3795Journal of Cell Science 108, 3795-3805 (1995)Printed in Great Britain © The Company of Biologists Limited 1995JCS4024

Expression of laminin isoforms in mouse myogenic cells in vitro and in vivo

Frank Schuler and Lydia M. Sorokin

Institute for Experimental Medicine, Connective Tissue Research, University of Erlangen-Nuremberg, Schwabachanlage 10, D-91054 Erlangen, Germany

The expression of laminin-1 (previously EHS laminin) andlaminin-2 (previously merosin) isoforms by myogenic cellswas examined in vitro and in vivo. No laminin α2 chain-specific antibodies react with mouse tissues, so rat mono-clonal antibodies were raised against the mouse laminin α2chain: their characterization is described here. Myoblastsand myotubes from myogenic cell lines and primarymyogenic cultures express laminin β1 and γ1 chains andform a complex with a 380 kDa α chain identified aslaminin α2 by immunofluorescence, immunoprecipitationand PCR.

PCR from C2C12 myoblasts and myotubes for thelaminin α2 chain gene (LamA2) provided cDNA sequenceswhich were used to investigate the in vivo expression ofmouse LamA2 mRNA in embryonic tissues by in situhybridization. Comparisons were made with specificprobes for the laminin α1 chain gene (LamA1). LamA2 butnot LamA1 mRNA was expressed in myogenic tissues of 14-

and 17-day-old mouse embryos, while the laminin α2polypeptide was localized in adjacent basementmembranes in the muscle fibres. In situ hybridization alsorevealed strong expression of the LamA2 mRNA in thedermis, indicating that laminin α2 is expressed other thanby myogenic cells in vivo. Immunofluorescence studieslocalized laminin α2 in basement membranes of basal ker-atinocytes and the epithelial cells of hair follicles, providingnew insight into basement membrane assembly duringembryogenesis.

In vitro cell attachment assays revealed that C2C12 andprimary myoblasts adhere to laminin-1 and -2 isoforms ina similar manner except that myoblast spreading was sig-nificantly faster on laminin-2. Taken together, the datasuggest that laminins 1 and 2 play distinct roles in myo-genesis.

Key words: laminin, mouse muscle cell, basement membrane

SUMMARY

INTRODUCTION

Myoblasts migrate in two successive waves from themyotomes of the somites to the limb anlagen and trunk duringmyogenesis in vivo. There they fuse to form the primary andthe secondary myotubes which will form the bulk of the maturemuscle (Rugh, 1968; Christ et al., 1986). Different factorsincluding cytokines (Florini et al., 1991) and the cell adhesionmolecules NCAM, M-cadherin, VCAM-1 and VLA4 (Rosenet al., 1992; Donalies et al., 1991; Hamshere et al., 1991) playcrucial roles in such processes. In addition, the results of invitro experiments suggest that extracellular matrix proteinsinfluence the migration of myoblasts and their development toterminally differentiated myotubes. Primary mouse myoblastspreferentially migrate on mixtures of laminin/collagen type IVcoated onto culture dishes, as opposed to fibronectin orcollagen type I substrates (Kühl et al., 1986; Öcalan et al.,1988; Goodman et al., 1989). Furthermore, on binding tolaminin, but not to fibronectin or collagen, murine myoblastsrapidly differentiate to myotubes (Kühl et al., 1986; von derMark and Öcalan, 1989).

However, an anomaly exists: the laminin isoform used inthese studies was extracted from the mouse Engelbreth-Holm-Swarm (EHS) tumour and is composed of α1, β1 and γ1 chains(nomenclature according to Burgeson et al., 1994; Timpl,

1989). The migratory and myogenic activity of this laminin-1isoform is contained in the α1 chain in a proteolytic fragmentknown as E8 (Deutzmann et al., 1988; Goodman et al., 1987,1989). However, in vivo the basement membranes of maturemuscle fibres do not contain the α1, but rather the laminin α2polypeptide (Leivo and Engvall, 1988; Sanes et al., 1990). Incombination with either β1 and γ1 or β2 and γ1 chains, α2characterizes laminins 2 (merosin) and 4 (S-merosin) (Leivoand Engvall, 1988; Paulsson and Saladin, 1989; Burgeson etal., 1994). The sequence identity of the laminin α2 and α1chains in human and mouse does not exceed 47%(Vuolteenaho et al., 1994; Bernier et al., 1995). The questiontherefore arises: what is the significance of in vitro assays usinglaminin-1 and myogenic cells, and what might be the functionof α2 containing laminins in myogenic tissues.

Myoblasts may migrate on and differentiate in response tothe α1 chain of laminin-1 synthesized by other cells in vivo.However, there is evidence that both during normal develop-ment (Hughes and Blau, 1990) and in muscle regeneration (e.g.muscular dystrophy or muscle damage; Brand-Saberi et al.,1989; Bischoff, 1990), myoblasts migrate along and traversemuscle fibre basement membranes lacking laminin α1. It isalso not clear whether α2-laminins are expressed duringmyogenic differentiation in vivo, in the myotome and limbanlage when myoblasts fuse to form myotubes. To investigate

3796 F. Schuler and L. M. Sorokin

its role in myogenesis, the expression of laminin α2 wasstudied in mouse myogenic cell lines and primary culturesduring differentiation. Laminin-2 was expressed by bothmyoblasts and myotubes in vitro. A 0.5 kb section of the 3′coding region of the gene coding for mouse laminin α2 chain(LamA2) was characterized and used to study the in vivoexpression of LamA2 mRNA in murine embryonic tissues.mRNA expression was compared to the distribution of thelaminin α2 polypeptide using monoclonal antibodies raisedagainst mouse laminin α2 chain.

MATERIALS AND METHODS

Cell culture Mouse myoblast cell lines (C2C12, G7 and G8) were from theAmerican Tissue Culture Collection (access numbers CRL1772,CRL1447, CRL1456). They had the spindle-shaped morphologytypical of myoblasts when maintained at subconfluency in Dulbecco’sminimal essential medium (DMEM) supplemented with 4.5 g/lglucose, 2 mM glutamine and 10% fetal calf serum (FCS). At con-fluency, the myoblasts spontaneously fused to form multinucleatemyotubes, a process promoted by changing the culture medium toDMEM supplemented with 5% horse serum (HS). The % fusion ratio(% nuclei contained in myotubes/total nuclei) was consistently highestin C2C12 (70%). This cell line was, therefore, used to obtain myotubecultures. In addition, primary cultures of myoblasts were preparedfrom the thigh muscle of three day post-natal Balb/c mice as previ-ously described (Kühl et al., 1982, 1986; Goodman et al., 1987). Cellswere grown in DMEM supplemented with 20% HS, 1% chick embryoextract; fusion was induced by reducing the content of HS to 5% andchick embryo extract to 0.5%. Chick embryo extract is a homogenateof 9- to 10-day-old chick embryos diluted 1:1 with serum free DMEMand subsequently centrifuged to pellet debris. A mouse cell linederived from a parietal yolk sac sarcoma (PYS), known to express thelaminin-1 isoform, was also employed (Wewer et al., 1987). TheRugli rat glioblastoma cell line (Goodman et al., 1987; Öcalan et al.,1988) was used in cell attachment assays. Both PYS and Rugli weremaintained in DMEM supplemented with 4.5 g/l glucose, 2 mMglutamine and 10% FCS. All cell lines and primary myogenic cultureswere maintained at 37°C and 7.5% CO2 in a water saturated envi-ronment.

