TISSUE-SPECIFIC STEM CELLS aRandall Division of Cell and Common and Distinctive Functions of the Hippo Molecular Biophysics, King’s College London, London, UK; Effectors Taz and Yap in Skeletal Muscle Stem Cell bSchool of Medicine, Medical Sciences & Nutrition, Function University of Aberdeen, Foresterhill, Aberdeen, a b b a Scotland, UK; cSystems CONGSHAN SUN, VANESSA DE MELLO, ABDALLA MOHAMED, HUASCAR P. O RTUSTE QUIROGA, c b d,e,f Biology Ireland, Conway AMAYA GARCIA-MUNOZ, ABDULLAH AL BLOSHI, ANNIE M. TREMBLAY, c g b c Institute, Dublin, Ireland; ALEXANDER VON KRIEGSHEIM, ELAINA COLLIE-DUGUID, NEIL VARGESSON, DAVID MATALLANAS, d b,h a Stem Cell Program, HENNING WACKERHAGE, PETER S. ZAMMIT Children’s Hospital, Boston, Massachusetts, USA; Key Words. Taz • Yap • Tead • Satellite cells • Muscle stem cells eDepartment of Stem Cell and Regenerative Biology, Harvard University, ABSTRACT Cambridge, Massachusetts, USA; fHarvard Stem Cell Hippo pathway downstream effectors Yap and Taz play key roles in cell proliferation and regen- Institute, Cambridge, eration, regulating expression especially via Tead transcription factors. To investigate their Massachusetts, USA; gCentre role in skeletal muscle stem cells, we analyzed Taz in vivo and ex vivo in comparison with Yap. for Genome Enabled Biology Small interfering RNA knockdown or retroviral-mediated expression of wild-type human or con- and Medicine, School of stitutively active TAZ mutants in satellite cells showed that TAZ promoted proliferation, a func- Medicine, Medical Sciences tion shared with YAP. However, at later stages of myogenesis, TAZ also enhanced myogenic and Nutrition, University of differentiation of myoblasts, whereas YAP inhibits such differentiation. Functionally, while mus- Aberdeen, Foresterhill, cle growth was mildly affected in Taz (gene Wwtr1–/–) knockout mice, there were no overt Aberdeen, Scotland, UK; effects on regeneration. Conversely, conditional knockout of Yap in satellite cells of Pax7Cre-ERT2/1: h Faculty of Sport and Health Yapfl8x/fl8x:Rosa26Lacz mice produced a regeneration deficit. To identify potential mechanisms, Sciences, Technical University microarray analysis showed many common TAZ/YAP target , but TAZ also regulates some of Munich, Munich, Germany genes independently of YAP, including myogenic genes such as Pax7, Myf5,andMyod1 (ArrayEx- Correspondence: Peter S. press–E-MTAB-5395). Proteomic analysis revealed many novel binding partners of TAZ/YAP in Zammit Ph.D., King’s College myogenic cells, but TAZ also interacts with distinct from YAP that are often involved in London, Randall Division of Cell myogenesis and aspects of cytoskeleton organization (ProteomeXchange–PXD005751). Neither TAZ and Molecular Biophysics, nor YAP bind members of the Wnt destruction complex but both regulated expression of Wnt London SE1 1UL, UK. Telephone: and Wnt-cross talking genes with known roles in myogenesis. Finally, TAZ operates through Tead4 144-207-848-8217; Fax: 144- to enhance myogenic differentiation. In summary, Taz and Yap have overlapping functions in pro- 207-848-6435; moting myoblast proliferation but Taz then switches to enhance myogenic differentiation. STEM e-mail: [email protected] CELLS 2017;35:1958–1972

Received June 14, 2016; accepted for publication January SIGNIFICANCE STATEMENT 7, 2017; first published online in STEM CELLS EXPRESS June 7, Hippo pathway effectors Yap and Taz play key roles in cell proliferation and tissue growth. We 2017. analyzed Taz in comparison with Yap in muscle stem cells. Taz promoted proliferation, a func- tion shared with Yap. However, Taz also enhanced differentiation into myotubes, unlike Yap. VC AlphaMed Press –/– 1066-5099/2017/$30.00/0 Muscle growth was affected in Taz (Wwtr1 ) knockout mice, while satellite cell-specific condi- tional knockout of Yap produced a regeneration deficit. Taz regulates some genes independently http://dx.doi.org/ of Yap, and Taz also interacts with proteins distinct from Yap, mainly involved in myogenesis/ 10.1002/stem.2652 cytoskeleton. In particular, Taz operates through Tead4 to enhance differentiation. In summary, This article was published online Taz and Yap have overlapping functions in promoting myoblast proliferation but later, Taz also on 28 June 2017. However, enhances myogenic differentiation. proof typesetting was subse- quently noticed to be incom- plete. This notice is included in the online and print versions to Many other signaling modules also regulate indicate that these have been INTRODUCTION Yap and Taz activity, such as mechanotransduc- finalized 5 July 2017. Transcriptional cofactors Yap (Yap1) and Taz tion [4], glucose-signaling [5], G- cou- pled receptors (GPCR) [6], Wnt [7–9], Smad This is an open access article (Wwtr1) mainly regulate by under the terms of the Creative binding Tead1–4 transcription factors. Together [10, 11], Notch [12], Pten-Akt-mechanistic Commons Attribution License, Yap, Taz, and Teads are the nexus of the Hippo target of rapamycin (mTOR) [13–16], and Lkb1- which permits use, distribution network that includes the Ampk [17–20]. Yap and Taz are inhibited by and reproduction in any Hippo cascade, comprising Mst1 phosphorylation of multiple HXRSS motifs by medium, provided the original work is properly cited. (Stk4), Mst2 (Stk3), Lats1, and Lats2 [1–3]. Lats1/2, which promotes localization to the

STEM CELLS 2017;35:1958–1972 www.StemCells.com VC 2017 The Authors STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 2017 Sun, De Mello, Mohamed et al. 1959 cytosol, 14–3-3 binding, and degradation [3, 21]. Yap Ser127 mice were donated by Fernando Camargo. Breeding/experi- and Taz Ser89 are key phosphorylation sites, and the mutation mental procedures were approved by the Ethical Review Pro- of Ser127 or Ser89 to alanine in Yap S127A and Taz S89A cess Committee of King’s College London, and performed respectively prevents phosphorylation at these residues and under the Animals (Scientific Procedures) Act 1986. results in constitutive activity. Signaling modules involving Yap/Taz control skeletal mus- Cell Culture cle myogenesis and adaptation to exercise. Tead transcription Satellite cells/myofibers were isolated as described previously factors bind CATTCC/GGAATG (MCAT or GTIIC motifs) found [41]. Briefly, murine extensor digitorum longus (EDL) myofib- in/near promoters and enhancers of cardiac and skeletal mus- ers were isolated by 0.05% type I collagenase for 2 hours. cle genes [3, 22–24]. However, Tead1 binds muscle genes Myofibers were washed in GlutaMax Dulbecco’s modified repressed in rhabdomyosarcoma [25], consistent with obser- Eagle’s medium (DMEM) (ThermoFisher Scientific (http://cor- vation that Yap and Taz can also repress gene expression [26]. porate.thermofisher.com/en/home.html), Waltham, Massachu- Satellite cells are responsible for postnatal skeletal muscle setts, USA) and plated on matrigel-coated plates in growth growth, hypertrophy, and repair/regeneration [27], and we medium (GlutaMax DMEM, 30% fetal bovine serum [FBS], 1% have shown that Yap promotes proliferation of myoblasts [28, chick embryo extract [CEE], 10 ng/mL basic fibroblast growth 29] but inhibits myogenic differentiation. YAP1 S127A expres- factor, 1% penicillin-streptomycin (Sigma-Aldrich (http://www. sion in activated, but not quiescent, satellite cells causes sigmaaldrich.com/united-kingdom.html), St. Louis, Missouri, embryonal rhabdomyosarcoma-like tumours with short latency USA.) at 378C with 5% CO2. After 72 hours, myofibers were and 100% penetrance [25]. Expression of YAP1 S127A in mus- removed, and satellite cells passaged and preplated for 30 cle fibers causes myopathy [30], whereas other types of Yap minutes to remove fibroblasts, before being plated on delivery or Yap mutants can cause hypertrophy [31, 32], with matrigel-coated plates. Differentiation was induced in differ- outcome likely dependent on Yap levels. entiation medium (GlutaMax DMEM, 2% horse serum, 1% Taz harbors the same functionally important WW and penicillin–streptomycin). Nonadherent cultures: myofibers Tead-binding domains as Yap and often acts as a paralogue, were cultured in GlutaMax DMEM, 10% horse serum, 0.5% but not always [2, 3]. This is demonstrated by Yap or Taz CEE, 1% penicillin–streptomycin. knockout mice: while Yap knockout causes early embryonic lethality [33], 50% of Taz-knockout (Wwtr1–/–) mice are viable Muscle Injury but develop glomerulocystic kidney disease [34]. Earlier To recombine via the loxP sites flanking Yap exons 1 and 2, reports suggest that Yap and Taz also have divergent functions 200 mg of Tamoxifen/gram body weight (Sigma T5648) was in the skeletal muscle lineage: both Yap and Taz promote skel- injected intraperitoneally in sunflower oil/5% ethanol for 3 etal muscle fiber hypertrophy and regeneration [35, 36], but consecutive days, followed by maintenance on a tamoxifen- only Taz promotes fusion into multinucleated myotubes [29, containing diet (Tekland). Injury was induced in tibialis ante- 37, 38]. rior (TA) by 30 mL intramuscular injection of 20 mM cardio- The distinct functions of YAP and TAZ are poorly under- toxin (CTX)/saline. stood, so we investigated their regulation, function, target Retroviral Expression and Small Interfering RNA genes, and binding partners in murine myoblasts. Taz and Yap have overlapping functions in promoting satellite cell prolifer- Wild-type (WT) TAZ, TAZ S89A, YAP S127A, or WT YAP was ation. However, later in myogenesis, Taz enhances myogenic subcloned into a pMSCV-IRES-eGFP retroviral expression back- differentiation and fusion, a function not shared by Yap. Taz- bone (Addgene Plasmids 24809, 24815, 17791 and 17790) knockout (Wwtr1–/–) mice had no overt muscle regeneration creating pMSCV-3xFlag TAZ-IRES-eGFP and pMSCV-3xFlag-TAZ phenotype, unlike mice with conditional knockout of Yap in S89A-IRES-eGFP [42]. Empty vector was negative control. Ret- satellite cells. To understand mechanism, we examined gene roviruses were packaged in HEK293T cells using standard expression, and found, for example, that YAP and TAZ both methods. Medium was changed 1 hour before transfection/ alter expression of multiple Wnt pathway genes. In addition transduction. Retroviral suspension diluted 1:4 with polybrene to common targets, TAZ also regulates a separate set of genes (4 lg/mL) was added for 6 h, before changing medium. from YAP during myogenesis, including myogenic factors Pax7, Taz small interfering RNA (siRNA) (Ambion (http:// Myf5, and MyoD. Using proteomics to identify binding part- www.ambion.com/), Foster City, California, USA, s97145) and ners of TAZ and YAP, we discovered both common and distinct Yap siRNA (Ambion, s202423) were used as per manufac- interactors, with TAZ binding proteins that regulate myogene- turer’s instructions. For plated satellite cells, 25 pmol of siRNA sis and aspects of cytoskeleton organization. Finally, we found with Lipofectamine RNAiMax (ThermoFisher Scientific) was that TAZ operates through Tead4 in myoblasts to enhance added to each well for either 6 hours (satellite cells) or 24 myogenic differentiation. hours (C2C12) before medium was changed. Real-Time Quantitative Polymerase Chain Reaction MATERIALS AND METHODS Total RNA was extracted with RNeasy (Qiagen (https://www.qia- gen.com/gb/), Manchester, United Kingdom) and reverse tran- Animals scribed using QuantiTect reverse transcription (Qiagen) as per Taz-knockout (Wwtr1–/–) (Jax, stock 011120) and Yapfl/fl mice manufacturer’s instructions. Real-time quantitative polymerase are described [39, 40]. Pax7Cre mice were purchased from The chain reaction (RT-qPCR) was performed with Brilliant II SYBR Jackson Laboratory (https://www.jax.org/), Sacramento, Cali- green reagents and a ROX reference dye (Agilent Technologies, fornia USA (stock 012476). Pax7Cre:Yapfl/fl:R26-floxstop-LacZ [9] (www.genomics.agilent.com), La Jolla, California, USA) using the www.StemCells.com VC 2017 The Authors STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 1960 Taz Enhances Proliferation and Differentiation

