Repair of Mybpc3 mRNA by 5'-trans-splicing in a Mouse Model of Hypertrophic Cardiomyopathy.

RNA trans-splicing has been explored as a therapeutic option for a variety of genetic diseases, but not for cardiac genetic disease. Hypertrophic cardiomyopathy (HCM) is an autosomal-dominant disease, characterized by left ventricular hypertrophy (LVH) and diastolic dysfunction. MYBPC3, encoding cardiac myosin-binding protein C (cMyBP-C) is frequently mutated. We evaluated the 5'-trans-splicing strategy in a mouse model of HCM carrying a Mybpc3 mutation. 5'-trans-splicing was induced between two independently transcribed molecules, the mutant endogenous Mypbc3 pre-mRNA and an engineered pre-trans-splicing molecule (PTM) carrying a FLAG-tagged wild-type (WT) Mybpc3 cDNA sequence. PTMs were packaged into adeno-associated virus (AAV) for transduction of cultured cardiac myocytes and the heart in vivo. Full-length repaired Mybpc3 mRNA represented up to 66% of total Mybpc3 transcripts in cardiac myocytes and 0.14% in the heart. Repaired cMyBP-C protein was detected by immunoprecipitation in cells and in vivo and exhibited correct incorporation into the sarcomere in cardiac myocytes. This study provides (i) the first evidence of successful 5'-trans-splicing in vivo and (ii) proof-of-concept of mRNA repair in the most prevalent cardiac genetic disease. Since current therapeutic options for HCM only alleviate symptoms, these findings open new horizons for causal therapy of the severe forms of the disease.Molecular Therapy-Nucleic Acids (2013) 2, e102; doi:10.1038/mtna.2013.31; published online 2 July 2013.

Introduction

In the last decade repair of mRNA by spliceosome-mediated RNA trans-splicing has raised interests as a novel therapeutic intervention (for reviews, see refs. [1,2]). Trans-splicing has many attractive features such as preservation of the endogenous regulation, replacement of selected portions of the target gene and, most importantly, corrections of dominant-negative mutations.[3,4] The 5′-, 3′- or even internal exons of a target pre-mRNA can be replaced by trans-splicing using engineered pre-trans-splicing molecules (PTMs). PTMs carry the wild-type (WT) sequence, a binding domain complementary to the endogenous target and an appropriate set of splicing elements. After nuclear import, PTMs are transcribed and can specifically hybridize the target mutant pre-mRNA via their binding domain, giving rise to a repaired mRNA molecule (). As a positive side effect of trans-splicing, cis-splicing should be reduced due to competition for access to the RNA splicing machinery. So far, successful trans-splicing between PTMs and endogenous targets has been described for different genetic diseases such as hemophilia A,[5] cystic fibrosis,[6] spinal muscular atrophy,[7,8] hyper-IgM-X-linked immunodeficiency,[9] frontotemporal dementia with Parkinsonism linked to chromosome 17,[10,11] epidermolysis bullosa with muscular dystrophy,[12] and Huntington's disease,[13] most of them being 3′-trans-splicing approaches.

To the best of our knowledge, no study has provided evidence for successful 5′-trans-splicing in vivo yet and this promising strategy has not been evaluated for cardiac genetic diseases. The aim of the present study was therefore to investigate this approach in hypertrophic cardiomyopathy (HCM). HCM is a myocardial disease mainly characterized by left ventricular hypertrophy (LVH) and diastolic dysfunction.[14,15,16] The clinical outcome of HCM is highly variable and ranges from an asymptomatic benign course to heart failure, atrial fibrillation and sudden cardiac death caused by arrhythmias.[14,15] HCM is a genetic disease transmitted as an autosomal-dominant trait and caused by mutations in genes encoding sarcomeric proteins.[17] Among them, mutations in MYBPC3 encoding cardiac myosin-binding protein C (cMyBP-C) are the most frequent ones.[18,19] cMyBP-C is a component of the thick filaments of the sarcomere, and plays important structural and functional roles.[18,20,21]

In the present study the feasibility of 5′-trans-splicing to repair HCM-mutant mRNA was assessed in isolated cardiac myocytes and in vivo in Mybpc3-targeted knock-in (KI) mice that have been generated previously.[22] KI mice carry a G>A transition on the last nucleotide of exon 6, which is associated with a severe phenotype and a poor prognosis in humans[19] and occurs in 13% of all HCM patients in Toscany.[23] KI mice exhibit LVH with systolic and diastolic dysfunction.[22,24]

Results

Design of PTMs

The G>A transition mutation leads to three different mutant mRNAs in homozygous KI mice (Supplementary Figure S1). Mutant-1 contains the mutation (missense), whereas mutant-2 (nonsense) and mutant-3 (deletion/insertion) are due to skipping of exon 6.