Proteins and peptidesLaminin-1 was isolated from the Engelbreth-Holm-Swarm (EHS)mouse tumour as previously described (Paulsson et al., 1987); recom-binant G1-5 domain of the mouse α1 laminin chain was kindlyprovided by R. Deutzmann (University of Regensburg, Germany); M.Paulsson (University of Cologne, Germany) supplied mouse heartlaminin-2; human laminin-2 was from Gibco.

Polyclonal sera and monoclonal antibodiesAnti-desmin polyclonal antibodies used for the identification ofmyogenic cells were from Eurodiagnostics (Amsterdam). An affinitypurified rabbit antibody against mouse laminin-1 which recognizesα1, β1 and γ1 chains (Klein et al., 1990) was employed in immuno-fluorescence, immunoprecipitation and immunoblotting experiments.Rat monoclonal antibodies raised against the E3 fragment of laminin-1 (nos 198, 200, 201) (Sorokin et al., 1992) were used as markers forthe laminin α1 chain.

To investigate the distribution of the laminin α2 polypeptide inembryonic mouse tissues monoclonal antibodies specific for themouse laminin α2 chain were raised. Lewis rats were immunized withlaminin-2 purified from mouse hearts (Lindblom et al., 1994).Hybridomas were produced as previously described (Sorokin et al.,

1992). Initial selection of hybridomas was based on specific reactionwith mouse laminin-2 and not laminin-1 in ELISA. Further charac-terization of monoclonal antibodies included immunoblotting of thelaminin α2 polypeptide and specific immunoprecipitation of 125I-labelled laminin-2.

ImmunohistochemistryCell lines grown on glass coverslips and 5 µm sections of embryonicor adult mouse tissues were fixed in methanol at −20°C, nonspecificprotein binding was saturated by PBS containing 1% bovine serumalbumin (BSA), and incubations with antibodies were subsequentlycarried out at room temperature. Bound antibodies were visualizedusing goat anti-rat IgG conjugated with FITC and goat anti-rabbit IgGconjugated with rhodamine second antibodies (Dianova GmbH,Hamburg, FRG). Slides were examined under a Zeiss Axiophotmicroscope equipped with epifluorescence.

To investigate the possibility that the epitopes recognized by themonoclonal antibodies specific for laminin α1 or α2 chains weremasked in vivo, the cells/sections were subjected to variousunmasking treatments following fixation. These included 0.1% aceticacid (Merck); 0.1% trypsin (Sigma) at pH 5; 2 ng/ml pepsin(Boehringer Mannheim) at pH 5; 1 mg/ml hyaluronidase (BoehringerMannheim) at pH 5; 0.118 mg/ml neuraminidase (BoehringerMannheim) at pH 5; or 1 U/ml heparitinase (Sigma) at pH 7.5.Immunofluorescence studies were subsequently carried out asdescribed above.

ELISAThe specificity of the newly produced monoclonal antibodies forlaminin-2 was tested in ELISA, as previously described (Dziadek etal., 1983).

ImmunoprecipitationCell cultures were incubated for 1 hour in methionine-free MEM andthen labelled for 4 hours (for immunoprecipitations from cell lysates)or overnight (for immunoprecipitations from culture media) with 50µCi/ml [35S]methionine (Amersham, specific activity 1190 Ci/mmol).Culture medium was collected and employed directly for immuno-precipitation experiments. To ensure efficient extraction of radio-labelled laminin complexes, cells were solubilized overnight in lysisbuffer (150 mM NaCl, 10 mM EDTA, 50 mM Tris-HCl, pH 7.4, 0.1%NP-40, 0.1% SDS, plus 1 mM N-ethylmaleimide and phenylmethyl-sulphonyl fluoride) (Paulsson et al., 1987). Characterization of themonoclonal antibodies raised against mouse laminin-2 also involvedimmunoprecipitations performed with 125I-labelled laminin-1 orlaminin-2. Proteins were labelled using the Iodogen-coated tubemethod (Fraker and Speck, 1978). 35S- or 125I-labelled proteins wereimmunoprecipitated using the affinity purified anti-laminin-1antiserum, laminin α1 or α2 chain specific monoclonal antibodies andProtein G-Sepharose (Pharmacia), and were subsequently analysed bySDS-polyacrylamide gradient gels under reducing conditions. All gelswere fixed with methanol/acetic acid and, in the case of immuno-precipitations of 35S-labelled proteins, gels were enhanced withAmplify TM (Amersham-Buchler), dried and exposed to Kodak XARfilms at −70°C.

ImmunoblottingCell lines were directly dissolved in an electrophoresis sample buffercontaining SDS and DTT (Laemmli, 1970). Approximately 4-5 µg oftotal protein extracted from PYS, myoblast or myotube cultures wasseparated on SDS-polyacrylamide gradient gels under reducing con-ditions and transferred to a nitrocellulose membrane (Towbin et al.,1979). Nonspecific binding to filters was blocked with 3% BSA con-taining 0.1% Tween-20. Filters were incubated for 2 hours at roomtemperature with anti-laminin-1 antiserum, α1 or α2 laminin chainspecific monoclonal antibodies. Bound antibodies were visualized bybiotinylated anti-rabbit IgG or anti-rat IgG, followed by peroxidase-

3797Laminin in muscle cells

coupled streptavidin (Amersham-Buchler) and subsequent chemi-luminescent reaction using the Amersham ECL system. 125I-strepta-vidin and autoradiography was used to visualize bound monoclonalantibodies.

Northern blot analysisTotal RNA was isolated from the myogenic cells as described byChirgwin et al. (1979). Poly(A)+ mRNA was isolated by two passagesthrough an oligo (dT)-cellulose column (Pharmacia mRNA purifica-tion kit, Uppsala, Sweden) and the final poly(A)+ fraction was 1-3%of the total RNA loaded; 15 µg of total RNA or 5 µg mRNA wasdenatured with formamide and electrophoresed on a 1% agarose gel(Maniatis et al., 1982). After transfer (Thomas, 1980) to Hybond-Nfilter (Amersham-Buchler) the filter was fixed by exposure to UVlight. Filters were preincubated for 2 hours at 42°C in 50%formamide, 2.5× Denhardt’s solution, 2× SSC, 0.1% SDS and 50µg/ml herring sperm DNA. Hybridizations were performed in thesame solution for 16-20 hours at 42°C with the following mousecDNA probes: a 1.5 kb cDNA covering nucleotides 7,786-9,286 inthe 3′ end of the coding region of the mouse LamA1 mRNA (LAC)(Deutzmann et al., 1988); a 1.1 kb cDNA corresponding tonucleotides 267-1,367 near the 5′ end of the LamA1 mRNA (cloneLA1) (Hartl et al., 1988) and a 4.5 kb cDNA covering nucleotides3,290-7,770 in the central coding region of the LamA1 mRNA (LA5)(Sasaki et al., 1988); a 1.3 kb cDNA for LamC1 mRNA coveringnucleotides 6,382-7,642 (Sasaki et al., 1988) and for LamB1 mRNA,a 2.7 kb cDNA covering nucleotides 2,580-5,300 (Oberbäumer,1986). Control hybridizations with the same filters were performedusing the 1.2 kb PstI fragment of murine β-actin cDNA clone pAL41(Minty et al., 1983). Probes were labelled to a specific activity of 0.5-1.0×109 cpm/µg cDNA by the primer extension method (Sambrooket al., 1989). Following hybridization, filters were washed severaltimes with a final stringency of 2× SSC, 0.1% SDS at 52-62°C, andexposed to Kodak XAR films at −70°C. mRNA sizes were estimatedusing DNA molecular mass markers (Boehringer).

