Bmp16 Regulates Arterial Valve Morphogenesis Through Modulation of Notch Signaling in Zebrafish

Abstract

Congenital valve defects account for a substantial proportion of cardiovascular malformations, yet the molecular mechanisms orchestrating cardiac valve development remain incompletely elucidated. While Bone morphogenetic protein (BMP) signaling is essential for valvulogenesis, the specific contributions of individual BMP ligands, particularly the teleost-specific bmp16, have not been characterized. Using the CRISPR/Cas9 system, we generated a bmp16 null knockout and delineated critical roles of this ligand in valvular morphogenesis. bmp16 knockout embryos display a significant reduction in Sox9-positive valvular cells and exhibit severely dysplastic arterial valves, characterized by increased interleaflet distance, thickened leaflets, and shortened leaflet lengths. These morphological abnormalities correlate with impaired valve function, culminating in progressive blood regurgitation, ventricular dilation, and pericardial edema. Mechanistically, loss of bmp16 or pharmacological inhibition of BMP signaling significantly downregulates notch1b expression in developing valves, while pharmacological activation of Notch signaling rescues the regurgitation phenotype in bmp16 mutants. Collectively, our findings establish bmp16 as a novel regulator of valve development and uncover a functional BMP-Notch signaling axis required for vertebrate valvulogenesis, providing new insights into the molecular mechanisms that govern cardiac valve formation and pathogenesis.

1. Introduction

Congenital heart disease (CHD) represents one of the most prevalent birth defects worldwide, affecting approximately 4–10 infants per 1000 live births [1]. Among these, defects in cardiac valves and associated structures constitute one of the most frequent subtypes, accounting for 25–30% of all cardiovascular malformations [2]. Beyond the pediatric population, valvular heart disease has emerged as a major contributor to the growing global burden of cardiovascular morbidity and mortality in adults [3]. Despite the clinical significance of valve diseases, therapeutic strategies remain largely limited to surgical intervention, highlighting an urgent need to delineate the molecular mechanisms underlying valve development and pathogenesis [3].

The aortic valve is located between the left ventricle and the aorta to ensure unidirectional blood flow from the ventricle to the systemic circulation [4]. During embryogenesis, cardiac valves originate from endocardial cushions that form in the atrioventricular canal (AVC) and outflow tract (OFT) via endothelial-to-mesenchymal transformation (EndoMT), a conserved process wherein a subset of endocardial cells delaminate, invade the underlying cardiac jelly, and differentiate into mesenchymal cells, which subsequently undergo extensive remodeling to form functional valve leaflets [5,6]. In zebrafish, valve development proceeds through similar cellular events and evolutionarily conserved molecular mechanisms. Recent studies have demonstrated that the zebrafish arterial valve primordium differentiates directly from second heart field (SHF) progenitors at the distal OFT transition zone [7]. Following the recruitment of SHF progenitors and elongation of the distal OFT, the arterial valve primordium emerges as multilayered protrusions [8]. Concurrently, endocardial cells within the proximal OFT undergo EndoMT, generating mesenchymal cells that differentiate into valve interstitial cells and populate the developing endocardial cushions. These structures subsequently elongate and remodel into mature valve leaflets, mirroring the later stages of mammalian valve morphogenesis [5].

Valve morphogenesis is a highly orchestrated developmental process that relies on precise spatiotemporal coordination of multiple evolutionarily conserved signaling pathways [2,5]. Disruption of these signaling pathways is closely associated with congenital valve defects and adult-onset valvular disorders. Among these conserved pathways, the Notch signaling pathway has emerged as a critical regulator of aortic valve development and postnatal valve homeostasis [9]. In humans, heterozygous mutations in the NOTCH1 gene confer a significantly increased risk of bicuspid aortic valve and subsequent progressive aortic valve calcification in adulthood [10]. Complementary genetic and functional studies in zebrafish have further validated the conserved role of Notch signaling in valvulogenesis and demonstrated that notch1b is essential for coordinating cellular responses to hemodynamic forces, regulating EndoMT in endocardial cushions, and ensuring proper valve leaflet formation and structural integrity [11,12]. These findings collectively establish Notch signaling as a core pathway governing valvular development across vertebrate species.

