Betaglycan expression is transcriptionally up-regulated during skeletal muscle differentiation. Cloning of murine betaglycan gene promoter and its modulation by MyoD, retinoic acid, and transforming growth factor-beta.

Betaglycan is a membrane-anchored proteoglycan co-receptor that binds transforming growth factor beta (TGF-beta) via its core protein and basic fibroblast growth factor through its glycosaminoglycan chains. In this study we evaluated the expression of betaglycan during the C(2)C(12) skeletal muscle differentiation. Betaglycan expression, as determined by Northern and Western blot, was up-regulated during the conversion of myoblasts to myotubes. The mouse betaglycan gene promoter was cloned, and its sequence showed putative binding sites for SP1, Smad3, Smad4, muscle regulatory factor elements such as MyoD and MEF2, and retinoic acid receptor. Transcriptional activity of the mouse betaglycan promoter reporter was also up-regulated in differentiating C(2)C(12) cells. We found that MyoD, but not myogenin, stimulated this transcriptional activity even in the presence of high serum. Betaglycan promoter activity was increased by RA and inhibited by the three isoforms of TGF-beta. On the other hand, basic fibroblast growth factor, BMP-2, and hepatocyte growth factor/scatter factor, which are inhibitors of myogenesis, had little effect. In myotubes, up-regulated betaglycan was also detectable by TGF-beta affinity labeling and immunofluorescence microscopy studies. The latter indicated that betaglycan was localized both on the cell surface and in the ECM. Forced expression of betaglycan in C(2)C(12) myoblasts increases their responsiveness to TGF-beta2, suggesting that it performs a TGF-beta presentation function in this cell lineage. These results indicate that betaglycan expression is up-regulated during myogenesis and that MyoD and RA modulate its expression by a mechanism that is independent of myogenin.

Skeletal muscle myoblasts are the precursors to skeletal muscle fibers. During development these cells are maintained in a proliferative and undifferentiated state until appropriate signals cause them to undergo conversion into multinucleated myotubes.
A network of muscle regulatory factors (MRFs), 1 some of which have been identified and characterized, governs this process. When myogenesis begins, myogenic regulatory genes encoding factors of the MyoD family are activated (myogenin, MRF4, and MRF5). These factors bind to specific DNA consensus sites called E boxes, which function as transcriptional enhancers of muscle differentiation genes (for review see Ref. 1). Factors of the MEF2 family are also required to specify muscle fate or to direct muscle differentiation (2). The expression and activity of these master regulatory genes are regulated by polypeptide growth factors, such as basic fibroblast growth factor (FGF-2) (3), transforming growth factor type ␤ (TGF-␤) (4), hepatocyte growth factor/scatter factor (HGF) (5), and insulin-like growth factor (6), as well as by retinoic acid (RA) (7).
The onset and progression of the differentiation process are controlled by a complex set of interactions between myoblasts and their environment. The presence of the extracellular matrix (ECM) is essential for normal myogenesis (8 -11). It has been demonstrated that the activities of FGF-2, HGF, and TGF-␤ can be regulated by binding to proteoglycans (2)(3)(4)(5)(6). Cell surface heparan sulfate proteoglycans, in particular, have been suggested to play a role in modulating the activities of heparinbinding growth factors (7,8). They modulate terminal myogenesis probably by acting as low affinity receptors for some growth factors such as FGF-2 (5,9) and HGF (10). We have shown that constitutive expression of syndecan-1, a heparan sulfate proteoglycan whose expression is down-regulated during skeletal muscle terminal differentiation, inhibits the differentiation of myoblasts in culture (5,11). In contrast, abolishment of the expression of syndecan-3, another heparan sulfate proteoglycan synthesized by myoblasts, results in acceleration of skeletal muscle differentiation by a mechanism being dependent on FGF-2 (9). Decorin is a dermatan sulfate proteoglycan that has the ability to bind TGF-␤ by its core protein (12,13). Decorin seems to be critical in C 2 C 12 cells to repress skel-etal muscle differentiation by a TGF-␤-dependent mechanism (14), and its expression is up-regulated during the process (15).
