SQSTM1 and LC3 during cardiac myofibril assembly and turnover

Developmental Biology 351 (2011) 46–61

Contents lists available at ScienceDirect

Developmental Biology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / d eve l o p m e n t a l b i o l o g y

Developmental regulation of MURF ubiquitin ligases and autophagy proteins nbr1, p62/SQSTM1 and LC3 during cardiac myofibril assembly and turnover Sue Perera, Mark R. Holt, Baljinder S. Mankoo, Mathias Gautel ⁎ King's College London BHF Centre of Research Excellence, Randall Division for Cell and Molecular Biophysics and Cardiovascular Division, New Hunt's House, Guy's Campus, London SE1 1UL, UK

a r t i c l e

i n f o

Article history: Received for publication 3 April 2010 Revised 14 December 2010 Accepted 15 December 2010 Available online 23 December 2010 Keywords: MURF2 isoforms Ubiquitin E3 ligase p62 SQSTM1 A170 nbr1 Myofibrils Microtubule dynamics Cardiac development titin

a b s t r a c t The striated muscle-specific tripartite motif (TRIM) proteins TRIM63/MURF1, TRIM55/MURF2 and TRIM54/ MURF3 can function as ubiquitin E3 ligases in ubiquitin-mediated muscle protein turnover. Despite their wellcharacterised roles in muscle atrophy, the dynamics of MURF expression in the development and early postnatal adaptation of striated muscle is largely unknown. Here, we show that MURF2 is expressed at the very onset of mouse cardiac differentiation at embryonic day 8.5, and represents a sensitive marker for differentiating myocardium. During cardiac development, expression shifts from the 50 kDa to the 60 kDa Aisoform, which dominates postnatally. In contrast, MURF1 shows strong postnatal upregulation and MURF3 is not significantly expressed before birth. MURF2 expression parallels that of the autophagy-associated proteins LC3, p62/SQSTM1 and nbr1. SiRNA knockdown of MURF2 in neonatal rat cardiomyocytes disrupts posttranslational microtubule modification and myofibril assembly, and is only partly compensated by upregulation of MURF3 but not MURF1. Knockdown of both MURF2 and MURF3 severely disrupts the formation of ordered Z- and M-bands, likely by perturbed tubulin dynamics. These results suggest that ubiquitin-mediated protein turnover and MURF2 in particular play an unrecognised role in the earliest steps of heart muscle differentiation, and that partial complementation of MURF2 deficiency is afforded by MURF3. © 2010 Elsevier Inc. All rights reserved.

Introduction The physiological turnover of muscle proteins is performed by predominantly two proteolytic systems, the ubiquitin–proteasome system (UPS) and the autophagy/lysosomal system (reviewed in (Sandri, 2008). Both degradation pathways remove misfolded and damaged proteins, but also such sarcomeric, contractile, metabolic, signalling and transcriptional proteins that need to be replaced by other isoforms during physiological muscle adaptation. Additionally, the cysteine-proteases of the calpain family aid selective protein turnover by the cleavage of a number of largely unidentified sarcomeric protein targets (Beckmann and Spencer, 2008; Willis et al., 2009). Rather than being a hallmark of extreme conditions like disuse atrophy or cachexia, controlled protein degradation is a requirement for muscle plasticity, which involves the removal and exchange of coordinated sets of proteins (Schiaffino et al., 2008). Common to both degradation pathways is the conjugation of the target proteins to the small protein modifier ubiquitin by an enzyme cascade whose specificity is achieved by dedicated ubiquitin E3ligases. In muscle, several tissue-specific E3 ligases have been identified as atrophy-related genes (“atrogenes”): the F-box protein ⁎ Corresponding author. King's College London, The Randall Division for Cell and Molecular Biophysics, London SE1 1UL, UK. Fax: + 44 207 848 6435. E-mail address: [email protected] (M. Gautel). 0012-1606/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2010.12.024

atrogin-1/MAFbx, and the MURF muscle-specific RING-finger proteins (reviewed in (Willis et al., 2009). The RING/B-box/coiled-coil or tripartite motif containing (TRIM) protein family of MURFs was identified initially by the interaction of MURF3 with the serum response transcription factor, SRF (Spencer et al., 2000). The three MURF genes (MURF1/TRIM63; MURF2/TRIM55; MURF3/TRIM54) encode highly homologous proteins that can homo- and heterodimerize via their coiled-coil domain (Centner et al., 2001), a defining feature of the TRIM protein family. Extensive differential splicing occurs in the C-termini of some MURF genes (Centner et al., 2001; Pizon et al., 2002), leading to isoforms with tissue-specific expression patterns in the case of MURF2 (Pizon et al., 2002). In contrast, MURF1 does not seem to be differentially spliced, whereas two human and one mouse isoform were reported for MURF3 (Centner et al., 2001; Spencer et al., 2000). Interactions of MURF1 and MURF2 with domains near the C-terminus of titin lead to association with the M-band (Centner et al., 2001; Pizon et al., 2002). However, both MURF1 and MURF3 have also been localised to the Z-band (Centner et al., 2001; Spencer et al., 2000), indicating that the sarcomeric targeting of MURFs is not only dependent on M-band titin interactions. MURF3 and MURF2 have additionally been found in association with glutamylated microtubules and nascent myosin filaments (Pizon et al., 2002; Spencer et al., 2000), for which the RING and B-box domains are crucial (Spencer et al., 2000). Myogenic differentiation is characterised by the transient formation of stable arrays of such

S. Perera et al. / Developmental Biology 351 (2011) 46–61

glutamylated tubulin, with a simultaneous reduction in the dynamic pool of tyrosinated tubulin (Gundersen et al., 1989). These observations suggest that MURFs could be crucially involved in regulating microtubule dynamics, similar to the related non-muscle TRIM protein MID1 (Berti et al., 2004). A knockout mouse model of MURF1, however, shows no signs of impaired myofibril assembly, but to the contrary, a resistance to both disuse- and steroid-induced atrophy (Bodine et al., 2001). Similarly, MURF1/3 double knockout animals do not seem to show defects in primary myofibrillogenesis, but rather a postnatal myosin storage myopathy due to disrupted myosin heavy chain turnover (Fielitz et al., 2007a). Additional roles in the regulation of energy metabolism likely stem from targeting multiple cytosolic and mitochondrial metabolic enzymes by MURF1 and 2 (Hirner et al., 2008; Koyama et al., 2008; Witt et al., 2008). These combined observations suggest that many MURF functions require synergistic action of more than one isogene. Similar to related TRIM proteins, where the RING domain is associated with ubiquitin modification (Jackson et al., 2000), the RING domain of MURFs has been implicated in posttranslational modification by the ubiquitin, but also tentatively by the ubiquitin-related SUMO systems (for a review, see e.g. (Gill, 2004). MURF1 and 3 have ubiquitin ligase activity in vitro (Bodine et al., 2001; Fielitz et al., 2007a) and MURF1 is upregulated in atrophic muscle similar to other ubiquitin E3 ligases (Glass, 2003). On the other hand, the RING domain of MURF1 (sometimes also called SMRZ) interacts with the small ubiquitin-like modifier Smt3b/SUMO3 (Dai and Liew, 2001), and all MURFs interact with the SUMO E2 transferase UBC9. Like other ubiquitin ligases, MURF1 and 2 can also appear in the nucleus (McElhinny et al., 2002; Pizon et al., 2002), specifically under conditions of muscle atrophy for MURF2 (Lange et al., 2005; Pizon et al., 2002). The interactions with the transcriptional cofactor glucocorticoid modulatory element binding protein-1 (GMEB-1; (McElhinny et al., 2002) and with SRF (Lange et al., 2005; Spencer et al., 2000; Willis et al., 2007) may thus be functionally related to the ubiquitin-associated activity of MURFs. Given the established E3 ligase activity for MURF1 and 3 in cooperation with the E2 ubiquitinconjugating enzymes UbcH5a, -b, and -c, and the near-identity of the RING/B-box domains between all MURFs, it is plausible to propose that nuclear ubiquitination may account for the changes in cellular localisation of SRF (Lange et al., 2005), similar to related nuclear ubiquitin signalling pathways (Salmena and Pandolfi, 2007; von Mikecz, 2006). Up-regulation of MURF1 also inhibited cardiomyocyte hypertrophy via protein kinase-C epsilon activity, by blocking kinase interaction with the scaffolding protein RACK1 (Arya et al., 2004). In the sarcomere, MURFs form an M-band associated kinase-ubiquitin signalling module with titin and the kinase-associated scaffolds nbr1 and p62/SQSTM1 (Lange et al., 2005), both of which show crosstalk to many other important signalling pathways (Gautel, 2008) as well as linking to the control of autophagy (Bjorkoy et al., 2005; Kirkin et al., 2009; Komatsu et al., 2007; Waters et al., 2009). In autophagy, proteins targeted for degradation by ubiquitin modification usually bear lysine-63 linked polyubiquitin chains, and are recruited to autophagosomes containing lipidated LC3 within their membranes by specific adapter proteins binding both ubiquitin and LC3 (Geng and Klionsky, 2008). Proteasomal degradation typically removes proteins labelled with lysine-48 linked chains. MURFs therefore not only associate with different cytoskeleton compartments such as microtubules, Z-bands and M-bands, as well as with nuclear proteins, but cooperate with diverse proteins implicated in selective protein degradation by the proteasome and autophagosome, and target proteins of metabolic regulation, sarcomere assembly, transcriptional regulation and possibly SUMO-mediated posttranslational regulation. To what extent the highly specialized but homologous MURF proteins are indeed redundant remains unclear. Conflicting information from studies in cultured cells and knockout animals is currently hard to reconcile, as the in vivo expression pattern of MURFs during

