TAZ exhibits phase separation properties and interacts with Smad7 and β-catenin to repress skeletal myogenesis

Development and maintenance of tissues and organs requires a delicate balance between proliferation, apoptosis and differentiation (Varelas, 2014). A search for mutations that cause tissue overgrowth in Drosophila melanogaster precipitated the discovery of a conserved kinase cascade comprising of Hippo and Warts kinase and adaptor proteins Salvador and Mob (Jia et al., 2003; Justice et al., 1995; Pantalacci et al., 2003; Tapon et al., 2002; Udan et al., 2003; Wu et al., 2003), collectively referred to as Hippo signaling. Activation of Hippo signaling promotes activation of Hippo kinases resulting in phosphorylation of a transcriptional regulator, Yorkie, and its subsequent sequestration in the cytoplasm, thus inhibiting its functional association with another transcription factor (Scalloped) (Huang et al., 2005). Yes-associated protein  (YAP; also known as YAP1), is the mammalian homolog of Yorkie in vertebrates, and MST1 and MST2 (STK4 and STK3), and LATS1 and LATS2 are homologs of the Drosophila Hippo kinases, indicating the evolutionary conservation of Hippo signaling (Sudol, 1994; Sudol et al., 1995). TAZ (also known as WWTR1), a paralog of YAP in vertebrates, was subsequently characterized as a novel transcriptional co-activator regulated by interactions with 14-3-3 and PDZ domain proteins (Kanai et al., 2000). TAZ phosphorylation on a conserved serine residue (Ser89 in human TAZ; equivalent to Ser127 in human YAP) results in inhibition of TAZ transcriptional co-activation through 14-3-3 protein-mediated cytoplasmic retention.

A number of studies have proposed extensive crosstalk between Hippo signaling and other cellular signaling pathways. Documentation of crosstalk between TGF-β and Hippo signaling was first reported in a yeast two-hybrid screen where YAP1 was identified as a novel Smad7-interacting protein (Ferrigno et al., 2002). In Drosophila, Yorkie interacts with and promotes Mad-dependent transcription in response to DPP signaling (Alarcon et al., 2009; Oh and Irvine, 2011). In another high-throughput screen aimed at identifying novel regulators of TGF-β signaling in human embryonic cells, TAZ was identified as a mediator of Smad (the vertebrate homolog of MAD) nucleocytoplasmic shuttling that is essential for TGF-β signaling. YAP was also shown to bind BMP-regulated Smad1, Smad5 and Smad8 and TGF-β regulated Smad2 and Smad3. The YAP–Smad1 complex is regulated by the mediator-associated kinases CDK8 and CDK9, which phosphorylate the Smad1 linker domain to expose a PPXY motif and promote binding to the YAP WW domains (Alarcon et al., 2009). Thus, there is extensive documented crosstalk between the Hippo and TGF-β pathways.

In the absence of Wnt signaling, components of the β-catenin destruction complex have been shown to sequester TAZ, promoting its degradation. β-catenin contributes to this by bridging TAZ to the ubiquitin ligase β-TrCP (Azzolin et al., 2014). Receptor activation by Wnt ligands, allows escape of β-catenin from its destruction complex, which concomitantly also stabilizes TAZ expression, thereby promoting nuclear accumulation of both β-catenin and TAZ (Azzolin et al., 2012). TAZ-mediated transcription and protein stabilization are a general feature of the Wnt response in a variety of cellular model systems suggesting that the β-catenin–TCF complex is not the sole modality by which the Wnt/β-catenin pathway can regulate gene expression. Hence, Wnt signaling can promote the nuclear accumulation of both β-catenin and TAZ. The consequences of crosstalk between these signaling pathways is of particular interest for skeletal muscle biology given that TGF-β, Wnt and Hippo signaling have all been implicated in regulating various aspects of skeletal muscle development and post-natal physiology (Rodriguez et al., 2014; Suzuki et al., 2015; Wackerhage et al., 2014; Watt et al., 2018).