We generated different PTMs encoding exons 1–6 of WT Mybpc3 under the control of a ubiquitous (cytomegalovirus) or cardiac myocyte-specific (TNNT2, human cardiac troponin T) promoter (). To specifically detect repaired Mybpc3 mRNA and protein, an N-terminal FLAG-tag was introduced in the coding sequence. In addition, an intron was inserted right after the promoter to increase mRNA stability and enhance expression of the constructs.[25,26] The splicing domain included a canonical 5′ splice donor site sequence followed by a downstream intronic sequence enhancer element, which has been shown to markedly increase the trans-splicing efficiency.[27] Importantly, the binding domain of the PTM is an essential part because it confers specificity to the target pre-mRNA.[28] To evaluate the feasibility and efficacy of 5′-trans-splicing, we designed several constructs differing only with respect to the length of the binding domain and to the target site in Mybpc3 intron 6 (Supplementary Table S1). The binding domains are complementary to this intron, but leave out its 3′ splicing elements. Moreover, to maintain the PTM in the nucleus and reduce its translation we deleted the SV40 polyadenylation (polyA) signal, which is known to contribute to mRNA stability and nuclear export ().[29] As negative controls we designed PTMs with reversed binding domains (PTM-R), which should not induce 5′-trans-splicing events. PTM-driven 5′-trans-splicing on the endogenous KI Mybpc3 pre-mRNA target should produce a full-length repaired Mybpc3 mRNA, in which the mutation is bypassed, resulting in a FLAG-tagged WT repaired cMyBP-C protein, and simultaneously cis-splicing should be reduced (,).

Evidence for repair of Mybpc3 mRNA by 5′-trans-splicing in vitro

To allow gene transfer in neonatal mouse cardiac myocytes (NMCMs), PTMs were packaged into self-complementary adeno-associated virus serotype 6 (AAV6), a serotype known to efficiently transduce cardiac myocytes in culture.[30,31] NMCMs were isolated from KI mice and transduced with AAV6-PTMs either with or without polyA signal (ΔpA) or AAV6-green fluorescent protein (GFP) as a control. After 4 days of transduction (multiplicity of infection (MOI): 3,000) about 80% of cells expressed GFP (Supplementary Figure S2). Using PCR primers that specifically amplify the repaired Mybpc3 mRNA (Supplementary Figure S1 and Supplementary Table S2), we obtained a specific signal in AAV6-PTM- and AAV6-PTMΔpA-transduced NMCMs, but not in untransduced or PTM-R-transduced NMCMs (). The absence of 5′-trans-splicing in AAV6-PTM-R-transduced NMCM excluded the possibility that recombination occurred between the highly homologous sequences of PTMs and endogenous Mybpc3. The amount of repaired Mybpc3 was higher in the absence than in the presence of the polyA signal in the PTM. To evaluate whether cis-splicing was reduced, we used Mybpc3 primers binding in exons 1 and 9, which amplify total (repaired plus mutant) Mybpc3 mRNA. Although no major difference was detected between samples, reduced signals were observed for certain Mybpc3 mRNA species in AAV6-PTM- and in AAV6-PTMΔpA-transduced NMCMs. This suggests a reduction in Mybpc3 cis-splicing when 5′-trans-splicing occurred (). Sequencing of repaired Mybpc3 mRNA amplicons confirmed the presence of the WT guanine (G) at the exon 6–exon 7 junction (). Conversely, sequencing of the upper 896-bp band of total Mybpc3 mRNA in AAV6-PTMΔpA- and AAV6-PTM-R-transduced NMCMs showed the presence of the mutant adenine (A) at the same position (). To estimate the amount of repaired Mybpc3 mRNA, we performed two rounds of PCR to amplify either total or only repaired Mybpc3 mRNA (). Comparison of amplicon intensities revealed that up to 33% of total Mybpc3 transcripts were repaired. To evaluate whether the efficiency of 5′-trans-splicing can be improved by increasing the dose of virus, we generated bicistronic recombinant adenovirus (AdV) encoding the PTMΔpA and GFP both under the control of the TNNT2 promoter. KI NMCMs were transduced with different MOI of AdV-PTMΔpA and analyzed 7 days after. Repaired Mybpc3 mRNA was detected in all transduced samples and its amount increased with increasing MOI (). The pattern of total Mybpc3 mRNA did not reveal major difference from one MOI to another, except at a MOI of 100 at which the intensity of the mutant-3 and mutant-2 mRNAs was lower than in untransduced cardiac myocytes (). Fluorescence analysis of AdV-GFP transduced cardiomyocytes confirmed a complete transduction with a MOI of 100 (Supplementary Figure S2). We further determined the efficiency of 5′-trans-splicing in several samples with AdV-PTMΔpA at a MOI of 100, and estimated by semi-quantitative analysis that 51 ± 7% of total Mybpc3 mRNA was repaired ().