Cell interaction with TGF-␤ is mediated by a complex group of proteins that includes the type I and II TGF-␤ receptors, which have signal transducing activity, and the membranebound proteoglycan betaglycan (16 -18). Betaglycan, a heparan/chondroitin sulfate proteoglycan also known as type III TGF-␤ receptor (T␤RIII), is present on the cell surface of several cell types (19,20) and has the ability to bind TGF-␤ through the core protein and FGF-2 through the heparan sulfate chains (21). Although betaglycan is not directly involved in the intracellular TGF-␤ signaling, it controls its access to the T␤RII receptor, thereby modulating its actions (6,22). In this study we show that expression of betaglycan is up-regulated during skeletal muscle differentiation of C 2 C 12 myoblasts, increasing its levels on the cell surface of the myotubes and in the ECM. Our data also indicate that the proximal region of the betaglycan gene promoter controls this process in a manner that is dependent of MyoD expression. Although it is not clear at this moment what is the function of betaglycan during the skeletal muscle differentiation process, the fact that its ectopic forced expression in myoblasts enhances their responsiveness to TGF-␤2 suggests a potential regulatory role.

EXPERIMENTAL PROCEDURES
Cell Culture-The mouse skeletal muscle cell line C 2 C 12 (ATCC, Manassas, VA) (23) was grown and induced to differentiate as described (15). Briefly, 3 days after plating (80% confluence), the medium was changed to Dulbecco's modified Eagle's medium supplemented with 5% horse serum to trigger myoblast differentiation. Two days later 0.1 mM cytosine-␤-D-arabinofuranoside was added to the culture medium. Thereafter the incubation medium was changed daily. Rat1 and L6E9 cells were routinely cultured in Dulbecco's modified Eagle's medium containing 10% or 20% fetal bovine serum, respectively, as described (19).
RNA Isolation and Northern Blot Analysis-Total RNA was isolated from cell cultures at the indicated times using Trizol (Invitrogen). The RNA samples were electrophoresed in 1.2% agarose/formaldehyde gels, transferred to Nytran membranes, hybridized with probes for mouse betaglycan (24), and exposed to Kodak x-ray film.
Generation of the Adenoviral Vector for the Expression of Wild Type Betaglycan (Adv-BG)-The DNA vectors and hosts employed were generous gifts of Bert Vogelstein (Johns Hopkins University Oncology Center), and the method employed has been described by He et al. (25). Briefly, to generate the appropriate transfer vector, a DNA fragment containing the 186 upstream bp of the 5Ј-untranslated region, the complete open reading frame, and the first 513 bp of the 3Ј-untranslated region of the wild type rat betaglycan cDNA (19) was subcloned at the NotI-HindIII sites of pAdenoTrack. The created transfer vector was recombined with pAdenoEasy-1, and the recombinant adenoviral genome was transfected in 293 cells to recover the recombinant viral particle (Adv-BG). Enrichment and purification of an Adv-BG viral stock was carried out according to standard methods.
Cloning of Murine Betaglycan Gene Promoter and Construction of Betaglycan Promoter Reporter Vector-A bacterial artificial chromosome mouse genomic library (Research Genetics, Huntsville, AL) was screened with the 190 nucleotides (bases 15-205) at the 5Ј end of the reported mouse betaglycan cDNA (24), and eight positive clones were analyzed by restriction mapping. A 6.5-kilobase pair HindIII fragment that hybridized to bases 88 -105 of the cDNA was recovered from bacterial artificial chromosome clone B70. A smaller HindIII-BamHI fragment containing the 2542 bp located at the 5Ј end was further isolated and fully sequenced at Instituto de Fisiologia Celular Molecular Biology Core Facility (deposited in the GenBank TM with accession number AF537325). For the construction of betaglycan reporter plasmid, pBG-Lux, the 2307-bp HindIII-NotI Klenow-filled-in fragment was subcloned in the SmaI of pGL3-basic (Promega, Madison, WI), and the proper orientation was verified by sequencing (see Fig. 2 for more details). The pMyo-Luc reporter construction was reported previously (14); pRLSV40 was from Promega; and pEMSV, pEMSV-MyoD, and pEMSV-myogenin were gifts from Dr. Stephen F. Konieczny (Purdue University).
