A total of 58 rats were used for animal experiments (all 8 weeks old; weight 200-230 g). A total of four female Fischer 344 rats, five male Fischer 344 rats, four female Sprague-Dawley rats and five male vasectomized Sprague-Dawley rats were purchased from Envigo for performing Kcnh2 knockout. A total of 20 wild type (WT) and 20 Kcnh2^+/− female rats were used for isolating embryos at different time points. The animal experiments were performed according to the investigator's protocols approved by the University of Southern California and Tongji University Institutional Animal Care and Use Committee (approval no. TJLAC-016-022) and complied with the AVMA euthanasia guidelines 2020 (https://olaw.nih.gov/policies-laws/avma-guidelines-2020.htm). All rats were raised under normal temperature (25°C), normal oxygen conditions (~21%) and normal relative humidity (50%) with a 12-h light/dark cycle. CO₂ with a flow rate of 9 l (30% of the chamber) per min was used to euthanize pregnant rats for isolation of blastocysts or embryos. The detail of the electrocardiograph, echocardiography and whole-cell patch-clamp can be found in the Data S1.

DAc8 rat ES cell line was used and cultured as previously described (16,17). Briefly, the rat ES cells were cultured in 2i medium (18,19) and transfected by electroporation of targeting vector. Puromycin (0.5 μg/ml; Beyotime Institute of Biotechnology) was used to screen of drug-resistant rat embryonic stem cell colonies. The genome of puromycin-resistant cell colonies was extracted with genomic DNA kit (Tiangen Biotech Co., Ltd.) and PCR was performed using LA Taq (Takara Biotechnology Co., Ltd.). The following thermocycling conditions were applied for 35 cycles: 98°C, 20 sec; Tm, 20 sec; 72°C, 1 kb/min. The products were separated in 1% agarose gel. Colonies that were positive in PCR were confirmed by Southern blotting. The genomic DNA of rat ES cells was digested by enzyme, separated by agarose gel electrophoresis and transferred to a piece of nylon (Beyotime Institute of Biotechnology). Then, 1.5 μg of probe was added to 30 ml of buffer for hybridization. F344 blastocysts were isolated from pregnant female Fischer 344 rats. Targeted-ES cell colonies were expanded and injected into F344 blastocysts. The obtained chimeric blastocysts were then transplanted into the uterus of pseudopregnant rats. Generally, 8-10 chimeric blastocysts were transplanted into each pseudopregnant rat. Male chimaeras were mated with SD female rats to generate germline offspring. Kcnh2 heterozygotes were mated with each other to produce homozygotes. Genomic DNA from rat tail tips were extracted and separated as described previously for genotyping. The sequences of primers are listed in Table I.

For PCR, 1 ml of Total RNA Extraction Reagent (Shanghai Yeasen Biotechnology Co., Ltd.) was used to extract the total RNA from 100 mg of tissue. The 1st Strand cDNA Synthesis kit, gDNA Digester Plus (Shanghai Yeasen Biotechnology Co., Ltd.) was applied for reverse transcription of total RNA. The reaction program involved heating to 95°C for denaturation and holding at 4°C for 40 cycles (95°C, 15 sec; 55-60°C, 10 sec; 72°C, 20 sec). RT-qPCR was performed using SYBR Green Master Mix (Shanghai Yeasen Biotechnology Co., Ltd.). All procedures were performed according to the manufacturer's protocols. GAPDH was applied to normalize the expression of mRNAs. The experiment was repeated 3 times for each sample. The Ct (threshold cycle) was defined as the number of cycles required for the fluorescence signal to exceed the detection threshold. The 2^−ΔΔCq method was used to calculate the differential expression of genes (20). The related primers are listed in Table SI.

Embryos from Kcnh2^−/− and WT rats were fixed in 4% paraformaldehyde for 1 h at room temperature. Tissue was dehydrated using ethanol at different concentrations (50, 75, 80, 95 and 100%) for 20 min each for histological analysis. Ethanol was exchanged with xylene twice and the xylene exchanged with paraffin twice, for 1-2 h per exchange. Finally, embryos were embedded in fresh paraffin. Tissue was rehydrated using 10% sucrose and 30% sucrose for 4 h each at 4°C for immunofluorescence. Equal volumes of OCT and 30% sucrose were incubated at room temperature for 30 min, and the tissues were transferred into mold. After paraffin/OCT-embedding, tissue sections were cut at 8-10 μm. The paraffin sections were stained with hematoxylin and eosin (MilliporeSigma). After permeabilization with 0.1% Triton X-100 and blocking, the OCT sections were stained with the indicated antibodies for 18 h at 4°C (see Table SII for antibody list).