Bone morphogenetic protein (BMP) signaling constitutes another evolutionarily conserved pathway that is essential for heart valve development across vertebrate species [5]. During early cardiogenesis, the canonical BMP ligands Bmp2 and Bmp4 are spatiotemporally expressed in the myocardium of the AVC and OFT, where they initiate the valve formation program [13,14]. Bmp4 promotes proliferation of OFT myocardium and growth of endocardial cushions between E9 and E10.5 in mice [15,16]. While Bmp6 and Bmp7 are individually dispensable for normal cardiogenesis, mice with combined knockout of Bmp6 and Bmp7 exhibit severe defects in valve morphogenesis and cardiac chamber septation [17]. These studies underscore the complexity of BMP ligand-specific functions and their context-dependent contributions to valvular development.

Among the BMP ligand family, Bmp16 represents a relatively understudied member. Phylogenetic and comparative genomic studies indicate that bmp16 is a novel paralog of the bmp2/4 subfamily which is uniquely retained in teleost genomes [18,19]. Recent studies have demonstrated that bmp16 is expressed in several embryonic structures, including the embryonic hearts [19,20]. Despite this intriguing expression profile and its evolutionary relationship to bmp2/4, the specific function of bmp16 in valvulogenesis has remained unexplored.

In this study, we investigate the function of zebrafish bmp16 in arterial valve development using a combination of genetic, pharmacological, and imaging approaches. We generate a bmp16 null knockout line and characterize the resulting valve phenotypes at cellular and functional levels. Furthermore, we explore the relationship between BMP and Notch signaling during valve formation. Our findings identify bmp16 as a novel regulator of arterial valve development and establish a functional BMP–Notch signaling axis essential for proper valve formation.

2. Results

2.1. Generation of a Bmp16 Null Knockout Zebrafish Line

Using a transgenic zebrafish line, Tg(BRE:EGFP), which expresses EGFP under the control of a conserved BMP-responsive element (BRE) [21], we observed active BMP signaling in zebrafish arterial valves from 72 h post fertilization (hpf) to 96 hpf (Figure S1). To determine whether bmp16 contributes to the arterial valvular BMP signaling, we performed whole-mount in situ hybridization (WISH) to examine the spatiotemporal expression pattern of zebrafish bmp16 gene. bmp16 transcripts were detected in the embryonic hearts as early as 24 hpf, with sustained expression through 96 hpf (Figure 1A).

To investigate the role of bmp16 in valvulogenesis, we generated a bmp16 null allele using the CRISPR/Cas9 system with two guide RNAs targeting the signal peptide-encoding region and the stop codon-proximal region, respectively (Figure 1B), which enabled the deletion of the entire genomic sequences encoding the secreted form of the Bmp16 protein. Sanger sequencing of the targeted locus confirmed that the mutant allele contains frameshift deletion that results in seven aberrant amino acids prior to encountering a cryptic stop codon within the 3′ untranslated region (UTR), lacking the functional Bmp16 peptide sequences (Figure 1C,D). Consistent with this design, bmp16 mRNA was undetectable in bmp16^−/− homozygotes at 72 hpf by WISH (Figure 1E). Quantitative reverse transcription PCR (qRT-PCR) analysis showed that bmp16 mRNA levels at 24 hpf were reduced to 65% in bmp16^+/− heterozygotes and to 5.8% in bmp16^−/− homozygotes relative to wild-type (bmp16^+/+) embryos (Figure 1F). In line with the bmp16 loss, expression levels of the well-known BMP signaling target genes, id1, id2a, and id2b [22,23,24,25], in bmp16^−/− embryos were significantly reduced to 52%, 74% and 65% of the wild-type levels at 24 hpf, respectively (Figure 1G–I). Collectively, these data demonstrate that the generated bmp16 knockout represents a bona fide null mutation that impairs BMP signaling in the zebrafish embryos.

2.2. Bmp16 Knockout Impairs Arterial Valve Development in Zebrafish

Given the active BMP signaling in the developing arterial valve and the expression of bmp16 in the embryonic heart, we next investigated the role of bmp16 in arterial valve morphogenesis. To visualize dynamic valve formation in vivo, we utilized a double-transgenic zebrafish line Tg(tp1:EGFP;kdrl:mCherry-CAAX), where EGFP labels Notch-responsive cells (including valvular interstitial cells) and mCherry-CAAX marks endothelial membranes [26,27]. The double transgenic zebrafish were crossed to bmp16 knockout and their F2 progeny embryos were imaged under a laser confocal microscope equipped with an ultrafast camera focusing on the maximal plane of the arterial valves to enable quantitative analysis of valve leaflet morphology within the cardiac cycle (Figure 2A,B).

