sign in sign up
<<
Home , Myogenin, Skeletal muscle, Type II keratin, Type I keratin

RESEARCH ARTICLE 513 α-Skeletal induces a subset of muscle independently of muscle differentiation and withdrawal from the cycle

Peter W. Gunning1,3,4, Vicki Ferguson1,3, Karen J. Brennan2,* and Edna C. Hardeman2,‡ 1Cell Unit and 2Muscle Development Unit, Children’s Medical Research Institute, Locked Bag 23, Wentworthville, NSW, 2145, Australia 3Oncology Research Unit, The New Children’s Hospital, PO Box 3515, Parramatta, NSW, 2124, Australia 4Department of Paediatrics and Child Health, University of Sydney, Sydney, NSW 2006, Australia *Present address: EMBL Heidelberg, Meyerhofstraße 1, Postfach 102209, D-69012 Heidelberg, Germany ‡Author for correspondence (e-mail: [email protected])

Accepted 14 November 2000 Journal of Cell Science 114, 513-524 © The Company of Biologists Ltd

SUMMARY

Muscle differentiation is characterized by the induction of characteristic of muscle differentiation were induced. genes encoding contractile structural and the Stable transfectants displayed a substantial reduction repression of nonmuscle isoforms from these families. in cell surface area and in the levels of nonmuscle We have examined the importance of this regulated order tropomyosins and β-actin, consistent with a relationship of by expressing the two sarcomeric muscle between the composition of the actin and cell characteristic of the differentiated state, i.e. α- surface area. The transfectants displayed normal cell cycle skeletal and α-cardiac actin, in C2 mouse myoblasts. progression. We propose that α-skeletal actin can activate Precocious accumulation of transcripts and proteins for a regulatory pathway linking a subset of muscle genes that a group of differentiation-specific genes was elicited by operates independently of normal differentiation and α-skeletal actin only: four muscle tropomyosins, two withdrawal from the cell cycle. muscle actins, and MyoD. The nonmuscle isoforms of and actin characteristic of the undifferentiated state continued to be expressed, and no Key words: Muscle, Actin, Differentiation, Contractile , heavy or light chain or transcripts Isoform

INTRODUCTION Munsterberg, 1994; Olson and Klein, 1994). The four members of this family, MyoD, Myf5, and MRF4, have At the onset of , mononucleated myoblasts been shown to be sufficient to induce gene withdraw from the cell cycle and fuse to form multinucleated transcription when ectopically expressed in a variety of cell myotubes. This profound change in cell shape is accompanied types. MyoD specifically has been shown to be involved in by a downregulation of the nonmuscle isoforms of the actin, myogenic differentiation in at least a two-step process. tropomyosin (Tm) and myosin genes and an induction of the Initially, MyoD causes myoblasts to withdraw from the corresponding skeletal muscle isoforms (Devlin and Emerson, cell cycle without initiating differentiation (Crescenzi et al., 1978; Buckingham and Minty, 1983; Wade and Kedes, 1989). 1990; Sorrentino et al., 1990). Its subsequent activation of the The induction of these contractile proteins drives the assembly cyclin-dependent kinase inhibitor p21 coordinates terminal of the sarcomeric structures that allow contraction of the withdrawal from the cell cycle, resulting in differentiation . The contractile proteins accumulate with a fixed (Halevy et al., 1995; Parker et al., 1995; Guo et al., 1995). The stoichiometry that reflects their physical relationships in the current model for myogenesis (Zhang et al., 1995), based on (Devlin and Emerson, 1978). Despite the extensive gene targeting of the bHLH factors in mice (Rudnicki et al., isoform switching in each contractile protein gene family that 1992; Braun et al., 1992; Rudnicki et al., 1993; Hasty et al., accompanies muscle maturation (Sutherland et al., 1991; Esser 1993; Nabeshima et al., 1993; Venuti et al., 1995), indicates et al., 1993), the relative total output between gene families that MyoD and Myf5 establish myoblasts and activate remains similar at different stages of maturation (Wade et al., myogenin. Myogenin in turn is involved in myotube formation 1990). This suggests that regulation of output between gene and activation of muscle structural genes. Therefore, data from families may contribute to the maintenance of stoichiometry. both cell culture studies and transgenic mice suggests that the Such a mechanism, however, may require communication bHLH factors function prior to the expression of muscle pathways to directly link expression between the different gene structural protein genes. families. These studies have provided a hierarchical view of Our understanding of muscle differentiation has been myogenesis in which the master regulators activate muscle dominated by the discovery of the myogenic family of genes, but do not respond to, or monitor, the function of these bHLH transcription factors (Weintraub, 1993; Lassar and muscle products. Studies of structural protein gene expression 514 JOURNAL OF CELL SCIENCE 114 (3) in a variety of systems including muscle provide convincing undifferentiated myoblasts. Recent evidence for the presence evidence that feedback regulatory pathways operate in these of actin in the nucleus and the potential roles it plays in nuclear gene families. Gene transfection studies have shown that functions suggests that specific actin isoforms can influence fibroblasts and myoblasts usually regulate nonmuscle actin gene expression (Rando et al., 2000). Alternatively, synthesis (Leavitt et al., 1987a; Ng et al., 1988) and mRNA transcriptional regulation by actin isoforms could occur via the levels (Lloyd et al., 1992) in order to maintain a constant actin G-actin pool; modulations in G-actin regulate a subset of SRF pool size. The nonmuscle actin monomer pool size has been target genes (Sotiropoulos et al., 1999). directly implicated as a regulator of actin gene expression (Bershadsky et al., 1995). Recently, the G-actin level was shown to control (SRF), a MATERIALS AND METHODS transcriptional regulator of nonmuscle actins, serum-inducible and muscle-specific genes (Sotiropoloulos et al., 1999). A DNA constructs α naturally occurring mutation in the α-cardiac actin (αca-actin) The human sk-actin gene was subcloned as a 9.5 kb HindIII fragment gene in some mouse strains reduces output from this in into pBR322 (Taylor et al., 1988). This DNA fragment contains 2 kb ′ ′ ′ the , which results in a compensating upregulation of the of 5 flanking and 4 kb of 3 flanking sequences. The 2 kb of 5 α-skeletal actin (α -actin) gene (Garner et al., 1986). In flanking region to the termination codon was isolated as a HindIII- sk XbaI fragment and ligated to an XbaI-BamHI fragment of the human transgenic studies, ectopic expression of γ-actin gene containing the 3′ half of the 3′UTR plus 5 kb of 3′ flank (MLC) 2 fast-elicited a reduction in transcript accumulation (Erba et al., 1988). The ligation product was inserted between the from the endogenous alleles (Shani et al., 1988). Forced HindIII and BamHI sites of pcDVI (Okayama and Berg, 1983). The expression of βTm in the heart resulted in the downregulation resulting construct, αsk-γ3′, is shown in Fig. 1. The HindIII-XbaI of αTm protein (Muthuchamy et al., 1995). Finally, the fragment of the αsk-actin gene was also cloned between the HindIII- finding that the relative mRNA output between contractile XbaI sites of pGEM-3, re-isolated as a HindIII-BamHI fragment and protein gene families is similar at different stages of cloned between the HindIII-BamHI sites of pcDVI. This resulted in maturation, independently of isoform switching, suggests the the placement of an SV40 3′UTR plus transcription termination signal α α ′ existence of some form of communication between these adjacent to the sk-actin termination codon as shown for sk-SV3 in genes (Wade et al., 1990). Therefore, it is conceivable that Fig. 1. An in-frame deletion between amino acids (aa) 22-367 was created by digestion of α -SV3′ with NaeI followed by religation to feedback regulatory pathways are a feature of structural sk produce αsk-SV3′ (Fig. 1). The αsk-actin promoter driving the CAT protein gene family expression. reporter gene was as described by Muscat and Kedes (Muscat and Actin gene transfections have also suggested the existence Kedes, 1987) and is shown in Fig. 1 (αsk-CAT). The human αca-actin of regulatory pathways between actin genes and genes gene was isolated as an EcoRI-BglII fragment containing 5 kb of 5′ associated with actin function. A mutant β-actin gene has been flanking and 1.5 kb of 3′ flanking sequence (Engel et al., 1982) and found to cause downregulation of synthesis of Tm isoforms cloned into the EcoRI-BamHI region of pSV2-neo (Southern and Tm1, 2 and 3 in fibroblasts (Leavitt et al., 1987b) and reduced Berg, 1982) (αca in Fig. 1). protein and mRNA levels of Tm2 and 3 in myoblasts Cell culture and transfection (Schevzov et al., 1993). In both studies, the Tm5 isoform was unaffected, demonstrating a clear isoform specificity of this C2 cells originally isolated by Dr D. Yaffe (Yaffe and Saxel, 1977) and subcloned in the laboratory of Dr H. Blau (Blau et al., 1983) were response. The normal γ-actin gene produced a similar result in β grown in DMEM medium (Gibco Laboratories, Grand Island, NY, myoblasts (Schevzov et al., 1993). Both the mutant -actin and USA) supplemented with 20% FCS (Commonwealth Serum Labs, normal γ-actin genes also impact on the protein and mRNA Melbourne, Australia) and 0.5% chicken extract (Flow levels of and , two proteins involved in anchorage Laboratories Australia Pty Ltd., North Ryde, Australia). of actin filaments to focal contacts (Schevzov et al., 1995). DNA transfection and isolation of transfected clones were These studies are compatible with the previous suggestion that performed as described by Graham and van der Eb (Graham and van some muscle structural gene products may influence der Eb, 1973). In brief, C2 cells (4-6×105 cells per 100 mm dish) were expression of other muscle structural genes (Sutherland et al., transfected with 10 µg of test plasmid plus 1 µg of pSV2-neo by the 1993). calcium phosphate precipitation technique. All DNA preparations We tested the ability of the sarcomeric actin genes, α-cardiac were twice purified on cesium chloride gradients. Cells were exposed α to the DNA precipitate for 24 hours and then split 1:10 into fresh and -skeletal, to influence the expression of muscle and growth medium. Selection was started 24 hours later with 450 µg nonmuscle contractile protein genes in myoblasts. Of the two, G418/ml medium (G418-geneticin, Sigma Chemical Company, St skeletal actin was able to elicit the expression of a specific Louis, MO, USA). As colonies were formed, they were seeded and subset of muscle gene products characteristic of the cultured in 24-well plates followed by 60 mm dishes and finally differentiated state. However, this occurred in the absence of expanded into 100 mm dishes. The constructs used for transfection normal muscle differentiation, including the cessation of cell are shown in Fig. 1. Control transfections substituted the plasmid division. Cell shape was profoundly affected and changes in pUC18 for the test plasmids. 24 clones were randomly selected from nonmuscle actin and Tm levels were consistent with a each transfection. All clones were analyzed for expression of the test relationship between isoform composition of microfilaments gene and only those expressing the exogenous gene at a significant and cell surface area. This provides strong evidence for level were chosen for further analysis. regulatory pathways that are independent of the differentiation RNA isolation process and link expression within and between contractile Total cellular RNA was isolated from 30% confluent cultures. In the protein gene families and cell shape. We propose that such a initial experiments, RNA was isolated by the method of Chomczynski loop operates in differentiated cells, but can only be dissected and Sacchi (Chomczynski and Sacchi, 1987). All analyses shown in from the entire differentiation program if elicited in this paper, however, were performed on RNA isolated using the Induction of muscle genes by α-skeletal actin 515