Therefore, in view of the putative connections between TAZ and both Smad7 and β-catenin, we questioned whether TAZ modulates the function of Smad7 and β-catenin in myogenic cells that we have previously reported (Tripathi et al., 2019). Here, we document a protein–protein interaction between Smad7 and TAZ in mouse muscle cells which functions to repress Smad7 and β-catenin function at muscle-specific promoters, such as the well-characterized muscle creatine kinase (Ckm) gene. In these studies, we also document that TAZ exhibits robust phase separation properties in myogenic cells. Collectively, these observations are consistent with the role of Hippo signaling in myogenic cells which, when off, promotes proliferation due to TAZ nuclear accumulation and activation of genes associated with proliferation, which is consistent with results from a previous study in which both Yap and TAZ were documented to enhance proliferation of satellite cells (Sun et al., 2017). We now propose that, as well as positively enhancing the expression of proliferative genes, nuclear TAZ also suppresses differentiation by acting as a co-repressor of the myogenic transcription complex containing Smad7 and β-catenin. Thus, our data indicate that, when Hippo signaling is activated, TAZ is phosphorylated and retained in the cytoplasm, which concomitantly de-represses the nuclear activity of Smad7 and β-catenin at muscle promoters. These data support a model in which TAZ acts as a fulcrum to balance the ratio of proliferation and differentiation in myogenic cells. These studies identify TAZ as a potential determinant of muscle differentiation, having implications for control of muscle mass during development, growth and regeneration.

Previously, the Hippo effector Yap was identified as a novel Smad7-interacting protein in Cos-7 cells (Ferrigno et al., 2002). We initially tested whether the YAP paralog TAZ also interacts with Smad7 since TAZ is expressed at a much higher level than YAP in muscle cells and, by immunofluorescence analysis, the nuclear accumulation of TAZ is reduced in differentiating cells (Fig. 1A,B). We have previously documented that Smad7 is expressed in muscle cells and plays an important role in muscle-specific gene transcription (Kollias et al., 2006; Miyake et al., 2010; Tripathi et al., 2019) and this has been confirmed in vivo (Cohen et al., 2015). Here, we show that TAZ is expressed in both proliferation and differentiation stages of cultured C2C12 muscle cells, and the nuclear accumulation of TAZ is reduced in differentiation conditions (Fig. 1A). To initially test for the interaction between TAZ and Smad7 biochemically, Flag–Smad7 was ectopically expressed in C2C12 cells, and total protein lysates were subjected to immunoprecipitation (IP) using anti-Flag magnetic beads. Lysate from ectopically expressed Flag–Trim28 was used as a negative control. Immunoblot analysis of the eluates from the Flag–Smad7 IP, but not from Flag–Trim28 IP revealed the presence of TAZ, indicating that they are potentially constituents of the same protein complex (Fig. 1B). Therefore, we hypothesized that TAZ modulates the myogenic transcriptional complex that we and others have previously documented (Rampalli et al., 2007; Tripathi et al., 2019; see schematic in Fig. 1C).

Previously, we documented the contribution of Smad7 to the transcriptional control of the muscle creatine kinase (Ckm) promoter in co-operation with MyoD (also known as MyoD1) and β-catenin (Fig. 1C; Tripathi et al., 2019). The Ckm promoter has been used in many studies in the myogenesis field as a paradigm of muscle-specific transcriptional control. We therefore tested whether TAZ could modulate this complex through its interaction with Smad7. Ectopic expression of TAZ alone or in combination with Smad7 demonstrated that TAZ potently inhibits the transcriptional induction of Ckm by Smad7 (Fig. 2A). To further validate these observations, we utilized two independent TAZ-directed siRNAs to deplete the levels of endogenous TAZ (Fig. 2B) and then performed the Ckm luciferase reporter assay in TAZ-depleted versus normal cells. Depletion of endogenous TAZ enhanced Ckm transcriptional activity by Smad7 (Fig. 2C,D). Moreover, depletion of endogenous TAZ resulted in an overall enhancement of Ckm transcriptional activity (Fig. 2C,D). In addition, ectopic TAZ expression resulted in a corresponding reduction in CKM protein level in a differentiation time course experiment (Fig. 2E; Fig. S2 documents total protein in this assay confirming the CKM downregulation in TAZ-expressing cells). It should be noted that the transfection efficiency for ectopic TAZ expression is not complete, and we suggest that the residual CKM expression is likely derived from the untransfected cells. Collectively, these data support a repressive role of TAZ during muscle cell differentiation. These data also confirm a previous study in which the repressive role of TAZ on myogenesis was observed (Huraskin et al., 2016).