ViiA7 qPCR system. Primer sequences were Yap (50-TGAGCC Microarray 0 0 0 CAAGTCCCACTC-3 ; R-5 -TGTGAGTGTCCCAGGAGAAA-3 ), Taz Total RNA was isolated from TAZ S89A or YAP S127A trans- 0 0 0 (5 -TATCCCAGCCAAATCTCGTG-3 , R-5 -TTCTGCTGGCTCAGGGTAC duced myoblasts after 24 or 48 hours using TRIzol (Thermo- 0 T-3 ) or as described [43]. Fisher Scientific) followed by purification and DNase digestion Immunolabeling and EdU Pulsing (RNeasy minikits, Qiagen). RNA quantification was performed on a Nanodrop spectrophotometer (ThermoFisher Scientific) Cells/myofibers were fixed with 4% paraformaldehyde (PFA)/ and quality tested on an Agilent Tapestation (RIN 7.6–9.8). phosphate-buffered saline (PBS) for 10 minutes, permeabilized Generation of sense strand cDNA, second strand synthesis, in with 0.5% Triton-X100/PBS and blocked with 10% goat serum/ vitro transcription cRNA synthesis, single stranded cDNA syn- PBS or 0.035% carrageenan/PBS followed by incubation with thesis and RNA hydrolysis, fragmentation, and labeling were antibodies overnight at 48C [41]. Antibodies: anti-Pax7 (Devel- as manufacturer’s instructions (GeneChip WT Plus reagent kit, opmental Studies Hybridoma Bank (DSHB) (http://dshb.biol- Affymetrix, now part of ThermoFisher Scientific). Hybridiza- ogy.uiowa.edu/), Iowa City, Iowa, USA); anti-myosin heavy tion, washing, staining, and scanning of microarrays were car- chain (MyHC) (MF20, DSHB); anti-myogenin (F5D, DSHB); anti- ried out on Affymetrix Mouse Gene 2.0 ST microarrays using MyoD (clone 5.8A, DakoCytomation, Glostrup, Denmark); anti- a GeneChip Fluidics station 450 and GCS3000 scanner Taz (HPA007415, Sigma); anti-Yap1 (2F12, Abnova (http:// R (AffymetrixV). www.abnova.com/), Taipei City, Taiwan); anti-Tead4 (M01, Data preprocessing/quality control were performed using Abnova). R R AffymetrixV GenechipV Expression Console v1.2. Probe cell Cryosections were fixed with 4% PFA/PBS followed by cooled intensity data (CEL files) were processed using the RMA16 methanol before antigen retrieval in heated citrate buffer [44] algorithm (Affymetrix), which fits a linear model at probe and blocking in 10% goat serum/PBS. Antibodies: anti-MyHC level by using background correction, quantile normalization Type I (BA-D5, DSHB), anti-MyHC Type IIa (A4.74, DSHB), and of log2 transformed data, and summarization, for primary QC anti-laminin (Sigma, L9393). Fluorochrome-conjugated second- analysis. Performed in triplicate at Centre for Genome ary antibodies were from ThermoFisher Scientific. 0 Enabled Biology and Medicine (University of Aberdeen). 5-Ethyl-2 -deoxyuridine (EdU) (10 mM) was added for 2 hours R R Data were analyzed in PartekV Genomics SuiteV, version before fixation and incorporation detected using Click-iT (Thermo- 6.6, build 6.15.0730 Copyright; 2014 (Partek Incorporated Fisher Scientific) according to manufacturer’s instructions. (http://www.partek.com/), St. Louis, Missouri, USA) using a Western Blotting Mouse Gene 2.0 ST annotation file from build mm10, MoGene-2.0-st-v1.na35.mm10.transcript. Affymetrix CEL files Western blotting was performed using Run Blue precast VR VR native Page gels (Expedeon (https://www.expedeon.com/) were imported to Partek Genomics Suite , data processed Over, Cambridge, United Kingdom). Protein transfer was per- using RMA normalization with RMA background correction formed with the XCell II blot module (ThermoFisher Scien- and quantile normalization of log2 transformed data and pro- tific). Polyvinylidene difluoride (PVDF) membranes were beset summarization by median polish. Two-way analysis of incubated with antibodies overnight/48C and visualized using variance (ANOVA) with time point (24 and 48 hours) and tran- fluorochrome-conjugated secondary antibodies (ThermoFisher scription factor (Control, TAZ S89A, YAP S127A) and Scientific) and digitally imaged. time 3 transcription interaction to evaluate significantly dif- ferentially expressed genes. Fold change in TAZ S89A or YAP Mass Spectrometry S127A compared with control vector as baseline at each time C2C12 cells were grown in DMEM (D5761) with 10% FBS and point calculated using geometric mean of samples in each 4 mM glutamine. Proliferating C2C12 cells were at 50% cell group with significance calculated by Fishers Least significant density. Confluent cultures were differentiated for 72 hours in difference. Fold change 1.3 and False Discovery Rate (FDR) DMEM, 2% horse serum, 4 mM glutamine. of 10% were evaluated. Microarray data available in the For immunoprecipitation, 80,000 C2C12s were seeded per ArrayExpress database (www.ebi.ac.uk/arrayexpress), accession 10 cm dish. The following day, fresh medium was added 1 number E-MTAB-5395. hour before addition of 1:5 diluted TAZ or YAP encoding retro- viral supernatant. The next day, cells were replated and trans- Histology duction confirmed by green fluorescent protein (GFP). Cells H&E and ATPase staining were described previously [30]. Mus- were washed on ice with PBS, and collected in lysis buffer cle sections were stained with ATPase at pH 4.47 to detect (150 mM NaCl, 20 mM Tris-HCl pH 7.5, 1% Triton 3100) with type I, and pH 10.5 to detect type II, myofibers. 1 mM sodium orthovanadate, protease inhibitor cocktail (Sigma, p8340), and Phenylmethylsulfonyl fluoride (PMSF) Image Acquisition and Analysis (Sigma-Aldrich). Lysates were incubated for 1 hour on ice and Images were obtained at room temperature on a Zeiss micro- centrifuged at 14,000 RPM at 48C for 5 minutes and superna- scope (Axiovert 200M) equipped with LD A-plan 203/0.85 tant incubated at 48C with anti-Flag M2 agarose beads (Sigma ph1, 103/0.30ph1 and 403/.075ph1 objectives or on a Zeiss Aldrich). Beads were washed three times with washing buffer Exciter laser scanning microscope (LSM) with 403/1.1 W Corr (150 mM NaCl; 20 mM Tris-HCL pH 7.5). Sample preparation/ LD C-Apochromat objective (Carl Zeiss Ltd (https://www.zeiss. mass spectrometry were as described [45], with proteomics co.uk/), Cambridge, United Kingdom). Images were quantified data deposited to the ProteomeXchange Consortium via the with image J software (NIH). Neonatal MyHC labeling was PRIDE [46] partner repository: identifier PXD005751. quantified with R software. Values are mean 6 SEM with a