TM TRIzoL Reagent (Life Technologies, Mt Waverley, Australia) for troponin (Tn) isoforms If, Is, Ic, Tf, Ts, Tc, Cf and Cs were as following the protocol supplied by the manufacturer. described (Sutherland et al., 1993). DNA probes MyoD gene family Human αsk-actin MyoD was detected using an EcoRI fragment from the expression Transcripts from the transfected αsk-actin gene constructs were construct pEMSVII containing the coding region (Davis et al., 1987) detected using a 190 bp DNA fragment containing the human αsk- provided by H. Weintraub. A 264 bp PstI fragment containing the actin 5′UTR. This probe is specific for the human gene and does not bHLH region of Myf5 (pchMyf5-P1) was provided by Richard hybridize to transcripts from the corresponding mouse gene as shown Harvey. Myogenin was detected using a 1486 bp fragment containing by Brennan and Hardeman (Brennan and Hardeman, 1993). the entire cDNA except for 22 bp at the 5′ end provided by W. Wright. A 1.3 kb EcoRI fragment containing the MRF4 cDNA was Human αca-actin provided by S. Koneiczny. Transcripts from the transfected α -actin gene were detected using a ca p21 probe corresponding to bp 1-101 of the human αca-actin 3′UTR, which hybridizes to the human, but not the corresponding mouse A 738 bp mouse cDNA probe from the coding region of WAF 1, transcript (Gunning et al., 1984). pCMW35, was a kind gift from B. Vogelstein.

Mouse αsk-actin Northern blots A probe containing the entire 3′UTR of human αsk-actin was used to Total cellular RNA was denatured and size-fractionated on 1% detect mouse αsk-actin transcripts (Brennan and Hardeman, 1993). agarose gels containing 2.2 M formaldehyde and transferred to This region was not present in any of the transfected gene constructs Hybond N (Amersham Australia, North Ryde, Australia) as described (Fig. 1) and, therefore, this probe could not detect transcripts from the (Maniatus et al., 1982). Probes were labelled by the random priming transfected genes. method (Feinberg and Vogelstein, 1983) and hybridized to RNA blots at 106 dpm/ml in a solution containing 4× SSC (1× SSC is 0.15 M Mouse αca-actin NaCl, 0.015 M sodium citrate), 50 mM NaH2PO4 (pH 7.0), 5× A probe was synthesized using a template corresponding to bp 1-100 Denhardts solution (Denhardt, 1966) and 10% (w/v) dextran sulphate of the mouse αca-actin 3′UTR and a primer recognizing bp 81-100 of at 65°C for 16 hours. After hybridization, the blots were washed at this sequences (Alonso et al., 1986). 65°C for 2 hours in 0.5× SSC, 0.1% SDS. Filters were exposed to Kodak XAR film for 1-14 days. To verify that equivalent amounts of Mouse α- actin RNA were transferred, the RNA blots were hybridized to an 18S A probe was synthesized using a template corresponding to the entire specific rRNA oligonucleotide probe under conditions of probe excess × 3′UTR of the rat αsm-actin and a primer recognizing the last 20 bp of and washed at 55°C in 4 SSC, 0.1% SDS. Levels of mRNA were the 3′UTR (McHugh and Lessard, 1988). quantitated using a Molecular Dynamics model 300 series (Mountain View, CA, USA) computing densitometer. Desmin α α A probe was generated by PCR amplification from rat Quantitative comparison of human sk- and ca-actin cDNA. The oligonucleotide primers 5′-CAACCTTCCCATCCAG- mRNAs ACC-3′ and 5′-GCTCGAGCACCTCGTGTTGT were used at an The levels of transcripts from the αsk-actin gene constructs were annealing temperature of 60°C to amplify a 166 bp fragment measured relative to the amount of αsk-actin mRNA in human corresponding to exons 7, 8 and 9. These oligonucleotides were gastrocnemius muscle. The level of αca-actin mRNA from the human synthesized from published hamster sequences (Quax et al., 1985) and αca-actin gene was measured relative to the amount of αca-actin sequencing confirmed the identity of the amplified product that was transcript in cultured human myotubes. The level of αca-actin mRNA cloned in pGEM3Z−. in cultured human myotubes is equal to the level of αsk-actin mRNA in human gastrocnemius muscle (Arkell, 1990). Tropomyosin A probe corresponding to exon 9a of the rat βTm gene (Gunning et al., 1990) was used to detect the two muscle-specific transcripts βTm (1.3 kb) and UTm (2.5 kb). These two mRNAs encode the same protein and differ only in the size of their 3′UTRs (Wang and Rubenstein, 1992). The 1.3 kb αTmf mRNA was detected using the rat αTmf exon 9b probe corresponding to the 3′UTR of this mRNA (Gunning et al., 1990). αTms was detected using the probe containing the 5′UTR plus aa 1-64 of the human αTms mRNA (Gunning et al., 1990). The nonmuscle isoforms Tm2,3, Tm4 and Tm5 were detected using the probes pHFTm3-PD, pRTm4-SR and pHMα-Tms-SA as previously described (Gunning et al., 1990). Tm5a and 5b transcripts were detected using an αΤmf exon 1b probe (Weinberger et al., 1993). Fig. 1. Maps of the transfected genes. Human αsk-actin constructs: Nonmuscle actins αsk-γ3′ contains 2 kb of 5′ flank through to the termination codon β- and γ-actin mRNAs were detected using mouse β- and γ-actin ligated to the 3′UTR of the human γ-actin gene; αsk-SV3′ contains 3′UTR probes as previously described (Lloyd et al., 1992). the SV40 3′UTR in place of the γ-actin 3′UTR; αsk-SV3′ contains an in-frame deletion between aa 22-367; αsk-CAT is the αsk-actin and promoter driving the CAT reporter gene. Human αca-actin construct Probes for MLC isoforms 1a, 1sa, 1sb, 1/3f, 2f, 2s and 2a were exactly (αca) contains 5 kb of 5′ flank through to 1.5 kb of 3′ flanking as described (Schevzov et al., 1995; Hailstones et al., 1992). Probes sequence. 516 JOURNAL OF CELL SCIENCE 114 (3)

Cell surface area was determined exactly as described (Schevzov et al., 1992). Antibodies The following monoclonal antibodies were purchased from Sigma Chemical Co: 5C5, which is specific for αsk- and αca-actin; 1A4, which recognizes αsm-actin; AC 15, which recognizes β-actin; CH1, which recognizes both αΤmf and βTm; and DE-U-10, which recognizes the muscle-specific intermediate filament protein desmin. The γ-actin antibody (Otey et al., 1988) was kindly supplied by Dr J. C. Bulinski. Rabbit antiserum directed against exon 9d of the αΤmf gene (Schevzov et al., 1997) reacts with Tm1, 2, 3, 6, 5a and 5b isoforms. MyoD antibody was kindly supplied by Dr A. John Harris (Koishi et al., 1995). Goat anti-rabbit IgG (H+L) alkaline and goat anti-mouse IgG (H+L) alkaline phosphatase conjugates were purchased from Bio-Rad Industries (Hercules, CA, USA). Fluorescein (DTAF)-conjugated goat anti-mouse IgG (H+L) and rhodamine (TRITC)-conjugated donkey anti-rabbit IgG (H+L) were obtained from Jackson Laboratories (West Grove, PA, USA).

