In this study, we demonstrate that NAT10 plays a crucial role in heart development and function by stabilizing the mRNAs of genes involved in fatty acid β-oxidation and heart contraction. NAT10 is associated with heart failure, and cardiomyocyte-specific deletion of Nat10 induces dilated cardiomyopathy (DCM), heart failure, and postnatal death. Restoration of NAT10 (WT) or NAT10 (G641E) (an N-acetyltransferase-inactive mutation), but not NAT10 (K290A) (a mutation that abolishes RNA binding), fully rescues the DCM, heart failure, and postnatal death phenotypes in Nat10-CKO mice. Importantly, the RNA-binding activity of NAT10 is essential for maintaining the mRNA stability of genes related to fatty acid β-oxidation and heart contraction, as NAT10 directly binds to these mRNAs. Interestingly, heart failure in both mice and humans triggers a compensatory increase in NAT10 expression. Together, these findings highlight the pivotal role of NAT10 in heart development and function, primarily through its ability to stabilize the mRNAs of genes critical for fatty acid β-oxidation and heart contraction.

Animal experiments were carried out in accordance with the Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee or Animal Experimental Ethics Committee of Harbin Institute of Technology (HIT/IACUC-2022066). Mice were housed under controlled light (12-h light/12-h dark cycle), temperature (24 ± 2 °C), and humidity (50 ± 10%) conditions and fed a normal chow diet with free access to water. Nat10 mice, in which the exon 4 of Nat10 gene was flanked by two loxP sites, were generated using the CRISPR-Cas9 technique. Nat10 mice were purchased from Cyagen. Nat10 mice were crossed with Myh6-Cre mice to generate cardiomyocyte-specific Nat10 knockout (Nat10-CKO) mice. The genotype of Nat10-CKO mice was Nat10 Myh6-Cre. Nat10 mice served as control. Male and female mice with different genotypes were randomly divided into different experimental groups. The age, strain, genotype and number of mice were described in the relevant figure legends of the manuscript. The number of mice was chosen based on expected variability of metabolic and cardiac parameters and on typical sample sizes reported in the literature [21,22,23,24]. No mice were excluded for analysis. Cardiac functions of neonatal mice were evaluated by echocardiography (Fujifilm VisualSonic Vevo 3100 MX700 linear array probe). The data were analyzed using Vevo LAB (version 3.2.0) software.

Primary rat cardiomyocytes were isolated from neonatal rats following a published protocol [24, 25]. The cells were cultured overnight in DMEM medium supplemented with 10% FBS, and then infected with equal amounts of Ad-βGal, Ad-NAT10 (WT), or Ad-NAT10 (K290A) overnight. One group of cells was used to measure gene expression via RT-qPCR. Another group of infected primary cardiomyocytes was treated with actinomycin D (200 ng/ml) for 0, 6, and 12 h. Cells were harvested at the indicated time, and RNA was isolated for RT-qPCR to measure mRNA stability. Additionally, primary cardiomyocytes were infected with Ad-shScramble or Ad-shNat10 for 48 h. One group of cells was used to measure gene expression via RT-qPCR. Another group of these infected primary cardiomyocytes were treated with actinomycin D (200 ng/ml) for 0, 6, and 12 h. Cells were harvested at the indicated time, and RNA was isolated for RT-qPCR to measure mRNA stability.

For immunoblotting, heart tissues were homogenized in RIPA lysis buffer and cells were harvested in RIPA lysis buffer. Tissue or cell extracts were immunoblotted with the indicated antibodies and visualized using the ECL. Antibody dilution ratios were as follows: NAT10 (13365-1-AP, Proteintech, 1:3000), FLAG (F1804, Sigma, 1:5000), β-actin (60008-1-lg, Proteintech, 1:5000), αTubulin (sc-5286, Santa Cruz, 1:5000), ACADS (Proteintech,16623-1-AP,1:3000), ACADM (Proteintech,55210-1-AP,1:3000), ACADVL (Proteintech,14527-1-AP,1:3000), CPT2 (Proteintech,26555-1-AP,1:3000), RYR2 (Proteintech,19765-1-AP,1:3000), KMT5A (Proteintech, 14063-1-AP,1:3000), SPEG (Sino biological 12472-RP02,1:3000).

The hiPSC line was a gift from Stanford Cardiovascular Institute iPSC Biobank with material transfer agreement. hiPSCs were routinely cultured in the mTeSR plus (STEMCELL Technologies) on Matrigel-coated (Corning) six-well plates. Confluent cells were passaged into six-well plates at a 1:10 ratio after a 5-min incubation with 0.5 mM EDTA (Thermo Fisher Scientific) at 37 °C. Cells were cultured in the mTeSR Plus medium supplemented with 10 μM Y-27632, ROCK inhibitor (Selleck Chemicals) for 24 h after re-plating, and the medium was then changed to the mTeSR Plus medium. A medium change was performed every 48 h. Cultures were maintained at 37 °C in a humidified incubator with 5% CO and 5% O.