Reverse transcriptase-polymerase chain reaction (PCR)analysisPCR analysis was employed to investigate the LamA2 mRNA andpossible low expression of LamA1: 5 µg total RNA from G7, G8,C2C12 and PYS was transcribed to cDNA using Moloney murineleukemia virus RNase H− reverse transcriptase (Gibco) and hexamericrandom primers. The cDNA was amplified by Taq DNA polymerase(Appligene) at 1.5 mM MgCl2 for LamA1 primers and at 7.5 mMMgCl2 for the LamA2 primers. The primers used for the amplificationwere constructed from: (1) sequences of the mouse LamA1 mRNAcovering nucleotides 6,430 to 6,907: DNA sense strand 5′CG-CGGTACCGTGTCTGCAGACAGAGACTGC3′; antisense strand5′CACAGGCCAATCCGATTTGCC3′; and (2) sequences of the ratLamA2 mRNA (Engvall et al., 1992): DNA sense strand 5′AAAG-TATCTGTGTCTTCAGGA3′; antisense strand 5′AGTGAATG-TAATCACACGTACAGC3′.

The LamA1 primers were used under the following conditions: 5cycles at 94°C for 0.7 minute, 50°C for 1 minute, 72°C for 1.5minutes, followed by 30 cycles at 94°C for 0.7 minute, 65°C for 1minute and 72°C for 1.5 minutes. LamA2 primers were used asfollows: 30 cycles at 94°C for 0.7 minute, 55°C for 1 minute and 72°Cfor 1.5 minutes. Control PCR reactions were performed using primersconstructed from the rat β-actin sequence. PCR products wereanalysed by electrophoresis in 1-2% agarose gels.

Sense primers contained an EcoRI site and the antisense primerscontained a BamHI site to allow cloning of PCR products into pUC18vectors for subsequent sequence analysis, or into pGEM3-blue vector(Promega Biotec) for in situ hybridization.

In situ hybridizationProbes used for in situ hybridization studies included the LamA2 PCR

product described above, and a 1.5 kb cDNA covering nucleotides7,786-9,286 in the 3′ end of the coding region of the mouse LamA1mRNA (LAC). Mid-sagittal frozen sections of mouse embryos wereemployed for in situ hybridization which was carried out accordingto the method of Aigner et al. (1993). No signals above backgroundlevel were observed in the routinely performed control hybridizationsemploying sense transcripts of comparable specific activity. Slideswere exposed to Ilford K2 emulsion for 10-15 days and after devel-opment were stained with 0.02% toluidine blue. Slides were examinedon a Zeiss Axiophot microscope using bright-field and dark-groundillumination.

Functional assaysTo investigate the roles of laminins 1 and 2 in myoblast attachment,in vitro cell attachment assays were performed as previouslydescribed (Goodman et al., 1987). The Rugli rat glioblastoma cellline, attachment of which on laminin-1 is well characterized, was alsoemployed in these experiments. Microtitre plates were coated withvarying concentrations (0.3-30 µg/ml) of mouse laminin-1, mouselaminin-2, human laminin-2, or BSA as described by Öcalan et al.(1988). After 90 minutes, the number of adhering cells was deter-mined by colorimetric analysis of lysosomal hexosaminidase(Landegren, 1984). The time course of adhesion to laminins wasstudied on plates coated with 10 µg/ml protein.

RESULTS

Characterization of laminin α2 chain-specificmonoclonal antibodiesTo investigate the expression of laminin α2 polypeptide in themouse, monoclonal antibodies were raised against mouselaminin-2. It was necessary to produce specific antibodies sincethose available do not react in mouse and function only inimmunoblots of the purified laminin-2 isoform. Based onspecific reaction with laminin-2 in ELISA, the 4H8-2, 4H8-4,4H8-9 and 8G11D10 hybridomas were chosen. ELISA resultsobtained with 4H8-2 are illustrated in Fig. 1. All monoclonalantibodies immunoprecipitated exclusively 125I-labelled-laminin-2 but not laminin-1, while the laminin α1 chainspecific monoclonal antibody, 198, immunoprecipitated onlylaminin-1 (Fig. 2A). Two antibodies, 4H8-2 and 8G1-D10,were functional in immunoblots of purified laminin-2 and gavedistinct patterns of results: 4H8-2 stained predominantly apolypeptide of 300 kDa, while 8G11-D10 recognized a 150kDa polypeptide (Fig. 2B). It has been shown that a 150 kDafragment is the major proteolytic fragment of the 300 kDalaminin α2 chain (Paulsson et al., 1991). The results, therefore,suggest that 8G11-D10 recognizes this 150 kDa fragment,while 4H8-2 reacts with the intact 300 kDa molecule. Bothantibodies reacted weakly or not at all on cell lysates ofmyoblasts or myotubes suggesting a low affinity for denaturedlaminin α2 chain (results not shown).

In vitro expression of laminin chains by myogeniccellsImmunofluorescence revealed that myoblasts derived from G7,G8, C2C12, and primary myogenic cultures express lamininextracellularly and intracellularly around the nucleus. A similarstaining pattern was seen with the laminin α2 chain specificmonoclonal antibodies, 4H8-2 and 8G11-D10, and is illus-trated in Fig. 3A. Myotubes generated from C2C12 and fromprimary myoblast cultures also express laminin but, in contrast

3798 F. Schuler and L. M. Sorokin

0,0

0,2

0,4

0,6

0,8 4H8-2

0,0

0,2

0,4

0,6

0,8

198

OD

450

nm

0,0

0,2

0,4

0,6

0,8

0,0

LM-1

LM-2G1-5

Polyclonal anti-Laminin-1

Ab Concentration (ug/ml)

1:101 1:100 1:10-1 1:10-2

Fig. 1. ELISA demonstrating specific binding of antibodies to mouselaminins 1 (s) and 2 (m), and to recombinant G1-5 domain of themouse α1 laminin chain (h). The laminin α2 chain specificmonoclonal antibody, 4H8-2, bound only laminin-2. Similar patternswere observed with 4H8-4, 4H8-9 and 8G11D10. Laminin α1 chainspecific monoclonal antibody, 198, bound only to laminin-1 and G1-5, while the polyclonal anti-laminin-1 serum recognized both lamininisoforms and recombinant G1-5.