RESULTS

αsk-actin constructs induce muscle-specific genes Transfection of the human αsk-actin gene containing its own 3′UTR into C2 myoblasts yielded 100+ clones, but none expressed the transfected gene at detectable levels. We had previously observed that the αsk-actin promoter was as active β Fig. 2. Expression of transfected genes in C2 myoblasts. Northern as the nonmuscle -actin promoter in C2 myoblasts (Gunning analysis of 5 µg of total RNA isolated from 30% confluent cultures et al., 1987). The failure of the intact human αsk-actin gene to of the highest expressing clones for four of the constructs described be expressed suggested the existence of a repressor element in Fig. 1 and control muscle (HG, HMT). mRNA levels were located in the body or 3′ end of the gene. In order for the human normalized to the corresponding 18S rRNA signals (data not shown). αsk-actin gene to express in myoblasts, we found it necessary The level of mRNA from the transfected gene is shown below each to replace the 3′UTR of this gene with the 3′UTR of SV40 (αsk- α lane as a percentage of control muscle. Note that ca B1 was SV3′, Fig. 1). Transfection of the truncated form of the αsk- significantly underloaded in the panel shown. (A,B) αsk-actin α ′ α actin gene, sk-SV3 , yielded expressing colonies, indicating expressing clones, (C) ca-actin expressing clones. HG, Human that the repressing element in the intact gene that operates in gastrocnemius; HMT, cultured human myotubes. myoblasts is within the 3′ end. All αsk-SV3′ transfectants were analyzed for expression and a northern blot of the four highest expressing clones is shown in Fig. 2A. The α -SV3′ H4 clone Cell cycle analysis sk was the highest expressor at 3.9% of the level of α -actin × 5 sk 5 10 cells were harvested at 30% confluency in PBS and treated mRNA in adult skeletal muscle. This is a relatively low level exactly as described (Smyth et al., 1993). Cells were analyzed using β γ a FACScanTM Analytical Flow Cytometer (Becton Dickinson, Lane of expression compared to that observed with the - and -actin Cove, Australia) and the SFIT modeling method (Dean et al., 1985). genes (Schevzov et al., 1992). All clones had similar doubling Results are expressed as a percentage of total events. Western blot analysis Table 1. Transcript levels of muscle-specific genes induced Total cellular protein was extracted from myoblasts grown to 30% by the human αsk-actin gene confluence or from myotube containing cultures 4 days post C2 αsk-γ3′αsk-SV3′ C2 differentiation. Protein concentrations were determined by the method of Minamide and Bamburg (Minamide and Bamburg, 1990). Gel mRNA Mb F3 C1 C4 H1 H4 MT electrophoresis and western blotting was as described (Gunning et al., UTm <1 22 65 45 68 96 100 1997). βTM <1 2 7 3 14 12 100 αTms 3 2214201412100 α Indirect immunofluoresence microscopy Tmf 3 7 17 10 40 23 100 α <1 47 34 35 28 20 100 Cells were cultured on -coated glass slides (type 1 from ca αsm 125 247 197 245 249 273 100 calf , Sigma Chemical Co), and fixed in 4% paraformaldehyde Desmin 16 79 97 96 84 71 100 for 20 minutes followed by cold methanol for 20 minutes. MyoD 238 521 724 930 1224 1366 100 Non-specific binding was blocked by incubation at room temperature in PBS containing 1% FCS for 30 minutes. Primary and Mb, myoblast; MT, myotube standards. secondary antibody incubations were each performed for 1 hour at Densitometry was performed on the autoradiographs shown in Fig. 3. RNA room temperature. Cells were examined on a Leica DMLB loadings were normalized to 18S rRNA and expression levels calculated as a microscope and photographed with Kodak T- 400 ASA film. percentage of that in myotube (set to 100%). See text for details. Induction of muscle genes by α-skeletal actin 517

Fig. 3. Induction of muscle-specific gene expression by αsk-actin constructs. Northern analysis of total RNA isolated from transfected clones described in Fig. 2. C2 cells transfected with the selection construct pSV2-neo alone provided the low confluence myoblast sample (Mb) and day 4 myotube (MT) standards. Probes detected the muscle Tm isoforms (UTm, βTm, αTmf, αTms), muscle actins (αca, αsm), the intermediate filament protein desmin and MyoD. 18S rRNA signals are representative of the detection procedure and do not correspond to all the structural protein transcripts shown. times and were capable of morphological differentiation into Induction is specific to the αsk-actin isoform myotubes. We tested the ability of the other major striated muscle actin, α ′ Clones expressing the sk-SV3 construct expressed αca-actin, to induce expression of muscle genes in C2s. A large particular muscle contractile protein mRNAs. RNA isolated portion of the α -actin gene (Fig. 1) was transfected into C2 α ′ ca from the four highest expressing sk-SV3 clones was myoblasts with no apparent toxic effect. The level of αca-actin hybridized to DNA probes specific for the muscle Tms and transcript accumulation was measured by northern blot actins (Fig. 3; quantitated in Table 1). Both muscle transcripts analysis and expressed relative to its level in human myotubes β βΤ from the Tm gene, UTm and m (Wang and Rubenstein, (Fig. 2C). Since the level of αca-actin mRNA in human 1992), were induced with the UTm transcript level in the H1 myotubes is equal to the level of αsk-actin mRNA in human and H4 clones similar to that in cultured myotubes. The fast gastrocnemius muscle, we can directly compare transcript α and slow isoforms of Tm were also induced in all clones with levels from the αsk-actin and αca-actin gene constructs. The the level of αTmf in the H1 clone approaching 50% the level level of α -actin mRNA in the four highest expressing clones α α ca in myotubes. ca-Actin mRNA was induced and the -smooth range above and below that of αsk-actin mRNA in the αsk-SV3′ muscle actin (αsm) mRNA was upregulated in all clones. The clones (Fig. 2A,C). We could therefore compare the impact of transcript level for desmin, the muscle intermediate filament these two genes at similar levels of expression. protein associated with Z disks, was upregulated compared The αca-actin gene failed to induce the expression of any with the level in myoblasts. muscle transcripts. RNA isolated from the two highest Although the level of expression of αsk-SV3′ differs expressing clones, B1 and C1, was hybridized to probes for the significantly between clones C1, C4, H1, and H4, the pattern of gene induction was similar. It appears that a low threshold level of αsk-SV3′ expression is sufficient to elicit the response. This is supported by observations in several subclones derived from the high expressing lines, H1 and H4. In these cells, the level of αsk-SV3′ transcript dropped to that observed in the C1 and C4 clones and yet the gene induction pattern was maintained (data not shown). The ability of αsk-SV3′ to induce expression of muscle genes is not due to the presence of the SV40 3′UTR. Cells were transfected with an αsk-actin gene construct in which the SV40 region was replaced by the 3′ end of γ-actin (Fig. 1, αsk-γ3′). As with the αsk-SV3′ gene, very few of the αsk-γ3′ transfected clones accumulated detectable transcript. The highest expressing clone, αsk-γ3′ F3, did express αsk- actin mRNA at a level comparable to that of α -SV3′ H1 sk Fig. 4. Nonmuscle actin and Tm gene expression are not repressed (Fig. 2A). This clone also induced expression of the four by α -actin contructs. Northern analysis of total RNA from α -γ3′ muscle Tms, two muscle actins and desmin (Fig. 3). sk sk F3 and αsk-SV3′ clones, hybridized to probes for nonmuscle Tm Therefore, we conclude that the SV40 3′ end in the αsk-γ3′ isoforms (Tm2,3, Tm4, Tm5, Tm5a, 5b) and the nonmuscle actin transfectants is not responsible for induction of the muscle isoforms (β, γ). Myoblast (Mb) and myotube (MT) standards are as transcripts. described in Fig. 3, as is 18S rRNA. 518 JOURNAL OF CELL SCIENCE 114 (3)