Proteins were extracted from 100 mg of tissue using 1 ml of ice-cold RIPA buffer containing fresh protease and phosphatase inhibitors (Beyotime Institute of Biotechnology) and the concentrations were measured using a BCA assay (Beyotime Institute of Biotechnology). Gels were prepared with a PAGE Gel Fast Preparation kit (Epizyme) according to the manufacturer's instructions. Then, 10 μg of protein was loaded per lane. Denatured protein samples were separated by 10% SDS-PAGE and transferred to a PVDF membrane (Bio-Rad Laboratories, Inc.). Then the membranes were blocked with 5% milk powder for 1 h at room temperature. Primary antibodies diluted in 5% BSA were used to probe the targeted proteins for 18 h at 4°C. The corresponding secondary antibody was used to enhance the signal for 1 h at room temperature. Finally, the blot bands were visualized using enhanced chemiluminescence (Pierce, Thermo Scientific). Details of antibodies are given in Table SII. ImageJ fiji software (version v1.53c; National Institutes of Health) was used for densitometric analysis.

Proteins were extracted from 100 mg of tissue using 1 ml of ice-cold RIPA lysis buffer containing fresh protease and phosphatase inhibitors (Beyotime Institute of Biotechnology; cat. no. P0013D) and concentrations measured. After treatment of samples, 500 μg of proteins diluted in 500 μl of RIPA lysis buffer were incubated with primary antibody (Integrin β1: 2 μg per IP reaction; KCNH2: 2 μg per IP reaction) and goat IgG isotype control (2 μg per IP reaction) separately, and then with rotation overnight at 4°C. Magnetic beads (Beyotime Institute of Biotechnology; cat. no. P2105) were added and the mixture was shaken for another 4 h. After incubation, the mixtures were placed on a magnetic stand (Invitrogen; Thermo Fisher Scientific, Inc.) for 10 sec at 4°C to separate the beads. The bead complexes were washed with RIPA lysis buffer and then the immunoprecipitants were mixed with SDS loading buffer, boiled for 10 min and subjected to western blotting as aforementioned (see Table SII for details of antibodies).

Apoptosis of rat embryos was detected by TUNEL assay (Roche Diagnostics GmbH). After fixation and permeabilization, TUNEL reaction mixture was added to each sample and the samples were incubated for 1 h at 37°C in dark. After adding Hoechst (cat. no. 14530-100MG; MilliporeSigma) and incubating the samples for 12 min at room temperature in the dark, the cells were rinsed with PBS and imaged with a Leica confocal microscope (Leica Microsystems GmbH). The normalized results were determined by comparing the TUNEL-positive (red) cells to the total nuclei in 3-6 randomly chosen microscopic fields.

The EGAS00001003996 dataset in the European Genome-phenome Archive (EGA) contains human embryonic heart small conditional RNA (sc)RNA-seq data of 6.5-7 weeks post-conception, including a total of 3,717 single-cells as determined by 10X Genomics (21).

Dimensionality reduction and clustering of the scRNA-seq data were performed using the Seurat package (version 4.1.1) of R (version 4.3.1; http://www.R-project.org/). The highly variable genes were selected using the FindVariableGenes function and clusters were identified using the FindClusters function with the same settings as previously reported (21). Different cell types were identified based on the abundance of cell clusters and the cardiac function represented by the cardiomyocyte clusters according to known marker genes. Uniform manifold approximation and projection (UMAP) was used to visualize single-cells.

The pseudotime trajectories of ventricular cardiomyocytes were constructed using the dynoverse dyno (version 0.1.2) set of R packages for trajectory inference (TI) with single-cell data (22). Single cell matrices were extracted from the Seurat object and wrapped into a dynwrap object using dynwrap wrap_expression. Then the immature cardiomyocytes, trabecular cardiomyocytes, inner compact zone (ICZ) cardiomyocytes and outer compact zone (OCZ) cardiomyocytes were respectively imported into a multiple spanning tree (MST). Pseudotime visualization was obtained using calculated pseudotimes on the MST model object.

Statistical analyses were performed using GraphPad Prism V software (GraphPad Software; Dotmatics). All assays were repeated independently at least three times. The data are represented as mean ± SD in the figures. Differences between the two groups were compared using unpaired Student's t-test. Differences between three or more groups were compared using one-way ANOVA with Tukey's post hoc test. P<0.05 was considered to indicate a statistically significant difference.