47

development is unknown. Therefore, we studied the expression of MURFs and the functionally associated components of the autophagy and proteasomal degradation machineries in cardiac development, and their mechanistic implications in myofibril assembly. Our results show, surprisingly, that MURF2 is the first MURF to be significantly expressed in the heart and to be required for myofibril assembly or maintenance, while MURF1 and -3 expression becomes significant only postnatally. Results The splicing pattern of MURF2-A isoforms is highly conserved between mouse and human MURF2 is extensively spliced to generate multiple tissue-specific isoforms in human striated muscle (Pizon et al., 2002) (Supplementary Fig. 1A). To determine whether a similar gene architecture is also applicable to mouse MURF2, cDNA from adult mouse left ventricles was amplified with p60A- and p50A-specific primer combinations (based upon the mouse gene structure, Supplementary Fig. 1B) to obtain the specific C-terminal regions, which produced bands of the predicted sizes (Supplementary Fig. 1C). Using primers in the flanking exons 1 and 11, full-length cDNA of p60A, p50A could be amplified; p27A message could be amplified with a nested PCR (Supplementary Figs. 1D–E). Sequencing confirmed that the exon–intron spliceboundaries of mouse MURF2 p60A, p50A and p27A correspond to those in human, and that transcripts analogous to the three known human MURF2-A isoforms – p60A, p50A and p27A – are present in adult mouse cardiac muscle (Supplementary Fig. 1) Overall, these data establish that the splicing pattern of MURF2-A isoforms is conserved between the human and mouse. Although the predicted mouse p60B isoform shows 90% amino acid identity to human, the primer pair used could not amplify it, which agrees with the skeletalmuscle expression described previously (Pizon et al., 2002). MURF2-A isoforms show developmental regulation in mouse cardiac muscle Given that multiple MURF2 transcripts are present in mouse left ventricles, we assumed that the equivalent encoding peptides should also be expressed. Western blotting of adult mouse cardiac muscle lysates using a commercially available goat anti-MURF2 antibody against the unique constitutive C-terminus of all human A-isoforms identified three bands migrating at approximately 70 kDa, 50 kDa and 27 kDa, corresponding to the predicted p60A, p50A and p27A isoforms respectively (Supplementary Fig. 2A). Further analysis was performed using two additional rabbit polyclonal antibodies against: (1) the constitutive C-terminus of human A-isoforms, and (2) the SerAla rich p60-specific domain of human MURF2 (Pizon et al., 2002). The results with these antibodies independently confirmed the identity of the bands and showed that all three MURF2-A species are expressed in adult mouse cardiac muscle (Supplementary Fig. 2A). Immunoblots employing different antibodies previously failed to detect the 27 kDa splice variant in rat heart and chick cardiomyocyte lysates in some studies (McElhinny et al., 2004). This is probably due to differences in antibody affinities and the much lower expression levels of the p27A isoform. While species-specific differences cannot be excluded, this seems unlikely given the conservation of MURF2 gene structure and splice pathways between human and mouse and the detection of p27 in adult mouse heart (Lange et al., 2005). Band intensities revealed that the expression levels of MURF2-A isoforms are approximately in the order p60A N p50A N p27A (Supplementary Fig. 2A). This, along with the observation that MURF2 isoforms may be variably expressed between foetal “broad band (55– 60 kDa)” and adult rat hearts (McElhinny et al., 2004), led us to

48

S. Perera et al. / Developmental Biology 351 (2011) 46–61

investigate its expression profile during murine cardiac development in the mouse. The principal MURF2 isoforms – p60A and p50A – undergo a major switch in expression levels during development of the mouse heart (Fig. 1A). From embryonic to early prenatal stages, p60A levels are considerably lower than those of p50A; however, from about P9 onwards, this ratio is changed and the two isoforms are equalised (Fig. 1A). Additionally, both p60A and p50A levels were downregulated as postnatal cardiac growth progressed (Fig. 1A, Supplementary Fig. 2B).

Given that MURF family members are proposed to regulate muscle protein degradation and turnover (Willis et al., 2009) the expression profile of MURF-associated proteins involved in the UPS and autophagy/lysosomal systems were also explored. LC3, nbr1 and p62/SQSTM1 exhibit parallel expression changes during murine heart development, with the highest protein levels being present at early stages whereas levels in adult muscle appear to be low (Figs. 1A–B). However, LC3 transcript levels increase postnatally, suggesting an increase in autophagic flux with concomitant reduced steady-state protein levels, while p62/SQSTM1 and nbr1 transcripts decrease

Fig. 1. Developmental expression of MURF-family members and related UPS/autophagy proteins in the heart. A. Western blot analysis of mouse heart lysates from embryonic (E) to postnatal (P) stages. The MURF2 isoforms are marked as p50A and p60A (here: detected with goat anti-MURF2) Full-length SRF (FL) and the repressive isoform SRF-Δ5. B. Plot of the densitometric analysis on Western blots normalised for GAPDH levels. C. RT-PCR on cDNA isolated from mouse hearts with the full-length MURF2 and MURF1 primers shows developmentally regulated levels and splice variant expression for MURF2. D. Plot of RT-PCR data obtained from 3 independent sets of experiments show the late-onset of MURF1 expression; error bars represent standard deviation. Data are normalised for GAPDH expression E. Western blots showing the differential expression of MURF2 p50A and p60A isoforms between the left ventricle (LV) and left atrium (LA) between postnatal day P9 and adult stages. E: embryonic day; P: postnatal day; Wk: week.

S. Perera et al. / Developmental Biology 351 (2011) 46–61

postnatally (Supplementary Figs. 2C–D). In comparison, MURF2 isoforms are consistently upregulated from E9.5 until birth, suggesting that ubiquitin-mediated protein turnover may be more prevalent than autophagy during embryonic cardiac remodelling. The expression of SRF, a key transcription factor required for cardiac differentiation and postnatal hypertrophic growth that is expressed in transcriptionally active and repressive isoforms (Belaguli et al., 1999), and a known target of MURF1 and MURF2 (Lange et al., 2005; Willis et al., 2007), also peaks at E9.5, then reduces rapidly, simultaneously with increasing levels of MURF2 (Figs. 1A–B). MURF3 protein expression is negligible during embryonic cardiac development and only upregulated postnatally (Fig. 1A). Due to the variable results obtained in Western blots with a panel of commercial antibodies that detect endogenous MURF1 in the mouse, semiquantitative RT-PCR was performed to compare the relative message levels of MURF1 and MURF2 (Figs. 1C–D), showing that MURF1 expression is upregulated significantly only postnatally. Altogether, these analyses confirm that MURF2 is the predominant, and the most dynamically regulated MURF-family member during embryonic development of the mouse heart. The most significant change in the postnatal heart involves the structural and functional remodelling of the ventricles, with the onset of pulmonary circulation and a strict separation of venous and arterial circulation and the cessation of hyperplastic heart growth. We therefore asked whether this functional remodelling would correlate with a postnatal isoform switch in the left atrium and left ventricle. At P9, although p50A is the predominant MURF2 isoform in both chambers, the relative amount of p60A is nevertheless noticeably higher in the left ventricle compared to the left atrium (Fig. 1E). In the adult, the ratio of the two isoforms is more equalised: whilst p60A levels are greater than those of p50A in the left ventricle, in the left atrium p50A levels remain, albeit marginally, greater than those of p60A (Fig. 1E). This suggests that MURF2 may contribute to the functional adaptation and differentiation of ventricular myocardium. To complete the picture of MURF2 isoforms in mouse cardiac muscle, the expression pattern of the 27 kDa splice variant was explored. It is first detected at very low levels at about three weeks of age and is readily detected in adult hearts (approximate molar ratios of p60A:p50A:p27A = 6:3:1), and thus appears to be strongly developmentally regulated and a marker of postnatal development (Supplementary Figs. 2A and E). MURF2 localisation during cardiac myofibrillogenesis in vivo To determine the spatial distribution of isoforms during cardiac morphogenesis, whole-mount specimens of mouse hearts were analysed by immunofluorescence techniques. MURF2 is widely expressed throughout the primary heart tube at E8.5 with the strongest signal observed in the ballooning regions that form the presumptive working myocardium (Fig. 2A). At the subcellular level, α-actinin, a Z-band marker, indicates the emergence of the first crossstriations of contractile sarcomeres (Fig. 2B). MURF2 shows no clear sarcomeric associations at this stage (Fig. 2B). At E9.5, when myofibrillogenesis in the beating heart is increasing, MURF2 is globally expressed in the looping heart tube (Fig. 2C). Its subcellular localisation is more organised and the first MURF2 crossstriations are visible as alternating bands between the α-actininpositive Z-bands, indicating that MURF2 is now M-band-associated (Fig. 2D). However, whilst Z-bands could be visualised with α-actinin, not all corresponding M-bands were positive for MURF2; additionally, in more immature areas of myofibrillogenesis, where Z-bands only start to assemble, MURF2 is more punctate, suggesting that the integration of MURF2 into M-bands occurs after Z-band assembly (Fig. 2D, inset top). At E10.5, MURF2 remains widely distributed (Fig. 2E), and higher magnification shows that the M-band-integration is more pronounced (Fig. 2F, bottom insert); nevertheless, this