To establish the NAT10 knockout hiPSC line, two gRNAs were designed based on a previous publication [26] and cloned into the PX458 plasmid, respectively (Addgene). Following plasmid construction, hiPSCs were electroporated with 2.5 μg of each plasmid. The electroporated hiPSCs were seeded onto Matrigel-coated plates with the treatment of 10 μM Y-27632 for 24 h. GFP-positive cells were sorted using fluorescence-activated cell sorting. 1K-positive cells were seeded into a well of Matrigel-coated six-well plate with the supplement of 5% CloneR (STEMCELL Technologies), and the medium was replaced with mTeSR Plus medium after 3 days. Once single-cell clones were formed, multiple clones were picked under a microscope, and the genotyping of each clone was determined using Sanger sequencing, followed by further validation using immunoblotting. sgRNA for NAT10 knockout were: NAT10-sgRNA-1 GAGTGCAACAGTACTCCTCA; NAT10-sgRNA-2 CACTCACAGCTGCTCGAGGA.

hiPSC-derived cardiomyocytes (hiPSC-CMs) differentiation followed a monolayer-based protocol adapted from a previous study [27]. Cells were passaged at a 1:6 ratio and grown for 3 days to 85-90% confluence. The medium was replaced with RPMI supplemented with B27 without insulin (Thermo Fisher Scientific) and 6 μM CHIR-99021 (Selleck Chemicals). Forty-eight hours later, the medium was changed to RPMI-B27 without insulin for 24 h, and then to RPMI-B27 without insulin supplemented with 5 μM IWR-1 (Selleck Chemicals) for 48 h. On day 5, the medium was changed back to RPMI-B27 without insulin for 48 h, and then switched to RPMI-B27 for 48 h. On day 9, the medium was transiently changed to RPMI-B27 without D-glucose (Thermo Fisher Scientific) for 72 h to allow metabolic purification of hiPSC-CM. Cells were then dissociated after an 8-min incubation with Accutase (Sigma Aldrich) at 37 °C followed by seeding into 6-well plates. A second metabolic purification was performed 48 h after re-plating as described above. After the second purification, hiPSC-CM were continuously cultured in RPMI-B27 until further use for experiments.

The Ca imaging protocol was adapted from a previous study [28]. Day 30 hiPSC-CMs were dissociated with TrypLE Select Enzyme (10X) (Thermo Fisher Scientific) and re-seeded in Matrigel-coated 35-mm glass-bottom culture dishes with 20-mm microwell (MatTek). After recovery for 4 days, cells were loaded with 2 µM Fura-2 AM with 0.1% F-127 for 10 min at RT in Tyrode's solution including NaCl (135 mM), KCl (5.4 mM), MgCl (1 mM), CaCl (1.8 mM), glucose (5 mM), and HEPES (10 mM). The hiPSC-CMs were paced using a field stimulation at 0.5 Hz, and Ca imaging was performed on the Ionoptix system using the appropriate filters and Dual OptoLED Power Supply (F340 and F380, Ionoptix). The results were analyzed using CytoSolver software (Ionoptix). "Peak height" was used as Ca transient amplitude; "Time to Baseline50" minus "time to peak" was used as half decay. For the rescue experiments, NAT10 KO hiPSC-CMs recovery for 4 days after seeding into the dish, and then were infected with Ad-GFP, Ad-NAT10 (WT), Ad-NAT10 (G641E), or Ad-NAT10 (K290A) adenovirus overnight. WT hiPSC-CMs were infected with equal amounts of Ad-GFP adenovirus serving as a control. After that, cells were loaded with 2 µM Rhod-2 AM for 25 min in the incubator, paced using field stimulation at 0.5 Hz, and Ca imaging was conducted on the Ionoptix system using the appropriate filters and LED Power Supply (F565, Ionoptix). The results were analyzed using CytoSolver software (Ionoptix).