Exons 7 and 8 of the rat Kcnh2 gene encode the pore region of the ether-a-go-go protein. To delete exons 7 and 8, the targeting vector was introduced into DAc8 rat ESCs via electroporation. After the recombination reaction, exons 7 and 8 were replaced with the positive selection cassette CAG-EGFP-IRES-Pac (Fig. 1A). For genotyping, the targeted rESCs were confirmed by PCR and Southern blot (Fig. 1B and C) and then expanded and microinjected into E4.5 Fisher344 blastocysts. Thereafter, the rESC-injected blastocysts were transferred into pseudopregnant SD rats. Male chimaeras (Fig. S1) were cross-bred with WT SD female rats to generate germline pups (Fig. 1D). Germline transmission of the Kcnh2 mutant allele was then verified based on the EGFP expression of F1 offspring and PCR genotyping (Fig. 1E and F). Kcnh2 heterozygous germline rats showed a normal growth rate and body weight and were fertile.

Mendelian ratios were counted for the hybrid offspring after crossing of heterozygous mutant rats (Table SIII). To our surprise, homozygous offspring were not found, which suggested that the Kcnh2^−/− genotype might induce embryonic lethality. To determine when the embryos stopped developing in the early stage, Kcnh2^−/− embryos were dissected at different time points during pregnancy. Prior to E10.5, no obvious differences were observed in body size at matched stages of development (Fig. 2A). After E10.5, the Kcnh2^−/− embryos developed slowly, showing a smaller body size. By E15.0, however, no live Kcnh2^−/− embryos were found (Fig. S2). Moreover, Kcnh2^−/− embryos at E11.0 exhibited with a massive pericardial effusion, which was very likely a consequence of heart failure with evidence of structural deformity in the bulbus cordis right ventricle and the outflow tract (Fig. 2B). Additionally, a reduction in myocardial thickness in embryonic hearts and dilation of the first branchial arch artery were identified and the length and structure of myofibrils were greatly affected (Fig. 2C). By contrast, Kcnh2^+/− embryos showed embryonic conformations similar to those of WT embryos.

In Kcnh2^−/− rats, no shorter variant blots were detected (Fig. S3), suggesting the channel protein was knocked out (13). To determine the potential role of KCNH2 in normal development, the KCNH2 expression status was assessed at different stages (Fig. 2D and E). KCNH2 expression gradually increased beginning at E10, peaked at E15 and then decreased progressively during later embryonic stages and the postnatal period. The distribution of KCNH2 at E10.5 was found to be very different, especially in cardiomyocytes (Fig. 2F), which was further verified by an expression assay of different embryonic tissues (Fig. 2G and H). In Kcnh2^−/− embryos at E10.5, the markers of angiogenesis and nervous system development were not significantly altered (Figs. S4 and 2I), while the expression of the related genes of both the first heart field, such as HCN4 and TBX5 and the second heart field, including MEF2C, ISL1 and Hand2, were significantly decreased (Fig. S5). These findings were consistent with the expression patterns of E11 embryos (Fig. 2J), indicating an abnormal status in the embryonic heart. In addition, to explore the influence of KCNH2 channel activity on embryonic and heart development, E4031, a pharmacological blocker of Kv11.1 ion-conducting properties was used to treat pregnant rats. The embryos treated with E4031 for 2 or 3 days after E9.5 showed abnormal phenotypes, similar to those of Kcnh2^−/− rats, such as significantly reduced body and heart sizes (Fig. S6A-C). Moreover, inhibition of KCNH2 at E10.5 or E11.5 caused embryonic lethality (Fig. S6D and E), which partly recapitulated the Kcnh2 knockout phenotype. Therefore, the activity of KCNH2 is essential for maintaining early embryonic development in rats.