Fig. 2. (A) Immunoprecipitation of 125I-labelled-laminin-1 (1) andlaminin-2 (2) by monoclonal antibodies specific for laminin α2(4H8-2, 4H8-4, 4H8-9, 8D11D10) or laminin α1(198), and bypolyclonal anti-laminin-1 (anti-LM-1). Polyclonal antibodyimmunoprecipitated laminin-1 (composed of 400 kDa α1 and 200kDa β1/γ1 chains) and laminin-2 (composed of 300 kDa α2, and 200kDa β1/γ1 chains); 400-600 kDa complexes represent β1 and γ1chain multimers in laminin-2 samples. 4H8-2, 4H8-4, 4H8-9,8D11D10 immunoprecipitated exclusively mouse laminin-2, while198 immunoprecipitated only laminin-1. (B) Immunoblots of 4H8-2and 8G10D11 on purified samples of laminin-1 (1) and laminin-2(2). For comparison, a sample of laminin-1 immunoblotted withpolyclonal anti-laminin-1 is shown (anti-LM-1) indicating positionsof β1 and γ1 chains at 200 kDa and the α1 chain at 400 kDa. 4H8-2reacted with the 300 kDa laminin α2 polypeptide, while 8G11D10showed predominant staining of the 150 kDa fragment of the lamininα2 chain. Neither 4H8-2 nor 8G11D10 reacted with laminin-1.

to myoblasts, staining was principally extracellular andoccurred in patches on the cell surface (Fig. 3C,D). Lamininα1 chain specific monoclonal antibodies (198, 200, 201) didnot stain myoblast or myotube cultures, regardless of variouspretreatments designed to unmask inaccessible epitopes (seeMaterials and Methods) (Fig. 3B).

To investigate the chain composition of the laminin isoformsexpressed by myoblasts and myotubes in vitro, immunoblottingwas performed on cell lysates with an anti-laminin-1 polyclonalantibody which reacts with α1, β1 and γ1 chains. Using equalamounts of protein, only polypeptides of approximately 200kDa were detected in myoblast and myotube cell lysates, andno 400 kDa laminin α1 chain was visible (Fig. 4). Controlimmunoblots of adult mouse skeletal and cardiac muscle

extracts using the same polyclonal antibody also revealed the200 kDa bands which probably represent the laminin β1 and γ1chains (results not shown). Polyclonal anti-laminin-1 immuno-precipitated a complex of [35S]methionine-labelled proteinsfrom myoblast and myotube cell lysates and con-ditioned mediawith molecular masses of approximately 380 kDa, 200 kDa and150 kDa (Fig. 5A). The same pattern was obtained with thelaminin α2 chain specific antibodies (shown in Fig. 5B forC2C12 myoblasts). Immunoprecipitations performed withlaminin α1 chain specific monoclonal antibody (198) revealedthe 400 kDa α1 chain and coprecipitated β1 and γ1 chains atapproximately 200 kDa from PYS cell lysates and from condi-tioned medium (Fig. 5C). In myoblast or myotube cell lysatesor conditioned media only a band non-specifically precipitatedat 250 kDa was evident which was shown to be fibronectin byimmunoblotting (results not shown). Fibronectin was also non-specifically precipitated with polyclonal anti-laminin-1 serumand anti-laminin α2 monoclonal antibodies (Fig. 5A,B).

3799Laminin in muscle cells

Fig. 3. Immunofluorescenceof C2C12 myoblast (A) andmyotube (C,D) and mixedmyoblast/myotube (B)cultures using laminin α2chain monoclonal antibody,4H8-2 (A,D), laminin α1chain monoclonal antibody,198 (B), and polyclonal anti-laminin-1 antibody (C). Bar,50 µm.

Northern blot analysis confirmed the expression of thegenes coding for laminin β1 (LamB1) and γ1 (LamC1) chainsin C2C12 myoblast and myotube cultures, and in the controlcell line PYS (Fig. 6). Expression of the laminin α1 chaingene (LamA1) was investigated with three different cDNAprobes which together code for the entire LamA1 mRNA;these failed to reveal a signal on myoblast or myotube mRNAdespite a strong reaction with the 10 kb LamA1 mRNA inPYS (Fig. 6).

Human LamA2 specific primers have been used in RT-PCRof rat cell lines (Engvall et al., 1992) and mouse thymussamples (Chang et al., 1993). RT-PCR with these primersrevealed the predicted 490 bp fragment with RNA from C2C12myoblast and myotube cultures, and new born mouse skeletalmuscle (Fig. 7). In contrast, the LamA2 primers gave no signalfrom the control PYS cell line (results not shown). LamA1primers amplified the predicted 480 bp fragment from PYS

Fig. 4.Immunoblotanalysis ofmyogenic celllysates usingpolyclonal anti-laminin-1.Samples areC2C12 myotubes(lane 1);myoblasts fromC2C12 (lane 2),G7 (lane 3) andG8 (lane 4), PYS

(lane 5), and purified mouse laminin-1 (lane 6). The polyclonalantibody recognizes α1, β1 and γ1 chains of PYS cell extracts andpurified laminin-1, but only 200 kDa β1 and γ1 chains in myoblastsand myotube cell lysates.

cells, but not from G7, G8 (results not shown) or C2C12myoblasts or myotubes (Fig. 7). Subsequent sequence analysisof the cDNA fragments obtained with LamA2 primers revealedhigher homology to the C terminus of the mouse (95%) andhuman (89%) LamA2 mRNAs (Vuolteenaho et al., 1994;Bernier et al., 1995) than to the mouse LamA1 gene (59%)(data not shown).

In vivo expression of laminin chains in myogenictissuesThe biochemical and molecular data described above show thatthe G7, G8, C2C12 cell lines, primary myoblast cultures, andmyotubes derived from C2C12 express the laminin-2 isoformconsisting of β1, γ1 and α2 chains. To investigate whether asimilar expression occurs in myogenic tissues in vivo, in situhybridization and immunofluorescence were employed. Myo-genesis occurs in two waves in the mouse, one at approxi-mately 10-14 days of embryonic development and anotherbetween day 16 of embryonic development and day 3 ofpostnatal life (Christ et al., 1986). Mid-sagittal sections ofmouse embryos of 14 and 17 days of gestation were thereforeexamined. The 490 bp PCR fragment corresponding to the 3′coding region of LamA2 mRNA was used for in situ hybridiz-ation. For comparison, in situ hybridization was also performedwith the LAC cDNA which codes for the equivalent region ofLamA1 mRNA (Sasaki et al., 1988).

In both day 14 and 17 embryos, LamA2 and not LamA1mRNA was expressed in the newly forming muscles of thelimbs and torso and in cardiac muscle. Fig. 8 illustrates thesimilar distribution pattern of desmin-positive myogenic cellsand LamA2 positive cells in the rib region of a day 17 embryo(compare Fig. 8B,E). The distribution of the laminin α2polypeptide closely paralleled this pattern as revealed byimmunofluorescent staining of muscle fibre basementmembranes by 4H8-2 (Fig. 8D), but not 198, 200 or 201 (Fig.