Transient Transfection-The cells were plated at a density of 8,000 cells/cm 2 in 6-well plates for 48 h and transfected for 5 h using with 4 g of pBG-Lux or 4 g of pMyoLuc and 0.01 g of pRLSV40 in each well mixed in Opti-MEM I with 4 l of PLUS reagent and 6 l of Lipo-fectAMINE (Promega). Following transfection, fetal bovine serum and chicken embryo extract were added to restore their complete growth medium concentrations, and the cells were cultured for a further 14 h. The cells were then trypsinized and plated in 24-well plates in normal growth medium. Under these conditions, FGF-2 (3 ng/ml) (Invitrogen), BMP-2 (1 nM) (generous gift from Genetic Institute, Cambridge, MA), HGF (3 ng/ml) (R & D Systems, Minneapolis, MN), TGF-␤ (1 ng/ml) (R & D Systems), and RA (1 M) (Sigma) were added at the indicated concentrations. After 24 h, the cells were harvested and assayed for dual luciferase activity (Promega). All of the transfections were performed at least three times.
TGF-␤ Affinity Labeling and Immunoprecipitation-TGF-␤ was radiolabeled with 125 I using the chloramine T, and affinity labeling and immunoprecipitations were done as described before (22).
Immunofluorescence Microscopy-Cells to be analyzed by immunofluorescence microscopy were grown on glass coverslips, as described previously (27). For cell surface staining the live unfixed cells were incubated in the first antibody solution 1:200 anti-betaglycan, 1:1000 anti-perlecan (27) in 50 mM Tris-HCl, pH 7.7, 0.1 M NaCl, 2% bovine serum albumin for 1 h on ice. After removal of this solution, the cells were fixed in 3% paraformaldehyde and permeabilized with 0.05% Triton X-100 and incubated with 1:100 monoclonal anti-myosin antibody (Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA). The bound antibodies were detected by incubating the cells with 1:100 affinity-purified fluorescein isothiocyanate-conjugated anti-rabbit IgG and 1:100 TRITC anti-mouse IgG (Pierce), respectively. After rinsing, the slides were viewed with a Nikon upright microscope equipped for epifluorescence.
Analysis of Creatine Kinase Activity-Creatine kinase activity was determined as previously described (9).
DNA and Protein Determination-DNA (28) and protein (29) were determined in aliquots of cell extracts as described.

Skeletal Muscle Myoblasts Up-regulate Betaglycan
Expression during Differentiation-To evaluate the expression of betaglycan during myoblast differentiation, the levels of its mRNA were determined in total RNA isolated from myoblasts and myotubes induced to differentiate for 2, 4, and 6 days. As shown in Fig. 1A, betaglycan mRNA was detected after 2 days of induced differentiation with an increase by day 4 and small decrease by day 6. At the same time, the level of cellular betaglycan during skeletal muscle differentiation was determined by Western blot analysis of detergent extracts treated previously with heparitinase and chondroitinase ABC, a procedure that reveals the total content of betaglycan as its core protein. Fig. 1B shows that cells allowed to differentiate for 4 days increased betaglycan expression significantly compared with the amount synthesized by myoblasts. Betaglycan has the ability to bind the three isoforms of TGF-␤ through its core protein. To evaluate this property of the betaglycan induced during muscle differentiation, affinity labeling experiments using 125 I-labeled TGF-␤2 were performed. Fig. 1C shows that the up-regulated betaglycan, displaying its characteristic pattern of migration in this kind of experiments, bound TGF-␤2 and that this binding increased during differentiation when the total extracts were analyzed. The specificity of this result was confirmed when immunoprecipitates from the total cell extract were analyzed using a specific anti-betaglycan antibody (Fig.  1D). Fig. 1E shows the densitometric analysis for the quantitative analyses of the data. As a test of concomitant differentiation, the level of creatine kinase activity, a specific skeletal muscle marker, was evaluated ( Fig. 1F). Together, these results demonstrate that the synthesis of betaglycan increased during skeletal muscle differentiation.
