Research ArticleCardiovascular Biology

The CCR4-NOT deadenylase complex controls Atg7-dependent cell death and heart function

See allHide authors and affiliations

Sci. Signal.  06 Feb 2018:
Vol. 11, Issue 516, eaan3638
DOI: 10.1126/scisignal.aan3638

Protecting the heart by destabilizing mRNA

The removal of polyadenylate tails from mRNAs by the CCR4-NOT complex marks these mRNAs for degradation. Yamaguchi et al. (see also the Focus by Das) found that this activity of this complex was required to prevent cell death in the heart. Mice deficient in a component of this complex suffered from cardiac dysfunction and died of heart failure due to cardiomyocyte death. The CCR4-NOT complex deadenylated Atg7 mRNA, which encodes a protein required for autophagy, a process by which cellular constituents and organelles are digested. The increase in Atg7 in the mutant mice resulted in activation of cell death–associated genes by the transcription factor p53. Drugs that increase autophagy have been explored for the treatment of various diseases, but the authors note that their results raise the possibility of cardiovascular side effects for such drugs.


Shortening and removal of the polyadenylate [poly(A)] tail of mRNA, a process called deadenylation, is a key step in mRNA decay that is mediated through the CCR4-NOT (carbon catabolite repression 4–negative on TATA-less) complex. In our investigation of the regulation of mRNA deadenylation in the heart, we found that this complex was required to prevent cell death. Conditional deletion of the CCR4-NOT complex components Cnot1 or Cnot3 resulted in the formation of autophagic vacuoles and cardiomyocyte death, leading to lethal heart failure accompanied by long QT intervals. Cnot3 bound to and shortened the poly(A) tail of the mRNA encoding the key autophagy regulator Atg7. In Cnot3-depleted hearts, Atg7 expression was posttranscriptionally increased. Genetic ablation of Atg7, but not Atg5, increased survival and partially restored cardiac function of Cnot1 or Cnot3 knockout mice. We further showed that in Cnot3-depleted hearts, Atg7 interacted with p53 and modulated p53 activity to induce the expression of genes encoding cell death–promoting factors in cardiomyocytes, indicating that defects in deadenylation in the heart aberrantly activated Atg7 and p53 to promote cell death. Thus, mRNA deadenylation mediated by the CCR4-NOT complex is crucial to prevent Atg7-induced cell death and heart failure, suggesting a role for mRNA deadenylation in targeting autophagy genes to maintain normal cardiac homeostasis.


Cardiovascular diseases are the leading causes of death in developed countries. Coordinated transcriptional and posttranscriptional regulation of gene expression is important to maintain normal heart physiology. Dysregulation in this coordination causes and/or accompanies multiple pathologies, such as cardiomyopathy and myocardial infarction. In posttranscriptional regulation, the exonuclease-mediated degradation of the mRNA polyadenylate [poly(A)] tail, a process called deadenylation, is a key step in regulated mRNA degradation, which contributes to determining the quality and quantity of translatable mRNAs (1, 2). Deadenylation is mediated by the CCR4-NOT (carbon catabolite repression 4–negative on TATA-less) complex, which is recruited to mRNA by RNA binding proteins (RBPs) or the microRNA repression complex, mainly by its scaffold subunit CNOT1. After recruitment, the CCR4-NOT complex catalyzes degradation of poly(A) through the two exonuclease subunits [CNOT6 (or CNOT6L) and CNOT7 (or CNOT8)], thereby regulating gene expression (14). We have previously identified CNOT3, a scaffold subunit of the CCR4-NOT complex, as a conserved regulator of heart function in Drosophila and mouse (5). Moreover, genome-wide association studies showed a strong association between single-nucleotide polymorphisms in CNOT1 or CNOT3 and prolonged QT intervals in humans (57). The underlying mechanisms of how the CCR4-NOT complex controls heart functions remained, however, elusive.

Autophagy is an evolutionally conserved mechanism in which lysosomes degrade cellular components and organelles, and this mechanism plays a crucial role in maintaining cellular energetics by recycling amino acids and fatty acids for energy production (8, 9). Autophagy can be protective and is generally beneficial in the heart under basal conditions and in response to stress, such as pressure overload and ischemic injury (10, 11). However, the activation of autophagy in some heart pathologies induces autophagic cell death or cell death through excessive autophagy (12, 13). There are several molecular and functional interactions between autophagy and apoptosis or necrosis/necroptosis (12, 14, 15). In Drosophila, genetic disruption of CPEB1, a CCR4-NOT–interacting RBP, enhances autophagic cell death in oocytes through impaired deadenylation and enhanced translation of ATG12 mRNA (16), suggesting functional interactions between mRNA deadenylation and autophagy.

The core autophagic machinery is composed of autophagy-related (ATG) components, which contribute to autophagosome formation and the subsequent fusion of autophagosomes with lysosomes to degrade substrates. Atg7 is a ubiquitin E1–like activating enzyme, which is involved in two ubiquitin-like conjugation systems: covalent attachment of Atg12 to Atg5 and Atg16l1 (Atg16-like 1) and of phosphatidylethanolamine to microtubule-associated protein 1 light chain 3 (Map1lc3 or LC3) in autophagosome formation (9). Several autophagy gene products also have nonautophagic functions, such as the proteolytic isoform of Atg5, which induces apoptosis (17), and the regulation of p53 transcriptional activity by Atg7 (18). Here, we report that in vivo inactivation of the CCR4-NOT complex resulted in altered mRNA deadenylation of the key autophagy regulator Atg7, resulting in cardiomyocyte death and heart failure.


Loss of muscle Cnot3 leads to lethal cardiomyopathy

We have previously generated whole-body Cnot3 mutant mice that exhibit embryonic lethality and reported that whole-body Cnot3 heterozygote mice develop heart failure (5). To directly examine the role of Cnot3 in cardiac muscle, we generated a Cnot3 floxed allele through a homologous recombination in embryonic stem (ES) cells (fig. S1A). After germline transmission, Cnot3 floxed mice were crossed with Ckmm-Cre transgenic (Tg) mice, which express Cre recombinase under the muscle creatine kinase promoter (Fig. 1A and fig. S1, B to D). Mice that lack Cnot3 gene expression in hearts and skeletal muscles are hereafter referred to as “Cnot3 mKO” mice. At around 3 weeks of age, Cnot3 mKO mice started to die, and all of these mice were dead by 30 days after birth (Fig. 1B). Despite the loss of Cnot3 protein in the skeletal muscle (fig. S1C), the skeletal muscles of Cnot3 mKO mice appeared normal, and the body weights of Cnot3 mKO mice were comparable to those of wild-type littermate mice (fig. S1E). However, Cnot3 mKO mice had substantially enlarged and dilated hearts (Fig. 1C) that weighed more (Fig. 1D).

Fig. 1 Severe heart failure by muscle-specific deletion of Cnot3 in mice.