Heart tissues in Nat10 and Nat10-CKO mice were collected at both 10 and 15 days of age. For rescue experiments, at day 1 after birth, Nat10 mice were injected with AAV9-cTnT-GFP (1 × 10 VP/mouse) subcutaneously, and Nat10-CKO mice were injected with equal amount (1 × 10 VP/mouse) of AAV9-cTnT-GFP, AAV9-cTnT-NAT10(WT), AAV9-cTnT-NAT10(G641E), and AAV9-cTnT-NAT10(K209A), respectively. Mice were sacrificed at days 12-14 after injection. Total RNA was extracted using TriPure Isolation Reagent (Roche, Mannheim, Germany) from mouse hearts. Three independent biological replicates for each group were used for RNA-seq. For 10-day samples, RNA-seq was performed by deep sequencing using an Illumina Novaseq X Plus. Paired-end clean reads were aligned to the mouse reference genome (grcm39) with Hisat2 (version 2.2.1), and the aligned reads were used to quantify mRNA expression by using featureCounts (version 2.0.6). For 15-day and rescue samples, RNA-seq was performed by deep sequencing using an Illumina Novaseq 6000 platform. Paired-end clean reads were aligned to the mouse reference genome (GRCm38/mm10) with Hisat2 (version 2.0.5), and the aligned reads were used to quantify mRNA expression by using featureCounts (version 1.5.0-p3) as described previously [29,30,31,32]. Differential expression analysis of two groups was performed using the DESeq2 R package (version 1.42.0 or 1.20.0). DESeq2 provide statistical routines for determining differential expression in digital gene expression data using a model based on the negative binomial distribution. The resulting P-values were adjusted using Benjamini and Hochberg's approach for controlling the false discovery rate. Transcriptional and posttranscriptional regulation were determined by exon-intron split analysis (EISA). EISA was analyzed by measuring changes in pre-mRNA (intron-containing transcripts) vs. mRNA in Nat10 and Nat10-CKO hearts by comparing the intronic and exonic reads in RNA-seq data, as described previously [33]. RNA sequencing, library construction, and RNA sequencing alignment were performed by technical staff at Novogene who were blinded to the experimental groups. RNA-seq data that support the findings of this study have been deposited in GEO under accession codes GSE274007, GSE274008, and GSE295332.

At day 1 after birth, WT mice were injected with AAV9-cTnT-FLAG-NAT10 (3 × 10VP/mouse) subcutaneously. Mice were sacrificed at day 20 after injection. Heart tissues (~60 mg) were homogenized in a lysis buffer (100 mM KCl, 5 mM MgCl, 10 mM HEPES [pH 7.0], 0.5% NP40, 1 mM DTT, 100 units/mL RNaseOut) prepared in DEPC HO. Ten percent of the tissue lysate was used as input, and the remaining lysate was immunoprecipitated with 50 μL anti-FLAG M2 magnetic beads (M8823, Millipore) at 4 °C for 2 h. The immunoprecipitated samples were washed four times with lysis buffer. Total RNA was extracted from both the immunoprecipitated and input samples using the TriPure Isolation Reagent (Roche, Mannheim, Germany). mRNA was purified using the Hieff NGS® mRNA Isolation Master Kit (12603, Yeasen Biotechnology). The fragment distribution and concentration of RNAs (after immunoprecipitation) were detected using an Agilent 2100 bioanalyzer (Agilent) and simpliNano spectrophotometer (GE Healthcare). The distribution and concentration of RNA fragments (post-immunoprecipitation) were analyzed using an Agilent 2100 bioanalyzer (Agilent) and a simpliNano spectrophotometer (GE Healthcare). The immunoprecipitated RNA or input was then utilized for library construction with the NEB NextR Ultra™ RNA Library Prep Kit (New England Biolabs). Library quality was evaluated on the Agilent Bioanalyzer 2100 system, and sequencing was performed on an Illumina Novaseq 6000 platform with a paired-end read length of 150 bp following standard protocols [32]. Raw data in fastq format were initially processed using fastp software, where clean data were obtained by removing reads containing adapters, poly-N sequences, and low-quality reads. Additionally, the Q20, Q30, and GC content of the clean data were calculated. All downstream analyses were conducted using these high-quality, clean data. Hisat2 was used to align reads to the genome (grcm39-110). DESeq2 was used to identify which RNAs are specifically enriched in the RIP sample compared to the input. The interaction between intracellular RNA and protein binding exhibited specific sequence preferences rather than being random. GO enrichment analysis was implemented with the GOseq R package, correcting for gene length bias, with GO terms having a corrected p-value of less than 0.05 considered significantly enriched by detected mRNAs. KEGG, a database resource, helps understand high-level biological functions and utilities from molecular-level information, particularly from large-scale molecular datasets generated by genome sequencing and other high-throughput experimental technologies (http://www.genome.jp/kegg/). KOBAS (3.0) software was used to test the statistical enrichment of detected mRNA-related genes in KEGG pathways. RIP-seq data that support the findings of this study have been deposited in GEO under accession codes GSE274009.