Fig. 5. MHC gene expression is not induced by αsk-actin constructs. Northern analysis of total RNA from transfected clones hybridized to a probe for MHC and then reprobed for 18S rRNA. Myoblast (Mb) and myotube (MT) standards are described in Fig. 3.

four muscle Tms, two muscle actins, and desmin (Fig. 3). In remaining isoforms from these gene families: MLC1sb, contrast to the αsk-actin clones, αsm-actin mRNA levels were MLC2s, TnTf, TnTs, TnTc, TnIf, TnCf (data not shown). substantially reduced in these αca-actin clones. This suggests Furthermore, the αsk-actin gene constructs do not induce that of the two major striated muscle actin genes, only the expression of the endogenous αsk-actin gene (Fig. 6). αsk-actin isoform is capable of eliciting this response from Therefore, the contractile protein genes activated by the αsk- other muscle genes. actin gene constructs are restricted specifically to the polymeric The activity of the αsk-actin promoter is not responsible for proteins of the actin filament (actin and Tm) and a protein the induction of muscle transcripts. A deletion construct of associated with the Z-disk (desmin) where the actin filaments αsk-SV3′ in which aa 22-367 were eliminated (Fig. 1, are anchored. We conclude that the phenotype displayed by the αsk-SV3′∆) was transfected into C2 myoblasts. The resulting αsk-actin gene transfectants does not reflect conventional clones accumulated truncated αsk-actin mRNA at levels above muscle differentiation. and below that obtained with the αsk-SV3′ gene (Fig. 2B). With the exception of low levels of UTm mRNA in the A2 clone MyoD is upregulated independent of withdrawal (Fig. 3), none of the four highest expressing αsk-SV3′ clones from the cell cycle induced any muscle Tm, actin nor desmin mRNA (Fig. 3; data The muscle , MyoD, is also upregulated in not shown). the αsk-actin transfectants. Both the αsk-SV3′ and αsk-γ3′ gene This experiment was repeated using the αsk-CAT construct (Fig. 1) and again, none of the four highest expressing clones induced any of the muscle mRNAs (data not shown). We conclude that the induction of muscle transcripts by the αsk- actin gene is isoform-specific and is not mediated by promoter activity per se. Induction of muscle transcripts is independent of nonmuscle transcripts The repression of nonmuscle actin and Tm mRNAs that normally accompanies muscle differentiation and induction of muscle isoforms does not occur in the αsk-actin transfectants. RNA was isolated from the αsk-actin transfectants and hybridized to the nonmuscle actin and Tm isoform probes (Fig. 4). The levels of β- and γ-actin, Tm2,3, Tm4 and Tm5 mRNAs are essentially identical to those in cultured myoblasts and are, in the case of Tms 2, 3, 4 and 5, in no way comparable to those in myotubes. This indicates that the coexpression of muscle and nonmuscle isoform transcripts in myogenic cells is not mutually exclusive and suggests that the αsk-actin transfectants are not differentiated.

αsk-actin constructs do not induce muscle differentiation The induction of muscle genes by αsk-actin gene constructs is not due to activation of the entire spectrum of muscle structural genes characteristic of muscle differentiation. The most widely used marker characteristic of muscle differentiation, sarcomeric myosin heavy chain (MHC), was not detectable at Fig. 6. The αsk-actin constructs do not induce the normal the mRNA level in any of the clones (Fig. 5). The MLC α γ ′ isoforms 1 and 2 , and the Tn isoforms I and C , are also differentiation pathway. Northern analysis of total RNA from sk- 3 sa f s s F3 and α -SV3′ H1 hybridized to probes for mouse α -actin, MLCs highly induced at the initiation of muscle differentiation sk sk α (1sa, 2f), Tns (Is, Cs), the MyoD family members (Myf-5, myogenin, (Sutherland et al., 1993). None of the sk-actin transfectants MRF4), and the cyclin dependent kinase inhibitor p21. Myoblast accumulate mRNA for any of these genes (Fig. 6). (Mb) and myotube (MT) standards are described in Fig. 3, as is 18S More extensive analysis failed to detect expression of the rRNA. Induction of muscle genes by α-skeletal actin 519

Fig. 7. Myf5 is induced in the αsk-γ3′ clone. Northern analysis of total RNA from transfected clones hybridized to a probe for Myf5 and reprobed for 18S rRNA. Myoblast (Mb) and myotube (MT) standards are described in Fig. 3. constructs induce MyoD mRNA accumulation to levels which found in cycling myoblasts (Fig. 6) indicating that the are three- to sevenfold higher than that found in nontransfected induction of this specific set of genes by αsk-actin bypasses this myoblasts and myotubes (Fig. 3). This demonstrates that regulatory pathway and is independent of the normal expression of a structural protein gene can impact on the differentiation pathway. The simultaneous expression of a expression of a transcription factor. It has been previously subset of muscle-specific contractile protein genes and the observed that ectopic overexpression of MyoD induces other normal complement of nonmuscle isoforms would seem to members of the myogenic bHLH factors (Olson and Klein, be incompatible with normal cell cycling. We therefore 1994). The other MyoD family member coexpressed with investigated how these αsk-actin transfected myoblasts MyoD in myoblasts, Myf5, shows a significant elevation in accommodate the coexpression of muscle and nonmuscle αsk-γ3′-F3, but not in αsk-SV3′-H1 (Fig. 6). Analysis of the isoforms. other αsk-SV3′ clones failed to detect significant elevation of Myf5 (Fig. 7). The elevation of Myf5, therefore is restricted to Muscle proteins accumulate in αsk-actin one clone. Neither myogenin nor MRF4 were induced (Fig. 6). transfectants The induction of MyoD is therefore quite specific and does not The induction of muscle-specific mRNAs is accompanied by involve obligatory coactivation of other myogenic bHLH the accumulation of the corresponding proteins (Fig. 8). The factors. The failure to detect myogenin further supports the sarcomeric Tm antibody was used to detect βTm and αfΤm on conclusion that these transfected cells have not initiated western blots of αsk- and αca-actin transfected clones. αfΤm conventional differentiation. was detected in all αsk-actin transfectants and βTm was The induction of MyoD occurs in the αsk-actin transfectants detected strongly in all four αsk-SV3′ clones and only weakly in the absence of any impact on the cell cycle. Previous studies in the αsk-γ3′ clone. Neither product was detected in αca-actin have suggested that elevated MyoD is compatible with transfectants or control myoblasts. A third product running withdrawal from the cell cycle (Crescenzi et al., 1990; between αfΤm and βTm was detected in all myoblast samples Sorrentino et al., 1990). We investigated this possibility by including controls and its exact composition is unknown. performing flow cytometry analysis on these transfectants. The Sarcomeric actin was detected in all αsk- and αca-actin percentage of cells in G1, S and G2/M was calculated for transfectants, but because this antibody detects both αsk- control and transfected clones and the results are shown in and αca-actin, we cannot determine which product(s) is Table 2. The majority of cells from each clone were in the S stage of the cell cycle with no significant difference between the transfectants and the parental cells. The dissociation of the induction of MyoD and other muscle genes from the cell cycle was confirmed further by analyzing the levels of p21 transcript. The expression of muscle structural genes is preceded by a MyoD-mediated elevation of p21, which results in terminal withdrawal from the cell cycle and differentiation (Halevy et al., 1995; Parker et al., 1995; Guo et al., 1995). In the αsk-actin transfectants, the p21 transcript level is not elevated above that

Table 2. Cell cycle analysis of clones expressing muscle- specific genes Cell cycle phase

Clone G1 SG2+M C2 29 53 18 SV2Neo 34 52 14 αsk-γ3′ F3 29 50 21 αsk-SV3′ H1 30 57 13 αsk-SV3′ H4 41 45 14