49

clear sarcomeric association is again absent in zones of immature myofibrils (Fig. 2F, top inset). MURF2 is preferentially associated with glutamylated tubulin during cardiac myofibrillogenesis in vivo MURF2 was shown to associate with microtubules in a skeletal muscle cell line, and these MURF2-decorated microtubules transiently associated with sarcomeric proteins (particularly myosin) prior to the formation of mature myofibrils (Pizon et al., 2002). This microtubule association is a feature shared with MURF3 (Spencer et al., 2000). Whilst MURF2 may influence myofibrillogenesis in multiple ways, one such avenue might be via a coordinated stabilisation or turnover of microtubules. To test whether microtubule association with specific subsets of microtubules occurs also in the heart in vivo, whole mount specimens of mouse hearts at the onset of cardiac myofibrillogenesis were analysed by immunofluorescence. Confocal microscopy at E8.5 revealed that tubulin and MURF2 expression overlapped predominantly in the areas lacking clear αactinin cross-striations (Fig. 3A). Zones with mature ≈2 μm α-actinin stripes at Z-bands display no such colocalisation (Fig. 3A, lower insert). At E9.5, a similar pattern is detected, as the majority of tubulin and MURF2 colocalisation is again limited to regions that do not show mature Z-band cross-striations (Figs. 3B–C, arrows). However, the overall degree of colocalisation observed was reduced at E9.5 compared to E8.5, suggesting that the interaction of MURF2 with microtubules may be transient and related to the extent of myofibrillogenesis. We next investigated the association of MURF2 with glutamylated tubulin, which corresponds to a major part of the stable detyrosinated tubulin accumulating during myofibrillogenesis (Gundersen et al., 1989). In E9.5 hearts, the location of MURF2 and glutamylated tubulin overlapped considerably, and this association was much more pronounced than previously noted with all tubulin populations (Fig. 3D). The highest degree of colocalisation was present in areas where the M-band cross-striations were emerging, but prior to achieving their highly regular periodicity (Fig. 3D, insets). To summarise, in the developing heart, MURF2 shows a transient and selective colocalisation pattern with glutamylated tubulin in areas of nascent myofibrils, suggesting that this association is significant for cardiac myofibrillogenesis in vivo. Isogene-specific knockdown of MURFs disrupts myofibril assembly and microtubule dynamics We next asked whether MURF2 isoforms show distinct functions that could be dissected by specific knockdown, and p50A- and p60Aspecific 19-mer siRNA constructs were therefore designed. The siRNA were expressed from a modified H1-GFP plasmid (Brummelkamp et al., 2002) that also encoded for GFP for identification of transfected cells. Efficiency of knockdown was initially tested in COS-1 cells by cotransfecting the siRNA along with murine GFP-tagged full-length p50A and p60A cDNA. The p50A-specific siRNA strongly reduced the levels of p50A protein, with a consistent down-regulation of approximately 70% (Supplementary Figs. 3A–B). The knockdown is specific for the targeted p50A isoform, as p60A was unaffected (Supplementary Figs. 3A–B). Similarly, the p60A-specific siRNA reduced p60A protein levels by approximately 79%, and the knockdown effect was again p60A-specific (Supplementary Figs. 3C–D). Analysis of neonatal rat cardiomyocyte (NRC) lysates confirmed that both MURF2 p50A and p60A are present in perinatal rat hearts (P0), and as expected, the levels of p60A are considerably lower than that of p50A (Supplementary Fig. 3E). Transfection of p50A- and p60A-specific siRNA resulted in reduced endogenous MURF2 expression in 72% and 67%, respectively, in neonatal rat cardiomyocytes after 48 h in culture in hypertrophy-

50

S. Perera et al. / Developmental Biology 351 (2011) 46–61

Fig. 2. Whole mount preparations of embryonic mouse hearts show in vivo organisation of MURF2 into M-bands. A–B. MURF2 is widely expressed in the ballooning regions (red arrowheads) of the primary heart tube at E8.5. While α-actinin staining at this stage shows organised Z-bands, MURF2 staining is not distinctly striated (inserts in B). C–D. At E9.5, MURF2 is strongly expressed in the looping heart (red arrowheads in C) and becomes M-band-associated, as seen by alternative staining between the α-actinin-positive Z-bands (D, bottom inset), although in regions where α-actinin staining is less organised, MURF2 expression is more punctate (D, top inset). E–F. MURF2 remains strongly expressed (E, red arrowheads) and is more generally sarcomeric in the E10.5 looping heart (F, bottom insert) with remaining non-striated patterns in areas with non-striated α-actinin (F, top insert). BA= branchial arches, A= future atrium, V= future ventricles, OFT= outflow tract, LA= left atrium, LV= left ventricle. In all panels of this figure, white arrows identify the region that is shown amplified 2-fold in the inset box.

S. Perera et al. / Developmental Biology 351 (2011) 46–61

51

Fig. 3. MURF2 associates with tubulin in vivo during cardiac myofibrillogenesis. A. MURF2 is colocalised with α-tubulin at E8.5 in areas of immature myofibrillogenesis, white arrows in upper magnified areas (amplified 2-fold). In zones with emerging regular α-actinin cross-striations, this colocalisation is lost (white arrow in lower magnified area). B. A similar pattern of MURF2 and all tubulin is observed at E9.5, and remains constrained to areas where α-actinin is being organised (C). D. MURF2 is colocalised with the stable, glutamylated microtubules set (here at E9.5, left magnified panel), particularly in zones of immature myofibrillogenesis (right magnified panel).

52

S. Perera et al. / Developmental Biology 351 (2011) 46–61

Fig. 4. siRNA knockdown of MURF2 p50A or p60A causes sarcomeric disruption in cardiomyocytes. Neonatal rat cardiomyocytes were transfected with MURF2 isoform-specific siRNAs expressed from a modified H1-GFP plasmid. Transfected cells were identified by expression of GFP. MURF2-A isoforms were detected with the HPC antibody against all Aisoforms. MURF2 p60 isoforms were detected specifically with the HP60 antibody. A–B. p50A-specific knockdown causes disorganisation of both the M-band (myomesin) and Zband (α-actinin) in NRC after 48 h of transfection, as shown by the loss of regular stripes (inserts). The α-actinin staining pattern is typical for nascent myofibrils. C–D. Similar disorganisation is observed with p60A-specific knockdown, but in fewer cells. White arrows in A–D identify the regions magnified in the insets.

inducing conditions. Examination of their sarcomeres revealed pronounced lack of organisation of both the M-bands and Z-bands (Figs. 4A–D), a phenotype seen in only about 25% of untransfected or control-siRNA transfected cells (Figs. 4E and 7D). Quantitation revealed that the p50A knockdown produces more severe and consistent sarcomeric disorganisation (Figs. 4A–B) than knockdown of the p60A isoform (Figs. 4C–D), in agreement with the endogenous levels of p60A being considerably lower than that of p50A (quantitation in Fig. 7D). In the culture maintenance conditions used here, NRC were kept in the presence of 5% horse serum and phenylephrine (Auerbach et al., 1999), inducing a hypertrophic phenotype with active myofibrillogenesis (Iwaki et al., 1990). The immature organisation of α-actinin in

stress-fibre like structures and the failure of M-band assembly suggest that these cells are not progressing through myofibrillogenesis and remain stuck at the nascent myofibril stage (Sanger et al., 2005). Given that MURF2 is developmentally associated with microtubules during myofibril assembly, we therefore hypothesised that the sarcomeric disorganisation observed could be at least in part due to defects of microtubule organisation during myofibril assembly. Immunostaining of NRC transfected with p50A-specific siRNA revealed that this was indeed the case, the stable glutamylated microtubule populations associated with myogenic differentiation were strongly reduced, whilst in contrast, the dynamic tyrosinated pool was increased (Figs. 5A–C). Quantitation of the mean intensity per transfected cells revealed that the changes for acetylated

S. Perera et al. / Developmental Biology 351 (2011) 46–61

53

Fig. 5. MURF2 p50A knockdown disrupts microtubule post-translational modifications in NRC. A. Loss of p50A leads to decreased levels of the stable, glutamylated microtubule population. B. Acetylated microtubules show little change. C. In contrast, the dynamic, tyrosinated microtubule pool is strongly upregulated in NRC with depleted p50A levels. Transfected cells are identified by GFP expression (red asterisk in separated channels). D. The mean intensity per transfected cell for the different tubulin stains was quantified for 50 cells per treatment (see Materials and methods). Error bars represent standard errors. One-way ANOVA was used for statistical analysis (p b 0.001 for glu-tub decrease and p b 0.006 for tyro-tub increase). Changes for acetyl-tub were not significant.

microtubules were not significant when compared to control siRNA transfected cells, in contrast to the highly significant decrease of glutamylated tubulin (p = 0.002) and concomitant increase in tyrosinated tubulin (p = 0.006; Fig. 5D). Taken together, these observations suggest that knockdown of the embryonically predominant MURF2 p50A isoform leads to disruption of myofibrillogenesis in cardiomyocytes under hypertrophic conditions in vitro, possibly due to deregulation of microtubule-modifying complexes. Our combined results support an important role of MURF2 in cardiac development, and yet MURF2 knockout animals appear not to show a striking basal phenotype. We therefore wondered whether partial compensation could be afforded by MURF1 or -3. Interestingly,

no upregulation of MURF1 is detected in MURF2 siRNA treated cardiomyocytes (Fig. 6B). However, we find that MURF3 levels increase in correlation to MURF2 siRNA expression (p b 0.0001; Figs. 6A, C, E). Mirroring the knockdown of MURF2, MURF3 knockdown by an isogene-specific siRNA leads to upregulation of MURF2 expression (Supplementary Fig. 5) and an increase in immature Z- and M-bands, as well as a significant decrease in glutamylated microtubules (p = 0.003; Supplementary Fig. 6), which is in excellent agreement with previous results in skeletal myoblasts (Spencer et al., 2000). As these findings imply that partial MURF isogene compensation can occur in vitro, we tested whether the combined knockdown of

54

S. Perera et al. / Developmental Biology 351 (2011) 46–61

Fig. 6. MURF-family members can show partial compensation in vitro. A. Knockdown of MURF2-A p50A in NRC leads to MURF3 upregulation. siRNA transfected cells are identified by GFP expression (red asterisk in separated channels) and show increased MURF3 levels. B. Reduced MURF2-A levels do not lead to increased MURF1 levels. C. Ratiometric imaging shows upregulation of MURF3 in MURF2 knockdown cells. D. Control siRNA constructs do not affect MURF3 ratios to myomesin. The separate channels for GFP, endogenous myomesin and MURF3, and the ratiometric image with overlaid GFP mask for the outline of the transfected cell are shown in C and D. The false-colour scale range indicator shows a ratio range of 0 (black) to 3 (white). Scale bar: 10 μm. E. Ratiometric analysis confirms that MURF3 upregulation under MURF2 knockdown is highly significant (p b 0.0001). Data are based on over 50 individual cardiomyocytes per treatment.