Loss of bmp16 resulted in persistent malformation of the arterial valves. The arterial valve interleaflet distance, measured between leaflet tips at valve closure (Figure 2B), was significantly increased in bmp16 knockout larvae by 45% at 72 hpf, 48% at 96 hpf, and 37% at 120 hpf compared to wild-type siblings (Figure 2C). In addition, the widths of the left and right leaflet bases were increased by 216% and 131%, respectively, compared to wild-type controls (Figure 2D). Conversely, the lengths of the left and right leaflets were significantly reduced by 52% in bmp16 knockout zebrafish (Figure 2E). These data indicate that bmp16 is essential for the proper elongation, shaping, and positioning of the arterial valve leaflets.

LDN-193189, a dorsomorphin derivative, is a selective inhibitor of BMP type I receptors, activin receptor-like kinase 2 (ALK2) and ALK3 [28]. LDN-193189 treatment starting from 10 hpf suppresses expression of id1, id2a, and id2b, as well as EGFP fluorescence in the arterial valves of Tg(BRE:EGFP) embryos without affecting gross morphology (Figure S2). Consistent with the bmp16 knockout phenotype, LDN treatment also increased the interleaflet distance by 36%, 42% and 125% at 72 hpf, 96 hpf and 120 hpf, respectively, relative to vehicle-treated controls (Figure 2F). Similarly, LDN-treated embryos exhibited significantly increased leaflet widths (left: 165% and right: 153%; Figure 2G) and decreased leaflet lengths (left 60% and 59%) compared to vehicle controls (Figure 2H). Collectively, these data demonstrate that both bmp16 knockout and pharmacological inhibition of BMP signaling disrupt the structural integrity of developing arterial valves, suggesting that bmp16 is essential for proper arterial valve formation in zebrafish.

2.3. Bmp16 Knockout Reduces Valvular Cell Proliferation and Population

Recent studies have established that zebrafish arterial valvular cells originate from SHF progenitors and maintain expression of the conserved transcription factor Sox9 in valvular progenitor cells and valvular interstitial cells [7,29]. To investigate the cellular basis underlying the arterial valve structural defects in bmp16 knockout, we crossed the bmp16 knockout line with the Tg(7×TCF:EGFP) reporter [30], a reporter line we previously found labels arterial valvular cells [31], and performed immunofluorescence at 96 hpf using an anti-Sox9 antibody (Figure 3A). In bmp16 knockout embryos, the numbers of Sox9⁺ and Sox9⁺; EGFP⁺ double positive valvular cells were reduced by 21% and 16%, respectively, compared to wild-type siblings (Figure 3B). Similarly, LDN-193189-treated embryos exhibited a 39% reduction in Sox9⁺EGFP⁺ valvular cells relative to vehicle-treated controls (Figure 3C). Collectively, these data demonstrate that loss of bmp16 and BMP inhibition reduces the Sox9-positive arterial valve cell population.

To determine whether the reduced valvular cell population results from impaired cell proliferation, we treated Tg(7×TCF:EGFP) embryos with LDN-193189 and performed 5-ethynyl-2′-deoxyuridine (EdU) incorporation assays to label proliferating cells at 72 hpf (Figure 3D). Quantitative analysis revealed that the number of EdU⁺; EGFP⁺ double-positive cells within the aortic valves of LDN-193189 treated embryos was decreased by 64% compared to vehicle-treated controls (Figure 3E), demonstrating that inhibition of BMP signaling suppresses valvular cell proliferation.

Collectively, these findings suggest that bmp16 loss and/or BMP inhibition reduce the proliferation of arterial valvular cells, leading to a reduced population of Sox9-positive valvular cells. This deficit in valvular cell number ultimately contributes to the structural malformations of the arterial valve.