(A) Western blot for CCR4-NOT (carbon catabolite repression 4–negative on TATA-less) complex subunits in the hearts of wild-type (WT) and Cnot3 muscle knockout mice (Cnot3 mKO) (representative of three mice per genotype in three independent experiments). (B) Postnatal survival curve for WT (n = 7) and Cnot3 mKO (n = 7) mice. Tissue samples were harvested at postnatal day 19 (arrowhead). (C) Macroscopic pictures and sections of the hearts of WT and Cnot3 mKO mice (representative of three mice per genotype). Scale bars, 1 mm. (D) Heart weights (HW) of WT (n = 7) and Cnot3 mKO (n = 7) mice. BW, body weight. (E and F) Heart function measurements of WT (n = 12) Cnot3 mKO (n = 9) mice. Representative M-mode echocardiography (E, left), left ventricular end diastolic diameter (LVEDD), left ventricular end systolic diameter (LVESD), percentage of fractional shortening (%FS) (E, right), representative electrocardiogram (ECG) chart (F, left), and corrected QT (QTc) interval (F, right) for WT and Cnot3 mKO mice at 19 days of age are shown. (G) Hematoxylin and eosin histology of hearts of WT and Cnot3 mKO mice (representative of three mice per genotype). Scale bars, 20 μm. (H) Myofibrils in WT and Cnot3 mKO mouse hearts (representative of three mice per genotype). Myofibrils (F-actin staining, red) and nuclei [4′,6-diamidino-2-phenylindole (DAPI), blue] are visualized. Scale bars, 20 μm. All values are means ± SEM. **P < 0.01, ***P < 0.001, unpaired two-tailed Student’s t tests.

Echocardiography of Cnot3 mKO mice showed severe cardiac contractility defects as assessed by fractional shortening (Fig. 1E), which was accompanied by long QT intervals and various arrhythmic changes as detected by electrocardiography (ECG) analysis (Fig. 1F and fig. S1F). Histological analysis of Cnot3 mKO mouse hearts revealed focal areas of dead cardiomyocytes with reduced cytoplasmic contents and vacuole formation (Fig. 1G). In addition, immunohistochemistry showed reduced myofibrils in Cnot3 mKO mouse hearts (Fig. 1H), consistent with our previous observation of myofibrillar disarray in the heart tubes of not3 RNAi (RNA interference) Drosophila lines (5). Cnot3 floxed mice were crossed with αMHC-MerCreMer Tg mice to induce cardiac muscle–specific deletion of Cnot3 at 3 months of age by tamoxifen treatment (Cnot3 cKO); adult mice with induced Cnot3 deletion developed lethal heart failure with the same structural and functional alterations as those in the young Cnot3 mKO mice (fig. S2, A to F). Thus, muscle-specific deletion of Cnot3 leads to structural and functional heart defects, ultimately resulting in lethal cardiomyopathy.

The CCR4-NOT complex is essential for cardiac homeostasis

Cnot3 is a subunit of the CCR4-NOT complex, and as expected from a previous in vitro study (19), it coimmunoprecipitated with Cnot1, Cnot6l, and Cnot7 from cardiac extract (Fig. 2A). In the hearts of Cnot3 mKO mice at around 3 weeks old, the protein abundance of Cnot1, a major scaffold for CCR4-NOT complex organization, was markedly decreased, whereas those of the deadenylase subunits Cnot6l and Cnot7 were apparently not changed (Fig. 1A), suggesting that Cnot3 is required to maintain Cnot1 protein stability and, as a consequence, the integrity of the CCR4-NOT complex in cardiomyocytes.

Fig. 2 CCR4-NOT depletion results in severe heart failure in mice.

(A) Coimmunoprecipitation of Cnot1, Cnot6l, and Cnot7 with Cnot3 from the lysates of mouse hearts (n = 2 independent experiments). (B) Western blot for Cnot1, Cnot3, Cnot6l, and Cnot7 in the hearts of Cnot1 muscle knockout (Cnot1 mKO) mice (n = 3 independent experiments). (C) Postnatal survival curve for WT (n = 6) and Cnot1 mKO (n = 15) mice. Tissue samples were harvested at postnatal day 9 (arrowhead). (D) Macroscopic pictures of the hearts (left) and heart weights (right) of WT (n = 13) and Cnot1 mKO (n = 12) mice. Scale bar, 1 mm. (E) Body weights (left) and skeletal muscle weights (right) of WT (n = 11) and Cnot1 mKO (n = 11) mice. Ga-MW, gastrocnemius muscle weight. (F and G) Heart function measurements of WT (n = 12) and Cnot1 mKO (n = 11) mice. Representative M-mode echocardiography (F, left), %FS (F, right), representative ECG chart (G, left), and QTc interval (G, right) for WT and Cnot1 mKO mice at postnatal day 9 are shown. (H) Hematoxylin and eosin histology of WT (representative of n = 3) and Cnot1 mKO (representative of n = 3) mouse hearts. Scale bars, 20 μm. All values are means ± SEM. **P < 0.01, unpaired two-tailed Student’s t tests.

To determine a potential role of Cnot1 in the heart, we next generated muscle-specific Cnot1 knockout mice by crossing Cnot1 floxed mice with Ckmm-Cre Tg mice (these mice are hereafter referred to as Cnot1 mKO mice) (Fig. 2B). Loss of Cnot1 in the heart muscle resulted in early lethality, with the mice dying at around days 10 to 15 after birth (Fig. 2C), which may be due to Cnot1 protein abundance being reduced more quickly in Cnot1 mKO mice than in Cnot3 mKO mice. Heart sizes and weights of the Cnot1 mKO mice were significantly increased (Fig. 2D), whereas the skeletal muscle weights and the body weights of Cnot1 mKO mice were unchanged (Fig. 2E). Echocardiography showed a marked decline of contractility in Cnot1 mKO mice at 10 days after birth (Fig. 2F), and ECG analysis demonstrated prolonged QT intervals in Cnot1 mKO mice (Fig. 2G). Histological analysis of the hearts of Cnot1 mKO mice revealed dying cardiomyocytes with reduced cytoplasmic contents and vacuole formation (Fig. 2H). These data show that genetic inactivation of the critical CCR4-NOT complex components Cnot1 and Cnot3 results in severe heart failure.

Autophagy protein expression is altered in Cnot3-depleted hearts, associated with cardiomyocyte death

Cnot3 depletion in mouse embryonic fibroblasts (MEFs) affects the expression of thousands of genes across the transcriptome (20). As a first step, we characterized cell death in Cnot3-depleted hearts. Although the number of terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL)–positive apoptotic cells was increased in the hearts of Cnot3 mKO mice compared with wild-type mice (fig. S3A), annexin V–positive apoptotic cells were not detectable among in vitro Cnot3 small interfering RNA (siRNA)–transfected cardiomyocytes (fig. S3B). However, the population of propidium iodide (PI)–positive dead cells was increased in cardiomyocytes with Cnot3 knockdown (fig. S3B), suggesting that Cnot3 depletion triggered the death of cardiomyocytes by necrosis/necroptosis. Transmission electron microscopy (TEM) analysis of Cnot3-depleted hearts showed focal areas of cardiomyocytes with disrupted actomyosin filaments, mislocalized mitochondria, lysis of cytoplasmic contents, and autophagic vacuoles (Fig. 3A), suggesting that Cnot3-deleted cardiomyocytes exhibit severe cell damage and potentially altered autophagy. The protein abundance of Ulk1, Pik3c3, Atg7, and p62 (also known as Sqstm1) was increased in Cnot3-deleted hearts, whereas LC3 (also known as Map1lc3b) protein abundance was decreased (Fig. 3B and fig. S4A). To address whether basal autophagy activity was altered by Cnot3 depletion, we treated Cnot3 mKO mice with bafilomycin A1, a protease inhibitor for autolysosomal protein degradation. Although the protein abundance of LC3-II, the activated form of LC3, was decreased in vehicle-treated Cnot3 mKO hearts compared with wild-type hearts, bafilomycin A1 treatment normalized the LC3-II abundance in Cnot3 mKO hearts to values comparable to wild-type hearts (Fig. 3C and fig. S4B). Consistently, the decrease in LC3 abundance in primary cardiomyocytes induced by Cnot3 depletion was restored by E64d and pepstatin A, a combination of protease inhibitors that prevents autolysosomal activation (Fig. 3D and fig. S4C). E64d and pepstatin A also restored LC3 abundance in MEFs in which Cnot3 was inducibly deleted by tamoxifen treatment (genotype: Cnot3flox/flox; CAG-Cre/Esr1Tg/+) (Fig. 3E and fig. S4, D and E). However, the increase in p62 abundance in Cnot3 mKO hearts was not affected by bafilomycin A1 treatment (Fig. 3C and fig. S4B). Thus, the loss of Cnot3 altered the protein abundance of several autophagy factors but had minor effects on autophagy flux.