Ribosome profiling sequence (Ribo-seq) assay was performed to identify ribosome-protected mRNA fragments (RPFs) associated with NAT10 in mouse hearts following a published method [34, 35]. Heart tissues in Nat10 and Nat10-CKO mice were collected at 15 days of age. Four hearts from each genotype were combined into a single sample. Pulverized tissue samples were lysed in lysis buffer containing 100 μg/mL cycloheximide on ice. Lysates were clarified by centrifugation and treated with RNase I to generate ribosome-protected fragments (RPFs). RPFs were isolated using the Epi™ Ribosome Profiling Kit (Epibiotek, R1814), followed by purification using the RNA Clean & Concentrator-5 kit (ZYMO, R1016). rRNA was removed using the Epi™ RiboRNA Depletion Kit (Human/Mouse/Rat) (Epibiotek, R1805). The raw sequencing data are first processed using Cutadapt (v3.7) to remove adapter sequences. Then, FastQC (v0.11.9) is used for sequencing quality assessment. The cleaned reads are aligned to an rRNA reference sequence using Bowtie (v1.3.1), and the aligned reads are removed. The remaining reads are then mapped to the reference genome (GRCm38/mm10) using STAR (v2.7.4a), with parameters suitable for short reads. To characterize ribosome occupancy and translation activity, P-site positions and ribosomal periodicity were estimated using the riboWaltz (v1.2.0) R package. For open reading frame (ORF) prediction, the GEDI suite was employed. Genomic indexing was performed with the IndexGenome function, and BAM files were converted to the CIT format using Bam2CIT. Candidate ORFs were identified with the Price module, and the resulting.orfs.tsv files were collected for downstream analysis. Read counts for each gene are obtained using HTSeq (v0.13.5). To accurately distinguish transcription-regulated genes (DTGs) from translation-regulated genes (DTEGs), we employed the delta TE analysis method described previously [36]. Using the R scripts provided in the article, we simultaneously analyzed both Ribo-seq and RNA-seq data. Ribo-seq data that support the findings of this study have been deposited in GEO under accession codes GSE295333.

RT-qPCR was performed as shown previously [29]. Briefly, total RNAs were extracted using TriPure Isolation Reagent (Roche, Mannheim, Germany), and the first-strand cDNAs were synthesized using random primers and M-MLV reverse transcriptase (Promega, Madison, WI). RT-qPCR was done using Roche LightCycler 480 real-time PCR system (Roche, Mannheim, Germany). The expression of individual genes was normalized to the expression of 36B4. Primers for real time RT-qPCR were Nat10-F: 5'-GAGCGGCAGAGGTCTCTTTT-3', Nat10-R: 5'-TTCTTCTGTAGCTGCCGCAT-3'; 36B4-F: 5'-AAGCGCGTCCTGGCATTGTCT-3', 36B4-R: 5'-CCGCAGGGGCAGCAGTGGT-3'; Mef2a-F: 5'-AAAAATAGCCCCGGTGTGGG-3', Mef2a-R: 5'-GCAATTCCGAGTCTCCGCTA-3'; Mef2c-F: 5'-GCACCAACAAGCTGTTCCAG-3', Mef2c-R: 5'-CTGAATCGTCTGCATCGGGA-3'; Ryr2-F: 5'-TGCCCTCTGAAGACCTGACA-3', Ryr2-R: 5'-GGACAGGGTTGGTCATGAGG-3'; Cpt2-F: 5'-CAAAAGACTCATCCGCTTTGTTC-3', Cpt2-R: 5'-CATCACGACTGGGTTTGGGTA-3'; Acadvl-F: 5'-TTCCGGATCTTTGAAGGGGC-3', Acadvl-R: 5'-CGATTCCTGTCCTCCGTCTC-3'; Acadm-F: 5'-GGGGAAAGGCCAACTGGTAT-3', Acadm-R: 5'-TCGCTGGCCCATGTTTAGTT-3'; Acads-F: 5'-TGACTTTGCCGAGAAGGAGT-3', Acads-R: 5'-ACTCAGCTCCTCTGGCACAT-3'; Speg-F: 5'-AGCCTAGCCTCAGTTGGTTC-3', Speg-R: 5'-CAGTCTCCGAGTCTCTGACC-3'; Kmt5a-F: 5'-TGAGCAGTTGGCTACAGTTGTT-3', Kmt5a-R: 5'-CTTGCACATCTTCCTGCCTCTA-3'.

Data were presented as means ± SEM. Differences between two groups were analyzed by two-tailed Student's t tests. Differences among three groups were analyzed by one-factor analysis of variance (ANOVA), and the least significance difference (LSD)-t test. In all analyses, P < 0.05 was considered statistically significant. Meanwhile, the Kolmogorov-Smirnov normality test with p > 0.1 suggested our samples followed a normal distribution. All statistical analyses were performed using GraphPad Prism version 6.02 (GraphPad Software Inc., San Diego, CA, USA). ANOVA and LSD-t test analyses were performed using SPSS 21.0 (SPSS Inc., Chicago, IL).