All animals and experiment protocols were approved by the Animal Ethical Committee of Peking University Health Science Center under the approval number LA2021034. Mice were housed at Peking University Health Science Center Department of Laboratory Animal Science. The homozygous Rosa^{CAG−LSL-Cas9−2A-tdTomato} mouse line was purchased from GemPharmatech (Strain No. T002249) (Chen et al., 2021).

In AAV injection, heart function analysis and tissue collection procedures, the mice were anesthetized with 2% isoflurane by inhalation, maintained asleep at 1.5% isoflurane, and euthanized through cervical dislocation if needed.

AAV-cTNT-Cre, AAV-U6sgRNA-U6sgRNA-cTNT-Cre, and AAV plasmids expressing sgRNAs targeting Jph2, Tead1 and Ryr2 were previously described (Guo et al., 2017). CRISPick (https://portals.broadinstitute.org/gppx/crispick/public) was used to design new sgRNAs targeting murine Atp2a2. The sequences of these sgRNAs were 5′-GCA​GAA​GCT​AGA​CGA​GTT​TG-3′ and 5′-CAA​GAT​CCG​GGA​TGA​AAT​GG-3’. To clone the AAV plasmid for Atp2a2 knockout, the two sgRNA sequences were synthesized in primers and amplified together with the scaffold and U6 promoter sequences by PCR using 2xEasyPfu PCR SuperMix (-dye) (AS211-02, TransGen, Beijing, China). The PCR product was next incorporated into the AAV-U6sgRNA-U6sgRNA-cTNT-Cre plasmid between AarI and SapI sites using the seamless cloning master mix (B632219-0040, BBI, China).

AAV9 was prepared as previously described with modifications (Guo et al., 2017; Guo et al., 2018). In brief, 70 μg AAV plasmids, 70 μg AAV9-Rep/Cap and 160 μg pHelper (pAd-deltaF6, Penn Vector Core) plasmids were produced by the EZgeneTM Plasmid ezFilter Megaprep 3 Kit (PD1611-01, Biomiga) and transfected into five 15-cm plates of HEK293T cells using the PEI transfection reagent (40816ES03, Yeasen, China). At 60 h after transfection, cells were scraped off from the plates, resuspended in lysis buffer (20 mM Tris pH 8, 150 mM NaCl, 1mM MgCl₂, 50 μg/ml benzonase (20156ES60, Yeasen, China)) and lysed by three freeze-thaw cycles. Simultaneously, AAV9 in the cell culture medium was precipitated by 40% PEG8000 (P2139, Sigma) in 2.5M NaCl solution, resuspended in lysis buffer and pooled with cell lysates. AAV was purified in an Optiprep (D1556-250ML, Sigma) density gradient by ultracentrifugation (XPN-100, Beckman) with a type 70Ti rotor. Next AAV was cleaned and concentrated in PBS with 0.001% pluronic F68 (PFL01-100ML, Caisson, China) using a 100 kDa centrifugal filter tube (UFC910096, Millipore). AAV titer was absolutely quantified by real-time PCR using primers amplifying a fragment of the cTNT promoter.

Echocardiography was performed with a VINNO6n machine (VINNO Corporation, Suzhou, China). The mice were shaved on the chest and the videos were acquired with a 23 MHz transducer along the parasternal long axis. The Lead II electrocardiogram was measured with a Softron ECG processor and analyzed by SP2006 software (Softron Biotechnology, Beijing, China). The animal treatment was blinded to the scientist who performed the measurement. Five consecutive heart beats were averaged to calculate the results for each animal.

Total RNA was purified from the heart apex using the TransZol Up Plus RNA Kit (ER501-01, TransGen, China). Genomic DNA removal and reverse transcription were performed using TransScript II One-Step gDNA Removal and cDNA Synthesis SuperMix (AH311-03, TransGen, China).

RT-PCR was performed using 2XTaq PCR Master Mix (KT211, TIANGEN). RT-PCR primers were designed as follow: 5′-TGC​CTT​GTT​GGG​GTG​ACT​G-3′ and 5′-CAC​CGT​TCC​ATT​GCT​GTG​C-3′ for Jph2; 5′-GCC​GTG​GTT​TGT​TTT​CTT​G-3′ and 5′-GCT​TGT​TGT​GGA​TGG​CAG​TAG-3′ for Tead1; 5′-TGA​GAA​CTG​ATG​ATG​AGG​TGG-3′ and 5′-GCT​TAG​AGG​CAG​GAT​GTA​TGG-3′ for Ryr2.