2.4. Bmp16 Knockout Impairs Arterial Valve and Cardiac Function in Zebrafish

The primary function of the arterial valve is to ensure unidirectional cardiac output by preventing blood backflow into the ventricle during diastole. Given the severe structural dysplasia of the arterial valve in bmp16 knockout larvae, we next assessed whether these morphological defects translate into functional insufficiency. To visualize blood flow dynamics in live embryos, we used the Tg(gata1:DsRed), which labels red blood cells with DsRed [32], and performed light-sheet fluorescence microscopy (LSFM) at 72, 96 and 120 hpf. In wild-type embryos, outflow blood cells remained confined within the bulbus arteriosus (BA) during ventricular diastole, with no evidence of retrograde flow into the ventricle. In contrast, bmp16 knockout embryos exhibited pronounced blood regurgitation from BA into the ventricle at all developmental stages examined, suggesting severe arterial valve incompetence (Figure 4A,B). Consistent with the genetic knockout phenotype, LDN-193189-treated embryos also displayed a significant increase in the number of regurgitated red blood cells compared to vehicle-treated controls (Figure 4C). These results indicate that loss of bmp16 or inhibition of BMP signaling disrupts arterial valve structural integrity, leading to valvular insufficiency and blood regurgitation.

Valvular insufficiency can alter cardiac hemodynamics, increase ventricular pressure load and ultimately induce pathological cardiac remodeling. To assess the downstream impact of BMP signaling inhibition on ventricular structure and cardiac function, we treated Tg(cmlc2:EGFP) embryos with LDN-193189 and performed LSFM at 72, 96, and 120 hpf to assess ventricular morphology. Consistent with compromised valvular function, LDN-193189-treated hearts exhibited significant enlargement of both end-diastolic and end-systolic ventricular chambers compared to vehicle-treated controls (Figure 4D). Quantitative analysis of maximal ventricular area revealed that end-diastolic ventricular area was increased by 41%, 66%, and 28% at 72, 96, and 120 hpf, respectively (Figure 4E), while end-systolic ventricular area was elevated by 32%, 54%, and 28% at 72, 96, and 120 hpf, respectively (Figure 4F). These findings suggest that LDN-193189 treatment causes progressive ventricular dilation. Consistent with compromised cardiac function, LDN-193189-treated embryos developed pronounced pericardial edema and exhibited reduced body length relative to controls (Figure 4G–I). Collectively, these results indicate that inhibition of BMP signaling disrupts both valvular integrity and overall cardiac function in zebrafish embryos.

2.5. Bmp16 Regulates Notch1b Expression During Arterial Valve Morphogenesis

The Notch signaling pathway is a key regulator of aortic valve development and postnatal valve homeostasis [9]. In zebrafish, notch1b is essential for sensing hemodynamic forces within the developing outflow tract and coordinating proper arterial valve leaflet formation [11]. To determine whether bmp16 knockout affects Notch signaling, we examined notch1b expression. WISH analysis revealed prominent notch1b expression in the arterial and atrioventricular valves of wild-type embryos at 72 hpf (Figure 5A). In contrast, notch1b expression was markedly reduced in bmp16 knockout embryos (Figure 5A). This reduction persisted into adulthood, with notch1b transcript levels decreased by 30% in adult bmp16^−/− hearts relative to wild-type controls (Figure 5B). Likewise, LDN-193189 treatment significantly reduced notch1b expression in the arterial valves at both 72 and 96 hpf (Figure 5C). Quantitative RT-PCR on dissected embryonic hearts confirmed a 43% reduction in notch1b transcript levels at 96 hpf following LDN-193189 treatment compared to vehicle-treated control (Figure 5D). These findings indicate that both genetic knockout of bmp16 and pharmacological inhibition of BMP signaling downregulate notch1b expression during zebrafish arterial valve development, supporting a functional BMP-Notch axis in this process.

To determine whether Notch signaling deficiency contributes to the arterial valve defects observed in bmp16 knockout embryos, we performed a rescue experiment using a selective Notch signaling pathway activator, Yhhu-3792, which can promote the expression of Notch target genes including Hes3 and Hes5 [33]. Wild-type and bmp16 knockout Tg(gata1:DsRed) embryos were treated with DMSO or 1 μM Yhhu-3792 from 24 to 72 hpf, and blood cell dynamics were monitored with LSFM at 72 hpf, 96 hpf, and 120 hpf to assess valve function (Figure 5E). As expected, wild-type embryos treated with either DMSO or Yhhu-3795 exhibited normal unidirectional blood flow, with red blood cells confined to the bulbus arteriosus during ventricular diastole (Figure 5F). In contrast, bmp16 knockout embryos displayed pronounced retrograde blood flow into the ventricle (Figure 5F,G). Strikingly, Yhhu-3792 treatment on bmp16^−/− embryos sharply reduced the number of regurgitated blood cells to levels statistically indistinguishable from those in wild-type embryos at all developmental stages examined (Figure 5G). This functional rescue demonstrates that Notch signaling deficiency is a critical downstream consequence of impaired BMP signaling in bmp16 knockout, and that restoring Notch pathway activity is sufficient to mitigate the valve defects caused by bmp16 loss. Collectively, these data establish a functional BMP-Notch signaling axis in embryonic arterial valve development, wherein bmp16 regulates Notch pathway activity via upregulation of notch1b to ensure proper valve morphogenesis and function.