3800 F. Schuler and L. M. Sorokin

Fig. 5. Immunoprecipitation of [35S]methionine labelledlaminin complexes from conditioned medium of PYS, andC2C12 myoblast (Mb) and myotube (Mt) cultures usingpolyclonal anti-laminin-1 (A), monoclonal antibodiesspecific for laminin α2, 4H8-2 (B), and α1, 198, (C), andwithout primary antibody (−). (A) Polyclonal anti-laminin-1 immunoprecipitated β1/γ1 laminin chains ofapproximately 200 kDa, and a 380 kDa polypeptide frommyogenic cells. The same serum precipitated α1, β1 andγ1 laminin chains from the control PYS cell line. (B) Immunoprecipitation of myoblasts with 4H8-2revealed β1/γ1 laminin chains and the 380 kDapolypeptide, confirming the identity of the 380 kDa bandwith laminin α2. The same results were found withmyotubes. For comparison, immunopreciptation oflaminin-1 from PYS using polyclonal anti-laminin-1 isshown. (C) 198 immunoprecipitated laminin α1, β1 andγ1 chains only from PYS and showed no specificimmunoprecipitation of laminin complexes frommyoblasts; the same pattern of results was observed withmyotubes. Non-specific precipitation of fibronectin (FN)from myoblast and myotube conditioned medium wasevident when immunoprecipitations were performed witheither polyclonal anti-laminin-1 or monoclonal antibodiesand is probably due to non-specific reaction with ProteinG-Sepharose.

8C). In both embryonic and mature muscle, the distribution ofthe laminin α2 polypeptide was similar and was restricted tobasement membranes of muscle fibres and peripheral nerves(Fig. 9A,B). Axons of peripheral nerves are ensheathed bymyelinating or nonmyelinating Schwann cells and surroundedby the endoneurial basement membrane. Bundles of axon-Schwann cell units are encircled by a multilamellar cellularsheath called the perineurium which is also bounded by abasement membrane. Laminin α2 chain specific monoclonalantibodies stained the endoneurial but not the perineurialbasement membranes of the peripheral nerves in immunofluo-rescence (Fig. 9B). No laminin α2 polypeptide was detected in

Fig. 6. Northern blotanalysis of genes codingfor laminin α1 (LamA1),β1 (LamB1), γ1 (LamC1)and β-actin in C2C12myoblasts (Mb), C2C12myotubes (Mt) and PYScells. The 10 kb LamA1mRNAwas detected withthe LAC probe in PYScells only. LamB1 andLamC1 specific probeshybridized to the expected6 kb and 8 kb mRNAs,respectively, in myogenicand PYS cells.

blood vessel basement membranes (Fig. 9C), as previouslyreported (Sorokin et al., 1994).

In situ hybridization also revealed expression of LamA2mRNA in various mesenchymal tissues (our unpublished data).In particular, LamA2 was strongly expressed in the mesenchy-mal cells of the dermis of the skin (Fig. 10A). At both stagesof development examined, this expression was widespreadthroughout the dermis with some concentration in mesenchy-

Fig. 7. PCR analysis of LamA1 and LamA2 gene expression inmyogenic cells. mRNA was isolated from C2C12 myoblasts (Mb)and myotubes (Mt), new born mouse skeletal muscle (M) and thecontrol PYS cell line. RT-PCR using primers specific for the genescoding for laminin α1 and α2 revealed only LamA2 gene sequencesin myogenic cells and LamA1 gene sequences only in the controlPYS cell line. Φ, DNA size markers.

3801Laminin in muscle cells

Fig. 8. In situ hybridization of LamA1 (A) and LamA2 (B) probes withembryonic mouse skeletal muscle and correspondingimmunofluorescent localization of laminin α1(C) and α2 (D)polypeptides, and the myogenic marker, desmin (E). Results shown arefor a day 17 mouse embryo, but the same pattern of results wasobserved for day 14 embryos. The LamA2 (B) and not the LamA1 probe(A) hybridized to the intercostal muscles between the ribs (r) but not tothe underlying liver (l). A similar distribution pattern was observed fordesmin (E). Immunofluorescence with 4H8-2 localized laminin α2polypeptide in basement membranes of muscle fibres (D), whilelaminin α1 specific monoclonal antibody, 198, showed no reaction inembryonic muscle (C). LamA1 or LamA2 specific probes did nothybridize to sections of mature mouse skeletal muscle. Bar, 20 µm.

mal cells at the base of hair follicles. Immunofluorescencelocalized the laminin α2 polypeptide in basement membranesunderlying the basal keratinocytes of the epidermis and sur-rounding the epithelial follicle cells (Fig. 10C). In contrast,LamA1 mRNA expression was weak in the skin of 14- and 17-day-old embryos and was restricted to the epithelial cells of theepidermis (Fig. 10B). It was often difficult to interpret whetherexpression of LamA1 occurred in the epidermal cells becausemost probes concentrated at the edge of the section. However,the laminin α1 polypeptide was found in basement membranesof the basal keratinocytes of the epidermis (Fig. 10D), sug-gesting that hybridization of the LAC probe to the epidermiswas indeed higher than background. In addition, laminin α1polypeptide was detected in basement membranes of theepithelial cells constituting hair follicles and glands, but not inendothelial cell basement membranes (results not shown).

Functional assaysSince laminin-1 plays a significant role in myoblast attachmentand differentiation, the ability of C2C12 and primarymyoblasts to attach to laminin-2 as compared to laminin-1 was

investigated using in vitro assays. The Rugli glioblastoma cellline was also employed since the attachment of this and similarglioblastoma cell lines (Engvall et al., 1992; Brown and Timpl,1994) to mouse laminin-1 and human laminin-2 substrates hasbeen characterized. Hence, Rugli in attachment assays servedas an internal control for the integrity of the laminin substrates.C2C12 myoblasts and primary myoblasts bound with similaraffinities to both laminin isoforms regardless of the species oforigin (Fig. 11A). In contrast, Rugli cells bound to mouselaminin-1 and human laminin-2 with higher affinity than tomouse laminin-2 (Fig. 11A). These differences may be due tospecies differences in laminin-2 substrate. The time course ofmyoblast attachment was comparable between all threelaminin sources, but the time required for cell spreading variedconsiderably according to substrate: Myoblasts spread within5-10 minutes of incubation on mouse or human laminin-2 (10µg/ml), while on the same concentration of mouse laminin-1they maintained a rounded morphology for up to 20-30 minutesof incubation (Fig. 11B). The time required for spreading ofRugli cells did not vary with laminin substrate, and rangedfrom 20-40 minutes of incubation.

3802 F. Schuler and L. M. Sorokin

Fig. 9. Immunofluorescence staining of maturemouse skeletal muscle sections by polyclonal anti-laminin-1 (A), or monoclonal antibodies specificfor laminin α2, 4H8-2 (B), laminin α1, 198 (C).Comparison of polyclonal anti-laminin-1 and 4H8-2 staining patterns localizes laminin α2polypeptide in basement membranes of musclefibres, and in endoneurial but not perineurial(arrows in A) basement membranes of peripheralnerves. No laminin α2 was detectable in basementmembranes of arteries (a) or veins (v). Bar, 50 µm.