RT-qPCR was performed using Perfect Start Green qPCR Super Mix (+DveII) (AQ602-24, TransGen, China) and the AriaMx Real-Time PCR System (Agilent Technologies). RT-qPCR primers were: 5′-CAA​CTC​CCT​CAA​GAT​TGT​CAG​CAA-3′ and 5′-GGC​ATG​GAC​TGT​GGT​CAT​GA-3′ for Gapdh; 5′- GGT​AGC​CAA​TGC​AAT​CGT​GG-3′ and 5′-CGT​TGC​ACA​CTC​TTT​CTG​TCC-3′ for Atp2a2.

The fresh heart apex was washed by ice-cold PBS prior to homogenization with an electric homogenizer (DREMEL F6/10) in RIPA Buffer (25 mM Tris PH7.0∼8.0, 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, 1% Triton X-100, protease inhibitors (Solarbio)). After 10-min lysis on ice, tissue lysates were centrifuged for 15 min at 12,000 rpm at 4°C. Then the supernatants were collected and the protein concentration was quantified using Easy II Protein Quantitative Kit (BCA) (DQ111-01, TransGen, Beijing, China). The protein concentration of the lysates was adjusted to the same level with RIPA Buffer and next diluted in 4x SDS sample buffer (P1016, Solarbio, China). After being boiled for 10 min, 20 μg proteins from each sample was separated on a 4–15% gradient gel (DG101-01, TransGen, China), transferred to a PVDF membrane and blocked for 1 h in 5% Non-Fat Milk/TBST at room temperature.

Primary antibodies (Ms-Anti-GAPDH, 1:5000, TransGen, HC301; Rb-Anti-SERCA2, 1:5000, Abcam, ab137020) were incubated with the membranes overnight at 4°C. After three washes with TBST, 5 min each, the membranes were incubated with HRP-conjugated secondary antibodies (HRP-Gt-Anti-Rb, 1:5000, D110058-0100, BBI; HRP-Gt-Anti-Ms, 1:5000, D110087-0100, BBI) for 1 h at room temperature. After three washes of TBST, 5 min each, the ECL Western blotting substrate (PE0010, Solarbio, China) was added to the membranes for chemiluminescence signal development. Images were acquired using an iBright CL1500 Imaging System (Thermo Fisher Scientific).

Cardiomyocytes were isolated by retrograde Langendorff perfusion according to an established protocol (Guo et al., 2017; Guo et al., 2018). In brief, the mice were injected with heparin and anesthetized with isoflurane. Then the hearts were removed and cannulated through the aorta onto the perfusion apparatus. Perfusion buffer (120.4 mM NaCl, 14.7 mM KCl, 0.6 mM KH₂PO₄, 0.6 mM Na₂HPO₄, 1.2 mM MgSO₄, 10 mM HEPES, 4.6 mM NaHCO₃, 30 mM Taurine, 10 mM BDM, 5.5 mM Glucose) was applied to flushing blood and equilibrating the heart. After 4-min perfusion, the fluid was switched to myocyte digestion buffer containing 2 mg/ml collagenase II (Worthington, LS004177) for a 10-min digestion at 37°C.

After digestion, the apical half of heart was next cut off, manually dissociated into single cardiomyocytes in myocyte stopping buffer (10% FBS/perfusion buffer) and filtered through a 100 μm cell strainer. The cell suspension was concentrated by 20 g centrifugation for 5 min and re-suspended in short-term culture medium containing DMEM (Gibco), 10% FBS (SE100-B, VISTECH), Penicillin/Streptomycin (Gibco) and 10 μM Blebbistatin. Cardiomyocytes were cultured on laminin-coated coverslips for less than 1 h at 37°C with 5% CO2 before fixation.