We then investigated whether the repaired Mybpc3 mRNA is translated into protein and whether the repaired cMyBP-C is properly incorporated into the sarcomere. The presence of the FLAG-tag allowed specific detection of repaired cMyBP-C. Whereas repaired cMyBP-C was not detected by standard western blot with the anti-FLAG antibody, it was detected at the correct molecular weight after FLAG-immunoprecipitation (,), confirming that 5′-trans-splicing occurred in cardiac myocytes. FLAG-immunoprecipitation of AAV6-PTM-R-transduced NMCMs did not show any band at 150 kDa, while FLAG-Mybpc3 transfected HEK293 cells, used as a positive control, did show it (). On the other hand, we detected a major FLAG-positive band around 35 kDa in AAV6-PTM- and AAV6-PTM-R-transduced NMCMs, which corresponds to the translated PTM transcripts (). This band was barely detected in AAV6-PTMΔpA-transduced NMCMs, supporting the view that the absence of the polyA signal prevented translation and putative accumulation of toxic PTM proteins in cells. Endogenous and/or repaired cMyBP-C, but not translated PTMs were stained with a specific cMyBP-C antibody, which recognizes the MyBP-C motif ( and ). To investigate whether the repaired cMyBP-C was incorporated into the sarcomere, we performed immunofluorescence analysis of transduced cardiac myocytes. About 9% of cMyBP-C-positive cells (= cardiac myocytes) were co-stained with the anti-FLAG antibody, and the repaired cMyBP-C showed the expected doublets in the A-band of the sarcomeres, indicating correct incorporation (). In contrast, the 35-kDa FLAG-PTM-R proteins showed a cellular and nuclear diffuse pattern without colocalization with endogenous cMyBP-C ().

Evidence for repair of Mybpc3 mRNA by 5′-trans-splicing in vivo

We next assessed the feasibility of PTM-driven 5′-trans-splicing in KI mice in vivo. The PTMΔpA and Renilla luciferase (RLuc) were inserted in the pdsAAV transfer vector under the control of the TNNT2 promoter and were packaged in AAV serotype 9 (AAV9), which has proven efficient cardiac transduction in mice in vivo.[32] AAV9 (mean dose 5.2 × 1012 vg/kg of body weight (BW)) was administered systemically into 7-week-old animals. Echocardiographic analysis performed during one month after injection did not display major differences in cardiac function between mice that received either AAV9 or NaCl (Supplementary Table S3). After 28 days, luciferase expression was evaluated by in vivo bioluminescence imaging and luminescence was recorded only in the heart of the AAV9-RLuc–injected mouse (Supplementary Figure S3). Accordingly, luciferase mRNA level was high in the heart and very low in the liver of the mouse that received AAV9-RLuc (), validating efficient and preferential cardiac transduction with AAV9. Importantly, the repaired Mybpc3 mRNA was detected in the heart of the mouse that received AAV9-PTMΔpA, but not in the others (). No effect on cis-splicing was discernible (). Semi-quantitative analysis showed that 0.05% of total Mybpc3 mRNA was repaired ().