Fig. 3 Autophagy protein abundance, but not autophagy flux, is altered in Cnot3-depleted heart.

(A) Transmission electron microscope (TEM) analysis for WT (representative of n = 3) and Cnot3 mKO (representative of n = 3) mouse hearts. Higher magnification of the colored rectangles (top) is shown in the bottom panel. Scale bars, 5 μm (top) and 1 μm (bottom). (B) Western blot for autophagy factors in the hearts of WT and Cnot3 mKO mice under fed and fasting conditions. AU, arbitrary units. Representative blots (left and fig. S3A) of n = 3 independent experiments are quantified for fed condition with WT (n = 5) and Cnot3 mKO (n = 5) mice (top right) and for fasting condition with WT (n = 3) and Cnot3 mKO (n = 3) mice (bottom right). (C) Western blot for LC3 (microtubule-associated protein 1 light chain 3) and p62 in the hearts of WT and Cnot3 mKO mice treated with or without bafilomycin A1 (Baf-A1). Representative blots (left and fig. S3C) of n = 3 independent experiments were quantified for WT mice treated with vehicle (n = 7), Cnot3 mKO mice treated with vehicle (n = 7), WT mice treated with Baf-A1 (n = 6), and Cnot3 mKO mice treated with Baf-A1 (n = 6). (D) Immunocytochemistry of LC3 in mouse cardiomyocytes transfected with Cnot3 (si-Cnot3) or control (si-Control) small interfering RNAs (siRNAs) and treated with or without E64d and pepstatin A (Pep) [n = 2 independent experiments with two different siRNAs for Cnot3 (fig. S3D)]. Scale bars, 20 μm. (E) Western blot for LC3 and p62 in WT and Cnot3 KO mouse embryonic fibroblasts (MEFs) treated with or without E64d plus pepstatin A (Pep) (n = 2 independent experiments). All values are means ± SEM. *P < 0.05, unpaired two-tailed Student’s t tests.

Poly(A) tail length and the stability of mRNAs encoding autophagy factors are regulated by Cnot3

Because the CCR4-NOT complex mediates mRNA deadenylation, we examined whether mRNAs encoding autophagy factors interacted with the CCR4-NOT complex using RNA immunoprecipitation and high-throughput sequencing (RIP-seq). RNA immunoprecipitation experiments with mouse heart lysates were performed with an antibody specific for Cnot3 and negative control immunoglobulin G (IgG), and RNA from total extracts (Input) and Cnot3-immunoprecipitated samples were sequenced (Fig. 4A and fig. S5, A to C). Differentially expressed gene (DEG) analysis revealed that 983 (of 24,421) protein-coding mRNAs were enriched only in Cnot3 RNA immunoprecipitates (data file S1). Gene ontology (GO) analysis showed that the mRNAs present only in Cnot3 RIP-seq DEGs were associated with the terms “transcription factor,” “histone modification,” “protein modification,” and “P-body” (Fig. 4B and data files S2 to S4). For mRNAs encoding autophagy factors, only Atg7 mRNA was detected in Cnot3 RNA immunoprecipitates (Fig. 4C). Quantitative polymerase chain reaction (qPCR) analysis showed that Atg7, Pik3c3, and Sqstm1 (encoding p62) mRNAs coimmunoprecipitated with Cnot3 (fig. S5D). qPCR analysis for mRNA expression showed that the mRNA abundance of Atg7, Atg5, and Sqstm1 was not changed in Cnot3 mKO hearts (Fig. 4D). By contrast, whereas Becn1, Atg12, and Map1lc3b (encoding LC3) mRNA abundance was decreased by Cnot3 depletion, Ulk1 and Pik3c3 mRNA abundance slightly increased (Fig. 4D). These results suggested that the protein abundance of Ulk1, Pik3c3, Atg7, and p62 was posttranscriptionally increased in Cnot3 mKO hearts.

Fig. 4 Poly(A) tail length and stability of autophagy factor–encoding mRNAs are regulated by the CCR4-NOT complex.

(A) RIP-seq (RNA immunoprecipitation and high-throughput sequencing) plot of the applied cutoff to identify significantly bound or unbound mRNAs by Cnot3 in the heart. (B) Gene ontology (GO) enrichment analysis for Cnot3 differentially expressed genes (DEGs). (C) Autophagy factor–encoding mRNAs in Cnot3 immunoglobulin G (IgG) (αCnot3) RIP or control IgG RIP, normalized to Input, for two biological replicates (Rep1 and Rep2). The symbol # indicates that Atg7 was selected as a gene enriched in Cnot3 RIP by R Bioconductor package edgeR with the criteria of false discovery rate (FDR) < 0.05, counts per million (cpm) for mean Input > 0.1, and αCnot3 RIP/Input > 20.5. (D) Quantitative reverse transcription polymerase chain reaction (qRT-PCR) to measure the expression of autophagy factor–encoding mRNAs in the hearts of WT (n = 6) and Cnot3 mKO (n = 6) mice (n = 2 independent experiments). (E) Polyadenylate [poly(A)] tail length measurements of autophagy factor–encoding mRNAs in hearts (left) and mouse cardiomyocytes (right). Total RNA was subjected to PCR-based poly(A) tail length analysis. Representative results for the hearts of WT (n = 3) and Cnot3 mKO (n = 3) mice at 18 days old (left) and cardiomyocytes transfected with si-Cnot3 or siRNA combinations for Cnot6, Cnot6l, Cnot7, and Cnot8 (si-Cnot6/6l/7/8) (right). Control PCR without poly(A) regions are shown in figs. S5E and S6E (n = 3 independent experiments). nt, nucleotide. (F) The stability of autophagy factor–encoding mRNAs in cardiomyocytes was analyzed after actinomycin D treatment (n = 3 independent experiments). The data obtained with a second set of siRNAs for Cnot3 and Cnot6/6l/7/8 are shown in fig. S6 (A to F). All values are means ± SEM. *P < 0.05, **P < 0.01, unpaired two-tailed Student’s t tests.