MURF2 and MURF3 would show an aggravated phenotype on myofibril assembly and microtubule modification. The siRNA targeted against both rat MURF2 and MURF3 (Supplementary Figs. 3G, H)

reduced the levels of both isogenes (Fig. 7A) and dramatically reduced the number of mature Z- and M-bands (Figs. 7B, C). Quantitation of the effects of the individual and double MURF2 and MURF3

S. Perera et al. / Developmental Biology 351 (2011) 46–61

55

Fig. 7. Double knockdown of both MURFs 2 and 3 produces widespread loss of mature sarcomeric structures. A. Neonatal rat cardiomyocytes expressing MURF2 + 3 specific siRNA have reduced expression of both MURF2 and -3. B. Double knockdown of both MURFs (imaged here: MURF2) severely disrupts M-band assembly. C. Z-bands are similarly disrupted in cells depleted of MURF2 and-3. D. Percentage of mature sarcomeric structures in siRNA-expressing cardiomyocytes compared to non-transfected and control-siRNA transfected cells; errors bars represent SE of 3 independent experiments with counts of over 50 cells per treatment. One-way ANOVA was used for statistical analysis (* = p b 0.05; ** = p b 0.01; *** = p b 0.001). There was no significant difference between non-transfected and control-siRNA transfected cells.

knockdown effects on sarcomeric structure revealed that while both MURF2 and MURF3 individually affected Z- and M-band formation in comparison to control siRNA or untransfected cells, the double knockdown of MURF2/MURF3 resulted in a striking loss of ordered

M- and Z-bands (Fig. 7D). The combined knockdown of MURF2 and MURF3 also affected the posttranslationally modified tubulin pools in a manner distinct from the individual knockdowns, perturbing the accumulation of all three assayed modified tubulin fractions (Fig. 8).

56

S. Perera et al. / Developmental Biology 351 (2011) 46–61

Fig. 8. Loss of both MURF2 and -3 causes disruption of different microtubule modifications. A–C. NRC expressing MURF2+3 specific siRNA have reduced glutamylated (Glu-tub), acetylated (Acetyl-tub) and tyrosinated (Tyr-tub) microtubules (marked by asterisks). D. Mean intensity quantification confirms that all sub-populations of posttranslationally modified microtubules are significantly reduced compared to controls, however glutamylated tubulin is most severely affected. Data represent the mean tubulin intensity of over 60 cells ± SE. One-way ANOVA was used for statistical analysis (* = p b 0.05; *** = p b 0.001).

This might suggest that the microtubular targets of different MURF isogenes are non-redundant. However, the effects of MURF2/3 knockdown could be partially rescued by co-transfection of a plasmid expressing myc-tagged human MURF2 (which is resistant to the siRNA) under control of the CMV promoter. This construct restored

mature Z-disks in about 66% (n = 50) and M-bands in 62% (n = 60) of double-transfected cells (Supplementary Fig. 7). Taken together, these results suggest that MURF2 and MURF3 act together and to a certain degree, functionally compensate for each other during sarcomere assembly in vitro, and that the aberrant

S. Perera et al. / Developmental Biology 351 (2011) 46–61

myofibrillar phenotype observed with knockdown of these isogenes is likely to be caused by deregulated microtubule dynamics. Discussion The regulation of muscle protein turnover involves both the UPS and the autophagy/lysosomal system. Until recently, both systems were thought to be primarily involved in the protein and organelle breakdown associated with catabolic loss of muscle mass. However, it is becoming increasingly clear that both the UPS and autophagy pathways are required for the maintenance and adaptation of muscle by the controlled removal of damaged, misfolded or unneeded proteins and organelles. Postnatal inhibition of autophagy in cardiac and skeletal muscle therefore leads to the accumulation of toxic proteins, and paradoxically does not protect from disuse atrophy, but instead promotes greater muscle loss (Masiero et al., 2009; Masiero and Sandri, 2010; Taneike et al., 2010). Similarly, chaperone-assisted selective autophagy involving the heat-shock protein-70 complex, the chaperone-associated ubiquitin ligase CHIP and p62/SQSTM1 play a recently recognised role in Z-band maintenance, and its inhibition leads to Z-band disintegration and progressive muscle weakness (Arndt et al., 2010). Targeting the first step of protein labelling for proteasomal or autophagic degradation, by ablation of ubiquitin E3ligases-like MURFs, similarly interferes with muscle maintenance and leads to a myosin storage myopathy in the case of the MURF1/3 double knockout model (Fielitz et al., 2007a). The role of both processes during the development of the vertebrate embryo however is poorly understood, although an essential role of autophagy in early embryo differentiation at the 1– 8 cell stages has recently been identified using Atg5−/− mice (Tsukamoto et al., 2008). Later stages of mammalian embryonic development also require autophagy for controlling cell numbers, as mice with a targeted deletion of the autophagy-related gene Beclin1 die between E7.5–8.5 and show abnormalities in cell growth in the visceral endoderm (Yue et al., 2003), probably due to a defect in programmed cell death. Lysosomal/autophagic programmed cell death (type II) has further been identified as important in tissue patterning (reviewed in (Cecconi and Levine, 2008; Penaloza et al., 2006). Knowledge about functions in embryonic cell remodelling is currently very fragmentary at best. During development, the heart grows by cell division, rather than by the hypertrophic cell growth observed postnatally. Cell division thus occurs in a functional, beating myocardium (Ahuja et al., 2004). This necessarily requires the controlled disassembly and proteasomal degradation of myofibrillar components in dividing cells, followed by the postmitotic re-synthesis of myofibrils. However, the identity of the E3 ligases involved in this continuous assembly–disassembly cycle remains unknown (Ahuja et al., 2004), and in the light of recent insight into the role of autophagy in myofibril turnover and maintenance, a participation of the autophagy system also seems possible. Here, we report the first comprehensive developmental analysis of the muscle-specific ubiquitin MURF E3-ligase family in the heart, and correlate their expression pattern with other components of the protein degradation machinery relevant for muscle, especially the autophagy-associated complex of nbr1, p62/SQSTM1 and LC3 that links to both autophagy and proteasomal protein degradation systems. We find that components of the UPS and autophagy systems are expressed from the earliest detectable stages of cardiac differentiation, with the MURF2 isoforms being the predominant MURF E3 ligases in the heart until shortly after birth. Expression of the autophagy proteins is high during foetal development and rapidly decreases after birth. MURF1 RNA expression is upregulated from late foetal to adult stages in the heart, with major increases occurring postnatally. The expression of MURF2, nbr1 and p62/SQSTM1 appears concerted, and correlates with the expression of LC3, the autophagy-

57

linked ligand of nbr1 and p62/SQSTM1 (Noda et al., 2010) that is also a microtubule-associated component. Only postnatally is there a discrepancy between LC3, nbr1 and p62/SQSTM1 transcript versus protein levels, presumably due to increased autophagic flux postnatally. We also confirm that the expression of SRF, a key transcription factor critical to cardiac differentiation and postnatal hypertrophic growth (reviewed in (Miano et al., 2004; Niu et al., 2007)), closely mirrors that of MURF2 throughout development. Given the association of MURF1, 2- and -3 with SRF (Lange et al., 2005; Spencer et al., 2000; Willis et al., 2007) and that over-expressed or nuclear targeted MURF2 leads to reduced nuclear SRF levels in cultured cardiomyocytes (Lange et al., 2005), one might speculate that MURF may participate in regulating appropriate levels and localisation of SRF during cardiac development. Interestingly, the postnatal isoform shift and overall downregulation in MURF2 cardiac expression also coincides with the onset of exclusively hypertrophic growth of the mouse heart (Leu et al., 2001). Additionally, many myofibrillar proteins show developmental isoform expression at this stage (Schiaffino and Reggiani, 1996). Known targets of MURFs, such as contractile proteins like the troponin complex (Kedar et al., 2004), and non-contractile sarcomeric proteins like EH-myomesin (Agarkova et al., 2000), change their isoform expression status early postnatally. These modifications are significant to the functional maturation of the mammalian myocardium, enabling it to adapt both metabolically and structurally to meet the demands of postnatal life. Given the developmental regulation of MURF2 isoforms from embryonic to postnatal stages, a role in disassembly and reassembly of sarcomeres in dividing cardiomyocytes, as well as in the regulation of hypertrophic growth programmes, is conceivable. The knockout of MURF1 shows no apparent baseline phenotype (Bodine et al., 2001), and the loss-of-function phenotypes for MURF2 and MURF3 are equally subtle (Fielitz et al., 2007b; Willis et al., 2007). However, the combined loss of MURF1 with either MURF2 (Willis et al., 2007; Witt et al., 2008) or MURF3 (Fielitz et al., 2007a) causes distinct postnatal phenotypes impinging on hypertrophic heart growth and energy metabolism, which suggests an obligatory interplay of the three MURF gene products for targeting specific subsets of muscle contractile, metabolic and regulatory proteins in an only partly redundant manner. However, no triple knockout animals have been reported so far, which might suggest that the total ablation of MURF genes is embryonic lethal. Surprisingly, given that it has become evident that MURFs are crucial for the selective turnover of many muscle proteins, the developmental role of MURFs has remained completely obscure, as all studies to date investigating the functions in muscle formation have been in cultured skeletal or cardiac myocytes. Ubiquitin conjugation is not only the mediator of proteasomal protein degradation, but lysine-63 linked polyubiquitination can also target proteins for autophagic degradation as well as playing a role in cellular signalling. Both MURF1 and 2 interact with p62/SQSTM1 (Lange et al., 2005; Witt et al., 2008), with MURF2 colocalising with p62/SQSTM1 in presumptive autophagosomes (Lange et al., 2005). This links MURFs also to the scaffolding functions of the nbr1-p62/ SQSTM1 complex that is implicated in shuttling polyubiquitinated proteins to the autophagosomes (Ichimura et al., 2008; Kirkin et al., 2009; Komatsu et al., 2007; Moscat and Diaz-Meco, 2009; Pankiv et al., 2007; Tan et al., 2007; Waters et al., 2009). The role of autophagy during muscle development is still unclear; however, our results suggest that this protein degradation pathway may play an important and unrecognised role during early heart and skeletal muscle differentiation. It is surprising therefore, that ablation of MURF2 (or any other single MURF knockout) causes no severe developmental cardiac and/or skeletal muscle phenotype; however, 74% of MURF1/2 double knockout mice die within the first two weeks of birth due to acute cardiac insufficiency and decompensation (Witt et al., 2008). Our experiments in NRC using MURF-specific siRNA