To augment the dose of virus and thus the 5′-trans-splicing efficiency, we performed experiments in neonates (). Longitudinal echo analysis in neonatal mice revealed that KI mice developed first systolic dysfunction, as shown by lower fractional area shortening than WT mice at day 2, followed by LVH, as shown by higher left-ventricular-mass-to-BW than WT mice at day 3 (). We then systemically administered AAV9-PTMΔpA into 1-day-old KI mice (3.4 × 1014 vg/kg BW). This dose of AAV9 resulted in an almost complete transduction of cardiac tissue at postnatal day 7 (Supplementary Figure S4). Although the dose per BW was ~65-fold higher than in the adult mouse, no beneficial effect on left ventricular mass/BW and on fractional area shortening were observed at day 4 and 7 () as well as 7 weeks after injection (Supplementary Table S3). Despite the absence of rescue, we evaluated the 5′-trans-splicing efficiency in one mouse 7 weeks after injection. The full-length repaired Mybpc3 mRNA was detected by reverse transcription-PCR only in the heart of the AAV9-PTMΔpA-injected mouse and represented 0.14% of total Mybpc3 transcripts (, and Supplementary Figure S5), which thus showed 2.8-fold higher 5′-trans-splicing event in the newborn than in the adult mouse. In addition, the repaired cMyBP-C was detected, although faintly after FLAG-immunoprecipitation ().

Discussion

RNA trans-splicing as a potential therapeutic technology has been applied to several diseases both in cell systems and in mouse models (for reviews, see refs. [1,2]). The present study provides the first evidence of successful 5′-trans-splicing both in cardiac myocytes and in the heart in vivo for the most prevalent cardiac genetic disease.

The percentage of total Mybpc3 mRNA that was repaired was estimated to be between 33 and 66% in transduced KI NMCMs. This is much higher than what has been reported in previous studies using endogenous targets.[5,33] However, despite the high efficiency of Mybpc3 5′-trans-splicing at the mRNA level, the amount of repaired cMyBP-C protein was rather low. This suggests a low efficiency of translation and underlines that mRNA copy number and protein levels do not need to be correlated.[34] On the other hand, the amount of total Mybpc3 mRNAs is 80% and the level of cMyBP-C protein 90% lower in KI than in WT mice.[22] Therefore, when 33% of total Mybpc3 mRNA is repaired, it represents less than 7% of the Mybpc3 mRNA amount and less than 4% of the cMyBP-C protein amount found in WT mice. This may well be under the limit of detection by western blot.

Our study provides additional evidence for removing the polyA signal in the PTM construct to prevent translation and therefore accumulation of PTM proteins that could exert a dominant-negative effect on the structure and/or function of cardiac myocytes.

Recently, in vivo 3′-trans-splicing has been shown to improve the phenotype of a mouse model of spinal muscular atrophy.[8] The present study provides the first evidence for successful mRNA repair by 5′-trans-splicing in vivo. The combination of AAV9 and TNNT2 promoter allowed efficient cardiac transduction in vivo. Although this resulted in detectable levels of repaired cMyBP-C, the amount was still too low to ameliorate the cardiac phenotype. Thus, further optimization of the technique is needed to increase the amount of “therapeutic” protein.

Among strategies that aim at specifically targeting mutant mRNA in dominant genetic disease, such as mutant-specific RNA interference,[35] trans-splicing has potential advantages for the therapy of HCM. First, it allows the repair of even complex consequences on RNA splicing that, as exemplified in the present HCM mouse model,[22] can result from a single point mutation and will be difficult/impossible to target with siRNA without affecting WT mRNA. Second, and in contrast to RNA interference therapies targeting a specific mutation, two different PTMs would be enough to treat the 40–60% of HCM patients who carry a MYBPC3 mutation[18,19,36,37,38]—one targeting the 5′ mutations and the other the 3′ mutations. Therefore, trans-splicing represents a promising, potentially causal therapy of severe forms of HCM.

Materials and methods

Animals. The investigation conforms to the guidelines for the care and use of laboratory animals published by the National Institutes of Health (Publication no. 85-23, revised 1985). The experimental procedures were in accordance with the German Law for the Protection of Animals and accepted by the Ministry of Science and Public Health of the City State of Hamburg, Germany (Nr. 69/10). Mybpc3-targeted KI mice were generated previously,[22] and maintained on a Black Swiss background.