VC 2017 The Authors STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press Figure 1. Taz levels increase during skeletal muscle regeneration and myogenic differentiation. (A): Yap and Taz mRNA expression in regen- erating tibialis anterior (TA) muscle collected at 1, 3, 5, 7, and 14 days post injury. Expression is fold change compared with control undam- aged TA (n 5 3). (B): Yap and Taz mRNA levels in C2C12 myoblasts at 24 and 48 hours in proliferation medium, and 24, 48, 72, and 96 hours in differentiation medium, expressed as fold change compared with 24 hours proliferation time-point (n 5 3). (C): Representative Western blots from biological replicate 4 (n 5 4) showing total Yap, Taz, and phosphorylated (P) Yap and Taz levels in C2C12 myoblasts at 24 and 48 hours in proliferation medium, and 24, 48, 72, and 96 hours in differentiation medium, with relevant a-Tubulin or Gapdh loading controls. (D): Taz mRNA increased through differentiation in plated satellite cell-derived myoblasts at 12, 24, and 48 hours. Data are mean 6 SEM from three experiments where an asterisk denotes significant difference (p < .05) from 0 hour using an unpaired Student’s t test. (E): C2C12 proliferating myoblasts coimmunolabeled for Yap and Taz. (F): Plated satellite cell-derived myoblasts coimmunolabeled with MyoD and Taz after 36 hours ex vivo. (G): Plated satellite cell-derived myotubes coimmunolabeled with myogenin and Taz after 48 hours ex vivo. (H): Confocal images of satellite cells coimmunolabeled with MyoD and Taz on isolated fibers cultured ex vivo for 72 hours; arrow points to Taz-containing nuclei. Scale bar 5 100 mm (F, G). Abbreviations: DAPI, 40,6-diamidino-2-phenylindole; P, Phosphorylation. 1962 Taz Enhances Proliferation and Differentiation

Student’s t test (paired or unpaired as appropriate) or ANOVA to mark myoblasts entering differentiation (Fig. 3B). TAZ signifi- with a post-hoc Tukey’s test for >2 groups. cantly decreased eGFP-containing self-renewing Pax7 1 ve cells (Fig. 3A), while TAZ or TAZ S89A significantly increased eGFP- containing differentiating myogenin1ve myoblasts (Fig. 3B), RESULTS compared with control. Thus, TAZ promotes entry into differen- tiation, at the expense of satellite cell self-renewal. Taz Is Dynamically Regulated during Myogenic To investigate effects of TAZ or TAZ S89A throughout myo- Progression genic differentiation, we used expanded, plated satellite cell- To assess Yap and Taz (Wwtr1) expression dynamics during derived myoblasts, which were first transduced and then muscle regeneration, we measured mRNA levels by RT-qPCR switched to differentiation medium 24 hours later. Self- in vivo and myogenic progression in vitro. Murine TA muscle renewal and differentiation were again assessed by coimmu- was injured by intramuscular injection of CTX and mRNA iso- nofluorescence for eGFP/Pax7 and eGFP/myogenin. Under dif- lated from regenerating muscle after 1, 3, 5, 7, and 14 days ferentiation stimulus, TAZ or TAZ S89A decreased the post-injury (dpi). Taz mRNA increased markedly at 3 dpi com- proportion of myocytes containing Pax7, while enhancing the pared with uninjured control, before dropping, while Yap was proportion expressing myogenin (Fig. 3C, 3D). At this early largely unchanged (Fig. 1A). Taz mRNA rose during differentia- stage of differentiation, myotubes were also more prominent tion of C2C12 myoblasts, while Yap mRNA remained relatively in myoblast cultures transduced with TAZ or TAZ S89A, than constant (Fig. 1B). Western blot analysis for total Yap and Taz control retrovirus, suggesting a time-dependent switch of TAZ protein in C2C12 reflected mRNA dynamics, with Taz levels function from pro-proliferation to pro-differentiation (Fig. 3E). rising during differentiation, while Yap again remained rela- At later stages of differentiation and fusion into multinucle- tively constant (Fig. 1C). Phosphorylated Yap and Taz levels ated myotubes, constitutive expression of TAZ or TAZ S89A both showed an increase during myogenic differentiation. Taz also increased incorporation of nuclei into myotubes (fusion mRNA also increased during ex vivo differentiation of plated index-Fig. 3F, 3G). siRNA-mediated Taz knockdown in murine satellite cell-derived myoblasts (Fig. 1D). expanded plated satellite cell-derived myoblasts did not signif- Coimmunolabeling C2C12 myoblasts for Yap and Taz icantly reduce (though p 5 .06) the fusion index to below con- revealed Taz localized to the cytoplasm and nucleus, while trol levels (Fig. 3H, 3I). Yap was nuclear (Fig. 1E), consistent with our previous obser- –/– vations [29]. Taz was similarly localized in satellite cells (Fig. Taz-knockout (Wwtr1 ) Myoblasts Differentiate Less 1F), but became cytoplasmic in mature multinucleated myo- Ex Vivo tubes (Fig. 1G). Coimmunolabeling of satellite cells on a myo- Skeletal muscles of Taz-knockout (Wwtr1–/–) mice have not fiber showed that Taz was expressed in satellite cells, as been well-characterized [40]. Consistent with the trend in shown by MyoD coexpression (Fig. 1H). siRNA data (Fig. 3H, 3I), satellite cells isolated and expanded from Taz-knockout (Wwtr1–/–) mice fused less into post- Taz Promotes Myoblast Proliferation mitotic myotubes compared with WT cells (Fig. 3J, 3K). To compare Taz with Yap function, we either increased Taz Thus, Taz promotes proliferation of myoblasts, but at later activity via retroviral-mediated expression of human TAZ or stages of myogenesis, Taz switches to enhance differentiation. In constitutively active TAZ S89A, or after decreasing Taz and/or contrast, as we reported previously, Yap continues to promote Yap levels using siRNA and performed proliferation and differ- proliferation while inhibiting differentiation at this stage [28, 29]. entiation assays. –/– The retroviral backbone contains an IRES-eGFP that allows Taz-knockout (Wwtr1 ) Mice Have Fewer Myofibers transduced cells to be identified by eGFP expression. Immuno- in Soleus labeling revealed that TAZ or TAZ S89A significantly increased In accordance with the role of the Hippo pathway in body/ the proportion of eGFP1 ve cells that had incorporated EdU, organ size regulation, 6-week-old Taz-knockout (Wwtr1–/–) indicating that TAZ increases proliferation rate (Fig. 2A, 2B). mice were lighter than WT mice (Fig. 4A) as were their TA, siRNA-knockdown of Taz, Yap, or both Taz/Yap simultaneoulsy, soleus, EDL, and gastrocnemius muscles (Fig. 4B–4E). How- however, significantly reduced EdU incorporation (Fig. 2C, 2D). ever, the number of satellite cells and myonuclei per EDL Thus, at this stage, TAZ promotes proliferation of myoblasts, myofiber was unchanged between Taz-knockout (Wwtr1–/–) as we reported previously for YAP [28, 29]. and WT mice (Fig. 4F, 4G). Myofiber type composition in RT-qPCR showed that siRNA-mediated knockdown of Taz soleus was analyzed using ATPase activity at pH 4.47 to detect suppressed Myf5 and myogenin expression, but enhanced Pax7. slow type I, or pH 10.5 to detect fast type IIa, myofibers. Conversely,knockdownofYapincreasedmyogenin expression, Soleus contained fewer muscle fibers in total, with both num- but reduced Pax7 and Myf5. Simultaneous knockdown of both bers of type I and type IIa myofibers significantly reduced in Taz and Yap rescued expression of myogenin (Fig. 2E). Taz-knockout (Wwtr1–/–) compared with WT (Fig. 4H, 4I), but the proportion of slow type I myofibers was significantly Taz and Yap Play Opposing Roles during Myogenic increased, while the proportion of type IIa were significantly Differentiation decreased, compared with control WT. Myofiber cross- We determined the effect of retroviral-mediated constitutive sectional area was unchanged though (Fig. 4J). expression of TAZ or TAZ S89A on early myogenic differentiation Next, we investigated muscle regeneration in the Taz-knock- in satellite cells cultured in their niche for 72 hours on an iso- out (Wwtr1–/–) mice using intra-muscular injection of CTX into lated myofiber, by coimmunolabeling for eGFP/Pax7 to mark TA. Cryosections of 5 dpi regenerating TA were immunolabeled undifferentiated/self-renewing cells (Fig. 3A) or eGFP/myogenin for neonatal MyHC to identify regenerating myofibers, but Taz-