We next measured the poly(A) tail length of autophagy factor–encoding mRNAs. The lengths of the poly(A) tails of Ulk1, Pik3c3, Atg5, Atg7, and Sqstm1 mRNAs were considerably longer in Cnot3-deleted hearts (Fig. 4E), and these longer poly(A) tails were more evident at postnatal day 18 than at postnatal day 8 (fig. S5E), suggesting that the deadenylase activity of the CCR4-NOT complex was impaired after Cnot3 deletion. To ask whether the longer poly(A) tail is also observed upon depletion of CCR4-NOT deadenylase subunits, we further investigated the poly(A) tail length and mRNA stability in primary mouse cardiomyocytes. siRNA knockdown of Cnot3 or all the deadenylase subunits (Cnot6, Cnot6l, Cnot7, and Cnot8) efficiently decreased the mRNA and protein abundance in cardiomyocytes (fig. S6, A and B). The mRNA abundance of Pik3c3, Atg5, and Sqstm1 was increased in Cnot3 siRNA–treated cardiomyocytes compared to those in control siRNA–treated cells, and Cnot6/6l/7/8-depleted cardiomyocytes showed similar or greater increase in Pik3c3, Atg5, and Sqstm1 mRNA abundance (fig. S6C). Similarly, the protein abundance of Atg5, Atg7, and p62 was increased in cardiomyocytes treated with either Cnot3 or Cnot6/6l/7/8siRNAs (fig. S6D). The poly(A) tails of Ulk1, Pik3c3, Atg5, Atg7, and Sqstm1 mRNAs were longer in cardiomyocytes treated with Cnot3 or Cnot6/6l/7/8siRNAs than those in cardiomyocytes that received control siRNA (Fig. 4E and fig. S6E). Furthermore, Ulk1, Pik3c3, Atg5, Atg7, and Sqstm1 mRNAs were more stable in cardiomyocytes treated with Cnot3 or Cnot6/6l/7/8siRNAs than in control siRNA–treated cardiomyocytes (Fig. 4F and fig. S6F). These results indicate that in Cnot3-depleted hearts, mRNAs encoding select autophagy factors have longer poly(A) tails and are stabilized because of decreased deadenylase activity of the CCR4-NOT complex, resulting in increased protein abundance.

Loss of Atg7 improves cardiac dysfunction in CCR4-NOT complex–depleted mice

Ulk1 and Pik3c3 are involved in autophagy initiation, and Atg7 executes autophagy by catalyzing the covalent attachment of Atg12 to Atg5 and that of phosphatidylethanolamine to LC3 with its ubiquitin E1–like enzyme activity (9). To ask whether the increased abundance of autophagy factors was functionally involved in heart failure of Cnot3 mKO mice, we investigated whether simultaneous knockout of Atg7 and Cnot3 genes might rescue the cardiac dysfunction of Cnot3 mKO mice. We crossed Atg7 floxed mice and Cnot3 floxed mice onto the Ckmm-Cre Tg background to delete both Atg7 and Cnot3 in cardiac muscle (Cnot3;Atg7 dmKO mice) (Fig. 5A). Loss of Atg7 impaired autophagosome formation as shown by a complete loss of LC3-II and an increase in p62 abundance (Fig. 5A). Although all of Cnot3 mKO mice died within 4 weeks after birth, Cnot3;Atg7 dmKO mice survived to ~7 weeks after birth (Fig. 5B). Deletion of Atg7 also significantly attenuated the increased heart weight of Cnot3 mKO mice to values comparable to control wild-type mice (Fig. 5C). The impaired heart contractility and the longer QT interval in Cnot3 mKO mice were also restored to wild-type values in Cnot3;Atg7 dmKO mice (Fig. 5, D and E). Histological analysis showed reduced cell death in Cnot3;Atg7 dmKO mice compared with Cnot3 mKO mice (Fig. 5F), and the reduced myofibril content in Cnot3 mKO mice was partially reversed in Cnot3;Atg7 dmKO mice (Fig. 5G). We next generated adult mice with a double knockout of Cnot3 and Atg7 (Cnot3;Atg7 dcKO mice) by crossing Atg7 floxed mice and Cnot3 floxed mice onto the αMHC-MerCreMer Tg background and treating the resulting mice with tamoxifen. Adult Cnot3;Atg7 dcKO mice survived for longer and showed increased heart contractility and reduced cardiomyocyte death compared to the single Cnot3 cKO mice (fig. S7, A to F) at 14 days after tamoxifen treatment. To clarify the effects of Cre overexpression, we generated the Cnot3;Atg7 dcKO mice homozygous for αMHC-MerCreMer Tg, namely, double-Cre Cnot3;Atg7 dcKO mice (genotype: Cnot3f/f;Atg7;f/fMHC-MerCreMerTg/Tg) and compared them with the mice heterozygous for αMHC-MerCreMer Tg; single-Cre Cnot3;Atg7 dcKO mice (genotype: Cnot3f/f;Atg7;f/fMHC-MerCreMerTg/+) (fig. S7D). The phenotypic rescue of cardiac dysfunction in adult Cnot3 cKO mice by double knockout of Cnot3 and Atg7 (Cnot3;Atg7 dcKO mice) was comparable between double-Cre and single-Cre expressing mice (fig. S7E), indicating that the overexpression of Cre recombinase did not affect the observed phenotype. In addition, the overexpression of ATG7 increased the population of PI-positive dead cells in Cnot3 siRNA–transfected cardiomyocytes but not in control siRNA–transfected cells (fig. S8, A and B). Furthermore, we deleted both Cnot1 and Atg7 on the Ckmm-Cre Tg background in cardiac muscle [Cnot1;Atg7 dmKO mice (genotype: Cnot1f/f;Atg7;f/f;Ckmm-CreTg/+)] (Fig. 6A). Cnot1;Atg7 dmKO mice survived longer than Cnot1 mKO mice (Fig. 6B). Deletion of Atg7 significantly attenuated the impaired heart contractility and longer QT interval in Cnot1 mKO mice to wild-type values (Fig. 6, C and D). Histological analysis showed reduced cell death in Cnot1;Atg7 dmKO mice compared with Cnot1 mKO mice (Fig. 6E).

Fig. 5 Atg7 promotes cardiac dysfunction in Cnot3 mKO mice.

(A) Western blot for autophagy proteins in the hearts. The hearts of WT, Cnot3 mKO, Cnot3;Atg7 double muscle knockout (Cnot3;Atg7 dmKO), and Cnot3;Atg5 double muscle knockout (Cnot3;Atg5 dmKO) mice at 18 days old under normal diet feeding were harvested. Gapdh was used as a loading control (n = 2 independent experiments). (B) Postnatal survival curve for WT (n = 6), Cnot3 mKO (n = 11), Atg7 mKO (n = 5), and Cnot3;Atg7 dmKO (n = 5) mice. (C) Macroscopic images of the hearts of WT (representative of n = 3), Cnot3 mKO (representative of n = 3), Atg7 mKO (representative of n = 3), Cnot3;Atg7 dmKO (representative of n = 3) mice (left). Heart weight–to–body weight ratios (HW/BW) of WT (n = 4), Cnot3 mKO (n = 4), Atg7 mKO (n = 3), and Cnot3;Atg7 dmKO (n = 4) mice at 18 days after birth (right). Scale bars, 2 mm. (D) Representative M-mode echocardiography (left) and %FS (right) at 18 days of age for WT (n = 5), Cnot3 mKO (n = 5), Atg7 mKO (n = 3), and Cnot3;Atg7 dmKO (n = 5) mice. (E) ECG measurements. Representative ECG chart (left) and QTc interval (right) of WT (n = 9), Cnot3 mKO (n = 9), Atg7 mKO (n = 3), and Cnot3;Atg7 dmKO (n = 5) mice at postnatal day 18 are shown. (F) Hematoxylin and eosin histology of the hearts of WT (representative of n = 3), Cnot3 mKO (representative of n = 3), Atg7 mKO (representative of n = 2), and Cnot3;Atg7 dmKO (representative of n = 3) mice. Scale bars, 20 μm. (G) Myofibrils of the heart sections of Cnot3 mKO (representative of n = 3) and Cnot3;Atg7 dmKO (representative of n = 3) mice at 18 days of age. FU, fluorescence units. Myofibrils (F-actin staining, red) and nuclei (DAPI, blue) were visualized. Scale bars, 20 μm. All values are means ± SEM. **P < 0.01, ***P < 0.001, unpaired two-tailed Student’s t tests.