Design of PTMs. The sequences of primers used are listed in Supplementary Table S2. The coding sequence of the PTMs was generated by PCR from WT Mybpc3 cDNA with a forward primer (PTM F) containing an XhoI restriction site, the ATG followed by the FLAG sequence and the first 20 nucleotides of Mybpc3 exon 1. The reverse primer (PTM R) contained a BamHI restriction site and the 5′ canonical splice donor site sequence followed by a downstream intronic splicing enhancer element/sequence from the rat fibroblast growth factor receptor 2 gene and last 23 nucleotides of Mybpc3 exon 6. The binding domains were obtained by PCR on genomic KI DNA using a forward primer (BD F) containing a BamHI restriction site, and 21 nucleotides of Mybpc3 intron 6. The reverse primer (BD R) contained a NotI restriction site and 28 nucleotides complementary to Mybpc3 intron 6. The reverse binding domain was amplified in the same way (primers BD-R F and BD-R R) but reverse complemented. PCR products were sequentially cloned into pdsAAV6-TNNT2 vector (human TNNT2 promoter) and accuracy of the insertion was verified by DNA sequencing analysis. The SV40 polyA signal was removed in one of the PTM plasmids by digestion with NotI and Mva1269I followed by religation of the plasmid.

Production and titration of AAV particles. AAV6 pseudotyped vectors were generated by cotransfection of HEK293-AAV cells (Biocat, Heidelberg, Germany) with the pdsAAV-TNNT2 transfer plasmid and the AAV packaging plasmid pDP6rs, which provides the AAV2 rep and AAV6 cap genes and adenoviral helper functions.[39] AAV9 pseudotyped vectors were generated by triple-transfection of pdsAAV-TNNT2 transfer plasmid with pAAV2/9 and pHelper encoding adenoviral helper functions (Biocat). Generation of recombinant AAV6 and AAV9 particles was carried out as described previously,[40] with some modifications. HEK293-AAV cells were cultivated in Dulbecco's modified Eagle's medium (high glucose) supplemented with 10% (vol/vol) heat-inactivated fetal calf serum, 0.1 mmol/l MEM non-essential amino acids, 2 mmol/l l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. Tissue culture reagents were obtained from Life technologies. Briefly, 1.5 × 107 HEK293-AAV cells were seeded on 15-cm plates and transfected with polyethylenimine. After 72 hours, cells were harvested, washed three-times with phosphate-buffered saline (PBS) and resuspended in PBS. After three freeze–thaw cycles, benzonase (Merck, Darmstadt, Germany; final concentration 250 U/ml) was added and the lysates incubated for 1 hour at 37 °C. Cell debris was pelleted and vector-containing lysates were purified using iodixanol step gradients.

The genomic titers of DNase-resistant recombinant AAV particles were determined by quantitative PCR using the SYBR Green qPCR Master MIX 2 (Fermentas, Darmstadt, Germany) and an ABI PRISM 7900HT cycler (Applied Biosystems, Foster City, CA). Vectors were quantified using primers specific for the TNNT2 promoter sequence. Real-time PCR was performed in a total volume of 10 μl with 0.3 μmol/l for each primer. pdsAAV-GFP plasmid was used as a copy number standard. A standard curve for quantification was generated by serial dilutions of the respective plasmid DNA. The cycling conditions were as follows: 50 °C for 2 minutes, 95 °C for 10 minutes, followed by 35 cycles of 95 °C for 15 seconds and 60 °C for 60 seconds. Calculations were done using the SDS 2.4 software (Applied Biosystems).

Generation of recombinant AdV. To generate the AdV-PTMΔpA under the control of TNNT2 promoter, we used the In-fusion kit (Clontech, St Germain-en-Laye, France) to fuse together the two cassettes into pShuttle85706. The pShuttle containing the TNNT2-PTMΔpA insert as well as the TNNT2-GFP in a bicistronic manner was electroporated into Escherichia coli BJ5183-D1 (Stratagene, Darmstadt, Germany) to produce adenoviral DNA through recombination. This DNA was used to transfect HEK293 cells and recombinant AdV was amplified using standard techniques.