Fig. 10. In situhybridization ofLamA2 (A) andLamA1 (B) probes toa section through thelimb of a 17-day-oldmouse embryo.LamA2 expressionwas detected inskeletal muscle layers(m), but also inmesenchymal cells ofthe dermis (d) of theskin. Note thatepithelial cells of hairfollicles and theepidermis do notexpress LamA2(arrowheads).Immunofluorescencelocalized the lamininα2 polypeptide (C) inbasement membranesof the basalkeratinocytes andepithelial cells of thehair follicles(arrowheads). TheLamA1 specific probe

hybridized only weakly to epidermal cells (B), while laminin α1 polypeptide occurred in basement membranes of the basal keratinocytes (D).The same pattern of results was observed for 14-day-old embryonic mice. Bar, 20 µm.

3803Laminin in muscle cells

Fig. 11. (A) In vitro attachment of C2C12 myoblasts and the Rugli glioblastoma cell line to varying concentrations of mouse laminins 1 (s)and 2 (m), and human laminin-2 (j). Cell attachment was measured after 90 minutes of incubation with the substrates. (B) The degree ofmyoblast spreading after 5 minutes incubation on 10 µg/ml mouse laminin-1, mouse laminin-2 or human laminin-2.

DISCUSSION

Immunofluorescence, immunoprecipitation and immunoblotexperiments revealed that the major laminin isoform expressedby both myoblasts and myotubes in vitro is laminin-2 (previ-ously merosin), whilst laminin-1 is not expressed. This wasconfirmed by northern blot analysis which demonstrated theexpression of the genes coding for laminin β1 (LamB1) and γ1(LamC1) chains, but not for the α1 chain (LamA1). Immuno-precipitation of radiolabelled laminins from myogenic celllysates and from conditioned medium revealed that laminin β1and γ1 chains were complexed with a 380 kDa α chain. Simi-larities in molecular mass suggest that this polypeptide is thelaminin α2 chain. The recent sequencing of the mouse lamininα2 gene (LamA2) revealed a mRNA coding for a protein witha predicted molecular mass of 390 kDa (Bernier et al., 1995)that may be cleaved into a 300 kDa N-terminal segment and a80 kDa C-terminal segment (Ehrig et al., 1990; Paulsson et al.,1991). The presence of laminin α2 in myogenic cells was sub-stantiated by the monoclonal antibodies 4H8-2 and 8G19-D10,raised against mouse laminin-2 and shown here to be specificfor the laminin α2 chain. Both monoclonal antibodies stainedmyoblasts and myotubes in vitro and immunoprecipitatedradiolabelled laminin complexes from these cells. In addition,RT-PCR of myoblast and myotube mRNA using primersspecific for the LamA2 gene resulted in a 490 bp fragment withhigher homology to the human (89%) and mouse (95%)LamA2 genes than to the mouse LamA1 gene (59%).

In situ hybridization demonstrated the expression of LamA2but not LamA1 mRNA in developing myogenic tissues of day

14 and 17 embryonic mice, while immunofluorescencelocalized the laminin α2 polypeptide in muscle fibre basementmembranes at both stages of development. In contrast, thelaminin α1 polypeptide was not expressed in developingmuscle of the embryos, indicating that the α2 and not the α1laminin chain is present when myogenic differentiation andmigration occurs in vivo. In skeletal or cardiac muscle of adultmice, however, expression of LamA2 mRNA was notdetectable. The corresponding expression pattern of lamininα2 polypeptide in these tissues revealed its localization in thebasement membranes of the mature muscle fibres and inendoneurial but not perineurial basement membranes ofperipheral nerves.

For want of appropriate tools, few studies have attempted toanalyse the chain composition of laminins expressed bymyogenic cells in vitro. Only Kroll et al. (1994) have made theattempt using MyoD-transfected C3H10T1/2 fibroblasts.Transfection with MyoD was associated with induction of cellsalong the myogenic lineage and, in contrast to the datapresented here, expression of a polypeptide designated lamininAc3H (350 kDa) which the authors state is distinct fromlaminin α2. In addition, the laminin α1 chain was shown to beexpressed by various clones of cells arising from MyoD-trans-fected cultures. A likely explanation for such discrepancies isdifferences in the types of cells used. It cannot yet be excluded,however, that the 350 kDa laminin chain expressed by theMyoD transfected cells is not the laminin α2 chain, sinceimmunoblots or PCR experiments were not performed toinvestigate the possible expression of the laminin α2 polypep-tide or mRNA.

3804 F. Schuler and L. M. Sorokin

The expression of laminin-2 by myoblasts and myotubes invitro is consistent with the predominant distribution of thislaminin isoform in mature myogenic tissues of man, rat, rabbitand guinea pig (Ehrig et al., 1990; Sanes et al., 1990). The invivo expression of laminin α2 in basement membranes ofmature muscle fibres and peripheral nerves observed here isalso comparable to those previously described in rabbit (Ehriget al., 1990; Sanes et al., 1990). However, the present study isfirst to correlate expression of the LamA2 mRNA and its cor-responding protein in embryonic muscle. Using an antibodydirected to the 80 kDa fragment of human α2 laminin chain(4F11), Leivo and Engvall (1988) investigated expression ofthe laminin α2 polypeptide in skeletal muscle of day 15embryos and new born rats. In contrast to the data presentedhere, laminin α2 polypeptide was detected only in new bornrat muscle. Such discrepancies may relate to species differ-ences or, more likely, to antibodies employed and the devel-opmental stages investigated. It is unlikely that the epitope rec-ognized by 4F11 is masked in tissues, since various unmaskingtechniques were used. Rather, the affinity of this anti-humanantibody for the rat epitope may be low so that significantstaining was observed only when sufficient laminin α2 chainwas deposited into the basement membranes at relatively latestages of development. We found here that the intensity ofimmunofluorescent staining for α2 laminin is indeed consid-erably weaker in embryonic than in mature muscle. Further-more, in vitro laminin is organized in patches on the myotubecell surface and only with fibre maturation does the basementmembrane form a continuous envelope (Kühl et al., 1982). Invivo, the muscle fibres continue development after birth (Christet al., 1986) and laminin is being deposited into the formingbasement membranes throughout fibre development.

The expression of α2-containing laminins in embryonicmuscle is not incongruous with a role for the laminin α1 chainin myogenesis, as suggested by in vitro studies (Öcalan et al.,1988; Kühl et al., 1986; Goodman et al., 1989). It is possiblethat myoblasts migrate to the sites in the body where they willeventually form muscle by adhering to α1 chain containinglaminins expressed by other cells. This is supported by theobservation made in the present study that both laminins 1 and2 support myoblast attachment to similar extents in in-vitroassays, and that the affinity of myoblast binding to bothisoforms is comparable. However, the observation thatmyoblasts spread more quickly on mouse or human laminin-2, as compared to laminin-1, may hint that the two isoformsare functionally distinct. Myoblasts bind to laminin-1 via theα7β1 integrin (von der Mark et al., 1991). At present, no con-clusions can be made on whether the same receptor mediatesbinding to laminin-2. However, the spreading results show thatthe intracellular signals which arise from myoblast adhesion tolaminins 1 and 2 differ. It has been shown that muscle fibrescan attach to the basement membrane via the laminin α2 chainand α-dystroglycan on the surface of muscle fibres, and theabsence of the laminin α2 chain from muscle fibre basementmembranes in congenital muscular dystrophies results inmuscle degeneration (Sunada et al., 1994; Xu et al., 1994,1995). These results suggest that laminin α2 interactions withα-dystroglycan are essential for muscle fibre stability and theconsequent maintenance of the myogenic phenotype.Therefore, laminin-2 may play a more significant role inmyoblast differentiation.