Fig. 6 Atg7 promotes cardiac dysfunction in Cnot1 mKO mice.

(A) qRT-PCR analysis of Cnot1 and Atg7 expression in the hearts of WT, Cnot1 mKO, Atg7 mKO, and Cnot1;Atg7 dmKO mice at 8 days old (n = 3 independent experiments). (B) Postnatal survival of WT (n = 18), Cnot1 mKO (n = 18), Atg7 mKO (n = 5), and Cnot1;Atg7 dmKO (n = 17) mice. (C) Representative M-mode echocardiography (left) and %FS (right) for WT (n = 13), Cnot1 mKO (n = 6), Atg7 mKO (n = 3), and Cnot1;Atg7 dmKO (n = 7) mice at 8 days old. (D) ECG measurements. Representative ECG chart (left) and QTc interval (right) of WT (n = 11), Cnot1 mKO (n = 6), Atg7 mKO (n = 3), and Cnot1;Atg7 dmKO (n = 5) mice at postnatal day 8 are shown. (E) Hematoxylin and eosin histology of the hearts of WT (representative of n = 3), Cnot1 mKO (representative of n = 3), and Cnot1;Atg7 dmKO (representative of n = 3) mice. Scale bars, 50 μm. All values are means ± SEM. *P < 0.05, **P < 0.01, unpaired two-tailed Student’s t tests.

To explore whether canonical autophagy was responsible for the rescue of heart failure in Cnot3 mutant mice, we also generated Atg5 floxed, Cnot3 floxed, and Ckmm-Cre Tg mice (Cnot3;Atg5 dmKO mice) (Fig. 5A and fig. S9A). Similar to Atg7, the loss of Atg5 also critically impaired autophagosome formation in the heart (Fig. 5A), which is consistent with a previous study showing essential roles of Atg5 and Atg7 in canonical autophagy (21). However, double knockout of Atg5 and Cnot3 (Cnot3;Atg5 dmKO) did not rescue the reduced survival, increased heart weight, impaired contractility, and conduction defects observed in the single Cnot3 mKO mice (fig. S9, B to E), indicating that autophagy per se was not responsible for the rescue of the Cnot3 KO heart phenotype. These data indicate that Atg7 promotes cardiac dysfunction of Cnot3 mKO mice independently of the canonical autophagy pathway.

Nuclear Atg7 regulates p53 activity to induce the expression of cell death genes in Cnot3-depleted cardiomyocytes

Consistent with previous studies on CCR4-NOT depletion in cancer cells and B cells (19, 22), p53 protein abundance was increased in the hearts of Cnot3 mKO mice (Fig. 7A). In response to starvation, Atg7 localizes to nucleus and promotes p53-mediated transcription of p21, which encodes a cell cycle inhibitor, in fibroblasts or HCT116 colon carcinoma cells (18). We thus examined the localization of Atg7 proteins in Cnot3 KO MEFs. Immunocytochemistry showed that Atg7 and p53 were detected in the nucleus of Cnot3 KO MEFs (Fig. 7B and fig. S10A), and subcellular fractionation also showed that the nuclear amounts of both Atg7 and p53 were increased in Cnot3 KO MEFs compared with control cells (Fig. 7C). In Cnot3 siRNA–transfected cardiomyocytes, the nuclear localization of Atg7 and p53 was also increased compared with control siRNA–transfected cardiomyocytes (Fig. 7D and fig. S10B). Atg7 and p53 coimmunoprecipitated from Cnot3 mKO heart lysates, indicating that Atg7 and p53 interacted in Cnot3-depleted hearts (Fig. 7E and fig. S10C). Moreover, simultaneous transfection of p53 siRNA or Atg7 siRNA with Cnot3 siRNA decreased the numbers of PI-positive dead cardiomyocytes (Fig. 7F and fig. S10, D to F), suggesting that p53 is involved in cell death in Cnot3-depleted hearts.

Fig. 7 Atg7 regulates p53 activity to induce expression of cell death–associated genes under Cnot3 depletion.

(A) Western blot for p53 in the hearts of WT and Cnot3 mKO mice. Each lane represents an individual mouse (three independent experiments). (B) Immunocytochemistry of Atg7 and p53 in WT and Cnot3 KO MEFs (n = 3 independent experiments). Scale bars, 20 μm. (C) Western blot for Atg7 and p53 in the nuclear fraction of Cnot3 KO MEFs. Representative images (left) and quantification results (right) of n = 6 independent experiments are shown.*P < 0.05, **P < 0.01, paired two-tailed Student’s t tests. (D) Immunocytochemistry of Atg7 in mouse cardiomyocytes. Cnot3 siRNA (si-Cnot3) or control siRNA (si-Control) was transfected into cardiomyocytes, which were immunostained for Atg7 (n = 2 independent experiments). Scale bars, 20 μm. (E) Coimmunoprecipitation of Atg7 and p53. Heart lysates from WT or Cnot3 mKO mice at 18 days old were immunoprecipitated with Atg7 IgG (αAtg7) and immunoblotted for Atg7 or p53 (n = 2 independent experiments). (F) Cell death assessed by propidium iodide (PI) uptake in mouse cardiomyocytes transfected with si-Cnot3 or si-Control in combination with siRNAs for Atg7 or p53. Scale bars, 50 μm. (G) qRT-PCR for mRNA expression of cell death–associated genes in hearts (n = 3 independent experiments). ns, not significant. (H) Chromatin immunoprecipitation (ChIP)–quantitative polymerase chain reaction for Atg7-bound promoter regions of Puma (left) and Ripk3 (right) in Cnot3 KO MEFs transfected with p53 siRNA or control siRNA (n = 3 independent experiments). All values are means ± SEM. #P < 0.1, *P < 0.05, **P < 0.01, unpaired two-tailed Student’s t tests, unless otherwise stated.

We next examined p53 target gene expression in Cnot3-deleted hearts. Although ATG7 induces p21 expression in nutrient-deprived cells (18), p21 expression was not significantly increased in Cnot3-depleted cardiomyocytes, which was also not affected by additional deletion of Atg7 (Fig. 7G). By contrast, genes encoding cell death factors, such as Puma, Bax, and Ripk3, showed increased expression in the hearts of Cnot3 mKO mice (Fig. 7G). The increased mRNA expression of Puma and Ripk3, but not Bax, was decreased to varying extents by double knockout of Cnot3 and Atg7 (Fig. 7G). Moreover, Puma protein abundance was consistently increased in Cnot3-deleted hearts but decreased by additional knockout of Atg7 (fig. S11A). Puma protein abundance is largely regulated at the transcriptional level through both p53-dependent and p53-independent mechanisms (23). Chromatin immunoprecipitation (ChIP) showed that in Cnot3 KO MEFs, Atg7 was bound to the genomic region close to the transcriptional start site of the Puma and Ripk3 genes (fig. S11B), overlapping the region where p53 binds and induces gene expression (fig. S11B) (23). Moreover, siRNA-mediated knockdown of p53 decreased the binding of Atg7 protein to the Puma and Ripk3 gene loci in Cnot3 KO MEFs (Fig. 7H and fig. S11C). Cnot3 RIP-seq analysis using mouse heart lysates (Fig. 4, A and B, and table S1) showed that Cnot3 protein bound to Puma mRNA, but not Bax, Ripk1, and Ripk3 mRNAs (fig. S11D). Puma mRNA expression was not stabilized but markedly increased by Cnot3 depletion in cardiomyocytes (fig. S11, E and F), suggesting that the increase in Puma expression in Cnot3-deleted hearts was primarily mediated through increased transcription. These results indicate that under CCR4-NOT–depleted conditions, Atg7 regulates p53 activity to induce expression of cell death–associated genes.