VC 2017 The Authors STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press Sun, De Mello, Mohamed et al. 1963

Figure 2. Taz controls satellite cell-derived myoblast proliferation. (A): Representative images of 5-ethyl-20-deoxyuridine (EdU) incorpo- ration in satellite cell-derived myoblasts 48 hours following transduction with control retroviral vector (RV), or retroviral constructs encoding wild-type human TAZ (TAZ RV) or constitutively active TAZ S89A (TAZ S89A RV), where green fluorescent protein (GFP) marks transduced cells. (B): Quantification of GFP1 cells with EdU incorporation shows that TAZ or TAZ S89A significantly increases EdU incor- poration, and so proliferation rate (n 5 5 mice). (C): Representative images of EdU incorporation in satellite cell-derived myoblasts fol- lowing Yap and/or Taz small interfering RNA (Si)-mediated knockdown using siControl, siYap, siTaz, or siYap 1 siTaz transfection. (D): Quantification of EdU incorporation shows that knockdown of Yap, Taz, or both Yap and Taz simultaneously reduces the proliferation rate (n 5 5 mice). (E): Relative gene expression measured by RT-qPCR in satellite cell-derived myoblasts compared with si Control under proliferation conditions following siControl, siYap, siTaz, or siYap 1 siTaz knockdown, normalized to Gapdh (n 5 5 mice except for myo- sin heavy chain/MyoD, where n 5 3). Data presented as mean 6 SEM, where an asterisk denotes significant difference from control RV/Si control (p < .05) using Student’s t test. Scale bar 5 100 mm. Abbreviations: DAPI, 40,6-diamidino-2-phenylindole; EdU, 5-ethyl-20- deoxyuridine; GFP, green fluorescent protein; RV, retroviral vector; Si, small interfering RNA. knockout (Wwtr1–/–) did not exhibit a reduction compared with Inactivation of Yap in Satellite Cells Impairs Muscle controls (Fig. 4K, 4L). At 10 dpi, the number of myofibers with Regeneration centrally located nuclei in TA of Taz-knockout (Wwtr1–/–)mice Administration of Tamoxifen to Pax7Cre-ERT2/1:Yapfl8x/fl8x: was also unchanged (Fig. 4M, 4N). Rosa26Lacz mice causes Pax7-driven Cre-ERT2-mediated www.StemCells.com VC 2017 The Authors STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 1964 Taz Enhances Proliferation and Differentiation

Figure 3. Taz regulates fusion of satellite cell-derived myoblasts. (A, B): Satellite cells on isolated myofibers were transduced with retro- viral vectors (RV) encoding wild-type (WT) human TAZ (TAZ RV), constitutively active TAZ S89A (TAZ S89A RV) or Control RV, and cul- tured ex vivo for 72 hours, before being coimmunolabeled for eGFP and Pax7, or eGFP and Myogenin. Quantification of the proportion of (A) green fluorescent protein (GFP)/Pax7 and (B) GFP/myogenin satellite cells in the ex vivo cultures after 72 hours (n 5 3 mice). (C, D): Quantification of the proportion of expanded plated satellite cell-derived myoblasts transduced with control RV, TAZ RV, or TAZ S89A RV and incubated in differentiation medium for 24 hours before being coimmunolabeled as myocytes for eGFP/Pax7 or eGFP/myogenin, illustrated in (E) (n 5 4 mice). (F): Representative images of expanded plated satellite cell-derived myoblasts transduced with TAZ RV, TAZ S89A RV, or Control RV and incubated in differentiation medium before being coimmunolabeled for eGFP/myosin heavy chain (MyHC). (G): Quantification of the proportion of nuclei within GFP/MyHC myotubes shows that TAZ or TAZ S89A increases myogenic fusion (n 5 5 mice). (H): Representative images of satellite cell-derived myoblasts transfected with small interfering RNA (siRNA) control or siRNA against Taz (SiRNA Taz) and immunolabeled for MyHC after 24 hours. (I): Quantification shows a trend toward reduced fusion (p 5 .06) with siRNA Taz (n 5 3 mice). (J): Representative images of myotubes formed from satellite cell-derived myoblasts isolated from WT (Wwtr11/1) or Taz-knockout (Wwtr1–/–) mice after 24 hours in differentiation medium. (K): Quantification reveals less fusion in the Taz-knockout (Wwtr1–/–) mice compared with control WT (n 5 4 per genotype). Data are mean 6 SEM, where an asterisk denotes significant difference (p < .05) using a Student’s t test from Control RV/siRNA control/Wwtr11/1 as appropriate. Scale bars 5 100mm. Abbreviations: GFP, green fluorescent protein; MyHC, myosin heavy chain; RV, retroviral vector; siRNA, small interfering RNA; WT, wildtype. deletion of Yap1 exons 1 and 2 in satellite cells, with simulta- revealed that there was a nonsignificant (p 5 .054) trend to neous induction of lacZ expression. Pax7Cre-ERT2/1:Yapfl8x/fl8x: reduced numbers of satellite cells in Tamoxifen-treated 5 dpi Rosa26Lacz and age-matched control Pax7Cre-ERT2/1:Yapfl8x/1: regenerating Pax7Cre-ERT2/1:Yapfl8x/fl8x:Rosa26Lacz muscle (Fig. Rosa26Lacz mice were given three intraperitoneal injections of 5D, 5F). Tamoxifen, and then maintained on tamoxifen-containing By 10 dpi, Pax7 immunolabeling revealed that satellite cell food. TA muscles were injured by intramuscular injection of number in tamoxifen-treated Pax7Cre-ERT2/1:Yapfl8x/fl8x:Rosa26Lacz CTX, and regenerating muscles cryosectioned at 5 and 10 dpi mice was as in control Pax7Cre-ERT2/1:Yapfl8x/1:Rosa26Lacz mice (Fig. 5A). (Fig. 5E, 5F). The number of myofibers with centrally located Regenerating TA muscles at 5 dpi from Tamoxifen-treated myonuclei, a hallmark of regenerated muscle, was also Pax7Cre-ERT2/1:Yapfl8x/fl8x:Rosa26Lacz mice coimmunolabeled for unchanged between Pax7Cre-ERT2/1:Yapfl8x/fl8x:Rosa26Lacz regen- neonatal MyHC and laminin had an approximate 30% smaller erating TA muscles and control (Fig. 5G, 5H). Thus, lack of Yap area containing neonatal MyHC than control Pax7Cre-ERT2/1: function in satellite cells slows, but does not prevent, muscle Yapfl8x/1:Rosa26Lacz (Fig. 5B, 5C). Immunolabeling for Pax7 regeneration.

VC 2017 The Authors STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press Sun, De Mello, Mohamed et al. 1965

Figure 4. Characterization of skeletal muscle and muscle regeneration in Taz-knockout (Wwtr1–/–) mice. (A): Body weight of 6-week-old Taz-knockout (KO) (Wwtr1–/–) mice (n 5 7) compared with wild-type (WT) Wwtr11/1 control (n 5 9). (B–E): Quantification of muscle weight of tibialis anterior (TA), Soleus, extensor digitorum longus (EDL), and Gastrocnemius from WT and Taz KO mice (n 5 3 mice). (F): Satellite cell and (G): myonuclei number per myofiber isolated from the EDL of WT or Taz KO mice (n 5 4 mice of each genotype). (H): Representative images of soleus from 6-week-old WT and Taz KO after ATPase histochemical staining at pH 4.47 to show slow type I (dark) and fast type IIa (light/no stain) muscle fiber types. (I): Quantification showing less total myofibers in the soleus of Taz KO mice, with numbers of both type I and type IIa reduced (n 5 3 mice of each genotype). (J): Quantification of cross section area of type I and type IIa reveals that myofibers from Taz KO are unchanged compared with controls (n 5 3 mice per genotype). (K): Representative images and (L): quantification of the proportion of the neonatal myosin heavy chain (Neo MyHC) area of regenerating TA myofibers coimmunolabeled for NeoMyHC/laminin at 5 days post injury (dpi) with CTX (n 5 4 mice of each genoptype). (M): Representative images of H&E staining at 10 dpi of TA from Taz KO and WT mice. (N): Quantification of the number of myofibers with a central nucleus per field of 10 dpi TA (n 5 3 mice per genotype). Data are mean 6 SEM where an asterisk denotes significant difference (p < .05) between WT and Taz KO using a Student’s t test. Scale bar 5 100 mm. Abbreviations: CSA, cross section area; CTX, cardiotoxin; EDL, extensor digitorum longus; KO, knockout; NeoMyHC, neonatal myosin heavy chain; TA, tibialis anterior; WT, wild type.