Given that our previous study reported a direct interaction between Smad7 and β-catenin and their functional co-operativity is required for their function on the Ckm promoter, we determined whether TAZ can directly target β-catenin activity using a more specific assay system. For this, we utilized the TOP flash reporter gene that contains multimerized binding sites for TCF/LEF, which serve as an indicator of β-catenin-mediated transcriptional activation. FOP flash contains mutated TCF/LEF-binding sites serving as a corresponding negative control in this assay (Anand et al., 2011). Ectopic expression of TAZ resulted in robust inhibition of β-catenin activity (Fig. 3A), while no appreciable effect of TAZ was observed on the FOP flash reporter gene (Fig. S1). Protein level expression of exogenous constructs was confirmed by western blot analysis (Fig. 3A, lower panels). These data indicate that TAZ specifically represses β-catenin-induced transcriptional activation. Next, we tested whether the TAZ S89A mutation inhibits β-catenin activity similar to wild type, which it did (Fig. 3B). The logic of using TAZ S89A was based on its property of evading the cytoplasmic degradation machinery, which has previously been shown to modulate the protein levels of both β-catenin and TAZ. These data indicate that the inhibitory effect of TAZ on β-catenin transcriptional activation does not appreciably involve the cytoplasmic degradation machinery affecting β-catenin levels. This is also supported in the protein blots, which show that ectopic expression of TAZ did not affect endogenous or exogenously transfected β-catenin levels (Fig. 3A). Next, since TAZ is perhaps better known for its role as a transcriptional co-activator, we tested whether TAZ enhances the well-characterized HIP/HOP reporter gene system in the same cellular system. HOP is comprised of multimerized TEAD-binding sites driving a luc reporter gene, which serves as a readout for YAP/TAZ activity (HIP is the same reporter gene, but with mutated TEAD sites serving as a negative control) (Kim and Gumbiner, 2015). In this assay, we observed that both wild-type TAZ and TAZ S89A robustly activate the HOP reporter gene under the same cellular conditions in which we documented repression of the β-catenin-dependent TOP flash reporter gene (Fig. 3C). Thus, these data indicate that whereas TAZ strongly represses β-catenin activity in muscle cells, it conversely potently activates the Hippo-responsive HIP reporter system under identical cellular conditions. We went on to further test whether TAZ S89A represses some other well-characterized myogenic reporter genes. For this, we chose to analyze the myogenin promoter (Myog-luc) and the Myod promoter regions (myod -2.5 kb/-275-luc). In Fig. 3D, the ability of MyoD to activate its own promoter was repressed by TAZ S89A. In Fig. 3E, we document that TAZ S89A also inhibits the activation of the Myog promoter in myogenic cells, and we report a corresponding decrease in myogenin protein level with ectopic TAZ expression (Fig. 3F). Collectively, these data indicate that TAZ potently represses a number of myogenic promoters (Ckm, Myog and Myod), while still retaining its potent co-activator potential at a TEAD responsive reporter gene (HIP/HOP system). In view of the potent repression of a number of critical myogenic regulators in our assays, we questioned whether TAZ might influence the commitment to the myogenic lineage. To do this we utilized the well-characterized myogenic ‘conversion assay’ in which MyoD expression converts 10T1/2 fibroblasts into myogenic cells. We used fluorescently tagged MyoD (GFP–MyoD) and TAZ S89A (mCherry–TAZ S89A) to assess the transfected cells in this assay. As depicted in Fig. 3G, we observed that, as expected, MyoD (GFP tagged) ‘converted’ cells into multinucleated myotubes and mCherry alone had no effect on myogenic conversion (lower panel). Conversely, when GFP–MyoD was co-transfected with mCherry–TAZ S89A there was no myogenic conversion detected (upper panel). We also assessed MyoD-mediated myogenic conversion in this assay system by measuring the cell lengths, based on the logic that elongated multinucleated muscle cells have a considerably greater length than the parental 10T1/2 cells (Fig. 3G, graph). In this alternative myogenic model system, these data indicate that TAZ robustly suppresses myogenic conversion. As shown in Fig 3H,I, we observed that whereas MyoD and β-catenin have no effect on TAZ activity using the HIP/HOP reporter system, TAZ potently represses β-catenin activity on the β-catenin-dependent TOP flash reporter gene. These data consistently support the idea that TAZ has an asymmetric function as both a transcriptional activator and repressor depending on the enhancer/promoter element context.