Here, we demonstrated that mRNA deadenylation of the autophagy regulator Atg7 through the CCR4-NOT complex was crucial to maintain cardiomyocyte survival, cardiac contractility, and proper QT intervals. Cnot3 interacted with nearly 1000 mRNAs including Atg7 and regulated poly(A) tail shortening and mRNA decay of Atg7 mRNA. Loss of Cnot3 led to increased expression and noncanonical activation of Atg7 to promote p53-induced expression of Puma and Ripk3, thereby accelerating cardiomyocyte death.

Although the loss of Cnot3 had minor effects on canonical autophagy process, our genetic data revealed a molecular connection between Cnot3 and Atg7, in which simultaneous deletion of Atg7 and Cnot3 markedly slowed the damage to Cnot3-deleted hearts. In autophagy flux measurements, we observed discrepant changes in autophagy markers. Decreased LC3-II abundance in Cnot3 mKO hearts was restored to control values with bafilomycin A1, which was indicative of increased autophagy flux, but counter to this observation was the lack of further increase in p62 abundance. These finding may point to previously unknown pathways and biology. In Cnot3-mutated cells, Atg7 bound to p53 in the nucleus where it localized to transcriptional start sites of Puma and Ripk3, close to the p53-binding motif. The interaction of Atg7 and p53 in CCR4-NOT–depleted cardiomyocytes was likely involved in inducing cell death, whereas the same interaction occurring in nutrient-deprived proliferating cells halts cell cycle progression and thereby prevent cell damage (18). In addition, inducible overexpression of Atg7 in the adult hearts does not induce overt cell death phenotypes (24). We thus anticipated that the overexpression of Atg7 alone was not sufficient to induce cardiomyocyte death and that Atg7-mediated cell damage required Cnot3 depletion. We assume that impaired mRNA deadenylation and/or dysregulation of other autophagy genes were necessary for Atg7-mediated cell damage. The mechanism by which CCR4-NOT depletion facilitates binding of Atg7 and p53 proteins to the genomic regions of Puma and Ripk3 is currently unknown. Because our TEM analysis also showed abnormal structures of nuclear membranes and RIP-seq analysis showed that GO terms for transcription factors were enriched in Cnot3-bound RNAs, CCR4-NOT depletion might alter chromatin architecture and accessibility to transcription factors. Nevertheless, our findings strengthen the biological relevance of Atg7 and p53 interactions and may warrant further studies on Atg7 functions.

Puma and Ripk3 promote necrosis and/or apoptosis of cardiomyocytes in mouse heart failure models (2527). Deletion of Puma attenuates cardiomyocyte apoptosis that is induced by pressure-overload stress (26). We did not find clear evidence that Cnot3 depletion increased cardiomyocyte apoptosis. On the other hand, reduced cytoplasmic contents and vacuole formation in cardiomyocytes in vivo and increased population of PI-positive cells in vitro suggested the involvement of necrosis and autophagic cell death in Cnot3-deleted cardiomyocytes. Thus, necrosis/necroptosis may be the primary cause of death in Cnot3-depleted cardiomyocytes. The necroptotic kinases Ripk1 and Ripk3 are involved in ischemia-reperfusion injury of the heart (25, 28), and in this study, we showed that Ripk3 expression was increased by Atg7 in Cnot3-depleted hearts, which is mechanistically distinct from the Cnot3-mediated decrease in Ripk1 expression through mRNA deadenylation in MEFs (20). These results suggest the existence of heart-specific mechanisms of cell survival. We have previously reported that in the heart failure model induced by pressure overload, Cnot3 heterozygous mice show enhanced cardiac fibrosis, which represents cardiomyocyte damage and tissue remodeling (5). Thus, we speculate that Cnot3 is crucial for survival of cardiomyocytes under pathological conditions. However, long QT interval and arrhythmic changes in Cnot3 mKO mice may or may not be secondary to the advanced myocardial dysfunction and cell death. Although detailed molecular mechanisms need to be further explored, our genetic models of Cnot1 or Cnot3 deficiency in hearts demonstrate the importance of the CCR4-NOT complex in cardiac homeostasis.

The CCR4-NOT complex regulates gene expression through both transcriptional and posttranscriptional mechanisms (1). Our data indicate that the expression of Ulk1, Pik3c3, Atg7, and Sqstm1 was posttranscriptionally suppressed by the CCR4-NOT complex through decreased mRNA stability and, possibly, translation suppression. Why Atg7, out of almost 1000 RNAs bound to Cnot3 and/or CCR4-NOT target genes, is important for cell death in Cnot3-depleted hearts is currently unknown, but the enrichment of transcription factors in Cnot3-bound mRNAs may be related to the connection between the phenotypes caused by the loss of Cnot3 and Atg7-regulated p53 transcriptional activity. In the mammalian CCR4-NOT complex, CNOT1 and CNOT3 form a stable core bound to other subunits but do not directly bind to mRNAs. The recognition of autophagy factor–encoding mRNAs by the CCR4-NOT complex is, hence, likely mediated through RBPs that interact with the core CCR4-NOT complex. Future work will be required to identify the RBPs that bind to mRNAs that encode autophagy factors and to the CCR4-NOT complex.

In summary, our findings linking mRNA deadenylation to Atg7 gene regulation uncover a cell survival pathway required for cardiac homeostasis. Modulating poly(A) mRNA tail length and/or targeting the prodeath effect of nuclear Atg7 might be candidate strategies for treating heart diseases. Furthermore, because compounds to activate canonical autophagy are being developed to treat various diseases, our results might serve as a cautionary warning for the potential side effects of such compounds and could contribute to the development of better autophagy-targeting therapeutics.


Cardiac gene knockout of Cnot3 or Cnot1 in mice

A targeting vector was constructed to flank exons 2 and 3 of the murine Cnot3 gene by loxP. The linearized construct was electroporated into A9 ES cells derived from 129/Ola and C57BL/6J hybrids. The correctly targeted ES cell clones were processed to blastocyst injection to generate chimeric mice, which were then crossed with FLPe Tg mice to delete Neo cassette and obtain Cnot3 floxed allele. Cnot3 floxed mice were further crossed with muscle creatine kinase promoter Cre Tg mice [Ckmm-Cre Tg mice (29)] to generate muscle-specific Cnot3 knockout (Cnot3 mKO) mice. Heart-specific tamoxifen-inducible Cnot3 knockout mice (Cnot3 cKO) were generated by crossing Cnot3 floxed mice with αMHC-MerCreMer Tg mice, and deletion of Cnot3 in adult mice was induced by five consecutive days of intraperitoneal injection of 4-hydroxytamoxifen (4-OHT) (20 mg/kg per day; H6278, Sigma-Aldrich), as described previously (30). For conditional deletion of Cnot3 in MEFs, we generated CAG promoter–driven tamoxifen-inducible Cnot3 knockout mice (Cnot3f/f; CAG-cre/Esr1*5AmcTg/+ mice) (31). Double-mutant mice carrying mutations in both Cnot3 and Atg7 (Cnot3;Atg7 dmKO or Cnot3f/f; Atg7f/f; Ckmm-Cre Tg mice) (Cnot3;Atg7 dcKO or Cnot3f/f; Atg7f/f; αMHC-MerCreMer Tg mice) or in both Cnot3 and Atg5 (Cnot3;Atg5 dmKO or Cnot3f/f; Atg5f/f; Ckmm-CreTg mice) were generated and intercrossed more than 10 times (21, 32). Cnot1 flox mice were generated by homologous recombination in ES cells, in which loxP sites flank exons 21 and 22 of the Cnot1 gene (accession no. CDB0916K, RIKEN), and muscle-specific Cnot1 knockout (Cnot1 mKO) mice were generated similarly by crossing Cnot1 flox mice with Ckmm-Cre Tg mice. A residual reactivity to the antibody or qPCR amplification of genes in the mKO hearts is likely derived from the nonmuscle cells in the heart (Figs. 2B and 6A). Mice were genotyped by PCR and Southern blotting and maintained at the animal facilities of Akita University Graduate School of Medicine. All animal experiments conformed to the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health (NIH Publication No. 85-23, revised 1996). Approvals for the experiments were granted by the ethics review board of Akita University.