3. Discussion

Valvulogenesis is a complex, spatiotemporally regulated process governed by the coordinated activity of multiple conserved signaling pathways, and dysregulation of these networks is a major cause of congenital valve defects in humans [34]. In the present study, we identify the teleost-specific BMP ligand bmp16 as a novel, non-redundant regulator of arterial valve morphogenesis in zebrafish. Through complementary genetic loss-of-function and pharmacological approaches, we demonstrate that bmp16-mediated BMP signaling is essential for the proliferation and maintenance of Sox9-positive valvular cells. Loss of bmp16 results in structural valve defects and functional insufficiency, leading to pathological blood regurgitation, ventricular dilation, and systemic edema. Mechanistically, we demonstrate that BMP signaling is required for normal expression of notch1b in developing valves, and that pharmacological activation of Notch signaling rescues the valve dysfunction in bmp16 knockout larvae. Collectively, these findings establish a functional BMP-Notch signaling axis essential for vertebrate valve formation and function, providing new insights into the molecular mechanisms underlying valvulogenesis and congenital valve diseases.

Although BMP signaling has long been recognized as essential for heart and valve development [35], the ligand-specific functions that govern valve morphogenesis remain incompletely characterized. In mammals, Bmp2 and Bmp4 are expressed in the AVC and/or OFT myocardium during valve differentiation and are required for valve primordium induction and endocardial cushion formation [13,15,36], while Bmp6 and Bmp7 exhibit functional redundancy in valve morphogenesis and chamber septation [17]. In zebrafish, bmp4 expression is restricted to the AVC and OFT myocardium during early cardiogenesis [31,37], but its role in arterial valve development has not been systematically investigated. Our study adds bmp16 to this growing repertoire of BMP ligands with specialized functions in cardiac valve morphogenesis. Notably, bmp16 is uniquely retained in teleost genomes [18,19], suggesting that it may have evolved to fulfill lineage-specific roles in finfish cardiac development. The valve defects observed in bmp16 knockout (Figure 2) indicate that bmp16 is not functionally redundant with other zebrafish BMP ligands (e.g., bmp2a/b, bmp4) during arterial valve formation, though partial redundancy cannot be excluded. Future studies employing double or triple knockout models will be critical to dissect the combinatorial roles of BMP ligands in valvulogenesis and determine whether bmp16 has acquired unique biochemical properties that distinguish it from its bmp2/4 paralogs.

Sox9 is an evolutionarily conserved transcription factor that plays a pivotal role in valvulogenesis by mediating cell fate specification, extracellular matrix (ECM) synthesis, and the maintenance of valvular progenitor cell identity [29,38]. Our findings demonstrate that loss of bmp16 or inhibition of BMP signaling reduces the number of Sox9⁺ valvular cells and suppresses valvular cell proliferation (Figure 3), indicating that bmp16-mediated BMP signaling is not only essential for valve induction and EndoMT, but also vital for the expansion of valvular progenitor cells and/or valvular interstitial cells. These findings align with the established role of BMP signaling in regulating cell proliferation during embryogenesis [39]. The hypoplastic valve phenotype observed in bmp16 knockout larvae likely stems from insufficient valvular cell proliferation, which fails to support the extensive tissue remodeling required for leaflet elongation and maturation after 72 hpf. Intriguingly, BMP signaling has also been shown to regulate Sox9 expression directly via Smad-dependent transcription in other tissues [40], raising the possibility that bmp16-mediated BMP signaling may directly activate sox9 transcription in valvular cells. Future studies employing chromatin immunoprecipitation and/or reporter assays will be necessary to determine whether Smad proteins bind to the sox9 promoter in valvular cells, and whether this interaction is dependent on bmp16.