Clones (Fig. 2) harvested at 30% confluency were analyzed by flow Fig. 8. αsk-actin clones express muscle-specific proteins. Western cytometry. The numbers of cell nuclei in G1, S and G2/M were calculated blot analysis of total protein from transfected clones immunostained from the DNA histograms and expressed as percentage of total events. with antibodies to sarcomeric Tm (β- and αfΤm), sarcomeric actin, C2, parental myoblasts; SV2Neo, C2 myoblasts transfected with the αsm-actin, desmin and MyoD. Myoblast (Mb) and myotube (MT) selection gene only. standards are described in Fig. 3. 520 JOURNAL OF CELL SCIENCE 114 (3) genes into fibroblasts and myoblasts has consistently demonstrated an impact on nonmuscle actin and Tm synthesis (Leavitt et al., 1987a; Ng et al., 1988; Lloyd et al., 1992; Leavitt et al., 1987b; Schevzov et al., 1993). These effects have not been attributable to promoter competition because the activity of a nonmuscle actin gene encoding an unstable mutant actin was incapable of eliciting the same impact on nonmuscle actin and Tm expression (Lloyd et al., 1992; Schevzov et al., 1993). In addition, nonmuscle actin gene transfections impact on the expression of two focal adhesion components, vinculin and talin (Schevzov et al., 1995). This has led us to propose that alterations in the composition of microfilaments change the stability of interaction with actin binding proteins and β precipitate feedback regulation of many of these proteins Fig. 9. Nonmuscle Tms and -actin proteins are downregulated in (Schevzov et al., 1993). α -actin clones. Western blot analysis of total protein from sk The changes observed in gene expression in this study transfected clones immunostained with antibody to αfΤm9d (Tm1, 2, 3, 5a, 5b, and 6) or β-actin. Myoblast (Mb) and myotube involve a number of mechanisms that operate at various levels (MT) standards are described in Fig. 3. of gene regulation. The induction of the slow isoform of αTm involves activation of the muscle-specific promoter adjacent to exon 1a, alternative splicing to select exons 6b (rather than 6a accumulating in the αsk-actin transfectants. Both αsm-actin and that is found in the nonmuscle isoform from the same gene and desmin accumulate to myotube levels in the αsk-actin is expressed in myoblasts) and 9a, and use of an alternative transfectants. MyoD protein displays a more variable response polyadenylation signal located in exon 9b (Gunning et al., between clones with clear elevation in the four αsk-SV3′ 1990; Pittenger et al., 1994). The induction of fast αTm, βTm clones. In multiple analyses, αsk-γ3′ F3 showed no evidence of and UTm all involve alternative splicing to select exons 6b elevated MyoD protein levels, which may be related to the (rather than 6a) and 9a, and use of an alternative induction of Myf5 mRNA only in this clone (Fig. 7). polyadenylation signal located in exon 9b. In the case of UTm, α the intron between exons 9a and 9b is now used as an exon sk-actin impacts on cell spreading and nonmuscle (Pittenger et al., 1994; Wang and Rubenstein, 1992). The isoforms decrease in protein levels of Tm 2,3 and β-actin do not occur The αsk-actin transfectants were greatly affected in their ability as a result of decreased mRNA levels and must, therefore, be to spread on a substratum. In contrast to control cells with an due either to reduced translation or to elevated protein turnover. 2 average surface area of 1100 µm , all αsk-actin transfectants The extent of these changes in gene expression suggest that that induced muscle proteins displayed greatly reduced surface αsk-actin has induced a regulatory pathway that acts at many areas from 280 to 420 µm2 (Table 3). The uniformity of this different steps or that some of these changes may be a response suggests that it is intrinsic to the presence of the consequence of the morphological impact of the αsk-actin gene α α sk-actin gene constructs. The surface area of the ca-actin constructs. The most likely explanation would be that αsk-actin transfectants was similar to controls, indicating that the induces changes in muscle gene promoter activity and muscle accumulation of a sarcomeric muscle actin is not enough to transcript processing. The changes in nonmuscle actin and compromise cell spreading. tropomyosin protein levels are more likely to be regulated by Previous studies revealed a relationship between reduced changes in the actin cytoskeleton. Previous studies have shown cell spreading and decreased expression of β-actin and the high that transfection of mutant and normal actins can impact on molecular Tms (Tm 1, 2, 3 and 6) (Schevzov et al., 1993; cytoskeletal organization and the levels of actin and Schevzov et al., 1992). The mRNA levels of Tm 2, 3 and β- tropomyosin isoforms. actin were unchanged in the αsk-actin transfectants (Fig. 4). We The notion of one structural protein regulating the synthesis measured the corresponding protein levels and the resulting of other structural proteins is not without precedent. Forced western blots are shown in Fig. 9. Whereas the αca-actin transfectants are identical to control myoblasts, the αsk-actin clones displayed variable downregulation of Tms 1, 2, 3 and Table 3. Surface areas of transfected cells 6, virtual elimination of Tms 5a and 5b, and reduction of β- Clone Surface area (µm2) actin to 20% of control cell levels. This supports a role for SV2Neo 1106±87 microfilament isoform composition in the regulation of cell αsk-γ3′ F3 275±21 α ′ spreading. sk-SV3 C1 424±94 αsk-SV3′ C4 383±21 αsk-SV3′ H1 352±83 αsk-SV3′ H4 285±25 DISCUSSION αca A1 1274±356 αca B1 1841±376 α Actin genes as regulators ca C1 1013±150 αca D1 1128±162 A number of studies have suggested that actin gene products are capable of regulating the expression of other genes. Cell surface areas were determined as described (Schevzov et al., 1992). Transfection of wild-type and mutant β- and wild-type γ-actin Values means ± s.d. Induction of muscle genes by α-skeletal actin 521 expression of a type II in fibroblasts induced the polymers with an ensuing generic regulatory signal. Indeed, the endogenous partner , both of which were found level of total actin in the transfectants is unchanged (Gunning in filamentous structures (Giudice and Fuchs, 1987). In et al., 1997), eliminating an effect on actin pool size as the contrast, the type I keratin could not induce type II keratin. This trigger. In addition, the substantial induction of αca-actin parallels the ability of αsk-actin to induce αca-actin and the which, based on transcript levels, must contribute the bulk of inability of αca-actin to induce αsk-actin. These studies suggest α-actin in these cells, cannot be the trigger because higher a heirarchy of control between different members of these levels of αca-actin expression in the cardiac actin transfectants multigene families. also fail to elicit this phenotype. Finally, the downregulation of β-actin is more likely to be a result of, rather than a cause of, Relationships between microfilament isoforms and these changes in gene expression because a similar cell shape downregulation of β-actin caused by γ-actin gene transfection There is increasing evidence that actin and Tm isoforms are of C2 myoblasts has no impact on muscle actins or functionally distinct and are partially responsible for the tropomyosins (P. W. Gunning, G. Schevzov and C. Lloyd, regulation of cell shape (Pittenger et al., 1994; Herman, 1993; unpublished observations). Although we cannot completely Janmey and Chaponnier, 1995; Gunning et al., 1998). The eliminate an effect on actin polymerization as the trigger observed changes in actin and Tm isoforms in the αsk-actin generated by αsk-actin, it appears unlikely at this stage. transfectants correlate very closely with the predicted roles for How then, could αsk-actin generate such an isoform-specific specific isoforms. First, the reduction in β-actin protein in our signal that relates to normal actin function? It may be that αsk- transfectants relates to the predicted role of β-actin in cell actin, unlike αca-actin, does not incorporate into a nonmuscle spreading and polarity. Experiments utilizing gene transfection cytoskeleton. In this situation, detection of ‘free’ αsk-actin may (Schevzov et al., 1992), disrupted localization of β-actin serve as a signal to induce the expression of muscle thin mRNA (Kislauskis et al., 1994) and correlations with cell filament proteins that should drive αsk-actin into polymeric behaviour (De Nofrio et al., 1989; Hoock et al., 1991; Hill and structures. The testing of this mechanism awaits the generation Gunning, 1993; Kislauskis et al., 1993; Hill et al., 1994; of antibodies that discriminate between αsk- and αca-actin. Latham et al., 1994; Yao et al, 1995) all suggest that β-actin It is possible that the initial impact of αsk-actin on some gene has an isoform-specific role in promoting cell spreading. products and/or cell architecture may produce secondary Second, the downregulation of the high molecular mass Tms compensating changes as the cell adapts to new structures. At (Tm 1, 2, 3 and 6) has also been both directly and indirectly this time, we cannot separate primary from secondary events. implicated in reduced cell spreading (Schevzov et al., 1993; Recent (Mounier et al., 1997), however, suggests that the Prasad et al., 1993; Boyd et al., 1995; Gimona et al., 1996). structural impact of αsk-actin may be a secondary consequence The function of Tm in these cells is likely to have been further of its expression. Mounier et al. found that transient compromised by the expression of the muscle isoforms. The transfection of nonmuscle actin gene constructs impacted nonmuscle Tms normally form homodimers; however, in the immediately on cell shape whereas the muscle actins had no presence of muscle isoforms they will form heterodimers initial impact. This may suggest that a muscle actin has to act (Gimona et al., 1995). through secondary mechanisms to impact on cell shape. The impact of αsk-actin on muscle and nonmuscle isoforms may involve both direct and indirect regulatory pathways. Regulatory pathways in myogenesis There is considerable evidence that feedback regulatory The detection of muscle mRNAs in proliferating myoblasts is mechanisms control the expression of many cytoskeletal incompatible with current models of muscle differentiation