58

S. Perera et al. / Developmental Biology 351 (2011) 46–61

knockdown suggest that at least on the cellular level, MURF2 or MURF3 levels are upregulated upon loss of the respective other isogene, and this may therefore afford partial compensation for the loss of specific MURF functions. This is supported by the aggravated effect of MURF2/3 double knockdown on myofibril assembly and posttranslational tubulin modification. Interestingly, no significant up-regulation of MURF1 was observed. A mouse deletion model of M-band titin, which also removes the protein kinase domain of titin and the MURF binding site (Weinert et al., 2006), causes early embryonic lethality. Analysis of the protein expression pattern in this model led the authors to conclude that neither nbr1 nor MURF2 were expressed at the onset of growth retardation and subsequent lethality at around E9. This is surprising, as Weinert et al. detected robust expression of the MURF2 message above MURF1 levels as early as E9, as well as expression of the nbr1 and p62/SQSTM1 messages (Weinert et al., 2006). In contrast, EScell derived cardiomyocytes express MURF2, nbr1 and p62/SQSTM1 (Musa et al., 2006). Our consistent results with both commercial and custom-generated antibodies demonstrate that the reported absence of these proteins (Weinert et al., 2006) is likely due to technical limitations by the unspecified antibodies used in that study. We show unequivocally that MURF2 is the dominant embryonic MURFfamily member, and that also nbr1 and p62/SQSTM1 are clearly expressed in the embryonic heart as early as E9.5, in agreement with the detection of their mRNAs. It is therefore possible that MURF2, nbr1 or p62/SQSTM1 might contribute in a hitherto unrecognised way to the lethal phenotype of the titin M-band deletion model of Weinert et al. Sarcomere assembly follows a distinct pattern, where Z-bands are organised first in nascent myofibrils, then M-band components are laid down (reviewed in (Ehler and Gautel, 2008; Sanger et al., 2005)). Our results indicate that MURF2 is localised at the M-bands in vivo only at the completion of their assembly. MURF2 was previously proposed to function as a transient adaptor in myofibrillogenesis in skeletal muscle in vitro (Pizon et al., 2002). Its function as an E3 ubiquitin ligase that has since emerged makes a passive adaptor role in microtubule-mediated sarcomere assembly unlikely. Our observations demonstrate that MURF2 participates in Z-band and M-band organisation and maturation also in cardiac muscle in vivo via its transient association with microtubules. Furthermore, using siRNA knockdown experiments, we show that MURF2 plays a crucial role in regulating the balance between dynamic and static pools of microtubules: siRNA-treated cardiomyocytes lose the pool of more stable glutamylated microtubules required for myofibril assembly (Gundersen et al., 1989) and accumulate dynamic, tyrosinated microtubules. This suggests that other MURF family members can only partly compensate for MURF2 functions on microtubules in cultured cells under short-term conditions, despite the upregulation of MURF3, and the similar role of MURF3 in microtubule posttranslational modification in vitro (Gregorio et al., 2005; Spencer et al., 2000). Partial compensation seems to occur, however, as the striking loss of mature sarcomeric structures by MURF2/MURF3 double knockdown demonstrates. The developmental coincidence of MURF2 p50A expression with intensive myofibrillogenesis, and the disruption of this process at the nascent myofibril stage we report here suggests that a role in myofibril assembly is a major function of this isoform in the developing heart, whereas the developmentally later isoforms are more likely to play regulatory roles in myofibril turnover, transcriptional regulation and metabolic adaptation. While it is important to remember that short-term cell culture effects may not be reflected in knockout animals, where compensatory mechanisms can act over days or weeks, the study of dual MURF knockout mice could be informative on whether the survival of MURF2 knockout animals is due to the compensatory upregulation of MURF1 or -3. Considering that MURF2 knockouts are functionally normal under unstressed situations, the degree of cooperation and redundancy between MURF-family members appears to be variable

between in vivo and in vitro conditions. The microtubule targets of MURF2 are yet unknown, and the change in posttranslationally modified tubulin in MURF2-depleted cells might be due to disrupted selective turnover of any member of the family of tubulin tyrosine ligase-like proteins that in mammals comprises 13 enzymes, of which 9 are polyglutamylases (Janke et al., 2008). The family of tubulin acetylating enzymes is only just emerging (Akella et al., 2010). The highly homologous MID1 E3 ligase (46% homology with MURFs over the first 380 residues) stabilises microtubules and is critical for the ubiquitin-specific modification of microtubule-associated protein phosphatase 2A in non-muscle cells. The analogy to MID1 might lie in the modification of enzyme activities associated with microtubule dynamics, as has been shown for mid1-mediated ubiquitination of PP2A, an enzyme required for stable microtubule maintenance (Gurland and Gundersen, 1993; Merrick et al., 1997). However, the situation for mid1 is more complicated, as MID1 both targets PP2A for ubiquitin-mediated degradation and thus regulated microtubule dynamics, while PP2A also regulates MID1 microtubular transport (Aranda-Orgilles et al., 2008; Berti et al., 2004; Trockenbacher et al., 2001). The targets via which MURFs regulate microtubule posttranslational modifications thus remain a matter of future research. In summary, we show that MURF ubiquitin E3 ligases are tightly developmentally regulated, and that MURF2 is the predominant MURF in differentiating myocardium. MURF2 isoforms undergo developmental switching in cardiac muscle and their expression correlates to that of proteasome- and autophagy-related ligands like p62/SQSTM1 and nbr1. Ablation of MURF2, and more so of MURF2/ MURF3 disrupts microtubule dynamics and myofibril assembly in cultured cardiomyocytes. MURF2 loss is partly compensated by an upregulation of MURF3, but not MURF1. This implies that in vivo compensatory up-regulation may confound the phenotypes of MURF knockout animals. Materials and methods cDNA cloning The predicted mouse MURF2 sequence was retrieved from the NCBI database (D830041C10Rik; accession number XM_355438) and compared against the human protein. The mouse exon–intron boundaries were predicted to be identical to that of the human due to the high sequence similarity. For RT-PCR cloning, a common primer pair to exons 1 and 11 was used that spans all exons of the A-isoforms (Supplementary Fig. 1B): MURF2_FOR 5′-CCGCTCGAGCCACCATGAGCACTTCTCTGAATT ACAAG MURF2_REV 5′-TTTTCCCGGGCACCTTCATTTAGGGAATTCAACC. Isoform-specific PCR primers were designed to analyse expression of individual A isoforms (Supplementary Table 1). Prediction of spliceboundaries, restriction sites and sequence alignments were performed using DNASTAR software. Analysis of MURF transcripts RNA was extracted from freshly isolated tissues of healthy C57BL/6 mice using the RNeasy Mini Kit (Qiagen, UK). cDNA was prepared in a total reaction volume of 20 μl using 2 μg of total RNA, utilising the Super-Script III RT-PCR Kit (Invitrogen, UK). 2 μl of cDNA was amplified in a final 50 μl reaction. The primers used were: MURF1 For 5′-CCGCTCGAGCCACCATGGATTATAAATCTAGCCTG, Rev 5′-TTTTTTGGATCCCCTTGGTGTTCTTCTTTACCCTC; MURF2 For 5′-CCGCTCGAGCCACCATGAGCACTTCTCTGAATTA CAAG, Rev 5′-TTTTCCCGGGCACCTTCATTTAGGGAATTCAACC; p62/SQSTM1 For 5′-GGAGGAGCTCGAGCCATGGCGTTCACGGTGAA, Rev 5′-TATTATTTTTGGATCCTTCAATGGTGGAGGGTGTTCG;

S. Perera et al. / Developmental Biology 351 (2011) 46–61

LC3B For 5′-CGGAGCTTTGAACAAAGAGTG, Rev 5′-GTCCCGAATGTCTCCTGCG; nbr1 For 5′-ATGGAACCACAGGTTACTCTA, Rev 5′-GCAGAAGAACATTGCTCTGG. To perform nested PCR for amplifying p27A, 1 μl of the first PCR reaction was used as a template. All PCR products were sequenced for verification. For analysis of relative transcript levels by semiquantitative RT-PCR, glyceraldehyde-3-phosphate (GAPDH) was used as a loading control, with For 5′-GGCACTGTCAAGGCTGAAAACG and Rev 5′-GGAGATGAGATGATACCACGCTTAG primers. All reactions were done in triplicate and relative expression was quantified by Image-J software (National Institute of Health, USA).

