibody S103L, which reacts specifically with an epitope in the chondroitin sulfate-rich domain of the chick cartilage chondroitin sulfate proteoglycan (CSPG) core protein, we have identified the predominant CSPG expressed by notochord. This large notochord CSPG is first detected immunohistochemically as early as stage 16, long before chondrogenesis occurs, and is expressed continuously during the time of active neural crest migration and through the onset of sclerotomal differentiation. Because of the cross-reactivity of both notochord and cartilage CSPGs with the S103L antibody, extensive molecular and biochemical analysis of the two CSPGs was carried out. Striking differences distinguish the notochord and cartilage (aggrecan) CSPGs at the level of posttranslational modification. Notably, cartilage aggrecan carries a significant content of keratan sulfate (KS) chains, while the notochord CSPG is devoid of KS. In contrast, cartilage aggrecan lacks the HNK-1 epitope, while the notochord CSPG has a high content of HNK-1. Three different approaches were used to establish the relationship of the two CSPGs at the molecular level. Northern blot analysis, using aggrecan probes, detected same-sized messages from notochord and cartilage RNA. Overlapping fragments, generated by RT-PCR using primers covering 98% of the entire coding sequence from the known cartilage structure, were of identical size in notochord and cartilage. Taking advantage of our recent studies, which demonstrated a single base change in the aggrecan gene resulting in conversion of Glu to a STOP codon in exon 12 of chick aggrecan as the molecular basis of the defect nanomelia, we demonstrated that the same mutation was present in notochord mRNA from nanomelic chicks. These results provide evidence that the chick aggrecan gene is expressed very early in development in notochord and confirm that the core proteins expressed in chick notochord and cartilage are derived from the same gene. These findings strongly support the hypothesis that the final structural characteristics of each proteoglycan are determined not only by the core protein but also by tissue-specific, developmentally regulated posttranslational mechanisms, functioning within the context of the requirement for specific extracellular matrices.

The chondrodystrophy, nanomelia: biosynthesis and processing of the defective aggrecan precursor.

The lethal chicken mutation nanomelia leads to severe skeletal defects because of a deficiency of aggrecan, which is the largest aggregating chondroitin sulphate proteoglycan of cartilage. In previous work, we have demonstrated that nanomelic chondrocytes produce a truncated aggrecan precursor that fails to be secreted, and is apparently arrested in the endoplasmic reticulum (ER). In this study, we investigated the biosynthesis and extent of processing of the abnormal aggrecan precursor. The truncated precursor was translated directly in cell-free reactions, indicating that it does not arise post-translationally. Further studies addressed the processing capabilities of the defective precursor. We found that the mutant precursor was modified by N-linked, mannose-rich oligosaccharides and by the addition of xylose, but was not further processed; this is consistent with the conclusion that it moves no further along the secretory pathway than the ER. Using brefeldin A we demonstrated that the defective precursor can function as a substrate for Golgi-mediated glycosaminoglycan chains, but does not do so in the nanomelic chondrocyte because it fails to be translocated to the appropriate membrane compartment. These studies illustrate how combined cell biological/biochemical and molecular investigations may contribute to our understanding of the biological consequences and molecular basis of genetic diseases, particularly those involving errors in large, highly modified molecules such as proteoglycans.

cDNA cloning of chick cartilage chondroitin sulfate (aggrecan) core protein and identification of a stop codon in the aggrecan gene associated with the chondrodystrophy, nanomelia.

We previously reported the cloning and sequencing of a 1.5-kilobase cDNA which encoded a portion of the chondroitin sulfate domain from the chick cartilage proteoglycan core protein and the localization of a species-specific monoclonal antibody epitope. Using polymerase chain reaction amplification and primer extension, cDNA clones which code for the entire proteoglycan core protein have now been obtained from a 10-day chick embryo cDNA library. The composite sequence is 6464 nucleotides long, coding for a protein of 2109 amino acid residues with a calculated M(r) = 223,500. The overall arrangement of globular and carbohydrate-attachment domains is similar to human and rat chondrosarcoma aggrecan, but there are significant differences in detailed homology between chick and mammalian core proteins. Most significantly a highly repetitive region (19 repeat units of 20 residues each), not found in either human or rat, enlarges one of the characteristic serine-glycine containing regions (designated CS-2) while the other serine-glycine containing domain (designated CS-1) is approximately one-fourth the length of the mammalian CS-1. Analysis of a polymerase chain reaction-amplified fragment encoding the chick-specific repeat region revealed a single base mutation at position 4553 (G to T transversion) that converted the codon GAA for glutamate at amino acid 1513 to TAA, a stop codon, in nanomelic chondrocytes. Genomic DNA from nanomelic liver was also digested with restriction enzyme BsaBI to verify the G to T transversion. This single mutation leads to a shortened core protein precursor with a calculated M(r) = 158,300. The resulting phenotype, nanomelia, arises because the truncated core protein is neither processed to a mature proteoglycan, nor secreted from the chondrocyte.