For CellROX staining, we incubated the isolated cardiomyocytes with 5 μM CellROX Green Reagent (Life technologies, United States) for 10min before immediately imaging the cells using a confocal microscope. For in situ T-tubule imaging, hearts were perfused with perfusion buffer added with 1 μg/ml FM 4–64 (21487, AAT Bioquest) at room temperature for 10 min. After labeling, the heart was removed from the perfusion system, positioned on a glass-bottom dish, and immediately imaged using an inverted confocal microscope (Leica SP8).

For immunofluorescence on isolated cardiomyocytes, the cells were fixed on coverslips by 4% paraformaldehyde/PBS for 10 min, permeabilized and blocked by 0.1% Triton-100 and 4% BSA/PBS at room temperature for 1 h. Primary antibodies (Ms-Anti-α-actinin, 1:500, abcam, ab9465) were incubated with the cells for 1 h at temperature. After 3-washes with PBS, 5 min each, fluorescent secondary antibodies (Alexa Fluor^® 488, 1:500, Abcam) and DAPI were incubated with the cells for 1 h at room temperature. After 3-washes with PBS, 5 min each, the cells were mounted with anti-fluorescence quenching mounting fluid (P0126-25ml, Beyotime) before confocal imaging.

For fluorescence detection on tissue sections, heart tissues were first immersed in 4% paraformaldehyde (PFA) at 4°C overnight and then dehydrated in sucrose solutions (15% followed by 30%) overnight. Then the tissue samples were embedded in OCT (Sakura) and frozen in -80 °C. Heart sections (7 μm) were cut on a cryostat microtome (CM 1950, Leica, Germany). For antibody staining, the tissue sections were incubated with primary antibodies (Rb-anti-Ki67, Cell Signaling Tech., 9129T) and then stained with the secondary antibodies. TUNEL staining was performed with the one-step fluorescence TUNEL kit (Beyotime, C1086). Or the sections were treated with wheat germ agglutinin (Alexa Fluor^® 488-conjugated, 1:500 dilution, AAT Bioquest) and DAPI for 10 min at room temperature. The slides were mounted with anti-fluorescence quenching mounting fluid (P0126-25ml, Beyotime) before confocal imaging.

The images were captured by a Leica SP8 confocal laser scanning microscope equipped with an APO 63X/1.4 oil objective. ImageJ was used for quantitative measurement. Cell size, cell shape and Z-line distance were manually measured.

Isolated cardiomyocytes were purified by a FACS Aria SORP (BD Biosciences) machine equipped with a 100 μm nozzle at Peking University Medical and Health Analysis Center. RNA was extracted from sorted cardiomyocytes with the RNAprep pure Micro Kit (Tiangen Biotech.). SMART-Seq was next performed at the Beijing Geekgene Company. First, the RNA was mixed with 0.1 μL rnase inhibitor (Clontech), 1.9 μL Triton X-100 solution (1%), 1 μL dNTP mix (10 mM), and 1 μL oligo-dT primer (5 μM). Reverse transcription was performed by further adding 0.5 μL SuperScript II reverse transcriptase (200 U/μL, Invitrogen), 0.25 μL rnase inhibitor (40 U/μL, Clontech), 2 μL Superscript II First-Strand Buffer (5×, Invitrogen), 0.5 μL DTT (0.1M, Invitrogen), 2 μL Betain (5 M, Sigma), 0.06 μL MgCl2 (1M, Sigma), and 0.1 μL TSO (100 μM). Reverse transcription was carried out at 25°C for 5 min, 42°C for 60 min, followed by 50°C for 30 min and 72°C for 10 min.

PCR preamplification was performed using KAPA HiFi HotStart Ready MIX (KAPA Biosystems) with 22 cycles of PCR and the IS PCR primer reduced to 50 nM (4 cycles at 98°C for 20 s, 65°C for 30 s, and 72°C for 5 min, followed by 18 cycles at 98°C for 20 s, 67°C for 15 s, and 72°C for 5 min, with a final cycle at 72°C for 5 min). Subsequently, Amplified samples were purified twice with 0.8X Ampure XP beads (Beckman, A63882).

Next, the enriched cDNA fragments were used to construct libraries using KAPA Hyper Prep Kits (KK8505). NEB U-shape adaptors were used for ligation. Libraries were sequenced to generate 150-bp paired-end reads on an Illumina X-Ten platform.