Cell–cell contact has been identified as a regulator of Hippo signaling. Activation of Hippo signaling at high cell densities results in a well-characterized phosphorylation and cytoplasmic sequestration of TAZ at Ser89, promoting its binding to 14-3-3 proteins in the cytoplasm (Wada et al., 2011). Therefore, we were interested to determine the levels of endogenous Ser89 phosphorylated TAZ during myogenic differentiation. Time course analysis of total protein lysates from differentiating cells indicate that TAZ levels remain relatively unchanged during differentiation, although TAZ is much more abundant than YAP in muscle cells. Importantly, hyperphosphorylation of TAZ was pronounced when the cells were differentiating, suggesting activation of Hippo signaling in differentiation conditions (Fig. 4A,B). Immunofluorescence analysis of TAZ and phospho-TAZ also indicates that TAZ localization shifts from being predominantly nuclear during the myoblast stage (low cell density) to a more diffuse cytoplasmic localization in differentiating myotubes (high cell density), although some TAZ is still retained in the nucleus. Phosphorylated TAZ is undetectable in mononucleated myoblasts (Fig. 4C) and primarily localized in the cytoplasm in differentiated myotubes (Fig. 4D). Taken together, these observations are consistent with a model in which the nuclear function of TAZ predominates in proliferating myoblasts. However, under differentiation conditions, activation of Hippo signaling results in TAZ Ser89 phosphorylation and increased retention in the cytoplasm. Based on the repressive function of TAZ at the Ckm gene that we have documented here, we propose that TAZ represses myogenic transcription complexes under proliferative conditions, and activation of Hippo signaling during differentiation relieves this repression by the phospho-dependent sequestration of TAZ in the cytoplasm. Analysis of proliferating C2C12 cells by immunofluorescence suggested that TAZ forms a pattern of speckled localization, which can be indicative of phase separation properties (Fig. 4E). These initial observations of the localization of endogenous TAZ in muscle cells led us to speculate that TAZ might form phase condensates, and we therefore subsequently proceeded to test this possibility.

Liquid–liquid phase separation (LLPS) of proteins and nucleic acids into condensates (also referred to as nuclear speckles, nuclear bodies or biomolecular condensates) is invoked by weak multivalent interactions between molecules (Banani et al., 2017; Shin and Brangwynne, 2017). Phase separation has recently been implicated in heterochromatin domain formation (Larson et al., 2017; Strom et al., 2017), transcriptional control through co-activator condensation (Boija et al., 2018; Sabari et al., 2018), and the clustering of nuclear enhancer elements and super-enhancers in gene control (Hnisz et al., 2017). Our observations indicate that TAZ displays a number of properties associated with phase separation that may underlie its diverse activities in both activation and repression of gene regulation.