Apart from its expression in myogenic cells, LamA2 mRNAis strongly expressed in mesenchymal cells, in particular thoseof the dermis of the skin. Clearly, non-myogenic cells alsohave the capacity to secrete α2 containing laminins. Thissupports the mRNA data reported by Vuolteenaho et al. (1994)for human fetal tissues. Similarly, Simon-Assman et al. (1994)have reported that laminin-2 is located in the basementmembrane of intestinal epithelial cells in the developing rat,surrounding the base of the crypts where cell proliferation anddifferentiation occur. Here we correlate the distribution of thelaminin α2 mRNA and polypeptide, a comparison whichshowed that although LamA2 mRNA is expressed by mes-enchymal cells, the protein is deposited into adjacent basementmembranes. This is true not only of the skin, but also of otherorgans including lung, kidney and testis (our unpublisheddata). Such observations provide new information on theassembly of basement membranes during embryogenesis,since similar data is available only for the laminin α1 polypep-tide and mRNA both of which are expressed principally byepithelial cells. The restricted deposition of laminin α2polypeptide in comparison to the widespread expression of itsmRNA suggests successful expression only in the vicinity ofthe basal keratinocytes and hair follicles where laminin α2 isdeposited into basement membranes. Alternatively, a balancebetween protein synthesis and degradation may be involved inestablishing the observed laminin α2 distribution pattern.

In conclusion, the data presented here show that laminin α2containing isoforms are expressed by myoblasts and myotubesin vitro and in vivo, and that laminins 1 and 2 play distinctroles in myogenesis in the mouse. In addition, the in vivoexpression of LamA2 mRNA and its polypeptide provides newinformation on mechanisms of basement membrane assemblyduring embryonic development.

This work was supported by Deutsche Forschungsgemeinschaftgrant So285/3-1 to L. M. Sorokin and would not have been possiblewithout the expert technical assistance of Barbara Dorn, and thesupport of Prof. K. von der Mark. The authors thank Dr M. Paulssonfor providing mouse laminin-2, Dr R. Deutzmann for recombinantG1-5, and Dr R. Hallmann and Dr E. Palombo-Kinne for criticalreading of the manuscript.

REFERENCES

Aigner, T., Bertling, W., Stoss, H., Weseloh, G. and von der Mark, K.(1993). Independent expression of fibril-forming collagens I, II, and III inchrondrocytes of human osteoarthritic cartilage. J. Clin. Invest. 91, 829-837.

Bernier, S. M., Utani, A., Sugiyama, S., Doi, T., Polistina, C. and Yamada,K. (1995). Cloning and expression of laminin α2 chain (M-chain) in themouse. Matrix Biol. 14, 447-455.

Bischoff, R. (1990). Interactions between satellite cells and skeletal musclefibres. Development 109, 943-952.

Brand-Saberi, B., Krenn, V. and Christ, B. (1989). The control of directedmyogenic cell migration in the avian limb bud. Anat. Embryol. 180, 555-566.

Brown, J. C. and Timpl, R. (1994). Protein binding and cell adhesionproperties of two laminin isoforms (AmB1eB2e, AmB1sB2e) from humanplacenta. J. Cell Sci. 107, 329-338.

Burgeson, R. E., Chiquet, M., Deutzmann, R., Ekblom, P., Engel, J.,Kleinman, H., Martin, G. R., Meneguzzi, G., Paulsson, M., Sanes, J.,Timpl, R., Tryggvason, K., Yamada, Y. and Yurchenco, P. D. (1994). Anew nomenclature for laminin. Matrix Biol. 14, 209-211.

Chang, A. C., Wadsworth, S. and Coligan, J. E. (1993). Expression ofmerosin in the thymus and its interaction with thymocytes. J. Immunol. 151,1789-1801.

3805Laminin in muscle cells

Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. and Rutter, W. J.(1979). Isolation of biologically active ribonucleic acid from sourcesenriched in ribonuclease. Biochemistry 18, 5294-5299.

Christ, B., Jacob, M., Jacob, H. J., Brand, B. and Wachtler, F. (1986). InSomites and Developing Embryos. pp. 261-276. New York: Plenum.

Deutzmann, R., Huber, J., Schmetz, K. A., Oberbäumer, I. and Hartl, L.(1988). Structural study of long arm fragments of laminin. Evidence forrepetitive C-terminal sequences in the A chain, not present in the B chains.Eur. J. Biochem. 177, 35-45.

Donalies, M., Cramer, M., Ringwald, M. and Starzinski-Powitz, A. (1991).Expression of M-cadherin, a member of the cadherin multigene family,correlates with differentiation of skeletal muscle cells. Proc. Nat. Acad. Sci.USA 88, 8024-8028.

Dziadek, M., Richter, H., Schachner, M. and Timpl, R. (1983). Monoclonalantibodies used as probes for the structural organization of the central regionof fibronectin. FEBS Lett. 155, 321-325.

Ehrig, K., Leivo, I., Argraves, W. S., Ruoslahti, E. and Engvall, E. (1990).Merosin: a tissue-specific basement membrane protein is a laminin-likeprotein. Proc. Nat. Acad. Sci. USA 87, 3264-3268.

Engvall, E., Earwicker, D., Day, A., Muir, D., Manthorpe, M. andPaulsson, M. (1992). Merosin promotes cell attachment and neuriteoutgrowth and is a component of the neurite-promoting factor of RN22Schwannoma cells. Exp. Cell Res. 198, 115-123.

Florini, J. R., Ewton, D. Z. and Magri, K. A. (1991). Hormones, growthfactors and myogenic differentiation. Annu. Rev. Physiol. 53, 201-216..

Fraker, P. J. and Speck, J. C. (1978). Protein and cell membrane iodinationswith a sparingly soluble chloroamide, 1,3,4,6-tetrachloro-3a,6a-diaphenylglycoluril. Biochem. Biophys. Res. Commun. 80, 849-857.

Goodman, S., Deutzmann, R. and von der Mark, K. (1987). Two distinct cellbinding domains in laminin can independently promote non-neural celladhesion and spreading. J. Cell Biol. 105, 589-598.

Goodman, S. L., Risse, G. and von der Mark, K. (1989). The E8 fragment oflaminin promotes the locomotion of myoblasts over extracellular matrix. J.Cell Biol. 109, 799-809.

Hamshere, M., Dickson, G. and Eperon, I. (1991). The muscle specificdomain of mouse N-CAM: structure and alternative splicing patterns. Nucl.Acid Res. 19, 4709-4716.

Hartl, L., Oberbäumer, I. and Deutzmann, R. (1988). The N-terminus oflaminin A chain is homologous to the B chains. Eur. J. Biochem. 173, 629-635.

Hughes, S. M. and Blau, H. M. (1990). Migration of myoblasts across basallamina during skeletal muscle development. Nature 345, 350-353.