Crosstalk between BMP and Notch pathways is a recurring theme in vertebrate organogenesis, including heart development [9]. We observed reduced notch1b expression in both bmp16 mutants and LDN-193189-treated embryos, while pharmacological activation of Notch signaling with Yhhu-3792 rescued the regurgitation phenotype in bmp16 mutants (Figure 5). These findings indicate that BMP signaling positively regulates Notch pathway activity in the developing arterial valve. This finding complements previous work in mice showing that endocardial Notch1 signaling induces Wnt4 expression, which in turn upregulates myocardial Bmp2 to promote EndoMT [41]. Together, these studies reveal a bidirectional regulatory loop between Bmp and Notch pathways: Notch signaling upregulates BMP ligand expression to initiate EndoMT, while BMP signaling maintains Notch pathway activity to support valvular cell proliferation and maturation. This feedback loop likely reinforces and stabilizes the valvular progenitor cell state, ensuring coordinated proliferation and differentiation during leaflet morphogenesis.

4. Materials and Methods

4.1. Zebrafish Strain and Maintenance

Zebrafish of the AB strain were used as the wild-type, the transgenic lines were Tg(BRE:EGFP)^pt509 [21], Tg(cmlc2:EGFP)^twu34, Tg(gata1:DsRed)^sd2 [32], Tg(kdrl:mCherry-CAAX)^s896 [27], Tg(tp1:EGFP)^um14 [26] and Tg(7×TCF-Xla.Siam:EGFP)^ia4 [30], and the knockout line was bmp16^sus206 (this study). They were maintained in a dedicated zebrafish breeding unit (ESEN Science & Technology, Beijing, China) following standard husbandry conditions. Embryos were obtained through natural mating and cultured at 28.5 °C. Embryos were treated with 0.2 mM 1-phenyl-2-thiourea (PTU, p110661–25g, Aladdin, Shanghai, China) to block pigmentation for further imaging analysis.

4.2. Whole-Mount In Situ Hybridization

Whole-mount in situ hybridization (WISH) was performed according to standard protocols, using PCR-amplified DNA fragments of the bmp16 and notch1b genes inserted into a customized vector and subsequently used as templates to transcribe a Digoxigenin-labeled RNA probe in vitro. The primer used to amplify probe templates from zebrafish cDNA library were listed as follows:

bmp16-probe-forward:5′-GCCCATACCATCCAGAACCT-3′

bmp16-probe-reverse:5′-TATCGACAGCCACATCCCTC-3′

notch1b-probe-forward:5′-TACGTGGTTCTGGCAGGATT-3′

notch1b-probe-reverse:5′-ATTCTGTCTTGCGCGATGTC-3′

4.3. Generation of Bmp16 Knockout Zebrafish

The bmp16 knockout allele was generated using CRISPR/Cas9-mediated genomic editing as described [42]. Briefly, guide RNAs targeting the signal peptide region (5′-GGCTCATCAGGTCACCAGGA-3′) and a site proximal to its stop codon (5′-GGCATGGTGGTGGAGGGATG-3′) of the bmp16 gene were generated. Subsequently, sgRNA was transcribed in vitro from sgDNA using the MAXIscript T7 Transcription Kit (AM1314, Thermo Fisher Scientific, Carlsbad, CA, USA) following the manufacturer’s instructions and co-injected with Cas9 protein into one-cell stage zebrafish embryos. Injected embryos were raised to adulthood and screened for germline transmission using PCR, followed by Sanger sequencing of amplicons from F1 animals to confirm the mutation. The bmp16 knockout used in this study was genotyped using the following primers, which can produce amplicons of 321 bp or 187 bp for wild-type or knockout alleles, respectively:

bmp16-1f: 5′-TTCTTTTACCTCTCTTTCCGCA-3′

bmp16-2f: 5′-TGCTTTCTTTTGCCTCGGAG-3′

bmp16-2r: 5′-TCGTGTCCATTTACCTCGCT-3′

4.4. Drug Treatment

LDN-193189 (S2618, Selleck, Houston, TX, USA) was dissolved in 0.01 M HCl to prepare a 3 mM stock solution, which was aliquoted and stored at −20 °C. For embryo treatments, the stock solution was thawed and diluted in egg water without methylene blue to a final concentration of 1 μM. Embryos were incubated in fresh egg water with or without LDN-193189 from 10 hpf until imaging.

Yhhu-3792 (HY-120782, MedChemExpress, Monmouth Junction, NJ, USA) was dissolved in DMSO to prepare a 10 mM stock solution, which was aliquoted and stored at −20 °C in dark. For embryo treatments, the stock solution was thawed and diluted in egg water without methylene blue to a final concentration of 1 μM. Zebrafish embryos were incubated with fresh egg water with DMSO or Yhhu-3792 from 24 to 72 hpf. Then the embryos were rinsed twice with fresh egg water to remove residual drug.