Cloning of Mouse Betaglycan Gene Promoter and Construction of Mouse Betaglycan Gene Reporter-To characterize the mechanism underlying the control of up-regulation of betaglycan mRNA, the murine betaglycan gene promoter was cloned. For that purpose, a bacterial artificial chromosome library was screened with the most 5Ј end sequences of the murine betaglycan cDNA, and the HindIII-BamHI 2542-nucleotide-long fragment ( Fig. 2A) containing the most upstream cDNA region was isolated and sequenced (Fig. 2B). This genomic DNA fragment had 86% sequence identity with the corresponding region of its rat ortholog (1). In addition, it contained a stretch of 127 nucleotides (2265-2391) that corresponded to bases 20 -141 of the published murine betaglycan cDNA and that was followed by a GT, presumably the splicing donor site of the first intron. Other relevant functional cis-regulatory elements, SP1 and GC boxes, including the putative BMP-2-responsive element, that have been experimentally determined in the rat gene by the Centrella and co-workers (1) were also conserved in the mouse promoter. These similarities strongly suggest that the cloned sequences contain the promoter and first exon of the mouse betaglycan gene. Importantly, the mouse sequences contained several MRFs binding sites: two MyoD-and five MEF2-binding sites, elements that have been shown to be essential for skeletal muscle differentiation (30). In addition, they also showed the presence of one RA receptor and four Smad4-and one Smad3-binding elements, factors that mediate responses to RA and TGF-␤, regulators of skeletal muscle differentiation.
To determine whether or not the isolated HindIII-BamHI 2542-bp fragment had promoter activity, it was used to construct a luciferase reporter plasmid, which was transiently transfected in cell lines capable of expressing the endogenous betaglycan. Fig. 2C shows that, as compared with the empty vector, the cloned HindIII-NotI fragment drove the expression of luciferase in Rat1 fibroblasts. Importantly, transfected L6E9 rat myoblasts, which do not express betaglycan (22), failed to express the reporter gene. Together, these findings confirmed that the cloned region contains the bona fide mouse betaglycan gene promoter.
Activity of Betaglycan Reporter Increases during Myogenesis-To characterize the mechanisms controlling the up-regulation of betaglycan during the myogenic differentiation, C 2 C 12 myoblasts were transiently transfected using the betaglycan reporter vector and induced to differentiate for 6 days. As expected from the results described above, Fig. 3 shows that the activity of the pBG-Lux increased 8.4-fold during skeletal FIG. 1. Betaglycan expression is upregulated during myoblasts differentiation. A, 15 g of total RNA isolated from myoblasts (lane 0) and myoblasts induced to differentiate for the indicated days were analyzed by Northern blot with 32 P-labeled betaglycan cDNA probe. The arrows indicate ribosomal RNA size, and the arrowhead indicates betaglycan mRNA (BG). B, cell extracts were prepared from myoblasts (lane 0), and cells were induced to differentiate for 4 days (lane 4) and then treated with heparitinase and chondroitinase ABC. The extracts were analyzed by SDS-PAGE. Betaglycan was visualized by immunoblotting with an antibody specific for betaglycan. The arrow indicates the core protein of betaglycan. C 2 C 12 myoblasts were culture under growth conditions for 48 h and switched to differentiate for several days. At the indicated times of differentiation the cells were affinity labeled with 100 pM of 125 I-labeled TGF-␤2. The total extracts were analyzed by SDS-PAGE (C) or by anti-betaglycan immunoprecipitation (D). Phosphorimages of the gels are shown indicating the migration of the molecular mass standards (kDa) and of the betaglycan-TGF-␤2 complex (bracket). E, the graphical representation of the phosphorimaging quantitation of the betaglycan-TGF-␤2 complex. Open bars correspond to total extracts, and striped bars correspond to the immunoprecipitates. F, cell extracts were prepared from myoblasts (0), cells were induced to differentiate for the indicated days, and creatine kinase activity was determined. For graphical representation, the enzymatic activity values were normalized by DNA content. When indicated the results are the means Ϯ S.D. of two independent experiments performed in triplicate.  (1) and by alignment to mouse betaglycan cDNA (24). B, nucleotide sequence of mouse betaglycan promoter and first exon. The positive numbers in the right margin correspond to the 2542-bp HindIII-BamHI sequence submitted to the GenBank TM (accession number AF537325). For description purposes, the nucleotide sequence of was also numbered taking the A that opens the reading frame in betaglycan cDNA as base ϩ1 (negative numbers in the left margin). The predicted first exon in the murine betaglycan gene spans from position Ϫ341 to Ϫ91 (underlined). Thus, the G at position 2391, which corresponds to position Ϫ91 and belongs to the 5Ј-untranslated region, is the last base in the putative first exon. The C at position 2195 (asterisk) is the first base that aligns with the published cDNA. The partial sequence of the putative first intron is shown in lowercase letters. Sequence analysis of the 2542-bp HindIII-BamHI fragment using Professional MatInspector software identified several potential cis-regulatory elements (consensus sequences higher than 0.9). The subset of these elements as well as restriction endonuclease sites that are relevant to the present work are shown in bold type. For clarity, if there is an overlap between two cis-elements, one of them is underlined, and the other is shaded. C, transactivating activity of murine betaglycan promoter. The 2307-bp HindIII-NotI fragment was used for the construction of betaglycan reporter plasmid, pBG-Lux, as described under "Experimental Procedures." Luciferase activity was measured in total cell extracts obtained from L6E9 or Rat1 cells after 48 h of transient transfection with control empty vector (pGL3-b) or pBG-Lux. muscle differentiation. In parallel, as a control of the differentiation process, the increase in expression of myogenin (40-fold) evaluated with pMyo-Luc reporter is shown.
BetaglycanExpressionIncreasesduringMyogenesisAreMyoDdependent-As indicated before, betaglycan promoter contain several DNA-binding sites for MyoD and MEF2 proteins. Upon induction of muscle differentiation, these proteins bind to the regulatory regions and activate the transcription of skeletal muscle genes, such as creatine kinase. Thus, we asked whether or not by forcing the expression of myogenin or MyoD in myoblasts the transcriptional activity of the betaglycan reporter could be enhanced. For that purpose, myoblasts cells were transiently co-transfected with an expression vector containing the full-length myogenin cDNA (pEMSV-Myog) or with an expression vector containing the full-length MyoD cDNA (pEMSV-MyoD), together with pBG-Lux and pRLSV40. The transfected cells were maintained under proliferating conditions, and dual luciferase activity was measured after 24 h. Fig.  4 (open bars) shows that MyoD overexpression produced an important increase of the betaglycan reporter activity. However, no effect on the betaglycan promoter activity was observed by myogenin overexpression. The same results were observed when the cells were triggered to differentiate for 24 h (Fig. 4, dark bars). The inset of Fig. 4 shows the protein levels of MyoD or myogenin in the corresponding transfected cells, evaluated by Western blot analyses, and is compared with the induction of those MRFs in myoblasts induced to differentiate. The levels of expression of MyoD or myogenin in the transfected cells were comparable with or even higher than the levels observed during normal differentiation. These results demonstrate that betaglycan expression is under the control of MyoD but not of myogenin activity.
Betaglycan Expression Is Inhibited by TGF-␤ and Stimulated by Retinoic Acid-The promoter sequence for betaglycan contains putative binding sites for Smads, the major transcriptional mediators of the TGF-␤ responses, as well as for RA receptor. Because the induction of myoblast differentiation is modulated by removal of diverse growth factors (26) or by the addition of RA to the medium (11), we decided to evaluate their role on the activity of betaglycan promoter. For that purpose, myoblasts were transiently transfected with pBG-Lux and cultured in the presence and absence of defined growth factors FGF-2, BMP-2, HGF, and TGF-␤ and growth medium (10% fetal bovine serum). Fig. 5 shows that treatment of the cells for 24 h with FGF-2, BMP-2, or HGF did not modify luciferase activity compared with growth medium (10% fetal bovine serum), whereas treatment with any of the three TGF-␤ mammalian isoforms resulted in a 50% inhibition. These results suggest that growth factors like FGF-2, BMP-2, and HGF are not necessary to maintain the expression of betaglycan, whereas TGF-␤ is inhibitory. On the other hand, RA, a vitamin A metabolite that plays a major role in skeletal muscle development (31) and differentiation, strongly stimulated the transcriptional activity of pBG-Lux (Fig. 5). This stimulation is observed under growth or differentiation conditions (data not shown), indicating that RA, an inducer of skeletal muscle differentiation, stimulates the expression of betaglycan.