During the early stage of embryonic development, trabecular cardiomyocytes are first to develop, followed by compact cardiomyocytes, which promote the thickening of the myocardium (23). KCNH2 affected the process in vivo (Fig. 2B and C) and has been found to mediate the differentiation of rESCs into cardiomyocytes in vitro (13), but the mechanism underlying the arrest of embryonic development in the absence of KCNH2 is unclear. To understand whether the distribution of KCNH2 in different types of cardiomyocytes is related to the formation of the compact myocardium, human embryonic hearts of 6.5-7 weeks post-conception and further subpopulations of the cardiomyocyte population were analyzed (21). According to the expression status of marker genes (Fig. 3A), eight clusters of the cells were classified, including both ICZ and OCZ cardiomyocytes, trabecular cardiomyocytes, atrial cardiomyocytes, cardiac conduction system (CCS) cardiomyocytes, outflow tract cardiomyocytes, immature cardiomyocytes and finally a cluster of smooth muscle cell/fibroblast-like cells. These clusters were combined and visualized through UMAP concerning cell identity (Fig. 3A and B). The analysis of the present study indicated that KCNH2 was expressed in compact cardiomyocytes, trabecular cardiomyocytes, atrial cardiomyocytes, CCS cardiomyocytes and outflow tract cardiomyocytes (Fig. S7A), but not in immature cardiomyocytes (Fig. 3C). To determine the progressive development of compact cardiomyocytes during the maturation process, a pseudotime analysis on ventricular cardiomyocytes based on ventricle-specific markers was performed to construct the trajectory of cardiomyocytes development. In this process, the pseudotime trajectory began with immature cardiomyocytes, moved to trabecular cardiomyocytes and ICZ cardiomyocytes and gradually progressing to OCZ cardiomyocytes (Fig. 3D). It was found that KCNH2 was upregulated along the trajectory (Fig. 3E), similar to the case for the maturation markers, creatine kinase-mitochondrial 2 and myosin heavy chain 7 (Fig. S7B and C). These results matured with the expression trend during early embryonic development (Fig. 4D) and suggested that KCNH2 might play a crucial role in the differentiation of compact cardiomyocytes involved in heart development.

KCNH2 has been shown to regulate apoptosis and proliferation to control cell fate in vitro (4). Although a previous study has shown that homozygous missense N629D Kcnh2 potassium channel mutation causes developmental defects in the right ventricle, outflow tract and embryonic lethality (24), the mechanism in embryonic development has not yet been confirmed. The immunostaining TUNEL assay showed that the apoptosis of Kcnh2^−/− embryos increased significantly (Fig. 4A). Consistently, the Kcnh2 mutation increased the cleaved caspase3/caspase3 ratio and enhanced the expression of the proapoptosis-related genes Bcl-2 like 11 (Bim) and p53 upregulated modulator of apoptosis (PUMA; Fig. 4B and C). FOXO3A, which can transcriptionally increase the expression of Bim and regulate embryo cell fate (25), was also upregulated. However, the proliferation marker Ki67 did not display any change at E10.5 embryos in the absence of KCNH2 (Fig. 4D). Furthermore, the expression of proliferation-related genes such as cyclin D1, proliferating cell nuclear antigen (PCNA) and VEGFR remained unchanged (Fig. 4E and F). These results suggested that KCNH2 is necessary to prevent apoptosis during embryo development.

Our previous findings have demonstrated that KCNH2 interacts with integrin β1 to activate the AKT signaling pathway during the differentiation of rESCs into cardiomyocytes (13). However, the relationship between the KCNH2/integrin β1 complex and the AKT pathway has not yet been confirmed in vivo. To explore how KCNH2 is involved in FOXO3A regulation, it was next assessed whether KCNH2 associates with integrin β1 to modulate AKT activity by mediating the phosphorylation of FAK (26). Immunofluorescence staining of E10.5 WT embryos showed relatively high expression of integrin β1, which was distributed along the cell membrane (Fig. 5A). Immunoprecipitation further confirmed that KCNH2 indeed interacted with integrin β1 (Fig. 5B). In addition, the integrin β1 protein, but not its mRNA (Fig. S8), was downregulated in Kcnh2 knockout embryos (Fig. 5C and D), suggesting that KCNH2 might play a role in the stability of integrin β1 rather than affecting its transcriptional expression. Consistent with the disruption of integrin β1 with KCNH2, the phosphorylation of FAK was also decreased in Kcnh2 depleted rats, which then induced a decrease in AKT activity (Fig. 5E and F). The reduction of AKT activity with this genotype also triggered increases in p-GSK-3β-S9 (Fig. 5E) and calcium/calmodulin dependent protein kinase II (CaMKII) (Fig. S9) levels, but a decrease in β-catenin at E10.5 embryos (Fig. 5E and G). Taken together, these results suggest that KCNH2 can directly interact with integrin β1 to facilitate the phosphorylation of AKT to prevent apoptosis in the rat heart during embryonic development.