Topography of glycosylation and UDP-xylose production.

In order to define the location and organization of the numerous reactions involved in polysaccharide assembly during synthesis of proteoglycans and glycoproteins, the topography of some of the glycosylation reactions in chondroitin sulfate synthesis was examined using a relatively new technique for generating permeable cells. Permeable chondrocytes were shown to directly take up nucleotide sugar precursors and incorporate them into chondroitin sulfate proteoglycan (CSPG), allowing specific labeling at each step in chondroitin sulfate synthesis. Subcellular fractionation following labeling with UDP-[14C]xylose, UDP-[14C]galactose, UDP-[14C]glucuronic acid, or [35S]PAPS localized the labeled CSPG to the compartment where each glycosylation reaction occurred. From these experiments it appears that xylose addition begins in the endoplasmic reticulum and continues in the Golgi apparatus where galactose, glucuronic acid, and sulfate are added. This conclusion was confirmed by direct visualization of xylose incorporation using electron microscopic autoradiography (Vertel, B. M., Walters, L. M., Flay, N., Kearns, A. E., and Schwartz, N. B. (1993) J. Biol. Chem. 268, 11105-11112). Further examination of xylose addition showed that permeable chondrocytes can utilize both exogenous UDP-xylose transported into the lumen and UDP-xylose generated from UDP-glucuronic acid within the lumen. The enzyme responsible for this reaction, UDP-glucuronate carboxy-lyase, co-localized with xylosyltransferase activity in subcellular fractions. Orientation toward the lumen in subcellular compartments was determined by trypsin sensitivity in the permeable chondrocytes. Therefore, we conclude that UDP-xylose can be produced in the lumen of the compartment where it is utilized in CSPG synthesis, obviating the need for a direct transport mechanism for this nucleotide sugar and providing close regulation of UDP-xylose and UDP-glucuronic acid levels.

Xylosylation is an endoplasmic reticulum to Golgi event.

The subcellular site of xylosylation, the first carbohydrate modification of the core protein that initiates glycosaminoglycan chain synthesis, was characterized in situ. Methods were developed to combine electron microscopic (EM) autoradiography and the radiolabeling of semi-intact chondrocytes. In the accompanying paper, Kearns et al. (Kearns, A. E., Vertel, B. M., and Schwartz, N. B. (1993) J. Biol. Chem. 268, 11097-11104) presented biochemical and subcellular fractionation studies that utilized semi-intact chondrocytes and radiolabeled UDP sugars to overcome obstacles to the direct analysis of xylosylation. The results suggested that xylosylation begins in the endoplasmic reticulum (ER) and continues in the Golgi. The site of xylosylation was not specified further due to the limitations of subcellular fractionation techniques. The studies described in this report were undertaken to localize these modifications directly in situ. Semi-intact cell preparations were optimized for ultrastructural preservation by modifications of permeabilization methods utilizing nitrocellulose filter overlays. Biochemical analysis demonstrated the exclusive incorporation of UDP-xylose into the cartilage chondroitin sulfate proteoglycan (aggrecan) core protein and 3'-phosphoadenosine 5'-phosphosulfate (PAPS) into the highly modified proteoglycan monomer. Immunolocalization studies showed the equivalence of cytoplasmic subcompartments in normal and semi-intact chondrocytes at the levels of light and electron microscopy. Once the biochemical and morphological equivalence of intact and semi-intact cells was established, EM autoradiographic studies were pursued using UDP-[3H]xylose and [35S]PAPS. Based on both qualitative and quantitative data, silver grains resulting from incorporated sulfate were concentrated in the perinuclear Golgi, while those resulting from incorporated xylose were found at the cis or forming face of the Golgi and in vesicular regions of the peripheral cytoplasm associated with the late ER. These data support the view that xylose addition begins in a late ER compartment and continues in intermediate compartments, perhaps including the cis-Golgi.