proteins (reviewed in Ben-Ze’ev, 1991). Transfection of actin (Olson and Klein, 1994; Zhang et al., 1995; Olson, 1992). The genes that alter cell shape have routinely impacted on Tm transcription factors myogenin and MRF4 that are responsible isoform expression in a variety of cell types (Leavitt et al., for muscle-specific gene expression are only expressed in 1987b; Schevzov et al., 1993). Actin is subject to feedback myoblasts that have exited the cell cycle (Olson and Klein, regulation (Leavitt et al., 1987a; Lloyd, 1992) at least partially 1994). This proposed linkage of cell cycle withdrawal and via a sensing of the state of actin assembly (Bershadsky et al., myogenin induction is mediated by the two muscle 1995). transcription factors MyoD and Myf5 that are expressed in Actin feedback regulation can show isoform-specificity, as proliferating myoblasts. MyoD-mediated activation of the was demonstrated by Lloyd et al. (Lloyd et al., 1992), where cyclin-dependent kinase inhibitor p21 coordinates withdrawal the nonmuscle actin isoforms β and γ, which differ by only four from the cell cycle and differentiation (Halevy et al., 1995; conservative amino acids, elicited quite different feedback Parker et al., 1995; Guo et al., 1995). Thus, these studies regulatory responses. Similarly, as shown in this study, the two suggest that muscle gene products cannot be induced in the sarcomeric actins differ by only four amino acids and yet αsk- absence of myogenin and/or MRF4 induction, which in turn actin induces and αca-actin represses expression of α-smooth cannot occur in cycling cells. actin (Figs 3, 4). One possible explanation for the specificity Our results indicate that muscle gene mRNAs can of these effects is that isoform-specific binding proteins accumulate in the absence of both withdrawal from the cell function as intermediaries that coordinate the expression of cycle and induction of either myogenin or MRF4. Because cell actin isoforms with the expression of those sets of gene cycle withdrawal is linked to MyoD activity, these results products that interact with actin. suggest that either the αsk-actin induction of muscle genes The low level of αsk-actin expression required to elicit this operates independently of the entire MyoD family of response suggests that the mechanism involves the generation transcription factors or that we have dissociated MyoD/Myf5- of an isoform-specific signal rather than a global effect on actin directed induction of some muscle genes from their effects on 522 JOURNAL OF CELL SCIENCE 114 (3) the cell cycle. If MyoD/Myf5 are acting directly on muscle We would like to thank P. Rowe for his support, V. Elsom for genes this would involve a novel mechanism since it is thought technical assistance, S. Hauschka, C. Ordahl, R. Schwartz, and J. that MyoD only acts on muscle genes via induction of Miller for their advice and comments and J. Smythe for critical myogenin (Hollenberg et al., 1993). Indeed, the failure of comments on the manuscript. We are particularly grateful to C. Smyth elevated MyoD levels to induce myogenin suggests that for providing FACS analysis of our cells. We appreciate the kind gifts MyoD-dependent regulatory pathways may not have been of probes from H. Weintraub, W. Wright, R. Harvey, S. Koneiczny, and B. Vogelstein. Supported by grants to E.H. and P.G. from the activated in these cells. It is clear from our results, however, NH&MRC. P.G. is a Principal Research Fellow of the NH&MRC. that cell cycling does not intrinsically provide an absolute inhibition of muscle gene expression. We therefore propose that a regulatory pathway that operates independently of REFERENCES myogenin and MRF4 can act on a subset of muscle genes in cycling cells. One candidate for such a pathway is the G-actin Alonso, S., Minty A., Bourlet, Y. and Buckingham M. (1986). Comparison mediated regulation of SRF, and subsequently, subsets of SRF of three actin-coding sequences in the mouse; evolutionary relationships between the actin genes of warm-blooded . J. Mol. Evol. 23, 11- target genes (Sotiropoulos et al., 1999). Since skeletal actin 22. is regulated by SRF, it is possible that this specific actin Arkell, R. (1990). Sketetal muscle maturation, an in vivo analysis. BSc isoform coordinates the expression of structurally associated Honours Thesis. University of Sydney, Sydney, Australia. sarcomeric genes via the actin monomer pool. Ben-Ze’ev, A. (1991). cell shape changes and gene expression. Several studies demonstrate that structural protein genes BioEssays 13, 207-212. Bershadsky, A. D., Gluck, U., Denisenko, O. N., Sklyarova, T. V., Spector, contribute to muscle gene regulation and differentiation. I. and Ben-Ze’ev, A. (1995). The state of actin assembly regulates actin and Inhibition of desmin synthesis blocks myoblast fusion and vinculin expression by a feedback loop. J. Cell. Sci. 108, 1183-1193. alters the expression of MyoD and myogenin in vitro (Li et Blau, H. M., Chiu, C. P. and Webster, C. (1983). Cytoplasmic activation of al., 1994). This suggests that desmin is required to permit human nuclear genes in stable heterocaryons. Cell 32, 1171-1180. ′ Boyd, J., Risinger, J. I., Wiseman, R. W., Merrick, B. A., Selkirk, J. K. progression of muscle differentiation. The 3 UTR sequences of and Barrett, J. C. (1995). Regulation of microfilament organization and several structural protein mRNAs were shown to complement anchorage-independent growth by tropomyosin 1 Proc. Natl. Acad. Sci. USA deficiencies in a mutant myoblast cell line that cannot complete 92, 11534-11538. differentiation (Rastinejad and Blau, 1993). Expression of the Braun, T., Rudnicki, M. A., Arnold, H. H. and Jaenisch, R. (1992). Targeted 3′UTR of the αTm mRNA has been shown to induce muscle inactivation of the muscle regulatory gene Myf-5 results in abnormal rib f development and perinatal death. Cell 71, 369-382. differentiation in embryonic chicken fibroblasts (L’Ecuyer et Bremer, J. W., Busch, H. and Yeoman, L. C. (1981). Evidence for a species al., 1995). Thus, both muscle proteins and sequences in their of nuclear actin distinct from cytoplasmic and muscles actins. Biochem. 20, mRNAs may be used to communicate regulatory information. 2013-2017. Brennan, K. J. and Hardeman, E. C. (1993). Quantitative analysis of the At this time we cannot say if αsk-actin protein is responsible human alpha-skeletal actin gene in transgenic mice. J. Biol. Chem. 268, 719- for inducing muscle genes, but we can certainly conclude that 725. the 3′UTR is not directly involved. Buckingham, M. E. and Minty, A. J. (1983). Contractile protein genes. In The ability of αsk-actin to induce specific muscle genes most Eukaryotic Genes (ed. N. MacLean, S. P. Gregory and R. A. Flavell), pp. likely reflects the action of a novel type of muscle regulatory 365-397. Butterworth, London. program. Regulatory pathways linking the expression of Chomczynski, P. and Sacchi, N. (1987). Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. muscle genes may serve as a back-up regulatory system. In Biochem.162, 156-159. multimeric polymer complexes such as a striated sarcomere, Crescenzi, M., Fleming, T. P., Lassar, A. B., Weintraub, H. and Aaronson, both the order and level of muscle gene expression may be S. A. (1990). MyoD induces growth arrest independent of differentiation in important for correct muscle function. Some of these normal and transformed cells. Proc. Natl. Acad. Sci. USA 87, 8442-8446. Davis, R. L., Weintraub, H. and Lassar, A. B. (1987). Expression of a single regulatory pathways may ensure that the correct combination transfected cDNA converts fibroblasts to myoblasts. Cell 51, 987-1000. of gene products is expressed at particular stages of Dean, P. N. (1985). Flow Cytometry: Instrumentation and Data Analysis. In development. This is supported by the observation that the only Anonymous Methods of Data Analysis in Flow Cytometry (ed. M. Van Dilla, contractile protein genes induced by αsk-actin encode proteins O. Learum and M. Melamed), pp. 195-225. Academic Press, London. that form a polymeric complex with α -actin. Given the Denhardt, D. T. (1966). A membrane-filter technique for the detection of sk complementary DNA. Biochem. Biophys. Res. Commun. 23, 641-646. potential complexity of regulatory programs running during DeNofrio, D., Hoock, T. C. and Herman, I. M. (1989). Functional sorting of myogenesis, it is not surprising that the existence of some of actin isoforms in microvascular . J. Cell. Biol. 109, 191-202. these pathways has only been revealed when they are triggered Devlin, R. B. and Emerson, C. P., Jr (1978). Coordinate regulation of outside of normal myogenesis. contractile protein synthesis during myoblast differentiation. Cell 13, 599- 611. Recent work has highlighted the presence of actin and actin Engel, J., Gunning, P. and Kedes, L. (1982). Human Cytoplasmic Actin binding proteins in the nucleus and the potential roles that it Proteins Are Encoded by a Multigene Family. Mol. Cell. Biol. 2, 674-684. plays in nuclear functions such as chromatin remodeling and Erba, H. P., Eddy, R., Shows, T., Kedes, L. and Gunning, P. (1988). splicing (Zhao et al., 1998; reviewed in Rando et al., 2000). Structure, location, and expression of the human gamma-actin There have been several reports that nuclei contain a novel gene, differential evolution, location, and expression of the cytoskeletal beta- and gamma-actin genes. Mol. Cell. Biol. 8, 1775-1789. form of actin that is more similar to the skeletal muscle actin Esser, K., Gunning, P. and Hardeman, E. (1993). -dependent and - isoforms than the cytoplasmic actin isoforms (Bremer et al., independent patterns of mRNA expression in regenerating skeletal muscle. 1981; Kumar et al., 1984). Recently, Wada et al. (Wada et al., Dev. Biol. 159, 173-183. 1998) identified functional nuclear export signals in all actin Feinberg, A. P. and Vogelstein, B. (1983). A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biol. 132, isoforms. This suggests that all actins are capable of shuttling 6-13. in and out of the nucleus and provides a possible mechanism Garner, I., Minty, A. J., Alonso, S., Barton, P. J. and Buckingham, M. E. that links gene regulation and cytoarchitectural remodeling. (1986). A 5′ duplication of the alpha-cardiac actin gene in BALB/c mice is Induction of muscle genes by α-skeletal actin 523