Betaglycan Enhances TGF-␤2 Potency in C 2 C 12 Myoblasts-Since the recent paper by Eickelberg et al. (32) reporting that depending on the type and quantity of the glycosaminoglycan attached to betaglycan, it may prevent the access of TGF-␤ to the signaling receptors, thereby acting as an inhibitor of the factor, the membrane-bound form of this co-receptor can not be assumed to be an enhancer of TGF-␤. Therefore, to determine whether or not betaglycan in the C 2 C 12 myoblasts was a TGF-␤ enhancer or inhibitor, we analyzed its effect on a prototypical TGF-␤ assay. For this purpose we forced the expression of betaglycan in C 2 C 12 myoblasts using the adenoviral vector Adv-BG, which drives the expression of the wild type receptor. Cells infected with Adv-BG and a control empty vector were then transiently transfected with p3TP-lux, a well known and widely used TGF-␤ responsive luciferase reporter (33). After another 24 h of incubation in growth medium, TGF-␤2 was added to the cells at the concentrations indicated in Fig. 6. The luciferase activity of these cells indicated that both the magnitude and the sensitivity of the response greatly increased in the Adv-BG infected cells as compared with the control infected cells. Importantly, the presence of betaglycan in the Adv-BGinfected cells was verified by Western blot (Fig. 6B). These results clearly show that betaglycan forced expression in the C 2 C 12 cells enhances TGF-␤ function at least in nondifferentiated myoblasts.
Betaglycan Increases on the Extracellular Matrix and on the Surface of Myotubes during Skeletal Muscle Differentiation-To further evaluate the expression and localization of betaglycan during skeletal muscle differentiation, nonpermeabilized cell cultures were stained with anti-betaglycan antibodies. Fig. 7D shows that myoblasts express betaglycan that can be detected on the cell surface by immunofluorescent staining. However, cells stained 6 days after initiation of differentiation (Fig. 7E) showed a much higher betaglycan immunostaining on the cell surface and on fibrillar structures characteristic of ECM components (34,35). Myotubes were revealed by staining using anti-myosin antibodies after permeabilization (Fig. 7, H and I). The ECM localization of perlecan, a heparan sulfate proteoglycan, is shown in Fig. 7F. Panels K and L of Fig. 7 correspond to the merge of the immunostaining for betaglycan or perlecan, respectively, with the immunostaining for myosin. These results indicate that up-regulated betaglycan is mainly localized to the myotubes cell surface and associated to the ECM. DISCUSSION The results presented in this paper demonstrate that the expression of betaglycan, a transmembrane heparan/chondroitin sulfate proteoglycan (19), increases during the differentiation of skeletal muscle cells. This conclusion is based on analysis of betaglycan mRNA and protein levels, activity of a reporter construct containing a portion of the mouse betaglycan promoter, analysis of functional betaglycan through binding, presentation of TGF-␤2 cells, and immunofluorescent staining during differentiation.