Isolation, culture, and transduction of neonatal mouse cardiac myocytes. We isolated and cultured NMCMs using a well-established protocol.[22] AAV6-mediated transductions of cardiac myocytes were performed for 30 minutes at 37 °C in suspension before plating (4.4 × 105 cells/well) at a MOI of 3,000 (AAV6-PTMΔpA, both RNA and protein analysis, AAV6-PTM and AAV6-PTM-R for protein analysis) or 30,000 (AAV6-PTM and AAV6-PTM-R for RNA analysis). Cardiac myocytes were kept in culture for 7 days at 37 °C and 10% CO2 before harvesting.

In vivo AAV9 administrations. Seven-week-old KI mice received AAV9-PTMΔpA (1.04 × 1011 vg), AAV9-RLuc (1.36 × 1011 vg) or NaCl via systemic administration into the tail vein with a 29-G needle. Intravenous injections of neonatal KI mice (postnatal day 1) with AAV9-PTMΔpA (4.7 × 1011 vg) or PBS were performed into the temporal vein using a 30-G needle.[41] All mice recovered quickly from the injection.

In vivo bioluminescence imaging. Luciferase activity in the mouse heart was non-invasively assessed by in vivo bioluminescence imaging 4 weeks after AAV9 injection. AAV9-Rluc–injected mouse and the NaCl-injected mouse were anesthetized with 1.8% isofluorane. Thereafter the substrate coelenterazine (Biosynth, Staad, Switzerland) dissolved in methanol and further diluted in sodium phosphate buffer pH 7, was injected intraperitoneally (i.p.) at a dose of 2.5 mg/kg body weight in both mice. The mice were then placed in the chamber of a Xenogen in vivo Imaging System under continuous anesthesia. The oxidation of coelenterazine by Renilla luciferase releases coelentarimide and blue light at 480 nm. This bioluminescence was recorded in a manually-selected region of interest centered over the mouse heart, using 3-minute scans.

Echocardiographic analysis. Transthoracic echocardiography was performed using the Vevo 2100 System (VisualSonics, Toronto, Ontario, Canada). KI mice were anesthetized with isofluorane (1–2%) and fixed to a warming platform in a supine position. B-mode images were obtained using a MS400 transducer for adult mice and a MS550 transducer for neonatal mice. Images were obtained in a parasternal short and long-axis view and dimensions of the left ventricle were measured in a short-axis view in diastole and systole.

Reverse transcription-PCR analysis. Total RNA was isolated from cultured NMCMs or ventricular tissue (30 mg) using the SV Total RNA Isolation System Kit (Promega, Madison, WI) according to the manufacturer's instructions. RNA concentration, purity and quality were determined using the NanoDrop ND-1000 spectrophotometer (Thermo Scientific, Darmstadt, Germany). Reverse transcription was performed from 150 to 200 ng RNA using oligo-dT primers (SuperScript-III kit; Life Technologies, Darmstadt, Germany). As a control for genomic contamination, a reaction without reverse transcription was performed in parallel. Touchdown PCR amplifications (65–60 °C) with different primer pairs (Supplementary Table S2) were performed using AmpliTaq Gold Polymerase (Applied Biosystems) in a total volume of 20 μl for 35 cycles. PCR products were visualized on 1.5% agarose gels. The full-length repaired Mybpc3 mRNA was amplified by touchdown PCR (67–62 °C) with Phusion Hot StartII High-Fidelity DNA polymerase (Biozym, Hessisch Oldendorf, Germany) for 31 cycles. For semi-quantitative analysis a touchdown PCR (65–60 °C) for 25 cycles was used to amplify either total (primers E1-F, E9-R) or repaired (primers FLAG, E9-R) Mybpc3 mRNA. PCR products of the first PCR round were purified on a column (QIAquick PCR Purification Kit; QIAGEN, Valencia, CA) prior a second touchdown PCR (65–60 °C, primers E1-F, E2-R) for 35 cycles.