YAP and TAZ Regulate the Hippo Negative Feedback Combination of 24 and 48 hours time points revealed that Loop, Myogenic Regulators, and Wnt Signaling TAZ S89A significantly changed expression of 860 genes, while To compare the effects of TAZ and YAP on gene expression in YAP S127A altered expression of 294 genes, with an additional the skeletal muscle lineage, we expressed TAZ S89A or 316 genes regulated by both TAZ S89A and YAP S127A (Fig. YAP1 S127A in proliferating murine satellite cell-derived myo- 6A; Supporting Information Data 1). Satellite cells were blasts for 24 and 48 hours and analyzed gene expression by already actively proliferating as cultured in high serum microarray (Supporting Information Data 1), comparing our medium, and serum is a potent YAP activator [47], likely observations with those obtained by induction of YAP1 S127A explaining why neither TAZ S89A nor YAP1 S127A significantly in myoblasts derived from a transgenic mouse model [29] and affected most mitotic genes, in contrast to earlier observa- from rhabdomyosarcomas driven by YAP1 S127A [25]. tions [25, 29]. www.StemCells.com VC 2017 The Authors STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 1966 Taz Enhances Proliferation and Differentiation

Figure 5. Yap conditional knockout in satellite cells affects muscle regeneration. (A): Time course for analysis of Yap conditional knock- out and control mice after tamoxifen administration via intraperitoneal injection and food, and cardiotoxin (CTX)-induced injury of the tibialis anterior (TA) muscle. (B, C): Representative images and quantification of TA cryosections 5 dpi after CTX injury of Pax7CreERT/1: Yapfl8x/fl8x:Rosa26Laz (Pax7Cre Yapfl/fl) and control Pax7CreERT/1:Yap fl8x/1:Rosa26Laz (Pax7Cre Yapfl/1) mice, coimmunolabeled for neona- tal myosin heavy chain (NeoMyHC)/laminin to identify regenerating myofibers (n 5 4 mice per genotype). (D–F): Representative images and quantification of cryosections of TA muscles after 5 or 10 dpi, immunolabeled for Pax7 to label satellite cells (5 dpi: n 5 4 per genotype; 10 dpi: n 5 4 Pax7Cre Yapfl/fl and n 5 3 Pax7Cre Yapfl/1). (G): Representative images of H&E staining and (H) quantification of centrally nucleated myofiber number per field of 10 dpi TA. Data are mean 6 SEM, where an asterisk denotes significant difference (p < .05) between Pax7Cre Yapfl/1 and Pax7Cre Yapfl/fl using a Student’s t test. Scale bar 5 100 mm. Abbreviations: CTX, cardiotoxin; DAPI, 40,6-diamidino-2-phenylindole; IP, intraperitoneal; NeoMyHC, neonatal myosin heavy chain; TA, tibialis anterior.

IncreasedTAZ/YAP-TeadactivitywasindicatedbybothTAZ Wnts are TAZ/YAP regulators, and TAZ/YAP regulation of S89A and YAP S127A inducing the Hippo negative feedback loop Wnt and Wnt-related proteins can inhibit Wnt signaling [48]. genes Amotl2, Frmd6 (Willin), and Lats2, which was presumably TAZ S89A and YAP S127A altered expression of key Wnt path- to limit Taz/Yap activity. TAZ S89A and YAP S127A also regulate way regulators, reducing expression of Wnt- proteins markers/regulators of the muscle lineage: Caveolin1 (Cav1) was Fzd7 and Lgr5, while TAZ S89A alone suppressed expression upregulated, while Six1 and Six4 were downregulated. In addi- of Fzd6, Lgr4, and Lgr6, suggesting that TAZ/YAP desensitize tion, TAZ and YAP regulated multiple members of the mTOR/ myoblasts to ligands that bind these receptors. Moreover, TAZ insulin-like growth factor (IGF) pathway, including Igfbp5,as S89A increased expression of Wnt4, and altered expression of well as Ndrg2 and Irs2, which regulate mTOR via Akt (Fig. 6B). other secreted Wnt-associated proteins and genes encoding

VC 2017 The Authors STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press Sun, De Mello, Mohamed et al. 1967

Figure 6. Genes regulated by TAZ S89A or YAP S127A and proteins bound by TAZ and/or YAP. (A): Venn diagram displaying the number of genes modulated by TAZ S89A and/or YAP S127A versus control retrovirus (up-reg: upregulated; down-reg: downregulated) at either 24 or 48 hours. (B): Expression fold change for genes that were regulated by both TAZ S89A and YAP S127A in transduced satellite cell- derived myoblasts versus control retrovirus at either 24 or 48 hours. (C): Expression fold change for genes that were regulated only by TAZ S89A at either 24 or 48 hours. (D): All TAZ and/or YAP interacting proteins identified by mass spectrometry in C2C12 myoblasts and myotubes. Those previously identified are in red [53] and novel proteins in black, with those of particular interest in blue. To minimize false positives, selected cut-off values were a ratio of 2 compared with empty vector and p < .05 Student’s t test between each data- set and control vector. Abbreviations: IGF, insulin-like growth factor; mTOR, mechanistic target of rapamycin. www.StemCells.com VC 2017 The Authors STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 1968 Taz Enhances Proliferation and Differentiation the structurally related proteins Wisp1 and Wisp2 [49]. Addi- levels increased upon entry into C2C12 myoblast differentia- tionally, both TAZ S89A and YAP S127A decreased Sox9 tion and were maintained (Fig. 7C). expression (Fig. 6B, 6C). As Tead4 is critical for myoblast fusion due to inducing myogenin by binding MCAT elements in its promoter [56], TAZ Regulates a Unique Set of Genes we further examined its function. Coimmunolabeling of Although the microarray was performed on proliferating myo- Tead4 and Taz showed both proteins present in the nucleus blasts, genes regulated by TAZ only might explain the func- of differentiating myoblasts, at a time when Tead4 levels rise tional switch from pro-proliferation to pro-differentiation (Fig. markedly (Fig. 7D). Thus, the TAZ-Tead4 interaction identified 6C). Both Pax7 and MyoD were downregulated by TAZ S89A in proliferating myoblasts (but not in mature myotubes) while Myf5 was upregulated, consistent with Taz siRNA knock- (Fig. 6D) could be maintained during induction and the early down (Fig. 2E). Upregulated cell junction proteins could con- phases of myogenic differentiation, not assayed by tribute to enhanced fusion capability in TAZ-overexpressing proteomics. myoblasts, since N-Cadherin is required for myoblast fusion We used retroviral-mediated expression to enhance TAZ [50]. TAZ overexpression also altered expression of olfactory activity and/or siRNA to reduce Tead4 levels in satellite cell- receptors Olfr311 and Olfr1086 (Fig. 6C), a family of genes derived myoblasts during myogenic differentiation, and with members engaged in controlling satellite cell self- assayed fusion index after coimmunolabeling for eGFP/MyHC. renewal [51]. Mki67 (protein Ki67) is involved in proliferation, TAZ S89A with control siRNA augmented myoblast fusion (Fig. but was downregulated by TAZ in myoblasts (Fig. 6C), again 7E, 7F), but TAZ S89A with concomitant Tead4 knockdown highlighting the switch from promoting proliferation toward reversed this enhanced fusion index (Fig. 7E, 7F). Knockdown cell cycle withdrawal/differentiation. of Tead4 alone resulted in shorter and thinner myotubes, reflected in a reduced fusion index, but expression of TAZ YAP and TAZ Bind Both Common and Unique Proteins S89A did allow recover of the fusion index back to control Apart from Tead1-4, Yap and Taz also bind a plethora of levels (Fig. 7E, 7F). upstream signaling proteins and downstream transcriptional To investigate the role of TAZ/Tead4 in entry into myo- regulators [52]. To identify binding partners in a non-biased genic differentiation versus fusion, we also transduced satel- way, we expressed flag-tagged WT human YAP or TAZ in pro- lite cell-derived myoblasts with TAZ S89A and/or transfected liferating murine C2C12 myoblasts and mature myotubes. with Tead4 siRNA and differentiated cells as myocytes. After Flag-immunoprecipitated YAP or TAZ complexes were identi- 24 hours in differentiation medium, coimmunolabeling fied using mass-spectrometry (Fig. 6D; Supporting Information revealed that TAZ S89A increased the proportion of eGFP1ve Data 2). cells containing myogenin, compared with controls. Knock- Our TAZ and YAP binding proteins were first validated down of Tead4 alone significantly reduced myogenin expres- against previously characterized binding partners in human sion (Fig. 7G, 7H). Simultaneous expression of TAZ S89A and embryonic kidney 293T cells [53]. Overlap included angiomo- Tead4 siRNA knockdown though, brought the proportion of tins, multiple 14-3-3 proteins, Tead1, Tead3, and Tead4 (red: GFP-expressing myoblasts containing myogenin back to con- Fig. 6D). However, we also identified many novel TAZ and/or trol levels (Fig. 7G, 7H). YAP interacting proteins in myogenic cells (Fig. 6D). TAZ bound many more proteins in myoblasts than YAP, or that were com- mon to both YAP and TAZ. In myotubes, however, YAP bound DISCUSSION many more proteins than TAZ or both TAZ and YAP, with only YAP binding Wnt regulators Dvl2 and Dvl3. Of 26 TAZ and YAP Yap, Taz, and Tead1-4 constitute a dynamic system in satellite binding proteins in mature myotubes (Fig. 6D), 10 are linked cells and the muscle lineage, with mRNA and protein levels of to actin cytoskeleton and the Chaperone assisted selective some changing extensively during muscle regeneration and autophagy (CASA) complex, including the WW-domain con- myogenic differentiation in vitro. Three are of note: Yap [29] taining Bag3. We also identified 20 proteins bound specifically and Taz only become detectable during satellite cell activa- by TAZ during differentiation, the majority associated with the tion, suggesting that Yap and Taz only operate during myogen- actin cytoskeleton and myogenesis including Unc45a and Fhl3 esis; mRNA/protein levels of some Teads also change, [54, 55]. MyoD was not among YAP and/or TAZ binding part- suggesting that Teads are not just static targets of Yap/Taz; ners, even at less stringent cut-offs. Taz and Tead4 increase dramatically during myogenic differen- tiation, while Yap expression remains fairly constant. During TAZ Operates through Tead4 to Control Myogenic regeneration, Taz and Tead4 levels peak around 3–5 dpi, as Differentiation myogenic differentiation is underway [57–59]. Our observa- We next assayed Tead 1–4 expression during myogenesis. TA tions are consistent with Taz and Tead4 as pro-differentiation muscle was injured using CTX and mRNA isolated from regen- factors [28, 29, 37, 56]. erating muscle after 1, 3, 5, 7, and 14 dpi. RT-qPCR showed Manipulation of Yap and Taz activity/levels via retroviral- that Tead2 and Tead4 exhibited similar expression profiles, mediated expression of human WT TAZ or TAZ S89A, or peaking around day 3, before falling to levels seen in undam- knockdown of Taz, revealed that TAZ promotes proliferation. aged muscle (Fig. 7A). Tead1-3 mRNA was at a constant level YAP [28, 29] and TAZ synergistically promote proliferation in in proliferating and differentiating C2 myoblasts, but Tead4 many cell types [60], with a recent ChIP-Seq study in breast expression increased robustly at the onset of myogenic differ- cancer cells showing that YAP and TAZ frequently operate entiation, being maintained at elevated levels through fusion through the same enhancer elements [24]. Later, in myogene- (Fig. 7B). Western blot revealed that Tead1 and Tead4 protein sis, however, TAZ switches to promote differentiation, in