These observations are of interest in the context of the regulation of the skeletal muscle formation by proteoglycans. Skeletal muscle cell lines provide a very informative model to study the involvement of proteoglycans and growth factors in this process. Heparin-binding growth factors such as FGFs, HGF, and TGF-␤ are among the principal known regulators of skeletal muscle regeneration and may be normally present in muscle tissue and released upon injury or secreted auto-or paracrinally by the cells during the regeneration process (36 -38). Quiescent and activated myoblasts express the c-Met receptor as well as FGF receptors (specially FGFR-1 and FGFR-4) (39, FIG. 6. Betaglycan enhances TGF-␤2 potency in C 2 C 12 myoblasts. A, C 2 C 12 myoblasts (control, closed circle) were infected with 100 plaque-forming units/cell of Adv-BG (closed triangle). After 24 h the myoblasts were transiently transfected with p3TP-Lux and pRL-SV40, an expression vector for Renilla reniformis luciferase, whose activity was used to normalize the transfection (pRL). After transfection the cells were incubated in growth medium for 24 h, and then increasing concentrations of TGF-␤2 were added. The cells were harvested after 24 h, and dual luciferase activity was determined. The results show the means of two independent experiments in triplicate. B, C 2 C 12 myoblasts were infected with the indicated plaque-forming units/cell (pfu) of Adv-BG. After 48 h, the cells were harvested, and betaglycan BG was determined by Western blot analyses as done for the experiment in  ). Growth factor responsiveness and consequently skeletal myogenesis are regulated by ligands and receptor availability (41), as well as by the presence of proteoglycans at the cell surface or at the ECM (either presenting or sequestering possible ligands). Thus, the expression, processing, and cellular localization of proteoglycans is critical to their modulatory growth factor function. As a matter of fact, we have shown before that perlecan, syndecan-1, and syndecan-3, all of them heparan sulfate proteoglycans, are regulated during in vitro skeletal muscle differentiation and play a role in modulating this process (5,9,11,27).
In the present work we provide evidence that betaglycan, another important proteoglycan, is modulated during the myogenic differentiation. Betaglycan is a versatile co-receptor capable of regulating the effects of several members of the TGF-␤ superfamily (42)(43)(44). In several cell lines, including rat L6E9 myoblasts lacking endogenous betaglycan, it significantly increases the affinity of T␤RII for TGF-␤, thereby enhancing the responses to TGF-␤ (22,45). It is postulated that this effect eliminates differences in efficacy among the three TGF-␤ isoforms. TGF-␤2, which binds to T␤RI and T␤RII with low affinity and has minimal consequence on growth inhibitory effects in the absence of betaglycan, behaves similarly to TGF-␤1 in the presence of betaglycan (22,45). Importantly, it has been shown that betaglycan, in the presence of the type II activin receptor, binds inhibin and thereby mediates its well known activin antagonism (42,43,46). Another proposed function of betaglycan derives from the existence of a soluble form of the receptor that is produced by the proteolytic cleavage of its ectodomain (47)(48)(49). Soluble betaglycan binds TGF-␤ with the same affinities as its membrane precursor, but contrary to it, this binding sequesters and neutralizes the factor (33,49). This capacity has been exploited therapeutically to prevent the deleterious effects of TGF-␤ in the progression of some kind of malignancies (50,51). Interestingly, a novel and surprising function of betaglycan has been shown recently. Eickelberg et al. (32) demonstrated that membrane betaglycan may inhibit TGF-␤ signaling. This inhibition is not caused by enhanced secretion of soluble betaglycan. It is proposed that in some cell lines, the type and amount of the glycosaminoglycan chains attached to betaglycan core protein, may sterically prevent the access of TGF-␤ to the T␤RI and T␤RII, thereby preventing its downstream signaling (32).
In the present work we show that the expression of betaglycan is regulated during the skeletal muscle differentiation and that this up-regulation is mainly accounted for at the level of transcription. Cloning of a genomic DNA containing the first exon and the first 2140-bp upstream region of mouse betaglycan gene allowed us to test its activity during differentiation.