As observed in fluorescence live-cell imaging analysis, ectopically expressed EYFP–TAZ wild-type (WT) and EYFP-TAZ S89A (nuclear TAZ) formed a similar localization pattern of spherical structures in the nucleus (Fig. 5A; Movies 1 and 2). Ectopic EYFP-tagged TAZ expression therefore allowed us to perform more detailed and higher resolution analysis of the TAZ nuclear speckles. Both EYFP–TAZ (WT) and EYFP–TAZ S89A formed dynamic spherical subcellular structures. We further observed that TAZ condensates were, in some cases, juxtaposed with heterochromatin, with projections extending into the heterochromatin domain (Fig. 5A; Movie 1). Conversely, we also noted patterns of localization within the same nuclei in which the TAZ condensates are not associated with heterochromatin. These differential associations may therefore reflect variant TAZ-containing complexes involved in gene activation or repression. Furthermore, fusion of the spherical structures was observed in time-lapse live-cell imaging analysis (Fig. 5B; Movie 2). These observations of the robust capability of TAZ to form condensates was recently reported under in vitro conditions and also in cultured MCF-10A cells (Franklin and Guan, 2020). In addition, YAP has also been shown to exhibit LLPS (Cai et al., 2019; Yu et al., 2021).

Next, we assessed whether β-catenin might also be localized in these phase-separated condensates with TAZ and Fig. 5C depicts that it is. We observed two types of phase condensates in these experiments. There was a TAZ-negative central region in the center of some spherical condensates (Fig. 5C, upper panels) while in others then whole condensate was TAZ positive (Fig. 5C, lower panels). At this point we cannot determine whether this is related to the expression level or the state of maturation of the TAZ condensates. To further test how robust this protein–protein interaction is, we used a nanotrap assay to capture EYFP–β-catenin at the interior of nuclear envelope using GBP–Lamin B1 as the tether. Fig. 5D indicates that this approach was successful in re-directing β-catenin to the nuclear envelope. Strikingly, mCherry–TAZ S89A was subsequently enriched at the nuclear envelope due to its robust protein interaction with β-catenin (Fig. 5D; see Fig. S4 for additional images). Finally, some endogenous TAZ and β-catenin foci were colocalized in the nuclei of mononucleated myogenic reserve cells, as revealed by immunofluorescence analysis (Fig. 5E). Collectively, these features of TAZ and β-catenin localization and dynamic assembly into condensates are consistent with LLPS properties recently reported for YAP, TAZ and other nuclear regulatory proteins (Boeynaems et al., 2018; Boija et al., 2018; Hnisz et al., 2017; Sabari et al., 2018).

The Hippo pathway has garnered widespread attention for its role in both physiological and pathological conditions in a variety of tissues. It has now been implicated in organ growth and development, stem cell biology, regeneration and tumorigenesis (Moroishi et al., 2015; Piccolo et al., 2014; Yu et al., 2015). The Hippo pathway is composed of a core kinase module and a transcriptional regulatory complex comprised of TAZ and/or YAP as the main effectors. To date, many cellular studies have documented how TAZ and YAP phosphorylation and nuclear localization are controlled. More recently, there have been efforts to reveal regulatory mechanisms governing how YAP and TAZ can exert selective functions on the transcriptional apparatus assembled at different cis-regulatory elements in the genome (Totaro et al., 2018). In this study, we document how nuclear TAZ inhibits myogenic differentiation by repressing components of a well-characterized muscle-specific transcription complex.