Western blot and immunoprecipitation analyses. Crude protein extract from cultured NMCMs or HEK293 cells were extracted in lysis buffer (30 mmol/l Tris base pH 8.8, 5 mmol/l EDTA, 30 mmol/l NaF, 3% SDS, 10% glycerol) and protein concentration was determined by Bradford protein assay (Bio-Rad, Hercules, CA). Total proteins (NMCMs 30 μg/lane, HEK293 2.5 μg/lane) were separated on 10% SDS-polyacrylamide (29:1) mini-gels (Bio-Rad) and transferred on polyvinylidene fluoride membranes by electroblotting. Membranes were stained overnight with primary antibodies directed against the FLAG epitope (1:5,000; Sigma, St Louis, MO) in 5% milk in TBS-T or against the MyBP-C motif (1:1,000). After incubation with anti-mouse (1:10,000; Sigma) or anti-rabbit (1:6,000; Sigma) peroxidase-conjugated secondary antibodies, proteins were visualized using Super Signal West Dura detection reagent (Thermo Scientific) and signals were detected with the ChemiGenius2 Bio Imaging System.

For immunoprecipitation AAV6-PTMΔpA-transduced NMCMs or ventricular tissue of AAV9-PTMΔpA-injected mouse were lysed in modified RIPA buffer (500 mmol/l NaCl, 1 mmol/l EDTA, 50 mmol/l Tris-HCl pH 7.4, 1% Triton X-100, protease inhibitors Complete Mini; Roche, Indianapolis, IN), sonicated 2× for 30 seconds and centrifuged for 10 minutes at full speed at 4 °C. The supernatant containing soluble proteins was diluted in 500 μl modified RIPA buffer and gently rolled overnight at 4 °C with or without 10 μg anti-FLAG antibody (Sigma). The immunocomplexes were recovered by incubation with 50 μl of protein A/G plus agarose (Santa Cruz Biotechnology, Santa Cruz, CA) for 4 hours at 4 °C. After three washing steps in RIPA buffer and one additional wash step in 1× PBS (200g, 3 minutes, 4 °C) the FLAG-tagged proteins were eluted in 1× Laemmli buffer (20 mmol/l Tris-HCl pH 6.8, 200 mmol/l DTT, 4% SDS, 0.02% bromophenol blue, 20% glycerol) and used for western blot as described above. Anti-mouse Exacta Cruz (1:2,000; Santa Cruz Biotechnology), which does not recognize the heavy and light chains of the immunoprecipitation antibody, was used as secondary antibody. The same immunoprecipitation protocol was applied to HEK293 cells (40 μg) transiently transfected with FLAG-WT-cMyBP-C used as control.

Immunofluorescence analysis. For immunofluorescence analysis, AAV6-transduced KI NMCMs were cultured for 7 days on coverslips. Cells were rinsed once with ice-cold 1× PBS and fixed 10 minutes at −20 °C in methanol/acetone (20/80). After two short washing steps in 1× PBS, cells were permeabilized 1 hour at room temperature in solution A (10% FCS, 1% BSA, 0.5% Triton X-100 in 1× PBS). Incubation with primary antibodies (anti-FLAG, 1:800; anti-MyBP-C motif, 1:500) was done in solution B (1% BSA, 0.5% Triton X-100 in 1× PBS) for 1 hour at room temperature. Cells were then rinsed twice in solution B and incubated for 1 hour at room temperature with secondary antibodies (anti-mouse IgG Alexa 488-conjugated, 1:800 and anti-rabbit IgG Alexa 546-conjugated 1:800; Molecular Probes, Darmstadt, Germany) diluted in solution B together with DRAQ5 (1:1,000; Biostatus, Leicestershire, UK) for nuclear staining. Coverslips were embedded in Mowiol and confocal images were acquired with a Zeiss LSM 710 system using a Zeiss Axiovert microscope (Zeiss, Jena, Germany) and a 40×-oil objective.

Figure S1. Schematic representation of the repaired and mutant Mybpc3 mRNA. Figure S2. Efficiency of AAV6- and AdV-mediated transduction in Mybpc3-targeted knock-in NMCMs. Figure S3. In vivo bioluminescence imaging. Figure S4. Efficiency of transduction after systemic administration of AAV9-GFP in a Mybpc3-targeted knock-in mouse. Figure S5. Detection of full-length repaired Mybpc3 mRNA. Table S1. Sequences of the binding domains. Table S2. Sequences of the PCR primers used in the present study. Table S3. Echocardiographic analysis.