Here we show unequivocally that the up-regulation of betaglycan is controlled at least by the expression of the MRF MyoD. In cells transiently transfected with MyoD, the activity of the betaglycan gene promoter, as measured by our reporter pBGlux, was up-regulated, even in the presence of high serum condition where muscle differentiation is inhibited (4). Interestingly, transient expression of myogenin, another MRF, had no effect on the activation of betaglycan transcription. We also found that treatment of myoblasts with any of the TGF-␤ isoforms was inhibitory. On the other hand, treatment of the cells with RA, an inducer of skeletal muscle differentiation in vivo (52) and in vitro (11), increased the activity of the betaglycan reporter. Furthermore, the increase in activity mediated by RA was observed independently of the presence of serum. This effect of RA is consistent with a report by Nakayama et al. (53), who found that RA increases the levels of betaglycan mRNA in osteoblast-like cells. It is particularly interesting that these growth factors and RA, which are known to be expressed in the vicinity of condensing mesenchymal cells and limb buds during early developmental stages (54 -57), modulate betaglycan expression. Interestingly, it has been described that RA induces differentiation only in MyoD-expressing muscle cells by the formation of the RA receptor-MyoD complex (58). Other growth factors such as FGF-2, BMP-2, or HGF did not significantly affect the promoter activity. This lack of BMP-2 effects contrasts with the findings of Ji et al. (1), who, using osteoblasts, demonstrated that BMP-2 inhibited the rat betaglycan gene promoter and identified a 180-bp region responsible of this silencing. The fact that we did not observe any BMP-2 effect may be explained simply by a difference in the cell line used in our experiments. Additionally, it may be possible that to detect the activity of the BMP cis-acting element it has to be removed form the basal promoter to reveal its silencing activity (1). Finally, although we have emphasized on the transcriptional control of betaglycan gene expression, we are aware that other levels of regulation may exist. This is particularly possible in view of the fact that our betaglycan reporter remains fully active at day 6, even when the levels of its mRNA and cell surface protein have decreased from their levels at day 4. Additionally, the fact that some of the up-regulated betaglycan is being localized to the ECM (Fig. 7) makes it very difficult to fully account for all of the betaglycan induced during differentiation.
The function of betaglycan in skeletal muscle cells and the consequences of betaglycan up-regulation during differentiation are not known and can only be determined experimentally. However, it is reasonable to guess that they will depend on the versatile ability of betaglycan as a co-receptor of diverse growth factors. The literature on the role of TGF-␤ in muscle differentiation is abundant but controversial. TGF-␤ has been shown to be a strong inhibitor of skeletal muscle differentiation (14,59,60) but also to be a promoter of terminal differentiation (61). However, the fact that the expression of a truncated dominant negative T␤RII receptor inhibits the C 2 C 12 differentiation suggests the need for an active TGF-␤ pathway for its differentiation (62). One clue to the role of betaglycan in C 2 C 12 differentiation can be anticipated from our finding that betaglycan performs a TGF-␤ presentation role in nondifferentiated C 2 C 12 myoblasts (Fig. 6). Reasoning based on these data would lead us to conclude that betaglycan induction may be acting to promote the effects of TGF-␤, and therefore it may act to promote the effects of this growth factor on the differentiation process. The regulation of TGF-␤ activity can be more intricate considering that TGF-␤ also binds to other proteoglycans. In myoblasts that do not express decorin, a proteoglycan with high affinity for TGF-␤ (63), we have observed an inhibition of TGF-␤ signaling (14) together with a significant increase in the binding of TGF-␤ to betaglycan, suggesting a competition between both proteoglycans for TGF-␤. 2 On the other hand, although it has not been demonstrated yet, the fact that betaglycan binds FGF-2 via its heparan sulfate chains makes it a candidate to regulate the association of this growth factor with its kinase receptor and therefore its actions (64). In this regard, the effects of betaglycan on muscle cell differentiation may resemble those of the other heparan sulfate proteoglycans that we have studied (5,9,11,27). Another fact to consider is the localization of the up-regulated betaglycan. It is possible to speculate that betaglycan is shed from the surface and incorporated into the ECM serving there as a TGF-␤ and/or FGF-2 reservoir. Thus, betaglycan roles during skeletal muscle differentiation are to be determined. Experiments to test the effects of constitutive betaglycan expression and inhibition of endogenous betaglycan expression in C 2 C 12 myoblasts are in progress.
In summary, we have found that the expression of betaglycan is up-regulated during skeletal muscle differentiation. This phenomenon is modulated by growth factors and RA and is MyoD-dependent. The regulatory sequences responsible for this modulation are contained within a 2100-bp segment of the gene that contains consensus binding sites for several transcriptions factor such as SP1, Smad3, Smad4, MyoD, MEF2, and RA binding proteins. These results would help to understand the complex regulation during development and differentiation of this key integral membrane heparan/chondroitin sulfate proteoglycan.