In particular, we characterize components of a pro-myogenic protein complex comprising Smad7 and β-catenin that is antagonized by TAZ protein interactions, thus serving to suppress transcriptional activation of muscle-specific genes, maintaining the cells in an undifferentiated, proliferative state. These data support a model for nuclear TAZ function in suppressing the myogenic differentiation program by targeting a key transcriptional regulatory complex assembled on muscle promoters. Phosphorylation-dependent cytoplasmic sequestration of TAZ by activation of Hippo kinases relieves this repression, providing a potential switch to allow the myogenic differentiation program to proceed. Interestingly, as reported by others recently (Lu et al., 2020), we observe TAZ in myonuclear speckles consistent with key features of phase-separated condensates, including high sphericity and fusion behavior. These studies outline a functional reciprocity between myogenic differentiation and Hippo signaling that may have important implications for our understanding of the co-ordination of cellular proliferation and differentiation in skeletal muscle growth control.

While YAP and TAZ are often considered functionally redundant, there are structural and functional observations indicating that they might also have non-overlapping roles (Plouffe et al., 2018). Despite their high sequence similarity, there are nevertheless some potentially important differences in their structure. Notably, although both proteins contain WW domains, YAP harbors two tandem WW domains, whereas TAZ has one. Also, YAP contains an SH3 binding domain, which is not present in TAZ. Of some interest for our study, in view of the TAZ connection with β-catenin, is the reported observation that GSK3β, a well-characterized regulator of β-catenin, can directly phosphorylate TAZ affecting its stability. Several reports in different model systems indicate that TAZ cannot compensate for YAP physiologically, supporting the idea that, while they are both targets of the Hippo kinase pathway, they also retain distinct roles (Hossain et al., 2007; Makita et al., 2008; Morin-Kensicki et al., 2006).

Given the structural and physiological differences between YAP and TAZ it is possible that their interactions with transcriptional complexes may differ. This is particularly highlighted in recent studies documenting differential properties of YAP and TAZ in promoting the formation of LLPS with different factors being sequestered within the condensates (Lu et al., 2020; Cai et al., 2019). It is now becoming increasingly apparent that the subcellular assemblage of membrane-less compartments through LLPS is an essential feature of temporal and spatial control of various biomolecular processes (Boeynaems et al., 2018; Boija et al., 2018; Cai et al., 2019; Hyman et al., 2014; Lu et al., 2019; Sabari et al., 2018). Here, we confirm and extend these observations in reporting that TAZ forms liquid–liquid phase-like condensates in cultured muscle cells. Moreover, hyperosmotic stress promotes the rapid assembly and fusion of spherical TAZ condensates leading to the compartmentalization of TAZ and its interacting partners. Our observations confirm a recent report documenting that TAZ forms liquid-like condensates that concentrate transcriptional regulatory proteins such as co-activators and the elongation machinery (Lu et al., 2020). One of the primary cellular functions of Hippo signaling is to control tissue growth by regulating the proliferative expansion of progenitor pools. Based on our data, we speculate that TAZ nuclear condensates sequester components of the myogenic nuclear machinery involved in cellular differentiation. Thus, TAZ-mediated ‘hijacking’ and repression of the pro-differentiation machinery might be deemed complementary to its pro-proliferative activities underlying growth. A pro-proliferative role of TAZ in muscle satellite cells was reported in a previous study employing an in vivo strategy in satellite cells (Sun et al., 2017), which is consistent with our observations; however, that study also proposed a pro-differentiation effect of TAZ in the later stages of differentiation, which is not congruent with our observations in myoblasts. This could potentially be explained by the enhanced ‘early’ proliferation promoted by YAP and TAZ leading to an enhanced accumulation of progenitors and a more robust eventual differentiated phenotype in vivo. We would also contend that this might be further complicated by the potential for Hippo pathway activation in the in vivo context, which would lead to TAZ cytoplasmic retention, and thus sequestration away from the differentiation machinery due to activation of Hippo kinases when the cells eventually commit to the differentiation program. This would be consistent with our observations, since we expect that the repressive effect of TAZ on the differentiation machinery would only be operative under growth/proliferative conditions. It will be interesting to further assess and clarify these possibilities regarding the physiological role of TAZ in different muscle contexts in future.