The statistical analysis was performed using GraphPad software. Median, 25 and 75% quartiles (horizontal lines in the box), individual data points and maximal/minimal values (whiskers) were shown in the Box-and-Whisker plots. Unless otherwise stated, data were analyzed using unpaired two-tailed student’s t-test. p < 0.05 was considered statistically significant.

The sequencing data for this study has been deposited at the BIG submission portal, Genome Sequence Archive (CRA005866), and are accessible at https://ngdc.cncb.ac.cn/gsa/s/81v8l014. Other materials and data sets are available upon reasonable requests.

The conventional CASAAV system uses the Rosa^{CAG−LSL-Cas9−2A−GFP} mouse in which Cre recombinase can recombine the LoxP-STOP-LoxP (LSL) site and activate CAG promoter-driven Cas9-2A-GFP expression (Guo et al., 2017; Vandusen et al., 2017). Although this mouse was proved efficient in Cas9-based mutagenesis, the GFP marker was not only relatively dim but also precluded downstream assays using green fluorescence (Guo et al., 2018). To solve these problems, we adopted a newly generated Rosa^{CAG−LSL-Cas9−2A-tdTomato} mouse line (Rosa^Cas−Tom) (Chen et al., 2021) in the CASAAV system, in which tdTomato is a red fluorescent protein that is much brighter than GFP (Shaner et al., 2004).

In the CASAAV system (Figure 1A), AAV vectors deliver Cre recombinase driven by a cardiomyocyte-specific cTNT promoter, thus Cas9-based mutagenesis is restricted to cardiomyocytes. The same AAV vectors also deliver two single-guide RNAs (sgRNAs) expressed by U6 promoters. The Cas9/sgRNA complexes target the coding sequences of the gene of interest, deposit small insertions and deletions to the genome, and eventually generate frame-shift mutations that silence the gene.

We first tested if the Rosa^Cas−Tom mice were suitable for CASAAV. We generated serotype-9 AAVs (AAV9) that were reported to efficiently knockout Jph2, Tead1 and Ryr2 in murine cardiomyocytes by CASAAV (Guo et al., 2017). We injected each AAV subcutaneously into homozygous postnatal day 1 (P1) Rosa^Cas−Tom pups at the dose of 5 × 10¹⁰vg/g (viral genome per Gram body weight). The heart ventricles were harvested at P7 for histological and molecular analysis. As compared to the control samples receiving no AAVs, all AAV-treated hearts exhibited bright red fluorescence (Figure 1B), indicating successful activation of the LSL and tdTomato components by the AAVs.

The sgRNAs in the above AAVs were purposely designed to target genomic regions proximal to each other (usually less than 300bp apart) so the DNA fragments between the two targeted loci would be deleted and readily detected by the size change of PCR products. We performed reverse-transcription PCR (RT-PCR) to amplify genomic regions including the sgRNAs targets. In control groups, single PCR products of the expected sizes were observed. However, in the AAV-treated groups, smaller PCR products corresponding to CASAAV-mediated DNA deletion were found (Figure 1C). These data indicate that CASAAV works in the Rosa^Cas−Tom mice.

We next designed sgRNAs targeting Atp2a2 and prepared AAV9 vectors carrying these sgRNAs for CASAAV-based knockout (Figure 2A). AAV vectors that do not express sgRNAs were used as the control. We injected 5 × 10¹⁰ vg/g control and knockout vectors into P1 pups and collected heart tissues at P14 to determine Atp2a2 knockout efficiency. Fluorescence imaging analysis of heart sections stained with wheat germ agglutinin (WGA), a marker of cell boundary, revealed 77% AAV transduction rate among cardiomyocytes (Figure 2B), thus we considered this “high-dose” AAV treatment a whole-heart experiment suitable for tissue level analysis. Real-time quantitative PCR (RT-qPCR) analysis showed profound reduction of Atp2a2 mRNA in the CASAAV-treated heart tissues (Figure 2C), which led to the reduction of SERCA2 protein levels as shown by western blot analysis (Figure 2D). To directly validate the Cas9-based genome mutagenesis, we performed RNA-Seq analysis on the AAV-transduced cardiomyocytes (see Section 3.4 for more details). We found the number of sequencing reads aligned to Atp2a2 exons was greatly reduced, presumably due to non-sense mediated mRNA decay (Supplementary Figure S1A). Among reads aligned to the sgRNA-targeted exon, we observed insertion and deletion mutations adjacent to the sgRNA targets, directly supporting the action of Cas9-based mutagenesis (Supplementary Figure S1B). Together, these data indicate the successful establishment of cardiomyocyte Atp2a2 knockout in mice.