VC 2017 The Authors STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press Figure 7. TAZ operates through Tead4 to control myogenic differentiation. (A): Tead2–4 mRNA expression in regenerating tibialis anterior (TA) muscle collected at 1, 3, 5, 7, and 14 days post injury with cardiotoxin. Expression is presented as fold change compared with control undamaged TA (n 5 3). (B): Tead1–4 mRNA levels in C2C12 myoblasts at 24 and 48 hours in proliferation medium, and 24, 48, 72, and 96 hours in differentiation medium, expressed as fold change compared with 24 hours proliferation time-point (n 5 3). (C): Representative Western blots of C2C12 myoblast lysate collected at 24 and 48 hours in proliferation medium and 24, 48, 72, and 96 hours in differentia- tion medium and immunoblotted for Yap, Taz, Tead1 (biological replicate 3), phosphorylated (P) Yap and Taz (biological replicate 2) and Tead4 (biological replicate 1) (n 5 4), with relevant a-Tubulin or Gapdh loading controls. (D): Representative images of plated satellite cell- derived myoblasts coimmunolabeled with Tead4 and Taz cultured under differentiation conditions for 24 hours. (E): Representative images of coimmunolabeling of myotubes formed from satellite cell-derived myoblasts following TAZ S89A RV or Control RV transduction and/or small interfering RNA (siRNA) knockdown of Tead4 or siRNA Control. Green fluorescent protein (GFP) labels transduced cells and myosin heavy chain (MyHC) identifies myotubes. (F): Comparison of the proportion of nuclei in GFP1ve/MyHC1ve myotubes under each condition shows that TAZ S89A requires Tead4 to enhance differentiation (n 5 3mice).(G): Representative images of coimmunolabeling of myocytes formed from satellite cell-derived myoblasts following TAZ S89A RV or Control RV transduction and/or siRNA knockdown of Tead4 or siRNA Control, and 24 hours in differentiation medium. GFP labels transduced cells and Myogenin identifies myocytes. (H): Comparison of the pro- portion of GFP1ve/Myogenin1ve myocytes under each condition shows that TAZ-induced myogenic differentiation operates through Tead4 (n 5 3 mice). Data are mean 6 SEM from three mice, where an asterisk denotes significant difference (p < .05) from Control RV/siRNA Control, or as indicated by bars, using a paired Student’s t test. Scale bar 5 100 mm. Abbreviations: DAPI, 40,6-diamidino-2-phenylindole; GFP, green fluorescent protein; MyHC, myosin heavy chain; P, Phosphorylated; RV, retroviral vector; siRNA, small interfering RNA; TA, tibialis anterior. 1970 Taz Enhances Proliferation and Differentiation accordance with an earlier reports in immortalized murine TAZ and YAP bind proteins in murine myoblasts/myotubes myoblasts [37, 38], unlike YAP, which inhibits differentiation that have been identified as binding partners in human 293T [28, 29]. TAZ actually augments the myogenic differentiation cells [53]. These generic YAP and TAZ binding proteins include program under differentiation stimuli, as shown by the preco- the angiomotins (Amot and Amotl1), Tead1 and Tead4 and cious expression of myogenin in many myocytes. Thus, the several 14-3-3 proteins (Ywhae, Ywhag, Ywhaz). Additionally, greater fusion index with TAZ overexpression is not merely we identified many novel TAZ and/or YAP binding partners in due to having more myoblasts available, due to the pro-prolif- the skeletal muscle lineage, including in myotubes, Bag3 [72] eration effects of TAZ. Consistent with its later pro- and syncoilin (Sync), an intermediate filament protein linked differentiation role, constitutive TAZ expression also reduced to the dystrophin-associated protein complex [73]. TAZ-specific self-renewal of satellite cells, while Taz-knockout (Wwtr1–/–) binding partners that could potentially explain pro- satellite cells exhibited reduced fusion ex vivo. differentiation effects include Fhl3, which binds MyoD to Taz-knockout (Wwtr1–/–) mice have functional and growth inhibit differentiation [55] and Unc45a, which increases mus- defects in kidneys and other organs, consistent with Taz as a cle cell proliferation and fusion, and when inhibited, blocks regulator of organ growth and transcriptional regulator of muscle cell fusion [54]. Runx1 bound to TAZ in myoblasts and genes involved in heart, lung, and bone growth [21, 61, 62]. downregulation of Runx1 leads to cell cycle exit and differen- Taz-knockout (Wwtr1–/–) mice were generally lighter, with TA, tiation [54]. Taz has been reported to bind MyoD using immu- soleus, EDL, and gastrocnemius muscle mass all reduced, but noprecipitation assays [37], but while we observed MyoD and with unchanged numbers of satellite cells and myonuclei per Taz in the same nuclei, we found no significant direct interac- myofiber. However, absence of Taz reduced the total number tion using mass spectrometry. of myofibers in soleus, with both fewer slow type I and fast Our combined gene expression and binding partner analy- Type IIa muscle fibers present, although there was a greater ses in myoblasts reveals that YAP or TAZ does not bind mem- proportion of slow type I myofibers. bers of the Wnt destruction complex [8]. YAP bound the Lack of an overt muscle regeneration phenotype in Taz- disheveled proteins Dvl2 and Dvl3 in myotubes, relevant as knockout (Wwtr1–/–) mice could be due to compensation by Dvl2 can affect satellite cell polarity and migration down- Yap in promoting satellite cell proliferation, as Yap or Taz stream of Wnt7a/Fzd7 [74]. TAZ S89A downregulates Fzd6 and knockdown did not affect expression of the other gene [63]. Fzd7, and Lgr4, Lgr5, and Lgr6, important Wnt receptors in In comparison, genetic deletion of Yap specifically in satellite developmental and regenerative myogenesis [75, 76], sugges- cells caused delayed regeneration, so Taz is unable to com- ting that TAZ desensitize cells to ligands that bind these pensate for Yap in early regeneration. This regeneration defect receptors. For example, quiescent satellite cells express high in tamoxifen-treated Pax7Cre-ERT2/1:Yapfl8x/fl8x:Rosa26Lacz mice levels of Fzd7 but not Taz or Yap, but during activation, the was transient although, maybe due to Taz promoting myo- increased abundance/activity of Yap/Taz could be involved in genic differentiation later in regeneration. Therefore, skeletal the down-regulation of Fdz7. muscle joins other systems where Yap is required for regener- Additionally, TAZ S89A increases expression of Wnt4, ation, for example, where defects in crypt regeneration are decreases Sox9 expression and alters expression of secreted observed when Yap is inactivated in epithelium of the small Wnt-associated proteins and genes encoding the structurally intestine, and in the skin [64–66]. related proteins Wisp1, Wisp2, Igfbp4, Cyr61, and Bmp4 [49]. We also examined gene expression in myoblasts in Many of these genes have been implicated in regulation of response to TAZ S89A or YAP S127A. As expected, YAP and developmental myogenesis, satellite cell function and muscle TAZ increased expression of Hippo marker gene Cyr61 and fibers. For example, over-expression of Wnt4 increased Pax7 Hippo negative-feedback loop genes Lats2, Frmd6 (Willin), and MyoD1 expression in chick embryos [77]. Bmp4 promotes and Amotl2: a typical response to increased Taz and Yap activ- satellite cell proliferation but inhibits differentiation [78]. ity [29]. Intriguingly, TAZ also downregulates Bag3,aWW Interestingly, Tead4 bound TAZ and YAP in proliferating domain-containing member of the CASA mechanosensitive myoblasts but not in mature multinucleated myotubes. Tead4 Z-disc linked complex [67] that binds to Yap and Taz, and increases dramatically as myoblasts enter myogenic differenti- Yes1, the earliest reported binding partner of Yap [68]. Bag3 ation and levels are then maintained. Thus the TAZ-Tead4 and Yes1 might represent additional members of the Hippo interaction identified in proliferating myoblasts by proteomics negative feedback loop in myoblasts. Several TAZ- and YAP- is likely maintained during induction and the early phases of regulated genes have been linked to myogenesis. For exam- myogenic differentiation, not assayed by proteomics. When ple, TAZ and YAP reduced expression of Six1 and Six4, two fac- Tead4 was knocked down, myogenic differentiation was sup- tors controling migration of muscle precursor cells [69] with a pressed, which recovered to controls levels when TAZ S98A role in muscle regeneration [70]. TAZ and YAP also regulated was also expressed. However, TAZ S89A could no longer many IGF/mTOR-related genes (Ndrg2, Igfbp5, Irs2, and enhance entry into the differentiation programme and aug- Igf2bp1), confirming crosstalk between Hippo and mTOR sig- ment fusion when Tead4 levels were reduced. Thus, the role naling [2, 3]. of TAZ in inducing and promoting myogenic differentiation is TAZ also regulates a unique cohort of genes, including linked to interaction with Tead4 [37, 56]. Pax7, whose expression was reduced by TAZ S89A, consistent with the switch toward differentiation. TAZ also modulates CONCLUSION multiple olfactory receptor genes, whose levels vary between mdx and control mice [51], and a family of genes with mem- In conclusion, in the skeletal myogenic lineage we demon- bers that can affect muscle regeneration [71]. strate similar roles of TAZ and YAP in promoting myoblast