Nanomelic chondrocytes synthesize, but fail to translocate, a truncated aggrecan precursor.

Cartilage extracellular matrix (ECM) is composed primarily of type II collagen and large, link stabilized aggregates of hyaluronic acid and chondroitin sulfate proteoglycan (aggrecan). Maturation and function of these complex macromolecules are dependent upon sequential processing events which occur during their movements through specific subcellular compartments in the constitutive secretory pathway. Failure to complete these events successfully results in assembly of a defective ECM and may produce skeletal abnormalities. Nanomelia is a lethal genetic mutation of chickens characterized by shortened and malformed limbs. Previous biochemical studies have shown that cultured nanomelic chondrocytes synthesize a truncated aggrecan core protein precursor that disappears with time; however, the protein does not appear to be processed by the Golgi or secreted. The present study investigates the intracellular trafficking of the defective aggrecan precursor using immunofluorescence, immunoelectron microscopy and several inhibitors. Results indicate that nanomelic chondrocytes assemble an ECM that contains type II collagen, but lacks aggrecan. Instead, aggrecan precursor was localized intracellularly, within small cytoplasmic structures corresponding to extensions of the endoplasmic reticulum (ER). At no time were precursor molecules observed in the Golgi. In contrast, normal and nanomelic chondrocytes exhibited no difference in the intracellular or extracellular distribution of type II procollagen. Therefore, retention of the aggrecan precursor appears to be selective. Incubation of chondrocytes at 15 degrees C resulted in the retention and accumulation of product in the ER. After a return to 37 degrees C, translocation of the product to the Golgi was observed for normal, but not for nanomelic, chondrocytes, although the precursors disappeared with time. Ammonium chloride, an inhibitor of lysosomal function, had no effect on protein loss, suggesting that the precursor was removed by a non-lysosomal mechanism, possibly by ER-associated degradation. Based on these studies, we suggest that nanomelic chondrocytes are a useful model for examining cellular trafficking and sorting events and the processes by which abnormal products are targeted for retention or degradation. Further investigations should provide insight into the mechanisms underlying chondrodystrophies and other related diseases.

Subcompartments of the endoplasmic reticulum.

The endoplasmic reticulum (ER) is the largest continuous endomembrane structure in the cytoplasm. It may be viewed as a series of unique subcompartments. In this review, we examine the rough ER, nuclear envelope and several smooth ER subcompartments. Consideration is given to the characteristic properties and functions of the ER and its domains, and to the formation and maintenance of subcompartments. Associations within the ER membrane bilayer, and with constituents of the cytoplasm and the ER lumen, contribute to the formation of domains and lead to the establishment of subcompartments that reflect specialized functions and vary according to the physiologic state and phenotype of the individual cell. Although the structural complexity of some ER subcompartments (such as the sarcoplasmic reticulum) is highly elaborate, the ER remains a dynamic organelle, subject to assembly and disassembly, capable of extensive remodelling and active in exchange with other organelles through mechanisms of membrane transport.

Precursors of chondroitin sulfate proteoglycan are segregated within a subcompartment of the chondrocyte endoplasmic reticulum.

Immunocytochemical methods were used at the levels of light and electron microscopy to examine the intracellular compartments of chondrocytes involved in extracellular matrix biosynthesis. The results of our studies provide morphological evidence for the compartmentalization of secretory proteins in the ER. Precursors of the large chondroitin sulfate proteoglycan (CSPG), the major proteoglycan species produced by chondrocytes, were present in the Golgi complex. In addition, CSPG precursors were localized in specialized regions of the ER. Link protein, a separate gene product which functions to stabilize extracellular aggregates of CSPG monomers with hyaluronic acid, was segregated similarly. In contrast, type II procollagen, another major secretory molecule produced by chondrocytes, was found homogeneously distributed throughout the ER. The CSPG precursor-containing ER compartment exhibits a variable tubulo-vesicular morphology but is invariably recognized as an electronlucent, smooth membrane-bounded region continuous with typical ribosome-studded elements of the rough ER. The observation that this ER structure does not stain with antibodies against resident ER proteins also suggests that the compartment is a specialized region distinct from the main part of the ER. These results support recent studies that consider the ER as a compartmentalized organelle and are discussed in light of the possible implications for proteoglycan biosynthesis and processing.

Nanomelic chondrocytes synthesize a glycoprotein related to chondroitin sulfate proteoglycan core protein.