associated with abnormal levels of alpha-cardiac and alpha-skeletal actin Transdifferentiation of chicken embryonic cells into muscle cells by the 3′ mRNAs in adult cardiac . EMBO J. 5, 2559-2567. untranslated region of muscle tropomyosin. Proc. Natl. Acad. Sci. USA 92, Gimona, M., Watakabe, A. and Helfman, D. M. (1995). Specificity of dimer 7520-7524. formation in tropomyosins: influence of alternatively spliced exons on Lassar, A. and Munsterberg, A. (1994). Wiring diagrams: regulatory circuits homodimer and heterodimer assembly. Proc. Natl. Acad. Sci. USA 92, 9776- and the control of skeletal myogenesis. Curr. Opin. Cell Biol. 6, 432-442. 9780. Latham, V. M. J., Kislauskis, E. H., Singer, R. H. and Ross, A. F. (1994). Gimona, M., Kazzaz, J. A. and Helfman, D. M. (1996). Forced expression Beta-actin mRNA localization is regulated by signal transduction of tropomyosin 2 or 3 in v-Ki-ras-transformed fibroblasts results in distinct mechanisms. J. Cell Biol. 126, 1211-1219. phenotypic effects. Proc. Natl. Acad. Sci. USA 93, 9618-9623. Leavitt, J., Ng, S. Y., Aebi, U., Varma, M., Latter, G., Burbeck, S., Kedes, Giudice, G. J. and Fuchs, E. (1987). The transfection of epidermal keratin L. and Gunning, P. (1987a). Expression of transfected mutant beta-actin genes into fibroblasts and simple epithelial cells: evidence for inducing a genes: alterations of cell morphology and evidence for autoregulation in type I keratin by a type II gene. Cell 48, 453-463. actin pools. Mol. Cell. Biol. 7, 2457-2466. Graham, F. L. and van der Eb, A. J. (1973). A new technique for the assay Leavitt, J., Ng, S. Y., Varma, M., Latter, G., Burbeck, S., Gunning, P. and of infectivity of human adenovirus 5 DNA. Virol 52, 456-467. Kedes, L. (1987b). Expression of transfected mutant beta-actin genes, Gunning, P., Mohun, T., Ng, S. Y., Ponte, P. and Kedes, L. (1984). Evolution transitions toward the stable tumorigenic state. Mol. Cell. Biol. 7, 2467- of the human sarcomeric-actin genes: evidence for units of selection within 2476. the 3′ untranslated regions of the mRNAs. J. Mol. Evol. 20, 202-214. Li, H., Choudhary, S. K., Milner, D. J., Munir, M. I., Kuisk, I. R. and Gunning, P., Leavitt, J., Muscat, G., Ng, S. Y. and Kedes, L. (1987). A Capetanaki, Y. (1994). Inhibition of desmin expression blocks myoblast human beta-actin expression vector system directs high-level accumulation fusion and interferes with the myogenic regulators MyoD and myogenin. J. of antisense transcripts. Proc. Natl. Acad. Sci. USA 84, 4831-4835. Cell Biol. 124, 827-841. Gunning, P., Gordon, M., Wade, R., Gahlmann, R., Lin, C. S. and Lloyd, C., Schevzov, G. and Gunning, P. (1992). Transfection of nonmuscle Hardeman, E. (1990). Differential control of tropomyosin mRNA levels beta- and gamma-actin genes into myoblasts elicits different feedback during myogenesis suggests the existence of an isoform competition- regulatory responses from endogenous actin genes. J. Cell Biol. 117, 787- autoregulatory compensation control mechanism. Dev. Biol. 138, 443-453. 797. Gunning, P., Ferguson, V., Brennan, K. and Hardeman, E. (1997). Impact Maniatus, T., Fritsch, E. F. and Sambrook, J. (1982). Molecular Cloning: of alpha-skeletal actin but not alpha-cardiac actin on myoblast morphology. A Laboratory Manual. Cold Spring Harbour Laboratory, New York, USA. Cell Struct. Funct. 22, 173-179. McHugh, K. M. and Lessard, J. L. (1988). The nucleotide sequence of a rat Gunning, P., Weinberger, R., Jeffrey, P. and Hardeman, E. (1998). Isoform alpha-actin cDNA. Nucleic Acids Res 16, 4167 sorting and the creation of intracellular compartments. Ann. Rev. Cell. Dev. Minamide, L. S. and Bamburg, J. R. (1990). A filter paper dye-binding assay Biol. 14, 339-372. for quantitative determination of protein without interference from reducing Guo, K., Wang, J., Andres, V., Smith, R. C. and Walsh, K. (1995). MyoD- agents or detergents. Anal. Biol. 190, 66-70. induced expression of p21 inhibits cyclin-dependent kinase activity upon Mounier, N., Perriard, J. C., Gabbiani, G. and Chaponnier, C. (1997). myocyte terminal differentiation. Mol. Cell. Biol. 15, 3823-3829. Transfected muscle and non-muscle actins are differentially sorted by Hailstones, D., Barton, P., Chan-Thomas, P., Sasse, S., Sutherland, C. cultured smooth muscle and non-muscle cells. J. Cell. Sci. 110, 839-846. Hardeman, E. and Gunning, P. (1992). Differential regulation of the atrial Muscat, G. E. and Kedes, L. (1987). Multiple 5′-flanking regions of the isoforms of the myosin light chains during striated muscle development. J. human alpha-skeletal actin gene synergistically modulate muscle-specific Biol. Chem. 267, 23295-23300. expression. Mol. Cell. Biol. 7, 4089-4099. Halevy, O., Novitch, B. G., Spicer, D. B., Skapek, S. X., Rhee, J., Hannon, Muthuchamy, M., Grupp, I. L., Grupp, G., O’Toole, B. A., Kier, A. B., G. J., Beach, D. and Lassar, A. B. (1995). Correlation of terminal cell cycle Boivin, G. P., Neumann, J. and Wieczorek, D. F. (1995). Molecular and arrest of skeletal muscle with induction of p21 by MyoD. Science 267, 1018- physiological effects of overexpressing striated muscle beta-tropomyosin in 1021. the adult murine heart. J. Biol. Chem. 270, 30593-30603. Hasty, P., Bradley, A., Morris, J. H., Edmondson, D. G., Venuti, J. M., Nabeshima, Y., Hanaoka, K., Hayasaka, M., Esumi, E., Li, S. and Nonaka, Olson, E. N. and Klein, W. H. (1993). Muscle deficiency and neonatal I. (1993). Myogenin gene disruption results in perinatal lethality because of death in mice with a targeted mutation in the myogenin gene. 364, severe muscle defect Nature 364, 532-535. 501-506. Ng, S. Y., Erba, H., Latter, G., Kedes, L. and Leavitt, J. (1988). Modulation Herman, I. M. (1993). Actin isoforms. Curr. Opin. Cell. Biol. 5, 48-55. of microfilament protein composition by transfected cytoskeletal actin Hill, M. A. and Gunning, P. (1993). Beta and gamma actin mRNAs are genes. Mol. Cell. Biol. 8, 1790-1794. differentially located within myoblasts. J. Cell. Biol. 122, 825-832. Okayama, H. and Berg, P. (1983). A cDNA cloning vector that permits Hill, M. A., Schedlich, L. and Gunning, P. (1994). Serum-induced signal expression of cDNA inserts in mammalian cells. Mol. Cell. Biol. 3, 280- transduction determines the peripheral location of beta-actin mRNA within 289. the cell. J. Cell. Biol.126, 1221-1229. Olson, E. N. (1992). Interplay between proliferation and differentiation within Hollenberg, S. M., Cheng, P. F. and Weintraub, H. (1993). Use of a the myogenic lineage. Dev. Biol. 154, 261-272. conditional MyoD transcription factor in studies of MyoD trans-activation Olson, E. N. and Klein, W. H. (1994). bHLH factors in muscle development, and muscle determination. Proc. Natl. Acad. Sci. USA 90, 8028-8032. deadlines and commitments, what to leave in and what to leave out. Genes Hoock, T.C., Newcomb, P. M. and Herman, I. M. (1991). Beta actin and its Dev. 8, 1-8. mRNA are localized at the plasma membrane and the regions of moving Otey, C. A., Kalnoski, M. H. and Bulinski, J. C. (1988). Immunolocalization cytoplasm during the cellular response to injury. J. Cell. Biol. 112, 653-664. of muscle and nonmuscle isoforms of actin in myogenic cells and adult Janmey, P. A. and Chaponnier, C. (1995). Medical aspects of the actin skeletal muscle. Cell Motil. Cytoskel. 9, 337-348. cytoskeleton. Curr. Opin. Cell. Biol. 7, 111-117. Parker, S. B., Eichele, G., Zhang, P., Rawls, A., Sands, A. T., Bradley, A., Kislauskis, E. H., Li, Z., Singer, R. H. and Taneja, K. L. (1993). Isoform- Olson, E. N., Harper, J. W. and Elledge, S. J. (1995). -independent specific 3′-untranslated sequences sort alpha-cardiac and beta-cytoplasmic expression of p21Cip1 in muscle and other terminally differentiating cells. actin messenger RNAs to different cytoplasmic compartments [published Science 267, 1024-1027. erratum appears in J. Cell Biol. 1993 Dec;123(6 Pt 2): following 1907]. J. Pittenger, M. F., Kazzaz, J. A. and Helfman, D. M. (1994). Functional Cell Biol. 123, 1.65-172. properties of non-muscle tropomyosin isoforms. Curr. Opin. Cell Biol. 6, Kislauskis, E. H., Zhu, X. and Singer, R. H. (1994). Sequences responsible 96-104. for intracellular localization of beta-actin messenger RNA also affect cell Prasad, G. L., Fuldner, R. A. and Cooper, H. L. (1993). Expression of phenotype. J. Cell Biol. 127, 441-451 transduced tropomyosin 1 cDNA suppresses neoplastic growth of cells Koishi, K., Zhang, M., McLennan, I. S. and Harris, A. J. (1995). MyoD transformed by the ras oncogene. Proc. Natl. Acad. Sci. USA 90, 7039-7043. protein accumulates in satellite cells and is neurally regulated in Quax, W., van den Broek, L., Egberts, W. V., Ramaekers, F. and regenerating myotubes and skeletal muscle fibers. Dev. Dyn. 202, 244-254. Bloemendal, H. (1985). Characterization of the hamster desmin gene, Kumar, A., Raziuddin, T. H., Finlay, J. O., Thomas and Szer, W. (1984). expression and formation of desmin filaments in nonmuscle cells after gene Isolation of a minor species of actin from the nuclei of Acanthamoeba transfer. Cell 43, 327-338. castellanii. 23, 6753-6757. Rando, O. J., Zhao, K. and Crabtree, G. R. (2000). Searching for a function L’Ecuyer, T. J., Tompach, P. C., Morris, E. and Fulton, A. B. (1995). for nuclera actin trends. J. Cell Biol. 92-97 524 JOURNAL OF CELL SCIENCE 114 (3)