The best-characterized partners for YAP and TAZ in transcriptional control are the TEAD1–TEAD4 factors, although to date the possibility of different partners for YAP and TAZ has not been exhaustively tested. Of particular relevance for the discussion of muscle transcription complexes is the previously reported interaction of TEAD factors with the MEF2 proteins, which are core regulators of the muscle transcription program (Chen et al., 2017; Maeda et al., 2002). It is perhaps coincidental, but noted, that p38 MAPKs are a potent and well-characterized regulator of MEF2 function, and it was recently documented that TEAD function is also targeted by p38 MAPKs (Lin et al., 2017). Moreover, our recent studies have documented an interaction between MEF2 and β-catenin that is p38 MAPK dependent (Ehyai et al., 2016), and a direct physical interaction between β-catenin and Smad7 (Tripathi et al., 2019), in a muscle context. Thus, the TAZ–Smad7 interaction reported here may intersect with these protein interactions. The schematic in Fig. 6 documents a working model of these interactions. Collectively, these observations support a model in which a core muscle transcriptional regulatory complex may be targeted by ancillary regulatory subunits, such as TAZ or β-catenin–Smad7, to repress or enhance its activity, respectively, depending on the cellular conditions. Collectively, these observations place TAZ in the cellular context of muscle-specific gene regulation. It will be of considerable interest to further examine the physiological role of TAZ in post-natal muscle growth and regeneration in health and disease.

C2C12 myoblasts and C3H10T1/2 cells were obtained from the American Type Culture Collection. Cells were cultured in growth medium (GM) consisting of high-glucose Dulbecco's modified Eagle's medium (DMEM, Gibco), 10% fetal bovine serum (FBS) and L-Glutamine (HyClone) supplemented with 1% penicillin-streptomycin (Invitrogen, Thermo Fisher Scientific). Myotube formation was induced by replacing GM with differentiation medium (DM), consisting of DMEM supplemented with 2% horse serum (Atlanta Biologicals) and 1% penicillin-streptomycin. Cells were maintained in an incubator at 95% humidity, 5% CO₂ and 37°C.

For ectopic protein expression, cells were transfected using the calcium phosphate precipitation method for transcription reporter assays. Cells were re-fed 16 h post-transfection and harvested. For small interfering RNA (siRNA) experiments, C2C12 myoblasts were transfected with Lipofectamine 2000 (Life Technologies) using instructions provided by the manufacturer and harvested 48 h later, unless otherwise indicated.

Knockdown of target genes was done using mission siRNA (Sigma-Aldrich). siTAZ#1 (SASI Mm01 00107364) and siTAZ#2 (SASI Mm01 00107368) and universal scrambled siRNA (SIC001) were used at 50 nM concentrations.

Expression plasmids for Myc-His-tagged full-length Smad7, FLAG-TAZ, β-catenin-Myc, transcription reporter assay constructs Ckm-luc and TOP/FOP flash have been described previously (Ehyai et al., 2016; Kollias et al., 2006; Miyake et al., 2010; Pagiatakis et al., 2017). HIP-Flash (#83466) and HOP-flash (#83467) were obtained from Addgene (Kim and Gumbiner, 2015). mCherry/EYFP tagged TAZ and β-catenin constructs were described previously (Ehyai et al., 2018; Pagiatakis et al., 2017). Myod (-2.5/-275)-luc was a kind gift from Dr Steven Tapscott, Fred Hutchinson Cancer Research Center (Seattle, USA) (Asakura et al., 1995). Myog-luc was described previously (Miyake et al., 2010).

Transcriptional reporter assays were performed using luciferase reporter plasmids along with expression constructs (indicated in figure legends) and a Renilla plasmid (pRL-Renilla, Promega) as an internal control. Cells were washed with 1× PBS and harvested in luciferase lysis buffer (20 mM Tris-HCl pH 7.4 and 0.1% Triton X-100). Enzymatic activity was measured in each sample on a luminometer (Lumat LB, Berthold) using luciferase assay substrate (E1501; Promega) or Renilla assay substrate (E2820; Promega). Luciferase activity values obtained were normalized to Renilla activity in the same cell extracts and expressed as fold activation to the control.