VC 2017 The Authors STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press Sun, De Mello, Mohamed et al. 1971

proliferation. However, during the later stages of myogenesis, AUTHOR CONTRIBUTIONS TAZ switches toward influencing satellite cell fate by promot- ing myogenic differentiation, in a process involving Tead4. C.S.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; V.D.M., A.M., A.v.K., E.C.-D., and D.M.: collection and/or ACKNOWLEDGMENTS assembly of data, data analysis and interpretation; H.P.O.Q., A.G.-M., and A.A.B.: collection and/or assembly of data; This work was supported by the Medical Research Council A.M.T.: conception and design; N.V.: data analysis and inter- grant (G11001931) (to H.W., and P.S.Z.); BIODESIGN (262948- pretation; H.W. and P.S.Z.: conception and design, data analy- 2) from European Union’s Seventh Framework Programme for sis and interpretation, manuscript writing. research, technological development and demonstration (to P.S.Z.); Sarcoma UK (award reference: SUK09.2015) (to A.M., H.W., and N.V.); BBSRC EastBio DTP Ph.D. studentship (to DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST V.D.M.); with additional support from Association Franc¸aise Contre les Myopathies (to P.S.Z.), and the Friends of Anchor. The authors indicate no potential conflicts of interest.

REFERENCES of Akt1. The EMBO Journal 2007;26:4523– effector of embryonal rhabdomyosarcoma 4534. formation. Cancer Cell 2014;26:273–287. 1 Harvey K, Tapon N. The Salvador-Warts- 15 Yuan Z, Kim D, Shu S et al. Phosphoino- 26 Kim M, Kim T, Johnson RL et al. Tran- Hippo pathway—An emerging tumour- sitide 3-kinase/Akt inhibits MST1-mediated scriptional co-repressor function of the hippo suppressor network. Nat Rev Cancer 2007;7: pro-apoptotic signaling through phosphoryla- pathway transducers YAP and TAZ. Cell 182–191. tion of threonine 120. J Biol Chem 2010;285: Reports 2015;11:270–282. 2 Hansen CG, Moroishi T, Guan KL. YAP 3815–3824. 27 Relaix F, Zammit PS. Satellite cells are and TAZ: A nexus for Hippo signaling and 16 Collak FK, Yagiz K, Luthringer DJ et al. essential for skeletal muscle regeneration: beyond. Trends Cell Biol 2015;25:499–513. Threonine-120 phosphorylation regulated by The cell on the edge returns centre stage. 3 Wackerhage H, Del Re DP, Judson RN phosphoinositide-3-kinase/Akt and mamma- Development 2012;139:2845–2856. et al. The Hippo signal transduction network lian target of rapamycin pathway signaling 28 Watt KI, Judson R, Medlow P et al. Yap in skeletal and cardiac muscle. Sci Signal is a novel regulator of C2C12 myogenesis. 2014;7:re4. limits the antitumor activity of mammalian 4 Dupont S, Morsut L, Aragona M et al. sterile 20-like kinase 1. J Biol Chem 2012; Biochem Biophys Res Commun 2010;393: Role of YAP/TAZ in mechanotransduction. 287:23698–23709. 619–624. Nature 2011;474:179–183. 17 Nguyen HB, Babcock JT, Wells CD et al. 29 Judson RN, Tremblay AM, Knopp P et al. 5 Enzo E, Santinon G, Pocaterra A et al. LKB1 tumor suppressor regulates AMP The Hippo pathway member Yap plays a key Aerobic glycolysis tunes YAP/TAZ transcrip- kinase/mTOR-independent cell growth and role in influencing fate decisions in muscle tional activity. EMBO J 2015;34:1349–1370. proliferation via the phosphorylation of Yap. satellite cells. J Cell Sci 2012;125:6009–6019. 6 Yu FX, Zhao B, Panupinthu N et al. Regu- Oncogene 2013;32:4100–4109. 30 Judson RN, Gray SR, Walker C et al. Con- lation of the Hippo-YAP pathway by G- 18 DeRan M, Yang J, Shen CH et al. Energy stitutive expression of Yes-associated protein protein-coupled receptor signaling. Cell 2012; stress regulates hippo-YAP signaling involving (Yap) in adult skeletal muscle fibres induces 150:780–791. AMPK-mediated regulation of -like muscle atrophy and myopathy. PLoS One 7 Rosenbluh J, Wang X, Hahn WC. Geno- 1 protein. Cell Rep 2014;9:495–503. 2013;8:e59622. mic insights into WNT/beta-catenin signaling. 19 Wang W, Xiao ZD, Li X et al. AMPK mod- 31 Watt KI, Turner BJ, Hagg A et al. The Trends Pharmacol Sci 2014;35:103–109. ulates Hippo pathway activity to regulate Hippo pathway effector YAP is a critical regu- 8 Azzolin L, Panciera T, Soligo S et al. YAP/ energy homeostasis. Nat Cell Biol 2015;17: lator of skeletal muscle fibre size. Nat Com- TAZ incorporation in the beta-catenin 490–499. mun 2015;6:6048. destruction complex orchestrates the Wnt 20 Mo JS, Meng Z, Kim YC et al. Cellular 32 Goodman CA, Dietz JM, Jacobs BL et al. response. Cell 2014;158:157–170. energy stress induces AMPK-mediated regula- Yes-Associated Protein is up-regulated by 9 Xin M, Kim Y, Sutherland LB et al. Regu- tion of YAP and the Hippo pathway. Nat Cell mechanical overload and is sufficient to lation of insulin-like growth factor signaling Biol 2015;17:500–510. induce skeletal muscle hypertrophy. FEBS by Yap governs cardiomyocyte proliferation 21 Kanai F, Marignani PA, Sarbassova D Lett 2015;589:1491–1497. and embryonic heart size. Sci Signal 2011;4: et al. TAZ: A novel transcriptional co-activator 33 Morin-Kensicki EM, Boone BN, Howell ra70. regulated by interactions with 14-3-3 and M et al. Defects in yolk sac vasculogenesis, 10 Varelas X, Samavarchi-Tehrani P, PDZ domain proteins. The EMBO Journal chorioallantoic fusion, and embryonic axis Narimatsu M et al. The Crumbs complex cou- 2000;19:6778–6791. elongation in mice with targeted disruption ples cell density sensing to Hippo-dependent 22 Mar JH, Ordahl CP. A conserved of Yap65. Mol Cell Biol 2006;26:77–87. control of the TGF-beta-SMAD pathway. Dev CATTCCT motif is required for skeletal 34 Hossain Z, Ali SM, Ko HL et al. Glomeru- Cell 2010;19:831–844. muscle-specific activity of the cardiac tropo- locystic kidney disease in mice with a tar- 11 Alarcon C, Zaromytidou AI, Xi Q et al. nin T gene promoter. Proc Natl Acad Sci U S geted inactivation of Wwtr1. Proc Natl Acad Nuclear CDKs drive Smad transcriptional acti- A 1988;85:6404–6408. Sci USA 2007;104:1631–1636. vation and turnover in BMP and TGF-beta 23 Galli GG, Carrara M, Yuan WC et al. YAP 35 Park GH, Jeong H, Jeong MG et al. pathways. Cell 2009;139:757–769. drives growth by controlling transcriptional Novel TAZ modulators enhance myogenic dif- 12 Camargo FD, Gokhale S, Johnnidis JB pause release from dynamic enhancers. Mol ferentiation and muscle regeneration. Br J et al. YAP1 increases organ size and expands Cell 2015;60:328–337. Pharmacol 2014;171:4051–4061. undifferentiated progenitor cells. Curr Biol 24 Zanconato F, Forcato M, Battilana G 36 Jeon YH, Park YH, Lee JH et al. Seleno- 2007;17:2054–2060. et al. Genome-wide association between protein W enhances skeletal muscle differen- 13 Tumaneng K, Russell RC, Guan KL. Organ YAP/TAZ/TEAD and AP-1 at enhancers drives tiation by inhibiting TAZ binding to 14-3-3 size control by Hippo and TOR pathways. oncogenic growth. Nat Cell Biol 2015;17: protein. Biochim Biophys Acta 2014;1843: Curr Biol 2012;22:R368–R379. 1218–1227. 1356–1364. 14 Cinar B, Fang PK, Lutchman M et al. 25 Tremblay AM, Missiaglia E, Galli GG 37 Jeong H, Bae S, An SY et al. TAZ as a The pro-apoptotic kinase Mst1 and its cas- et al. The Hippo transducer YAP1 transforms novel enhancer of MyoD-mediated myogenic pase cleavage products are direct inhibitors activated satellite cells and is a potent differentiation. FASEB J 2010;24:3310–3320.