59

Samples were washed 3 × 10 min in PBS and digested for 30 min at room temperature in 1 mg/ml hyaluronidase (Sigma, USA)/PBS to dissolve the cardiac jelly and allow antibodies access to the inner myocardial wall (Tokuyasu and Maher, 1987). Samples were then permeabilised with 0.2% Triton X-100 in PBS for 1 h; followed by blocking in 5% normal horse serum (NHS) in 1% BSA/Gold buffer solution for 1 h. Primary antibody was diluted in 1% BSA/Gold buffer and incubated with samples at 4 °C overnight on a rotating platform. Samples were then washed 10 × 15 min with 0.002% Triton X-100 in PBS at room temperature. Secondary antibody was also diluted in 1% BSA/Gold buffer and incubated at 4 °C overnight in a rotating platform. Samples were mounted as previously described (Messerli et al., 1993).

Western blotting Snap-frozen tissues were homogenised by freeze-slamming, resuspended in Urea-SDS-sample buffer (4 M Urea, 134.6 mM Tris, 5.4% SDS, 2.3% NP-40, 4.45% β-mercaptoethanol, 4% glycerol, 6 mg/100 ml bromophenolblue, pH 6.8) and heated to 95 °C for 2 min. Preparations were run on 10% or 12% polyacrylamide minigels and separated proteins were either immediately visualised by Coomassie blue staining (loading) or blotted onto nitrocellulose Protran BA-83, (Schleicher & Schuell, Germany) overnight in standard transfer buffer at 60 mA in a wet-blot transfer unit (Biorad, UK). Protein transfer was confirmed by Ponceau-S staining. The membrane was blocked by incubating for 1 h in isotonic phosphate buffer supplemented with 5% non-fat milk powder. The membrane was incubated in primary antibody for 2 h at room temperature or overnight at 4 °C, processed with horseradish peroxidase (HRP)-conjugated secondary antibody and enhanced chemoluminescence (ECL) visualised on Hyperfilm ECL (Amersham, USA). Densitometry was performed using ImageJ software (National Institute of Health, USA). Cell culture and immunofluorescence Neonatal rat cardiomyocytes (NRC) isolated from neonatal rat ventricles and maintained as previously described (Lange et al., 2002), were kindly provided by Dr E Ehler. Cells were transfected with 2 μg of plasmid DNA using 3.5 μl Escort III transfection reagent (Sigma, USA), and fixed with 4%-paraformaldehyde (PFA)/PBS after 3 days. NRC used for microtubule staining were washed twice with microtubule protection (MP) buffer (65 mM PIPES, 25 mM HEPES, 10 mM EGTA, 3 mM MgCl2, pH 6.9), and fixed with 4% PFA/MP buffer. Samples were permeabilised with 0.2% Triton X-100 in PBS for 5 min. Primary and secondary antibodies were diluted in Gold buffer (20 mM Tris-base; 155 mM NaCl; 2 mM EGTA; 2 mM MgCl2 pH 7.5) containing 5% BSA. Cells were incubated in primary antibodies for 1 h at room temperature or overnight at 4 °C, then in secondary antibodies for 1 h at room temperature. Specimens were mounted in 0.1 M Tris–HCl pH 9.5-glycerol (3:7) including 50 mg/ml n-propyl-gallate as antifading reagent (Messerli et al., 1993) and viewed on a Zeiss LSM Meta 510 confocal microscope. Transfection of COS-1 cells was performed in a similar manner using the transfection reagent Escort IV (Sigma, USA), and cells were lysed 2 days later with 80–100 μl of SDS-sample buffer for Western blotting. For confocal microscopy competition assay in NRCs and image ratiometry, NRCs were isolated from rats, transfected with GFP coexpressing siRNA constructs, and cultured essentially as described in (Fukuzawa et al., 2008). Ratiometric image analysis was performed as in (Fukuzawa et al., 2008; Pernigo et al., 2010). Whole-mount immunofluorescence Embryonic mouse hearts were dissected in MP buffer and fixed for 1–3 h (depending on developmental stage) in 4% PFA/MP buffer.

siRNA knock-down and expression plasmids Isoform-specific 19-mer siRNA were designed against mouse- and rat-specific sequences (identical for both species) that bridged the unique exon–exon boundaries for MURF2 p50A and p60A, which corresponded to exons 8–11 and exons 9–11 respectively, as obtained from the NCBI database. For p60A knockdown, the primers below were used for siRNA plasmid construction: P60A_FOR 5′ gatccccGCTACCTCTCAGATTGGATttcaagagaATCC AATCTGAGAGGTAGCtttttggaac-3′ P60A_REV 5′-gttccaaaaaGCTACCTCTCAGATTGGATtctcttgaaATCC AATCTGAGAGGTAGCggggatc-3′ were used. For p50A knockdown, P50A_FOR 5′-gatccccCTGGTGACACAGATTGGATttcaagagaATCC AATCTGTGTCACCAGtttttggaac-3′ P50A_REV 5′-gttccaaaaaCTGGTGACACAGATTGGATtctcttgaaATCC AATCTGTGTCACCAGggggatc-3′ For MURF3 knockdown, the following sequences were used: MURF3_FOR 5′ gatccccACAGGAGTCCTCCCGGCCAttcaaga gaTGGCCGGGAGGACTCCTGtttttggaac-3′ MURF3_REV 5′ gttccaaaaaA CAGGAGTCCTCCCGGCCAtctcttgaaTGGCCGGGAGGACTCCTGggggatc-3′ For simultaneous knockdown of both MURFs 2 and 3, a region common to both isogenes was selected from rat genomic sequences (nucleotides 161–179 according to NCBI accession nos. NM_001012218 and NM_001013217 respectively) and the primers below were used: MURF2+3_FOR 5′ gatccccTCTCCAGGCCTCTAACCCttcaagagaGGGT TAGAGGCCTGGAAGAtttttggaac-3′ MURF2+3_REV 5′ gttccaaaaaTCTC CAGGCCTCTAACCCtctcttgaaGGGTTAGAGGCCTGGAAGAggggatc-3′ For controls, the scrambled sequences were used. The H1-GFP vector used is based on pSuper (Brummelkamp et al., 2002) and pEGFP (Clontech; EGFP under the control of the CMV promoter). This ensures that the siRNA is expressed concurrently with GFP, which allows for easy identification of transfected cells. The forward and reverse primers for siRNA were dissolved in 50 μl of water; and self-annealed using 2 μl of each oligonucleotide and 46 μl of annealing buffer (100 mM potassium acetate, 30 mM HEPES–KOH pH 7.4, 2 mM magnesium acetate) to give double-stranded constructs. Construct specificity and knockdown efficiency was tested by transfecting the siRNA into COS-1 cells, along with a plasmid that expressed either GFP-mouse MURF2 p60A or p50A, and assessing the reduction of the specific target protein by Western blotting Supplementary Fig. 3). For expression of these GFP-fusion proteins, fulllength mouse MURF2 cDNA was PCR-amplified as explained previously, and the products were cloned into the pEGFP-C1 vector (Clontech). For expression of MURF2 in rescue experiments, human MURF2 p50A was cloned into a modified pCMV vector introducing a C-terminal c-myc tag. In NRC, the control siRNAs were not associated with higher than 25% immature sarcomeres (Supplementary Fig. 4), which was observed for mock-transfections alone. NRC were transfected with the construct as described (Lange et al., 2002), yielding between 10 and 20% transfected cells, of which 10 to 50

60

S. Perera et al. / Developmental Biology 351 (2011) 46–61

(depending on transfection efficiency) were randomly chosen, imaged and analysed. Antibodies For Western blotting, the commercial goat anti-MURF2-A (Abcam ab4387, 1:500), rabbit anti-HPC (1:2000) and rabbit anti-HP60 (1:2000, (Pizon et al., 2002) antibodies were used. Additionally, rabbit anti-p62/Sequestosome1 (Abcam ab56416, 1:500), rabbit antiLC3 (Cell Signalling #2775, 1:500), mouse monoclonal anti-nbr1 (Abcam ab55474, 1:500), rabbit anti-SRF (Santa-Cruz SC13029, 1:500), goat anti-MURF3 (Santa-Cruz SC50252, 1:500), rabbit antiGAPDH (Abcam ab9485, 1:1000) and rabbit anti-actin (Sigma A2066, 1:1500) antibodies were used. As secondary antibodies, HRPconjugated rabbit anti-goat IgG and goat anti-rabbit IgG from Calbiochem (1:1000); and HRP-conjugated rabbit anti mouse IgG (Dako, 1:1000) were used. For immunofluorescence, MURF2 was visualised with the goat anti-MURF2-A (1:100), rabbit anti-HPC (1:100), and rabbit anti-HP60 (1:200) antibodies. Additionally, rabbit anti-sarcomeric alpha-actinin, rat anti-all tubulin (Abcam, clone YOL1/34, 1:100), mouse antiglutamylated alpha-tubulin (Synaptic Systems, clone 1D5, 1:200), rat anti-tyrosinated tubulin (Abcam, clone YL1/2, 1:200), mouse antiacetylated tubulin (Abcam ab24610, clone 6-11B-1, 1:100) and mouse anti-multi ubiquitin (Stressgen SPA-205E, clone FK2, 1:100) were used. The mouse anti-all MyHC A4.1025 and anti-slow MyHC A4.840 antibodies were a gift from Dr. S Hughes, while the mouse antimyomesin B4 antibody was a gift of Dr. E Ehler (Grove et al., 1984). For triple staining, combinations Alexa488, Cy3 and Cy5-conjugated secondary antibodies were used. Cy3 anti-goat IgG, Cy3 anti-mouse IgG, Cy5 anti-mouse IgG, Cy5 anti-rabbit IgG, Cy3 anti-rabbit IgG, and Cy5-anti rat IgG were from Stratech Scientific, USA and the subclass specific FITC-anti mouse IgM (for A4.840) was from Sigma, UK. Supplementary materials related to this article can be found online at doi:10.1016/j.ydbio.2010.12.024. Acknowledgments We thank Birgit Brandmeier for excellent technical assistance, and Elisabeth Ehler for NRC preparations. This work was supported by the EU MYORES network, and the Medical Research Council of Great Britain. MG holds the BHF Chair of Molecular Cardiology. References Agarkova, I., Auerbach, D., Ehler, E., Perriard, J.C., 2000. A novel marker for vertebrate embryonic heart, the EH-myomesin isoform. J. Biol. Chem. 275, 10256–10264. Ahuja, P., Perriard, E., Perriard, J.-C., Ehler, E., 2004. Sequential myofibrillar breakdown accompanies mitotic division of mammalian cardiomyocytes. J. Cell Sci. 117, 3295–3306. Akella, J.S., Wloga, D., Kim, J., Starostina, N.G., Lyons-Abbott, S., Morrissette, N.S., Dougan, S.T., Kipreos, E.T., Gaertig, J., 2010. MEC-17 is an alpha-tubulin acetyltransferase. Nature 467, 218–222. Aranda-Orgilles, B., Aigner, J., Kunath, M., Lurz, R., Schneider, R., Schweiger, S., 2008. Active transport of the ubiquitin ligase MID1 along the microtubules is regulated by protein phosphatase 2A. PLoS ONE 3, e3507. Arndt, V., Dick, N., Tawo, R., Dreiseidler, M., Wenzel, D., Hesse, M., Furst, D.O., Saftig, P., Saint, R., Fleischmann, B.K., Hoch, M., Hohfeld, J., 2010. Chaperone-assisted selective autophagy is essential for muscle maintenance. Curr. Biol. 20, 143–148. Arya, R., Kedar, V., Hwang, J.R., McDonough, H., Li, H.-H., Taylor, J., Patterson, C., 2004. Muscle ring finger protein-1 inhibits PKC{epsilon} activation and prevents cardiomyocyte hypertrophy. J. Cell Biol. 167, 1147–1159. Auerbach, D., Bantle, S., Keller, S., Hinderling, V., Leu, M., Ehler, E., Perriard, J.C., 1999. Different domains of the M-band protein myomesin are involved in myosin binding and M-band targeting. Mol. Biol. Cell 10, 1297–1308. Beckmann, J.S., Spencer, M., 2008. Calpain 3, the “gatekeeper” of proper sarcomere assembly, turnover and maintenance. Neuromuscul. Disord. 18, 913–921. Belaguli, N.S., Zhou, W., Trinh, T.H., Majesky, M.W., Schwartz, R.J., 1999. Dominant negative murine serum response factor: alternative splicing within the activation domain inhibits transactivation of serum response factor binding targets. Mol. Cell. Biol. 19, 4582–4591.