We next assessed the impact of whole-heart Atp2a2 knockout on heart functions. The high dose-treated P1-injected mice undergoing Atp2a2 CASAAV knockout died in the third week after birth (Figure 3A). Echocardiogram analysis demonstrated reduced fraction shortening (FS), increased left ventricular internal diameters at the end of diastole (LVIDd) and systole (LVIDs), as well as a decrease of left ventricular posterior wall thickness at the end of diastole (LVPWd) (Figure 3B; Supplementary Figure S2) in the P14 mutant hearts. Thus, whole-heart Atp2a2 knockout led to reduced cardiac contractile functions and ventricular dilatation.

We also performed electrocardiogram analysis and observed prominent ST segment elevation in P14 mutant hearts (Figure 3C). This phenotype was specifically associated with increased QT intervals while no other measured electrocardiac parameters were changed (Figure 3D). Therefore, Atp2a2 knockout perturbed cardiac electrophysiology particular in the ventricular repolarization phase.

The prominent phenotypes observed in the P14 mutant hearts strongly implied the presence of cardiomyocyte maturation defects. However, the whole-heart cardiac dysfunction could deposit secondary effects that confound further cell maturity analysis (Guo et al., 2017; Guo and Pu, 2018). To solve this problem, we next titrated down AAV doses and performed a genetic mosaic analysis.

We serially diluted the knockout AAV and injected 5 × 10⁹ vg/g and 1 × 10⁹ vg/g to P1 Rosa^Cas−Tom pups. These two treatments were 1:10 and 1:50 diluted as compared to the high-dose treatment, and thus referred as “mid-dose” and “low-dose” treatments in the following experiments. In both treatments, animal deaths were circumvented within the cardiac maturation time window (Figure 4A). The ST segment elevation phenotypes were attenuated in the mid-dose treatment and barely detectable in the low-dose group (Figure 4B). Low-dose AAV treated Atp2a2 knockout hearts displayed nearly normal contractile functions and ventricle sizes (Figure 4C; Supplementary Figure S2). In this experiment, fluorescence imaging revealed about 23% tdTomato positive cardiomyocytes in the heart sections (Figure 4D). Thus, we successfully generated a genetic mosaic model to study Atp2a2 knockout cardiomyocytes in vivo without the concern that the phenotypes were secondary to overall cardiac dysfunction. This model was next used in all following assays.

At the cellular level, we first wondered if SERCA2 regulated T-tubule, the key ultrastructural hallmark of calcium handling maturation. We harnessed Langendorff retrograde perfusion to load the whole heart with a cell membrane dye FM 4–64 and performed confocal optical section to directly observe T-tubule in situ (Chen et al., 2015). This experiment was performed on 4-week old hearts when T-tubule maturation just finished (Chen et al., 2013). In tdTomato-negative cardiomyocytes that were not transduced by AAV, transversely-aligned T-tubules could be observed throughout the cytoplasm. However, in tdTomato-positive Atp2a2 knockout cardiomyocytes, no sign of T-tubule formation was detected (Figure 5A). AutoTT software (Guo and Song, 2014) was used to quantify T-tubule contents and corroborated the observation (Figure 5A).

Sarcomere expansion is another hallmark of cardiomyocyte maturation. To determine the impact of Atp2a2 knockout on sarcomere expansion, we isolated cardiomyocytes from 4-week-old mosaic hearts and immunolabeled sarcomeres with the Z-line marker ACTN2 (α-actinin-2). This work demonstrated reduced ACTN2 signals in the mutant cardiomyocytes as well as the decreased spacing between adjacent Z-lines (Figure 5B). We also measured the projected cell area, cell length and cell width and we found these parameters were all reduced in Atp2a2 knockout cardiomyocytes (Figure 5C). However, the length: width ratios of the mutant cardiomyocytes were maintained, thus SERCA2 promoted the hypertrophic growth of cardiomyocytes by proportionally increasing cell length and width.