www.StemCells.com VC 2017 The Authors STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 1972 Taz Enhances Proliferation and Differentiation

38 Mohamed A, Sun C, De Mello V et al. 51 Pallafacchina G, Francois S, Regnault B potential of an intestinal regeneration pro- The Hippo effector TAZ (WWTR1) transforms et al. An adult tissue-specific stem cell in its gram. Genes Dev 2010;24:2383–2388. myoblasts and TAZ abundance is associated niche: A gene profiling analysis of in vivo qui- 65 Lee MJ, Ran Byun M, Furutani-Seiki M with reduced survival in embryonal rhabdo- escent and activated muscle satellite cells. et al. YAP and TAZ regulate skin wound heal- myosarcoma. J Pathol 2016;240:3–14. Stem Cell Res 2010;4:77–91. ing. J Invest Dermatol 2014;134:518–525. 39 Schlegelmilch K, Mohseni M, Kirak O 52 Tremblay AM, Camargo FD. Hippo sig- 66 Imajo M, Ebisuya M, Nishida E. Dual et al. Yap1 acts downstream of alpha-catenin naling in mammalian stem cells. Semin Cell role of YAP and TAZ in renewal of the intesti- to control epidermal proliferation. Cell 2011; Dev Biol 2012;23:818–826. nal epithelium. Nat Cell Biol 2015;17:7–19. 144:782–795. 53 Wang W, Li X, Huang J et al. Defining 67 Ulbricht A, Eppler FJ, Tapia VE et al. Cel- 40 Tian Y, Kolb R, Hong JH et al. TAZ pro- the protein-protein interaction network of lular mechanotransduction relies on tension- motes PC2 degradation through a SCFbeta- the human hippo pathway. Mol Cell Proteo- induced and chaperone-assisted autophagy. Trcp E3 ligase complex. Mol Cell Biol 2007; mics 2014;13:119–131. Curr Biol 2013;23:430–435. 27:6383–6395. 54 Price MG, Landsverk ML, Barral JM 68 Sudol M, Bork P, Einbond A et al. Char- 41 Moyle LA, Zammit PS. Isolation, culture et al. Two mammalian UNC-45 isoforms are acterization of the mammalian YAP (Yes-asso- and immunostaining of skeletal muscle fibres related to distinct cytoskeletal and muscle- ciated protein) gene and its role in defining a to study myogenic progression in satellite specific functions. J Cell Sci 2002;115:4013– novel protein module, the WW domain. cells. Methods Mol Biol 2014;1210:63–78. 4023. J Biol Chem 1995;270:14733–14741. 42 Zammit PS, Relaix F, Nagata Y et al. Pax7 55 Cottle DL, McGrath MJ, Cowling BS 69 Kawakami K, Sato S, Ozaki H et al. Six and myogenic progression in skeletal muscle et al. FHL3 binds MyoD and negatively regu- family genes–structure and function as tran- satellite cells. J Cell Sci 2006;119:1824–1832. lates myotube formation. J Cell Sci 2007;120: scription factors and their roles in develop- 43 Figeac N, Serralbo O, Marcelle C et al. 1423–1435. ment. Bioessays 2000;22:616–626. ErbB3 binding protein-1 (Ebp1) controls pro- 56 Benhaddou A, Keime C, Ye T et al. Tran- 70 Liu Y, Chakroun I, Yang D et al. Six1 reg- liferation and myogenic differentiation of scription factor TEAD4 regulates expression ulates MyoD expression in adult muscle pro- muscle stem cells. Dev Biol 2014;386:135– of myogenin and the unfolded protein genitor cells. PloS One 2013;8:e67762. 151. response genes during C2C12 cell differentia- 71 Griffin CA, Kafadar KA, Pavlath GK. 44 Ortuste Quiroga HP, Goto K Zammit PS. tion. Cell Death Differ 2012;19:220–231. MOR23 promotes muscle regeneration and Isolation, cryosection and immunostaining of 57 Goetsch SC, Martin CM, Embree LJ et al. regulates cell adhesion and migration. Dev skeletal muscle. Methods Mol Biol 2016; Myogenic progenitor cells express filamin C Cell 2009;17:649–661. 1460:85–100. 72 Arndt V, Dick N, Tawo R et al. Chaperone- 45 Turriziani B, Garcia-Munoz A, Pilkington in developing and regenerating skeletal mus- assisted selective autophagy is essential for mus- R et al. On-beads digestion in conjunction cle. Stem Cells Dev 2005;14:181–187. with data-dependent mass spectrometry: A 58 Millay DP, Sutherland LB, Bassel-Duby R cle maintenance. Curr Biol 2010;20:143–148. shortcut to quantitative and dynamic interac- et al. Myomaker is essential for muscle 73 Moorwood C. Syncoilin, an intermediate tion proteomics. Biology (Basel) 2014;3:320– regeneration. Genes Dev 2014;28:1641–1646. filament-like protein linked to the dystrophin 332. 59 Ciciliot S, Schiaffino S. Regeneration of associated protein complex in skeletal mus- 46 Vizcaino JA, Csordas A, Del-Toro N et al. mammalian skeletal muscle. Basic mecha- cle. Cell Mol Life Sci 2008;65:2957–2963. 2016 update of the PRIDE database and its nisms and clinical implications. Curr Pharm 74 Bentzinger CF, von Maltzahn J, Dumont related tools. Nucleic Acids Res 2016;44: Des 2010;16:906–914. NA et al. Wnt7a stimulates myogenic stem 11033. 60 Yu FX, Zhao B, Guan KL. Hippo pathway cell motility and engraftment resulting in 47 Plouffe SW, Meng Z, Lin KC et al. Char- in organ size control, tissue homeostasis, and improved muscle strength. J Cell Biol 2014; acterization of Hippo pathway components cancer. Cell 2015;163:811–828. 205:97–111. by gene inactivation. Mol Cell 2016;64:993– 61 Hong JH, Hwang ES, McManus MT et al. TAZ, 75 von Maltzahn J, Chang NC, Bentzinger 1008. a transcriptional modulator of mesenchymal stem CF et al. Wnt signaling in myogenesis. Trends 48 Park HW, Kim YC, Yu B et al. Alternative cell differentiation. Science 2005;309:1074–1078. Cell Biol 2012;22:602–609. Wnt signaling activates YAP/TAZ. Cell 2015; 62 Murakami M, Nakagawa M, Olson EN 76 Rudnicki MA, Williams BO. Wnt signaling 162:780–794. et al. A WW domain protein TAZ is a critical in bone and muscle. Bone 2015;80:60–66. 49 Holbourn KP, Acharya KR, Perbal B. The coactivator for TBX5, a 77 Takata H, Terada K, Oka H et al. Involve- CCN family of proteins: Structure-function implicated in Holt-Oram syndrome. Proc Natl ment of Wnt4 signaling during myogenic pro- relationships. Trends Biochem Sci 2008;33: Acad Sci USA 2005;102:18034–18039. liferation and differentiation of skeletal 461–473. 63 Kim J, Jo H, Hong H et al. Actin remod- muscle. Dev Dyn 2007;236:2800–2807. 50 Mege RM, Goudou D, Diaz C et al. N- elling factors control ciliogenesis by regulat- 78 Ono Y, Calhabeu F, Morgan JE et al. cadherin and N-CAM in myoblast fusion: ing YAP/TAZ activity and vesicle trafficking. BMP signalling permits population expansion Compared localisation and effect of blockade Nat Commun 2015;6:6781. by preventing premature myogenic differenti- by peptides and antibodies. J Cell Sci 1992; 64 Cai J, Zhang N, Zheng Y et al. The Hippo ation in muscle satellite cells. Cell Death and 103: 897–906. signaling pathway restricts the oncogenic Differentiation 2011;18:222–234.

See www.StemCells.com for supporting information available online.

VC 2017 The Authors STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press