Berti, C., Fontanella, B., Ferrentino, R., Meroni, G., 2004. Mig12, a novel Opitz syndrome gene product partner, is expressed in the embryonic ventral midline and cooperates with Mid1 to bundle and stabilize microtubules. BMC Cell Biol. 5, 9. Bjorkoy, G., Lamark, T., Brech, A., Outzen, H., Perander, M., Overvatn, A., Stenmark, H., Johansen, T., 2005. p62/SQSTM1 forms protein aggregates degraded by autophagy and has a protective effect on huntingtin-induced cell death. J. Cell Biol. 171, 603–614. Bodine, S.C., Latres, E., Baumhueter, S., Lai, V.K., Nunez, L., Clarke, B.A., Poueymirou, W.T., Panaro, F.J., Na, E., Dharmarajan, K., Pan, Z.Q., Valenzuela, D.M., DeChiara, T.M., Stitt, T.N., Yancopoulos, G.D., Glass, D.J., 2001. Identification of ubiquitin ligases required for skeletal muscle atrophy. Science 294, 1704–1708. Brummelkamp, T.R., Bernards, R., Agami, R., 2002. A system for stable expression of short interfering RNAs in mammalian cells. Science 296, 550–553. Cecconi, F., Levine, B., 2008. The role of autophagy in mammalian development: cell makeover rather than cell death. Dev. Cell 15, 344–357. Centner, T., Yano, J., Kimura, E., McElhinny, A.S., Pelin, K., Witt, C.C., Bang, M.L., Trombitas, K., Granzier, H., Gregorio, C.C., Sorimachi, H., Labeit, S., 2001. Identification of Muscle Specific Ring Finger Proteins as Potential Regulators of the Titin Kinase Domain. J. Mol. Biol. 306, 717–726. Dai, K.S., Liew, C.C., 2001. A novel human striated muscle RING zinc finger protein, SMRZ, interacts with SMT3b via its RING domain. J. Biol. Chem. 276, 23992–23999. Ehler, E., Gautel, M., 2008. The sarcomere and sarcomerogenesis. Adv. Exp. Med. Biol. 642, 1–14. Fielitz, J., Kim, M.S., Shelton, J.M., Latif, S., Spencer, J.A., Glass, D.J., Richardson, J.A., BasselDuby, R., Olson, E.N., 2007a. Myosin accumulation and striated muscle myopathy result from the loss of muscle RING finger 1 and 3. J. Clin. Invest. 117, 2486–2495. Fielitz, J., van Rooij, E., Spencer, J.A., Shelton, J.M., Latif, S., van der Nagel, R., Bezprozvannaya, S., de Windt, L., Richardson, J.A., Bassel-Duby, R., Olson, E.N., 2007b. Loss of muscle-specific RING-finger 3 predisposes the heart to cardiac rupture after myocardial infarction. Proc. Natl Acad. Sci. USA 104, 4377–4382. Fukuzawa, A., Lange, S., Holt, M.R., Vihola, A., Carmignac, V., Ferreiro, A., Udd, A.B., Gautel, M., 2008. Interactions with titin and myomesin target obscurin and its small homologue, obscurin-like 1, to the sarcomeric M-band: implications for hereditary myopathies. J. Cell Sci. 121, 1841–1851. Gautel, M., 2008. The sarcomere and the nucleus: functional links to hypertrophy, atrophy and sarcopenia. Adv. Med. Biol. Exp. 642, 176–191. Geng, J., Klionsky, D.J., 2008. The Atg8 and Atg12 ubiquitin-like conjugation systems in macroautophagy. ‘Protein modifications: beyond the usual suspects’ review series. EMBO Rep. 9, 859–864. Gill, G., 2004. SUMO and ubiquitin in the nucleus: different functions, similar mechanisms? Genes Dev. 18, 2046–2059. Glass, D.J., 2003. Signalling pathways that mediate skeletal muscle hypertrophy and atrophy. Nat. Cell Biol. 5, 87–90. Gregorio, C.C., Perry, C.N., McElhinny, A.S., 2005. Functional properties of the titin/ connectin-associated proteins, the muscle-specific RING finger proteins (MURFs), in striated muscle. J. Muscle Res. Cell Motil. 26, 389–400. Grove, B.K., Kurer, V., Lehner, C., Doetschman, T.C., Perriard, J.C., Eppenberger, H.M., 1984. Monoclonal antibodies detect new 185, 000 dalton muscle M-line protein. J. Cell Biol. 98, 518–524. Gundersen, G., Khawaja, S., Bulinski, J.C., 1989. Generation of a stable, posttranslationally modified microtubule array is an early event in myogenic differentiation. J. Cell Biol. 109, 2275–2288. Gurland, G., Gundersen, G.G., 1993. Protein phosphatase inhibitors induce the selective breakdown of stable microtubules in fibroblasts and epithelial cells. Proc. Natl Acad. Sci. USA 90, 8827–8831. Hirner, S., Krohne, C., Schuster, A., Hoffmann, S., Witt, S., Erber, R., Sticht, C., Gasch, A., Labeit, S., Labeit, D., 2008. MuRF1-dependent regulation of systemic carbohydrate metabolism as revealed from transgenic mouse studies. J. Mol. Biol. 379, 666–677. Ichimura, Y., Kominami, E., Tanaka, K., Komatsu, M., 2008. Selective turnover of p62/ A170/SQSTM1 by autophagy. Autophagy 4, 1063–1066. Iwaki, K., Sukhatme, V.P., Shubeita, H.E., Chien, K.R., 1990. Alpha- and beta-adrenergic stimulation induces distinct patterns of immediate early gene expression in neonatal rat myocardial cells. fos/jun expression is associated with sarcomere assembly; Egr-1 induction is primarily an alpha 1-mediated response. J. Biol. Chem. 265, 13809–13817. Jackson, P.K., Eldridge, A.G., Freed, E., Furstenthal, L., Hsu, J.Y., Kaiser, B.K., Reimann, J.D., 2000. The lore of the RINGs: substrate recognition and catalysis by ubiquitin ligases. Trends Cell Biol. 10, 429–439. Janke, C., Rogowski, K., van Dijk, J., 2008. Polyglutamylation: a fine-regulator of protein function? Protein Modifications: beyond the usual suspects' review series. EMBO Rep. 9, 636–641. Kedar, V., McDonough, H., Arya, R., Li, H.-H., Rockman, H.A., Patterson, C., 2004. Musclespecific RING finger 1 is a bona fide ubiquitin ligase that degrades cardiac troponin I. PNAS 0404341102. Kirkin, V., Lamark, T., Sou, Y.S., Bjorkoy, G., Nunn, J.L., Bruun, J.A., Shvets, E., McEwan, D.G., Clausen, T.H., Wild, P., Bilusic, I., Theurillat, J.P., Overvatn, A., Ishii, T., Elazar, Z., Komatsu, M., Dikic, I., Johansen, T., 2009. A role for NBR1 in autophagosomal degradation of ubiquitinated substrates. Mol. Cell 33, 505–516. Komatsu, M., Waguri, S., Koike, M., Sou, Y.S., Ueno, T., Hara, T., Mizushima, N., Iwata, J., Ezaki, J., Murata, S., Hamazaki, J., Nishito, Y., Iemura, S., Natsume, T., Yanagawa, T., Uwayama, J., Warabi, E., Yoshida, H., Ishii, T., Kobayashi, A., Yamamoto, M., Yue, Z., Uchiyama, Y., Kominami, E., Tanaka, K., 2007. Homeostatic levels of p62 control cytoplasmic inclusion body formation in autophagy-deficient mice. Cell 131, 1149–1163. Koyama, S., Hata, S., Witt, C.C., Ono, Y., Lerche, S., Ojima, K., Chiba, T., Doi, N., Kitamura, F., Tanaka, K., Abe, K., Witt, S.H., Rybin, V., Gasch, A., Franz, T., Labeit, S., Sorimachi,