Echocardiography and ECG measurements

Echocardiographic measurements were performed as described previously (33). Briefly, mice were anesthetized with 1% isoflurane/oxygen, and echocardiography was performed using Vevo 770 equipped with a 30-MHz linear transducer. Fractional shortening was calculated as follows: [(LVEDD − LVESD)/LVEDD] × 100. We used two-dimensional (2D)–guided M-mode measurements to determine the percentage of fractional shortening (%FS). The heart was first imaged in 2D mode in the parasternal short-axis view. From this view, an M-mode cursor was positioned perpendicular to the interventricular septum and posterior wall of the left ventricle at the level of the papillary muscles. M-mode images were obtained for measurement of wall thickness and chamber dimensions with the use of the leading-edge convention adapted by the American Society of Echocardiography. For measurements of ECG, the anesthetized mice were placed on a heating pad with continuous monitoring of body temperature for three-lead ECG measurements in lead(II) for over 10 min using pad electrodes and a PowerLab 26T system (ADInstruments). Recordings (16 bit, 2 kHz/channel) were analyzed using the LabChart v7.0 program (ADInstruments) and filtered between 0.5 and 500 Hz. Corrected QT intervals were obtained using the formula QT/(RR/100)1/2, as described previously (34).


For histology, hearts were arrested with 1 M KCl, fixed with 10% formalin, and embedded in paraffin. Sections (5 μm thick) were then cut and stained with hematoxylin and eosin. To detect apoptotic cells, we performed TUNEL assay using ApopTag Peroxidase In Situ Apoptosis Detection Kit (Chemicon). To detect fibrotic areas, we stained sections with Masson’s trichrome stain. To visualize filamentous actin (F-actin) as myofibrils, we cryosectioned frozen hearts with a thickness of 8 to 10 μm and probed them with Alexa Fluor 546–labeled phalloidin and 4′,6-diamidino-2-phenylindole (DAPI) (Molecular Probes). For TEM analyses, heart tissues were fixed by a conventional method. Fixed samples were embedded in Epon 812, and thin sections were then cut and stained with uranyl acetate and lead citrate for observation under a Jeol-1010 electron microscope (Jeol) at 80 kV (35).

RNA analyses

Tissue RNA was extracted using TRIzol reagent (Invitrogen), and RNA from cells was extracted with RNeasy Mini Kit (Qiagen). For real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR), cDNA was synthesized using the PrimeScript RT reagent kit (TAKARA), and real-time PCR was run in 96-well plates using a SYBR Premix Ex Taq II (TAKARA) according to the manufacturer’s instructions. Relative gene expression was quantified using the Thermal Cycler Dice Real Time System II software (TAKARA). All primers used in qRT-PCR are listed in table S5. To assess mRNA stability, we treated cells with actinomycin D (2.5 μg/ml; Wako). Total RNA was extracted at the indicated time points after actinomycin D treatment and subjected to qRT-PCR. Measurement of the poly(A) length was performed as described previously (36). Briefly, heart RNAs (150 ng) were subjected to reverse transcription with oligo(dT) anchor primer (5′-GCGAGCTCCGCGGCCGCGTTTTTTTTTTTT-3′) using Cloned AMV First-Strand cDNA Synthesis Kit (Invitrogen). The 3′ ends of synthesized cDNAs were amplified with oligo(dT) anchor and gene-specific primers.

RNA immunoprecipitation and high-throughput sequencing

RNA-CNOT3 complexes were immunoprecipitated from mouse heart lysates using anti-CNOT3 antibody (20) or control IgG (MBL) antibodies, as described previously (37). RNAs in immunoprecipitates were purified with RIP-Assay Kit (MBL). Total RNA extracted from heart lysates (Input) or immunoprecipitated with control IgG or anti-CNOT3 IgG was used for RNA-seq library preparation with TruSeq Stranded mRNA Sample Prep kit (Illumina). Single-end read RNA-seq [36 base pair (bp)] was performed with HiSeq3000 (Illumina). For data analysis, we used a conventional method commonly applied to RNA-seq data to process RIP-seq data (38). Raw FASTQ files were assessed to remove low quality reads by Trimmomatic version 0.3.6 (39). High-quality reads were subsequently aligned to UCSC mm10 as the reference genome by Bowtie2 version 2.2.5 with TopHat version 2.1.0. Raw read count data were extracted, normalized, and analyzed to obtain DEGs by R Bioconductor version 3.5 packages Rsubread and edgeR. The tools used have been described previously (40). The read count data showed sufficient sequencing depth in terms of the total number assigned reads and good correlations among technical replicates (fig. S5, B and C). Genes were detected as differentially expressed if a false discovery rate calculated by the Benjamini-Hochberg method was less than 0.05, gene expression for the mean value of Input samples (two replicates) measured by counts-per-million value is greater than 0.1, and the fold change over Input is greater than 20.5 = 1.414. GO enrichment analysis was implemented by R Bioconductor package clusterProfiler (41). GO terms were selected by P values with a cutoff threshold of 0.01. The RIP-seq data generated from Cnot3 RIP (n = 2 independent experiments), control IgG RIP (n = 2 independent experiments), and RNA from total extracts (n = 2 independent experiments) have been deposited at the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) database ( under the accession number GSE103629.

Cell cultures

Primary cardiomyocytes were isolated from prenatal mouse hearts of wild-type mice, as described previously (42). Briefly, hearts were excised and rapidly minced into three or four pieces and digested with collagenase (Wako) for 45 min at 35°C. Cardiomyocytes were collected, preplated to exclude noncardiomyocytes, and plated on gelatinized culture dishes or plates with siRNAs for control, Cnot3, Atg7, p53, or combinations of Cnot6, Cnot6l, Cnot7, and Cnot8 siRNAs. siRNA target sequences are listed in table S6. Twenty-four hours after plating, control plasmid or pCMV-hATG7 (43) was transfected. At 72 hours after plating, cardiomyocytes were subjected to various assays. For inducible deletion of Cnot3 in MEFs, MEFs were isolated from 13.5 days post-coitum embryos from the crossing of Cnot3f/f;CAG-Cre/Esr1*5AmcTg/+ mice and Cnot3f/f mice, and the genotypes were determined by PCR. When plating, the Cnot3f/f;CAG-Cre/Esr1*5AmcTg/+ MEFs (passages 4 to 7) were treated with 2 μM 4-OHT to obtain Cnot3 KO MEFs, and vehicle-treated Cnot3f/f;CAG-Cre/Esr1*5AmcTg/+ MEFs or Cnot3f/f MEFs served as wild-type MEFs. At 48 or 72 hours after plating, MEFs were harvested for various assays.