Chicken embryos homozygous for the autosomal recessive gene nanomelia exhibit cartilage defects, synthesize low levels of cartilage chondroitin sulfate proteoglycan (CSPG), and are missing the CSPG core protein (Argraves, W. S., McKeown-Longo, P. J., and Goetinck, P. F. (1981) FEBS Lett. 131, 265). In our studies of nanomelic chondrocytes in culture, we detected neither sulfate-labeled CSPG nor its Mr 370,000 core protein. However, in immunoprecipitation reactions using both polyclonal and monoclonal antibodies directed against the cartilage CSPG core protein, we identified a protein of Mr 300,000 that contains an epitope found in the hyaluronic acid-binding region of the normal core protein. This protein was also detected among products synthesized by chondrocytes obtained from phenotypically normal embryos resulting from matings between parents heterozygous for nanomelia. Sensitivity to endoglycosidase H indicated that the product is a glycoprotein with attached mannose-rich oligosaccharides. Pulse-chase studies revealed the disappearance of the glycoprotein after 6 h of chase, but no detectable formation of proteoglycan. Our results suggest that although nanomelic chondrocytes are deficient in the production of normal CSPG and its core protein, they do synthesize a smaller, immunologically related glycoprotein that does not undergo the post-translational processing characteristic of the normal cartilage core protein.

Biosynthetic precursors of cartilage chondroitin sulfate proteoglycan.

Early steps in the biosynthesis of chondroitin sulfate proteoglycan (CSPG) and collagenous cartilage matrix molecules were examined by the comparison of products translated in mRNA-directed cell-free reactions and those synthesized by intact cartilage cells. RNA isolated from embryonic chicken sterna was used to direct cell-free translation reactions. Chicken sternal chondrocytes in culture were pulse-labeled with [35S]-methionine. The CSPG core protein was identified by immunoprecipitation. The Mr of the cartilage cell-synthetized core protein was determined to be 370K, approximately 10-15K greater than that of the comparable cell-free translation product. Experimental results strongly support the view that the observed difference in Mr reflects the cotranslational addition of mannose-rich, N-asparagine-linked oligosaccharides to the cell-synthesized core protein: 1) the cell-synthesized product was labeled with [3H]-mannose and precipitated by concanavalin A-sepharose beads; 2) the incorporated [3H]-mannose could be subsequently removed by digestion with endoglycosidase H (Endo H); 3) the Mr of the cell-synthesized core protein was reduced by Endo H digestion to that of the comparable cell-free translation product; 4) the core protein synthesized by tunicamycin-treated chondrocytes (inhibited in their ability to add N-asparagine-linked mannose-rich oligosaccharides to proteins) was comparable in electrophoretic mobility to that of the core protein cell-free translation product; and 5) the core protein translated in microsome-coupled cell-free reactions had an Mr 8-10K greater than that of the core protein translated in the absence of microsomes. For the purpose of examining biosynthetic intermediates, chondrocytes were labeled continuously or pulse-chase labeled for varying times. No biosynthetic CSPG intermediates migrating between the core protein and the CSPG monomer were detected. However, a band of 355Kdal appeared to share certain characteristics with the 307Kdal core protein (including its immunoprecipitability with CSPG antibodies), and a 340Kdal band was noted. Type II procollagen and other collagenase-sensitive products of 205Kdal and 110Kdal were observed among translation and chondrocyte-synthesized products. In chondrocytes, all three products exhibited labeling or chase time-dependent increases in Mr which were accelerated by ascorbate supplements and inhibited by the addition of alpha, alpha'-dipyridyl. These results suggest that the observed time-dependent increases in Mr are a consequence of collagen hydroxylation. The 110Kdal and 205Kdal collagenous proteins may be related to the minor collagens recently described in cartilage.

Immunofluorescence studies on cartilage matrix synthesis. The synthesis of link protein, chondroitin sulfate proteoglycan monomer and type II collagen.