S. Perera et al. / Developmental Biology 351 (2011) 46–61 H., 2008. Muscle RING-finger protein-1 (MuRF1) as a connector of muscle energy metabolism and protein synthesis. J. Mol. Biol. 376, 1224–1236. Lange, S., Auerbach, D., McLoughlin, P., Perriard, E., Schafer, B.W., Perriard, J.C., Ehler, E., 2002. Subcellular targeting of metabolic enzymes to titin in heart muscle may be mediated by DRAL/FHL-2. J. Cell Sci. 115, 4925–4936. Lange, S., Xiang, F., Yakovenko, A., Vihola, A., Hackman, P., Rostkova, E., Kristensen, J., Brandmeier, B., Franzen, G., Hedberg, B., Gunnarsson, L.G., Hughes, S.M., Marchand, S., Sejersen, T., Richard, I., Edstrom, L., Ehler, E., Udd, B., Gautel, M., 2005. The kinase domain of titin controls muscle gene expression and protein turnover. Science 308, 1599–1603 Epub ahead of print March 31. Leu, M., Ehler, E., Perriard, J.C., 2001. Characterisation of postnatal growth of the murine heart. Anat. Embryol. (Berl) 204, 217–224. Masiero, E., Agatea, L., Mammucari, C., Blaauw, B., Loro, E., Komatsu, M., Metzger, D., Reggiani, C., Schiaffino, S., Sandri, M., 2009. Autophagy is required to maintain muscle mass. Cell Metab. 10, 507–515. Masiero, E., Sandri, M., 2010. Autophagy inhibition induces atrophy and myopathy in adult skeletal muscles. Autophagy 6, 307–309. McElhinny, A.S., Perry, C.N., Witt, C.C., Labeit, S., Gregorio, C.C., 2004. Muscle-specific RING finger-2 (MURF-2) is important for microtubule, intermediate filament and sarcomeric M-line maintenance in striated muscle development. J. Cell Sci. 117, 3175–3188. McElhinny, M.S., Kakinuma, K., Sorimachi, H., Labeit, S., Gregorio, C.C., 2002. Musclespecific RING finger-1 interacts with titin to regulate sarcomeric M-line and thick filament structure and may have nuclear functions via its interaction with glucocorticoid modulatory element binding protein-1. J. Cell Biol. 157, 125–136. Merrick, S.E., Trojanowski, J.Q., Lee, V.M., 1997. Selective destruction of stable microtubules and axons by inhibitors of protein serine/threonine phosphatases in cultured human neurons. J. Neurosci. 17, 5726–5737. Messerli, J.M., Eppenberger-Eberhardt, M.E., Rutishauser, B.M., Schwarb, P., von Arx, P., Koch-Schneidemann, S., Eppenberger, H.M., Perriard, J.C., 1993. Remodelling of cardiomyocyte cytoarchitecture visualized by three-dimensional (3D) confocal microscopy. Histochemistry 100, 193–202. Miano, J.M., Ramanan, N., Georger, M.A., de Mesy Bentley, K.L., Emerson, R.L., Balza Jr., R.O., Xiao, Q., Weiler, H., Ginty, D.D., Misra, R.P., 2004. Restricted inactivation of serum response factor to the cardiovascular system. Proc. Natl Acad. Sci. USA 101, 17132–17137. Moscat, J., Diaz-Meco, M.T., 2009. p62 at the crossroads of autophagy, apoptosis, and cancer. Cell 137, 1001–1004. Musa, H., Meek, S., Gautel, M., Peddie, D., Smith, A.J.H., Peckham, M., 2006. Targeted homozygous deletion of M-band titin in cardiomyocytes prevents sarcomere formation. J. Cell Sci. 119, 4322–4331. Niu, Z., Li, A., Zhang, S.X., Schwartz, R.J., 2007. Serum response factor micromanaging cardiogenesis. Curr. Opin. Cell Biol. 19, 618–627. Noda, N.N., Ohsumi, Y., Inagaki, F., 2010. Atg8-family interacting motif crucial for selective autophagy. FEBS Lett. 584, 1379–1385. Pankiv, S., Clausen, T.H., Lamark, T., Brech, A., Bruun, J.-A., Outzen, H., Overvatn, A., Bjorkoy, G., Johansen, T., 2007. p62/SQSTM1 binds directly to Atg8/LC3 to facilitate degradation of ubiquitinated protein aggregates by autophagy. J. Biol. Chem. 282, 24131–24145. Penaloza, C., Lin, L., Lockshin, R.A., Zakeri, Z., 2006. Cell death in development: shaping the embryo. Histochem. Cell Biol. 126, 149–158. Pernigo, S., Fukuzawa, A., Bertz, M., Holt, M., Rief, M., Steiner, R.A., Gautel, M., 2010. Structural insight into M-band assembly and mechanics from the titin-obscurinlike-1 complex. Proc. Natl. Acad. Sci. USA 107, 2908–2913.

61

Pizon, V., Iakovenko, A., Van der Ven, P.F.M., Kelly, R.A., Fatu, C., Fürst, D.O., Karsenti, E., Gautel, M., 2002. Transient association of titin and myosin with microtubules in nascent myofibrils directed by the MURF2 RING-finger protein. J. Cell Sci. 115, 4469–4482. Salmena, L., Pandolfi, P.P., 2007. Changing venues for tumour suppression: balancing destruction and localization by monoubiquitylation. Nat. Rev. Cancer 7, 409–413. Sandri, M., 2008. Signaling in muscle atrophy and hypertrophy. Physiology (Bethesda) 23, 160–170. Sanger, J., Kang, S., Siebrands, C., Freeman, N., Du, A., Wang, J., Stout, A., Sanger, J., 2005. How to build a myofibril. J. Muscle Res. Cell Motil. 26, 343–354. Schiaffino, S., Mammucari, C., Sandri, M., 2008. The role of autophagy in neonatal tissues: just a response to amino acid starvation? Autophagy 4, 727–730. Schiaffino, S., Reggiani, C., 1996. Molecular diversity of myofibrillar proteins: gene regulation and functional significance. Physiol. Rev. 76, 371–423. Spencer, J.A., Eliazer, S., Ilaria Jr., R.L., Richardson, J.A., Olson, E.N., 2000. Regulation of microtubule dynamics and myogenic differentiation by MURF, a striated muscle RING-finger protein. J. Cell Biol. 150, 771–784. Tan, J.M., Wong, E.S., Dawson, V.L., Dawson, T.M., Lim, K.L., 2007. Lysine 63-linked polyubiquitin potentially partners with p62 to promote the clearance of protein inclusions by autophagy. Autophagy 4. Taneike, M., Yamaguchi, O., Nakai, A., Hikoso, S., Takeda, T., Mizote, I., Oka, T., Tamai, T., Oyabu, J., Murakawa, T., Nishida, K., Shimizu, T., Hori, M., Komuro, I., Shirasawa, T., Mizushima, N., Otsu, K., 2010. Inhibition of autophagy in the heart induces agerelated cardiomyopathy. Autophagy 6. Tokuyasu, K.T., Maher, P.A., 1987. Immunocytochemical studies of cardiac myofibrillogenesis in early chick embryos.I. Presence of immunofluorescent titin spots in premyofibril stages. J. Cell Biol. 105, 2781–2793. Trockenbacher, A., Suckow, V., Foerster, J., Winter, J., Krauss, S., Ropers, H.H., Schneider, R., Schweiger, S., 2001. MID1, mutated in Opitz syndrome, encodes an ubiquitin ligase that targets phosphatase 2A for degradation. Nat. Genet. 29, 287–294. Tsukamoto, S., Kuma, A., Murakami, M., Kishi, C., Yamamoto, A., Mizushima, N., 2008. Autophagy is essential for preimplantation development of mouse embryos. Science 321, 117–120. von Mikecz, A., 2006. The nuclear ubiquitin–proteasome system. J. Cell Sci. 119, 1977–1984. Waters, S., Marchbank, K., Solomon, E., Whitehouse, C., Gautel, M., 2009. Interactions with LC3 and polyubiquitin chains link nbr1 to autophagic protein turnover. FEBS Lett. 583, 1846–1852. Weinert, S., Bergmann, N., Luo, X., Erdmann, B., Gotthardt, M., 2006. M line-deficient titin causes cardiac lethality through impaired maturation of the sarcomere. J. Cell Biol. 173, 559–570. Willis, M.S., Ike, C., Li, L., Wang, D.Z., Glass, D.J., Patterson, C., 2007. Muscle ring finger 1, but not muscle ring finger 2, regulates cardiac hypertrophy in vivo. Circ. Res. 100, 456–459. Willis, M.S., Schisler, J.C., Portbury, A.L., Patterson, C., 2009. Build it up-Tear it down: protein quality control in the cardiac sarcomere. Cardiovasc. Res. 81, 439–448. Witt, C.C., Witt, S.H., Lerche, S., Labeit, D., Back, W., Labeit, S., 2008. Cooperative control of striated muscle mass and metabolism by MuRF1 and MuRF2. EMBO J. 27, 350–360. Yue, Z., Jin, S., Yang, C., Levine, A.J., Heintz, N., 2003. Beclin 1, an autophagy gene essential for early embryonic development, is a haploinsufficient tumor suppressor. Proc. Natl Acad. Sci. USA 100, 15077–15082.