Total cellular protein extracts were prepared in NP-40 lysis buffer [0.5% NP40 (v/v), 50 mM Tris-HCl pH 8, 150 mM NaCl, 10 mM sodium pyrophosphate, 1 mM EDTA pH 8, and 0.1 M NaF supplemented with 1× protease inhibitor cocktail (P-8340; Sigma) and 0.5 mM sodium orthovanadate]. Protein concentrations were determined by a standard Bradford assay. Equivalent amounts of protein were denatured in SDS loading buffer at 100°C for 5 min and then run in SDS-polyacrylamide gels, followed by electrophoretic transfer onto Immobilon-FL PVDF membrane (Millipore) as directed by the manufacturer. Blots were incubated with blocking buffer that consisted of 5% milk in Tris-buffered saline with Tween 20 (TBS-T; 10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% Tween 20) prior to the incubation with primary antibody (please see below for the dilution of each antibody) at 4°C overnight with gentle agitation. After three washes with TBS-T, appropriate HRP-conjugated secondary antibodies (BioRad, 1:2000), were added for 2 h at room temperature. Protein–antibody immuno-complexes were detected with enhanced chemiluminescence western blotting substrate (Pierce, ThermoFisher).

Rabbit monoclonal for TAZ (mAb8418, 1:1000), phospho-TAZ (mAb59971, 1:1000) and β-catenin (pAb9562, 1:2000) were purchased from Cell Signaling. Monoclonal anti-Flag antibody (F1804, 1:1000) was from Sigma. Antibodies against Myc (9E10, 1:20), HA (12CA5, 1:20), MyHC (MF20, 1:20) and myogenin (F5D, 1:20) were purchased from DSHB. Antibodies against Actin (I-19; sc-1616, 1:1000), MyoD (C-20; sc-304, 1:2000), and CKM (G-9; sc-365046, 1:200) were purchased from Santa Cruz Biotechnology.

For co-immunoprecipitation assays (co-IP), C2C12 myoblasts were transfected with Flag-tagged Smad7 or Trim28 using Lipofectamine 2000. Cells were re-fed 16 h post-transfection and total protein lysates were harvested 24 h after the medium change. Proteins were extracted as described above. Immunoprecipitation was performed using anti-Flag M2 magnetic beads (Sigma, M8823) according to the manufacturer's instructions. Eluates were analyzed by western blotting as described above.

C2C12 cells were seeded on the glass-bottom dishes (MatTek) and fixed with 4% PFA. Cells were washed (three times) with 1× PBS and permeabilized with ice cold 90% Me-OH for 1 min. Cells were washed again (three times) with 1× PBS and incubated with blocking buffer (5% FBS in 1× PBS) for 2 h at room temperature, and then incubated with an indicated primary antibody in the blocking buffer overnight at 4°C. After removal of the unbound primary antibody, the cells were incubated with Alexa Fluor-conjugated secondary antibody (Life Technologies) in the blocking buffer at room temperature for 1.5 h. After washing with 1× PBS three times, the cells were subjected to imaging with a Zeiss Observer Z1 confocal fluorescent microscope equipped with Yokogawa CSU-X1 spinning disk. Images were recorded by AxioCam MRm camera (Zeiss) and processed by Zen 2.5 (blue edition) software (Zeiss).

C2C12 cells were seeded on polymer-coated glass-bottom dishes (#81158, ibidi). The cells were transfected with expression plasmids encoding the indicated fluorescently tagged proteins. The cell culture media was exchanged for FluoBrite (A1896701, ThermoFisher) with 10% FBS (HyClone). The cells were maintained in a moisturized environmental chamber (5% CO2) on the Zeiss Observer Z1 confocal fluorescent microscope platform during live cell imaging.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.259097.