4.5. Imaging

Embryonic zebrafish were carefully excised using fine-tip tweezers and subsequently mounted on the bottom of glass-bottom Petri dishes containing 0.8% low-melting-point agarose (A600015, Sangon Biotech, Shanghai, China). Bright-field images were acquired using a Axio Zoom V16 stereomicroscope (Zeiss, Oberkochen, Germany). A LiTone XL light-sheet microscope (Light innovation Technology, Hong Kong, China) and an A1R HD25 confocal microscope (Nikon Instruments, Tokyo, Japan) were used to visualize the phenotype. Acquired light-sheet and confocal images were processed via ImageJ software (version 1.54f) and/or Imaris X64 software (version 9.0.1).

4.6. RNA Extraction and Quantitative Reverse Transcription PCR

Total RNA was extracted from 30 zebrafish embryos, 300 embryo zebrafish hearts, or 6 adult zebrafish hearts per biological replicate using Trizol (R401-01-AA, Vazyme, Nanjing, Jiangsu, China). First-strand cDNA was synthesized using HiScript II Q RT SuperMix for qPCR (R223-01, Vazyme, Nanjing, Jiangsu, China). Real-time quantitative polymerase chain reaction (RT-qPCR) was performed using 2× SYBR Green qPCR Master Mix (B1202, Selleck, Houston, TX, USA). Gene-specific primers were used for quantitative polymerase chain reaction (qPCR) as stated below, and the housekeeping gene eef1a1l1 was used as an internal control for the qPCR analysis. Gene expression was calculated using the 2^−ΔΔCT method. Three biological replicates were performed, each consisting of 150 pooled samples from individual hearts. The mean relative expression of each gene between groups was used for statistical analysis.

id1-forward:5′-GCAAACTGAAGGAGCTGGTG-3′

id1-reverse:5′-CGGCACATGATCCGGTCATC-3′

id2a-forward:5′-CGATAACCAGCCTGCATCAC-3′

id2a-reverse:5′-GTCCTGCTGTCCTCTGTGAT-3′

id2b-forward:5′-TTGGACCTTCAGATCGCACT-3′

id2b-reverse:5′-AACGAGACAGGGCTATGAGG-3′

bmp16-forward:5′-CAGTCCGATGCTTTCACCAC-3′

bmp16-reverse:5′-TGAGAACCTCCACTTGACCC-3′

notch1b-forward:5′-GTCAATGGCTTCACTTGTCTGT-3′

notch1b-reverse:5′-TGAGAACCTCCACTTGACCC-3′

4.7. Immunofluorescence and EdU Staining

Zebrafish embryos were fixed overnight at 4 °C in freshly prepared 4% paraformaldehyde (PFA), and permeabilized in pre-chilled acetone. Following blocking, embryos were incubated in primary antibody solution at 4 °C overnight. Embryos were then washed and incubated in secondary antibody solution overnight at 4 °C. DAPI (C1002, Beyotime, Shanghai, China) was used for counterstaining. Primary antibodies used were rabbit anti-Sox9 (1:500, ab185230, Abcam, Cambridge, UK) and chicken anti-GFP (1:200, ab13970, Abcam, Cambridge, UK). Secondary antibodies were goat anti-chicken Alexa Fluor 488 (1:500, ab150173, Abcam, Cambridge, UK) and goat anti-rabbit Alexa Fluor 594 (1:500, ab150084, Abcam, Cambridge, UK).

For EdU labeling, embryos were injected intrapericardially with 1 nL of 1 mM EdU (ST067, Beyotime, Shanghai, China) and incubated for 4 h at 28.5 °C prior to fixation in 4% PFA for 2 h at room temperature. EdU signals were stained following standard procedure with Alexa Fluor 647 Azide (A10277, Thermo Fisher Scientific, Carlsbad, CA, USA) and nuclei were counterstained with DAPI.

4.8. Statistical Analysis

All quantitative data are presented as mean ± standard deviation (SD). Statistical analyses were performed using unpaired Student’s t-test, one-way or two-way ANOVA as appropriate to determine statistical significance. The statistical significance is denoted as follows: ns, not significant (p > 0.05); *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.