Rastinejad, F. and Blau, H. M. (1993). Genetic complementation reveals a Hardeman, E. C. (1991). Coordination of skeletal muscle gene expression novel regulatory role for 3′ untranslated regions in growth and occurs late in mammalian development. Dev. Biol. 146, 167-178. differentiation. Cell 72, 903-917. Sutherland, C. J., Esser, K. A., Elsom, V. L., Gordon, M. L. and Rudnicki, M. A., Braun, T., Hinuma, S. and Jaenisch, R. (1992). Hardeman, E. C. (1993). Identification of a program of contractile protein Inactivation of MyoD in mice leads to upregulation of the myogenic HLH gene expression initiated upon skeletal muscle differentiation. Dev. Biol. gene Myf-5 and results in apparently normal muscle development. Cell 71, 196, 25-36. 383-390. Taylor, A., Erba, H. P., Muscat, G. E. and Kedes, L. (1988). Nucleotide Rudnicki, M. A., Schnegelsberg, P. N., Stead, R. H., Braun, T., Arnold, H. sequence and expression of the human skeletal alpha-actin gene: evolution H. and Jaenisch, R. (1993). MyoD or Myf-5 is required for the formation of functional regulatory domains. Genomics 3, 323-336. of skeletal muscle. Cell 75, 1351-1359. Venuti, J. M., Morris, J. H., Vivian, J. L., Olson, E. N. and Klein, W. H. Schevzov, G., Lloyd, C. and Gunning, P. (1992). High level expression of (1995). Myogenin is required for late but not early aspects of myogenesis transfected beta- and gamma-actin genes differentially impacts on myoblast during mouse development. J. Cell Biol. 128, 563-576. cytoarchitecture. J. Cell Biol. 117, 775-785. Wada, A., Fukuda, M., Mishima, M. and Nishida, E. (1998). Nuclear export Schevzov, G., Lloyd, C., Hailstones, D. and Gunning, P. (1993). Differential of actin: a novel mechanism regulating the subcellular localization of a regulation of tropomyosin isoform organization and gene expression in major cytoskeletal protein. EMBO J. 17, 1635-1641. response to altered actin gene expression. J. Cell Biol. 121, 811-821. Wade, R. and Kedes, L. (1989). Developmental regulation of contractile Schevzov, G., Lloyd, C. and Gunning, P. (1995). Impact of altered actin gene protein genes. Annu. Rev. Physiol. 51, 179-188. expression on vinculin, talin, cell spreading, and motility. DNA Cell Biol. Wade, R., Sutherland, C., Gahlmann, R., Kedes, L., Hardeman, E. and 14, 689-700. Gunning, P. (1990). Regulation of contractile protein gene family mRNA Schevzov, G., Gunning, P., Jeffrey, P. L., Temm-Grove, C., Helfman, D. pool sizes during myogenesis. Dev. Biol. 142, 270-282. M., Lin, J. J., Weinberger, R. P., Sotiropoulos, A., Gineitis, D. J., Wang, Y. C. and Rubenstein, P. A. (1992). Choice of 3′ cleavage/ Copeland and Treisman, R. (1997). Tropomyosin localization reveals polyadenylation site in beta-tropomyosin RNA processing is differentiation- distinct populations of microfilaments in neurites and growth cones. Signal- dependent in mouse BC3H1 muscle cells. J. Biol. Chem. 267, 2728-2736. regulated activation of serum response factor is mediated by changes in actin Weinberger, R. P., Henke, R. C., Tolhurst, O., Jeffrey, P. L. and Gunning, dynamics. Mol. Cell. Neurosci. 8, 439-454. P. (1993). Induction of neuron-specific tropomyosin mRNAs by nerve Shani, M., Dekel, I. and Yoffe, O. (1988). Expression of the rat myosin growth factor is dependent on morphological differentiation. J. Cell Biol. light-chain 2 gene in transgenic mice: stage specificity, developmental 120, 205-215. regulation, and interrelation with the endogenous gene. Mol. Cell. Biol. 8, Weintraub, H. (1993). The MyoD family and myogenesis, redundancy, 1006-1009. networks, and thresholds. Cell 75, 1241-1244. Smyth, C. M., Helmer, M. A., Dalla, P. L. and Rowe, P. B. (1993). Flow Yaffe, D. and Saxel, O. (1977). Serial passaging and differentiation of cytometric DNA analyses of frozen samples from children’s solid tumors. myogenic cells isolated from dystrophic mouse muscle. Nature 270, 725- Pathol. 25, 388-393. 727. Sorrentino, V., Pepperkok, R., Davis, R. L., Ansorge, W. and Philipson, L. Yao, X., Chaponnier, C., Gabbiani, G. and Forte, J. G. (1995). Polarized (1990). Cell proliferation inhibited by MyoD1 independently of myogenic distribution of actin isoforms in gastric parietal cells. Mol. Biol. Cell. 6, 541- differentiation. Nature 345, 813-815. 557. Sotiropoulos, A., Gineitis, D., Copeland, J. and Treisman, R. (1999). Zhang, W., Behringer, R. R. and Olson, E. N. (1995). Inactivation of the Signal-regulated activation of serum response factor is mediated by changes myogenic bHLH gene MRF4 results in up-regulation of myogenin and rib in actin dynamics. Cell 98, 159-169. anomalies. Genes Dev. 9, 1388-1399. Southern, P. J. and Berg, P. (1982). Transformation of mammalian cells to Zhao, K., Wang, W., Rando, O. J., Xue, Y., Swiderek, K., Kuo, A. and antibiotic resistance with a bacterial gene under control of the SV40 early Crabtree, G. R. (1998). Rapid and phosphoinositol-dependent binding of region promoter. J. Mol. Appl. Gen. 1, 327-341. the SWI/SNF-like BAF complex to chromatin after T lymphocyte Sutherland, C. J., Elsom, V. L., Gordon, M. L., Dunwoodie, S. L. and signaling. Cell 95, 625-636.