A comparison of the synthesis and deposition of fibrous type II collagen and the constituents of chondroitin sulfate proteoglycan (CSPG) aggregates, CSPG monomer and link protein, was made for chicken sternal chondrocytes in culture, using simultaneous double immunofluorescence and lectin localization. Chondrocytes deposited only CSPG constituents--and not type II collagen--into the extracellular matrix (ECM). Intracellular precursors of CSPG monomer were localized primarily in perinuclear regions, but were observed in other cytoplasmic vesicles as well. Link protein antibodies stained the same intracellular structures, but stained the perinuclear cytoplasm less intensely. In contrast, type II procollagen was distributed in vesicles throughout the cytoplasm and was clearly absent from the distinctive, CSPG precursor-containing vesicles. Fluorescence-labelled lectins were used to further identify intracellular membrane compartments. Wheat germ agglutinin (WGA) and Ricinus lectins (which recognize carbohydrates added in the Golgi) stained the perinuclear cytoplasm, while concanavalin A (conA) (which recognizes mannose-rich oligosaccharides added co-translationally) stained vesicles throughout the rest of the cytoplasm and not the perinuclear cytoplasm. The distinctive CSPG-containing vesicles were not stained with WGA or Ricinus agglutinins. Data presented elsewhere demonstrate that the vesicles do not react with monoclonal antibodies which recognize chondroitin sulfate (CS) or keratan sulfate (KS) determinants. Thus, we conclude that the vesicles accumulate CSPG precursors which have not been modified by Golgi-mediated processes. The data indicate that matrix molecules may be segregated selectively prior to transit through the Golgi complex. The co-distribution of link protein and CSPG monomer precursors in vesicles prior to further, Golgi-mediated modification may reflect an as yet undetermined function of these vesicles in the processing or assembly of CSPG.

Intracellular features of type II procollagen and chondroitin sulfate proteoglycan synthesis in chondrocytes.

The intracellular compartments of chondrocytes involved in the synthesis and processing of type II procollagen and chondroitin sulfate proteoglycan (CSPG) monomer were investigated using simultaneous double immunofluorescence and lectin localization reactions. Type II procollagen was distributed in vesicles throughout the cytoplasm, whereas intracellular precursors of CSPG monomer were accumulated in the perinuclear cytoplasm. In this study, cytoplasmic vesicles that stained intensely with antibodies directed against CSPG monomer but did not react with type II collagen antibodies, also were observed. A monoclonal antibody, 5-D-4, that recognizes keratan sulfate determinants was used to identify the Golgi complex (the site of keratan sulfate chain elongation). Staining with 5-D-4 was restricted to the perinuclear cytoplasm. The vesicles outside the perinuclear cytoplasm that stained intensely with antibodies to CSPG monomer did not react with 5-D-4. Fluorescent lectins were used to characterize further subcellular compartments. Concanavalin A, which reacts with mannose-rich oligosaccharides, did not stain the perinuclear region, but it did stain vesicles throughout the rest of the cytoplasm. Because mannose oligosaccharides are added cotranslationally, the stained vesicles throughout the cytoplasm presumably correspond to the rough endoplasmic reticulum. Wheat germ agglutinin, which recognizes N-acetyl-D-glucosamine and sialic acid (carbohydrates added in the Golgi), stained exclusively the perinuclear cytoplasm. By several criteria (staining with the monoclonal antibody 5-D-4 and with wheat germ agglutinin), the perinuclear cytoplasm seems to correspond to the Golgi complex. The cytoplasmic vesicles that react with anti-CSPG monomer and not with anti-type II collagen contain precursors of CSPG monomer not yet modified by Golgi-mediated oligosaccharide additions (because they are not stained with wheat germ agglutinin or with the anti-keratan sulfate antibody); these vesicles may have a unique function in the processing of CSPG.

Stimulation and inhibition of myoblast differentiation by hormones.

The growth and differentiation of L6 myoblasts are subject to control by two proteins secreted by cells of the Buffalo rat liver line. The first of these, rat insulinlike growth factor-II (formerly designated multiplication stimulating activity) is a potent stimulator of myoblast proliferation and differentiation, as well as associated processes such as amino acid uptake and incorporation into protein, RNA synthesis, and thymidine incorporation into DNA. In addition, this hormone causes a significant decrease in the rate of protein degradation. All of these actions seem to be attributable to a single molecular species, although their time courses and sensitivity to the hormone differ substantially. The second protein, the differentiation inhibitor (DI), is a nonmitogenic inhibitor of all tested aspects of myoblast differentiation, including fusion and the elevation of creatine kinase. Indirect immunofluorescence experiments demonstrated that DI also blocks accumulation of myosin heavy chain and myomesin. Upon removal of DI after 72 h incubation, all of these effects were reversed and normal myotubes containing the usual complement of muscle-specific proteins were formed. Thus, this system makes it possible to achieve specific stimulation or inhibition of muscle cell differentiation by addition of purified proteins to cloned cells in serum-free medium.