In addition to the above structural parameters, cardiomyocyte maturation was also characterized by profound gene expression changes (Guo and Pu, 2020). To investigate the impact of Atp2a2 knockout on cardiomyocyte maturation, we next performed RNA sequencing analysis. The genetic mosaic approach adds technical difficulties in enriching AAV-treated cardiomyocytes for bulk analysis. To solve this problem, we developed a protocol to perform fluorescence-activated cell sorting (FACS) of isolated cardiomyocytes from mice and extracted the small amount of mRNA from the sorted cells (Guo et al., 2018). We also adopted the SMART-Seq (Switching Mechanism At the end of the 5′-end of the RNA Transcript) technique (Picelli et al., 2013) to perform transcriptomic analysis from the low input of mRNA (Figure 6A).

We analyzed three biological replicates of control and Atp2a2 knockout cardiomyocytes that were collected at P14 and observed clear separation between the two groups of data in a principal component analysis (PCA) (Figure 6B). Using adjusted p value (Padj) less than 0.05 as the cutoff, we identified 69 upregulated genes and 160 down-regulated genes, among which Atp2a2 counts were one of the most significantly reduced (Figure 6C). Thus, this FACS and SMART-Seq platform provided a reliable method to uncover major transcriptional changes in cardiomyocytes and to understand the cell-autonomous functions of SERCA2 in cardiomyocyte maturation.

We performed gene set enrichment analysis (GSEA) to uncover the major biological functions associated with the differentially regulated genes in the Atp2a2 knockout cardiomyocytes. Among the Hallmark gene sets, oxidative phosphorylation and fatty acid metabolism genes appeared most enriched in the down-regulated genes (Figure 7A). This data indicated cardiomyocyte maturation defects since oxidative respiration using fatty acids as substrates was a major metabolic feature for mature cardiomyocytes (Guo and Pu, 2020). Inflammation-related gene sets such as Interferon and TNFa responses were up-regulated (Figure 7B). Both oxidative respiratory and inflammatory changes were known to associated with reactive oxygen species (ROS) elevation. Agreeing to this notion, we stained isolated cardiomyocytes with a ROS dye CellROX and observed a significant increase of the ROS signal in the Atp2a2 mutant cells (Figure 7C,D).

A hallmark of murine cardiomyocyte maturation is the isoform switching from the fetal expression of Myh7 and Tnni1 to the adult expression of Myh6 and Tnni3 (Hunkeler et al., 1991; Ng et al., 1991). In the SERCA2-depleted cardiomyocytes, we observed decreased Myh6:Myh7 and Tnni3:Tnni1 ratios (Figure 7E), further supporting the presence of cardiomyocyte maturation defects. Another important hallmark of cardiomyocyte maturation is the withdrawal from cell cycle. Thus, we performed GSEA of the cell proliferation gene set and observed weak but significant enrichment in the upregulated genes in the SERCA2-depleted cardiomyocytes (Supplementary Figure S3A). However, immunostaining of Ki67 in the Atp2a2 mutant cardiac sections did not show increase of proliferation in the AAV-infected tdTomato + cardiomyocytes (Supplementary Figure S3B, arrow heads). To assess gene expression changes at the transcriptome level, we harnessed a published RNA-Seq data comparing P6 vs P30 cardiomyocytes using the FACS-SMART-Seq experimental setup identical to this study (Guo et al., 2021). GSEA revealed P30-specific genes were significantly enriched among the down-regulated genes in the Atp2a2 knockout cardiomyocytes (Figure 7F). Thus, SERCA2 globally regulated the transcriptome changes during cardiomyocyte maturation.

Because calcium is well-known to regulate cardiomyocyte survival versus death, thus we also examined the impact of SERCA2 depletion on cell death. We performed GSEA and observed significant upregulation of a cell death gene set in the Atp2a2 knockout cardiomyocytes (Supplementary Figure S3C). The terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) technique further validated the presence of dying cardiomyocytes in the high-dose AAV-treated Atp2a2 mutant hearts (Supplementary Figure S3D), while no such signal was detected in the control heart sections. This result suggested that the cardiomyocyte maturation defects of the Atp2a2 knockout cardiomyocytes were partly associated with cell death.