Klein, G., Ekblom, M., Fecker, L., Timpl, R. and Ekblom, P. (1990).Differential expression of laminin A and B chains during development ofembryonic mouse organs. Development 110, 823-837.

Kroll, T. G., Peters, B. P., Marzias Hustad, C., Jones, P. A., Killen, P. D.and Ruddon, R. W. (1994). Expression of laminin chains during myogenicdifferentiation. J. Biol. Chem. 269, 9270-9277.

Kühl, U., Timpl, R. and von der Mark, K. (1982). Synthesis of type IVcollagen and laminin in culture of skeletal muscle cells and their assembly onthe surface of myotubes. Dev. Biol. 93, 344-354.

Kühl, U., Öcalan, M., Timpl, R. and von der Mark, K. (1986). Role oflaminin and fibronectin in selecting myogenic verus fibrogenic cells fromskeletal muscle cells in vitro. Dev. Biol. 118, 628-635.

Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly ofthe head of bacteriophage T4. Nature 227, 680-685.

Landegren, U. (1984). Measurement of cell numbers by means of theendogenous enzyme hexosaminidase. Applications to detection oflymphokines and cell surface antigens. J. Immunol. Meth. 67, 379-388.

Leivo, I. and Engvall, E. (1988). Merosin, a protein specific for basementmembranes of Schwann cells, striated muscle, and trophoblast, is expressedlate in nerve and muscle development. Proc. Nat. Acad. Sci. USA 85, 1544-1548.

Lindblom, A., Engel, J. and Paulsson, M. (1994). Characterization of nativelaminin from bovine kidney and comparisons with other laminin variants.Eur. J. Biochem. 219, 383-392.

Maniatis, T., Frisch, E. F. and Sambrook, J. (1982). Molecular Cloning: ALaboratory Manual. New York, Cold Spring Harbor: Cold Spring HarborLaboratory Press.

Minty, A. J., Alonso, S., Guenet, J. L. and Buckingham, M. E. (1983).Number and organization of actin-related sequences in the mouse genome. J.Mol. Biol. 167, 77-101.

Oberbäumer, I. (1986). New pUC-derived expression vectors for rapidconstruction of cDNA libraries. Gene 49, 81-91.

Öcalan, M., Goodman, S. L., Kühl, U., Hauschka, S. D. and von der Mark,K. (1988). Laminin alters cell shape and stimulates motility and proliferationof murine skeletal myoblasts. Dev. Biol. 125, 158-167.

Paulsson, M., Aumailley, M., Deutzmann, R., Timpl, R., Beck, K. andEngel, J. (1987). Laminin-nidogen complex: extraction with chelatingagents and structural characterization. Eur. J. Biochem. 166, 11-19.

Paulsson, M. and Saladin, K. (1989). Mouse heart laminin. J. Biol. Chem.264, 18726-18732.

Paulsson, M., Saladin, K. and Engvall, E. (1991). Structure of lamininvariants. The 300kDa chains of murine and bovine heart laminin are relatedto the human placental merosin heavy chain and replace the A chain in somelaminin variants. J. Biol. Chem. 266, 17545-17551.

Rosen, G. D., Sanes, J. R., LaChance, R., Cunningham, J. M., Roman, J.and Dean, D. C. (1992). Roles for the integrin VLA-4 and its counterreceptor VCAM-1 in myogenesis. Cell 69, 1107-1119.

Rugh, R. (1968). The Mouse: Its Reproduction and Development. Burgess,Minneapolis, MN.

Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989). Molecular Cloning: ALaboratory Manual. 2nd edition. Cold Spring Harbor, New York: ColdSpring Harbor Laboratory Press.

Sanes, J. R., Engvall, E., Butkowski, R. and Hunter, D. D. (1990). Molecularheterogeneity of basal laminae: Isoforms of laminin and collagen IV at theneuromuscular junction and elsewhere. J. Cell Biol. 111, 1685-1699.

Sasaki, M., Kleinman, H., Huber, H., Deutzmann, R. and Yamada, Y.(1988). Laminin, a multidomain protein. The A chain has a unique domainand homology with the basement membrane proteoglycan and the laminin Bchains. J. Biol. Chem. 263, 16536-16544.

Simon-Assman, P., Duclos, B., Orian-Rousseau, V., Arnold, C., Engvall, E.and Kedinger, M. (1994). Differential expression of laminin isoforms andthe α6β4 integrin subunits in the developing human and mouse intestine.Dev. Dynamics 201, 71-85.

Sorokin, L. M., Conzelmann, S., Ekblom, P., Aumailley, M., Battaglia, C.and Timpl, R. (1992). Monoclonal antibodies against laminin A chainfragment E3 and their effects on cell- and heparin-binding, and on kidneydevelopment. Exp. Cell Res. 201, 137-144.

Sunada, Y., Bernier, S. M., Kozak, C. A., Yamada, Y. and Campbell, K.(1994). Deficiency of merosin in dystrophic dy mice and genetic linkage oflaminin M chains gene to dy locus. J. Biol. Chem. 269, 13729-13732.

Thomas, P. S. (1980). Hybridization of denatured RNA and small DNAfragments transferred to nitrocellulose. Proc. Nat. Acad. Sci. USA 77, 5201-5202.

Timpl, R. (1989). Structure and biological activities of basement membraneproteins. Eur. J. Biochem. 180, 487-502.

Towbin, H., Staehelin, T. and Gordon, J. (1979). Electrophoretic transfer ofproteins from polyacrylamide gels into nitrocellulose sheets. Procedure andsome applications. Proc. Nat. Acad. Sci. USA 76, 4350-4354.

von der Mark, K. and Öcalan, M. (1989). Antagonistic effects of laminin andfibronectin on the expression of the myogenic phenotype. Differentiation 40,150-157.

von der Mark, H., Dürr, J., Sonnenberg, A., von der Mark, K., Deutzmann,R. and Goodman, S. (1991). Skeletal myoblasts utilize a novel β1-seriesintegrin and not α6β1 for binding to E8 and T8 fragments of laminin. J. Biol.Chem. 266, 23593-23601.

Vuolteenaho, R., Nissinen, M., Sainio, K., Byers, M., Eddy, R., Hirvonen,H., Shows, T., Sariola, H., Engvall, E. and Tryggvason, K. (1994). Humanlaminin M chain (merosin): Complete primary structure, chromosomalassignment, and expression of the M and A chain in human fetal tissues. J.Cell Biol. 124, 381-394.

Wewer, U. M., Damjanov, A., Paulsson, M. and Damjanov, I. (1987).Distinct antigenic characteristics of murine parietal yolk sac laminin. Dev.Biol. 121, 397-407.

Xu, H., Christmas, P., Wu, X. R., Wewer, U. and Engvall, E. (1994).Defective muscle basement membrane and lack of M-laminin in thedystrophic dy/dy mouse. Proc. Nat. Acad. Sci. USA 91, 5572-5576.

Xu, H., Wu, X. R., Wewer, U. and Engvall, E. (1995). Murine musculardystrophy caused by a mutation in the laminin α2 (LamA2) gene. HumanGenet. 8, 297-302.

(Received 6 July 1995 - Accepted 8 September 1995)