Immunoprecipitation and Western blotting

Heart proteins were extracted with a TNE lysis buffer [50 mM tris, 150 mM NaCl, 1 mM EDTA, 1% NP-40, protease inhibitor (cOmplete Mini, Roche), 100 mM NaF, and 2 mM Na3VO4], as described previously (42). Heart lysates were precleared with Protein G–Sepharose (GE Healthcare) for 1 hour at 4°C, and proteins in the supernatant were immunoprecipitated with anti-CNOT3 (clone 4B8; Abnova), anti-ATG7 (43), anti-p53 (1C12, Cell Signaling Technology), or control IgG at 4°C overnight. Immune complexes were washed five times with the TNE lysis buffer followed by mixing with LDS sample buffer (Invitrogen). After sonication and denaturation with the LDS sample buffer (Invitrogen) at 70°C, proteins were electrophoresed on NuPAGE bis-tris precast gels (Invitrogen) and transferred to nitrocellulose membranes (0.2 μm pore; Invitrogen). Membranes were probed with the following antibodies: CNOT1, CNOT3, CNOT6L, CNOT7, and ATG7, as described previously (4345), and commercially obtained CNOT3 (clone 4B8; Abnova), Ulk1 (D8H5, Cell Signaling Technology), Pik3c3 (3811, Cell Signaling Technology), Becn1 (D40C5, Cell Signaling Technology), ATG5 (A0731, Sigma-Aldrich), hnRNPC (R5028, Sigma-Aldrich), LC3 (2775, Sigma-Aldrich), Puma (14570, Cell Signaling Technology), p62 (PM045, MBL), α-tubulin (T5168, Sigma-Aldrich), and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (14C10, Cell Signaling Technology). The bands were visualized with ECL reagent (GE Healthcare) using ChemiDoc Touch Imaging System (Bio-Rad). Image Lab software was used to quantify band intensity.


Mouse cardiomyocytes on LabTek chambers (177437, Thermo Fisher Scientific) or MEFs on coverslips were fixed with 4% paraformaldehyde and incubated with antibodies for ATG7 (43), p53 (1C12, Cell Signaling Technology), LC3 (2775, Cell Signaling Technology), or p62 (PM045, MBL) and then incubated with appropriate secondary antibodies. LabTek chambers were mounted with mounting medium containing DAPI. Apoptotic cells were detected using Annexin V-FITC Apoptosis Detection Kit (BioVision) according to the manufacturer’s instructions. Necrotic/necroptotic cells were detected by PI incorporation. Cells were treated with PI (5 μg/ml; Sigma-Aldrich) and Hoechst (100 μg/ml; Thermo Fisher Scientific) for 10 min at room temperature and analyzed using fluorescence microscopy. The cell death rate was calculated as ratio of PI-to-Hoechst incorporation in nuclei. Images were analyzed using multiphoton laser microscopy (A1R MP, Nikon) or fluorescence microscopy (BZ9000, Keyence).

ChIP assay

Wild-type and Cnot3 KO MEFs at 72 hours after 4-OHT treatment were cross-linked with 1% paraformaldehyde for 5 min and quenched with glycine, cell lysates were harvested, and chromatin DNAs were shared to a size of 300 to 600 bp by using the Picoruptor (Diagenode). ChIP assays were performed using the ChIP-IT (Active Motif), and IgG (Active Motif) was used as a negative control. Anti-Atg7 or anti-p53 antibodies were used to immunoprecipitate the DNA-protein complex. Cross-link–reversed samples were treated with proteinase K, and the DNA was purified and analyzed by qPCR. The qPCR primers were designed in the region of Puma genomic locus (+253 ~ +485 from the transcriptional start site) and Ripk3 genomic locus (−698 ~ −543 from the transcriptional start site).

Measurement of autophagy flux

Cardiac autophagy flux was determined as described previously (46). Briefly, mice were intraperitoneally injected with bafilomycin A1 (6 μmol/kg; LC Laboratories) 30 min before sacrifice. To determine autophagy flux in primary cardiomyocytes and MEFs, we treated cells with E64d (10 μg/ml) and pepstatin A (10 μg/ml; Peptide Institute) for 24 hours at 37°C. LC3-II and p62 protein abundance and LC3 puncta were detected by Western blotting and immunocytochemistry, respectively, as markers of autophagy flux.

Statistical analyses

Data are means ± SEM. Normally distributed data were analyzed by an unpaired t test. Data not normally distributed were analyzed using the Mann-Whitney test. P < 0.05 was considered significant.


Fig. S1. Generation of Cnot3 muscle knockout (Cnot3 mKO) mice.

Fig. S2. Inducible cardiac-specific deletion of Cnot3 in adult mice (Cnot3 cKO).

Fig. S3. Cnot3 depletion increased apoptosis and necrosis in mouse cardiomyocytes.

Fig. S4. Cnot3 depletion altered autophagy protein abundance without changing autophagy flux.

Fig. S5. Cnot3 RIP-seq analysis and poly(A) tail length measurements of autophagy factor–encoding mRNAs in hearts.

Fig. S6. Poly(A) tail length and stability of autophagy factor–encoding mRNAs are regulated by the CCR4-NOT complex in cardiomyocytes.

Fig. S7. Atg7 promotes cardiac dysfunction in adult Cnot3 cKO mice.

Fig. S8. Atg7 promotes cell death in Cnot3-depleted cardiomyocytes.

Fig. S9. No phenotypic rescue of Cnot3 mKO mice by double knockout of Cnot3 and Atg5.

Fig. S10. Cnot3 depletion enhances the interaction of Atg7 with p53 to induce the expression of cell death–associated genes.

Fig. S11. Expression of cell death-associated genes in Cnot3 mKO mice.

Table S1. Primer list.

Table S2. siRNA list.

Data file S1. mRNAs significantly enriched in Cnot3 RIP-seq.

Data file S2. GO biological process terms.

Data file S3. GO cellular component terms.

Data file S4. GO molecular function terms.


Acknowledgments: We thank all members of our laboratories for technical assistance and helpful discussions. Funding: K.K. is supported by the Funding Program for Next Generation World-Leading Researchers (grant number LS015), Japan Society for the Promotion of Science (JSPS) Grants-in-Aid for Scientific Research (KAKENHI) (grant numbers 30733422, 16K19013, and 17H04028), and Japan Science and Technology Agency (JST) PRESTO (grant number JPMJPR13MD). Y.S. is supported by the Platform for Advanced Genome Science of JSPS KAKENHI (grant number 16H06279). S.N. is supported by JST PRESTO (grant number JPMJPR16E9) and JSPS KAKEN (grant numbers 15KT0147 and 16K05265). K.K., A. Kimura, and S. Shimizu are supported by Nanken-Kyoten, Tokyo Medical and Dental University. S. Shimizu is supported by JSPS KAKEN (grant numbers 17H01533 and 15K19004). Y.I. is supported by JSPS KAKEN (grant number 17H06179). Author contributions: K.K. and Y.I. designed the project. T. Yamaguchi, T. Suzuki, and T. Sato performed the experiments with assistance from A. Kadowaki, M.N., and Y.K. K.K., H. Inagaki, A.T., and T. Yamamoto generated knockout mice. S. Arakawa and S. Shimizu performed TEM analysis. S.N. and Y.S. performed RIP-seq analysis. H.W., S. Seki, S. Adachi, A.F., T.F., T.N., M.K., A. Kimura, H. Ito, and J.M.P. provided unpublished new experimental materials. K.K. and T. Yamaguchi analyzed the data. K.K. organized the project and wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The RIP-seq data have been deposited in the NCBI GEO database ( under the accession number GSE103629.
View Abstract

Navigate This Article