Abstract
The sinoatrial node (SAN) is the primary pacemaker in the heart. During cardiogenesis, Shox2 and Nkx2-5 are co-expressed in the junction domain of the SAN and regulate pacemaker cell fate through a Shox2-Nkx2-5 antagonism. Cx40 is a marker of working myocardium and an Nkx2-5 transcriptional output antagonized by Shox2, but the underlying regulatory mechanisms remain elusive. Here we characterized a bona fide myocardial-specific Gja5 (coding gene of Cx40) distal enhancer consisting of a pair of Nkx2-5 and Shox2 co-bound elements in the regulatory region of Gja5. Transgenic reporter assays revealed that neither element alone, but the conjugation of both elements together, drives myocardial-specific transcription. Genetic analyses confirmed that the activation of this enhancer depends on Nkx2-5 but is inhibited by Shox2 in vivo, and its presence is essential for Gja5 expression in the myocardium but not the endothelial cells of the heart. Furthermore, chromatin conformation analysis showed an Nkx2-5-dependent loop formation between these two elements and the Gja5 promoter in vivo, indicating that Nkx2-5 bridges the conjugated activation of this enhancer by pairing the two elements to the Gja5 promoter.
Introduction
The morphogenesis and physiological maturation of the cardiac conduction system are complicatedly coupled processes, which require a precise spatiotemporal expression of various transcription factors, including Shox2 and Nkx2-5 (Espinoza-Lewis et al., 2009; Hatcher and Basson, 2009; Pashmforoush et al., 2004; van Weerd and Christoffels, 2016). Nkx2-5 is essential for murine cardiogenesis (Lyons et al., 1995) but its expression was thought to be excluded from the SAN (Espinoza-Lewis et al., 2009), and ectopic expression of Nkx2-5 in the developing SAN inhibits pacemaker cell differentiation (Espinoza-Lewis et al., 2011). In contrast, Shox2 is essential for SAN development and represses Nkx2-5 expression in the developing SAN head domain (Espinoza-Lewis et al., 2009; Liu et al., 2014), thus ensuring the absence of Nkx2-5 in the SAN head from the surrounding Nkx2-5+ working myocardium. It was recently demonstrated the existence of a Shox2 and Nkx2-5 co-expression domain in the developing SAN, genetically dividing the SAN into a Shox2+/Nkx2-5− “head” domain and a Shox2+/Nkx2-5+ “jFunction” domain (Li et al., 2019; Ye et al., 2015). In the Shox2+/Nkx2-5+ SAN junction domain, a Shox2-Nkx2-5 antagonistic mechanism appears to direct the pacemaker cell fate (Ye et al., 2015).
Gap junction alpha-5 (Gja5)/connexin 40 (Cx40), encoded by Gja5 in mice, plays a critical role in mediating electrical conduction and diffusion of cellular substances in the heart (Bagwe et al., 2005; de Wit et al., 2003; Simon and McWhorter, 2002). Mutations in GJA5 in humans are associated with atrial fibrillation (AF) (Bai, 2014; Firouzi et al., 2004; Gollob et al., 2006; Juang et al., 2007) and tetralogy of Fallot (TOF) (Greenway et al., 2009; Guida et al., 2013). During murine cardiogenesis, Gja5 is selectively expressed in the atrial myocardium, the ventricular conduction system, and the arterial endothelial cells (Beyer et al., 2011; Christoffels and Moorman, 2009; van Weerd and Christoffels, 2016). Cardiac Gja5 expression depends on Nkx2-5 (Dupays et al., 2005; Espinoza-Lewis et al., 2011; Linhares et al., 2004) but is repressed by Shox2 (Blaschke et al., 2007; Espinoza-Lewis et al., 2009), and is controlled tightly by the Shox2-Nkx2-5 antagonism in the Nkx2-5+/Shox2+ domains (Li et al., 2019; Ye et al., 2015). However, the underlying mechanisms utilized by Shox2 and Nkx2-5 to control Gja5 expression remain unknown.
The precise spatiotemporal control of transcriptional networks is critical for cardiac development (Bruneau, 2008; Olson, 2006). Whereas proximal promoters provide a basis for gene transcription, the tissue-specific gene expression is dependent mainly on the interactions between existing tissue-specific transcription factors and a variety of regulatory DNA sequences classified as distal enhancers (Levine, 2010; Nord et al., 2013; Visel et al., 2009). In the heart, numerous cardiac-specific distal enhancers have been characterized by comparative genomic analyses searching for evolutionarily conserved elements and genome-wide profiling of active enhancer associated epigenetic marks or co-activator protein binding sites in heart tissues (Blow et al., 2010; Dickel et al., 2016; He et al., 2011; May et al., 2012; Narlikar et al., 2010; Paige et al., 2012; Wamstad et al., 2012). Although the combinatorial binding of cardiac transcription factors at proximal promoters and their physiological significance in transcription regulation have been studied extensively (Bruneau, 2002), the functions of cardiac-specific distal enhancers and how transcription factors regulate them remain largely unexplored.
In this study, we report the characterization of a myocardial-specific Gja5 distal enhancer, which is formed by a pair of Shox2 and Nkx2-5 co-bound elements (named as Gja5-S1 and Gja5-S2) spaced by a 12-Kb non-coding region downstream of Gja5. Neither Gja5-S1 nor Gja5-S2 alone, but the conjugation of both elements termed as Gja5-eh, displayed robust enhancer activity recapitulating the endogenous Gja5 expression pattern in the Nkx2-5+ domain of the developing and adult heart. Genetic analyses showed that the Gja5-eh enhancer activity is indeed dependent on Nkx2-5 but is inhibited by Shox2 in vivo. Mice bearing the ablation of both Gja5-S1 and Gja5-S2 exhibit drastically reduced endogenous Gja5 expression only in the myocardium but not the arterial endothelial cells, illustrating that these two elements are indispensable for Gja5 expression specifically in the myocardium in vivo. Moreover, chromatin conformation analysis of mouse embryonic hearts showed that both Gja5-S1 and Gja5-S2 have frequent contacts with the Gja5 promoter in an Nkx2-5-dependent manner, indicating that Nkx2-5 bridges the enhancer activation by conjugating Gja5-S1 and Gja5-S2 together with the Gja5 promoter.
Results and Discussion
Identification of Gja5-S1 and Gja5-S2 and generation of Gja5-eh-LacZ reporter mouse line
To understand the functional mechanisms underlying the Shox2-Nkx2-5 antagonism that regulates cell fate in the developing SAN, we revisited the genomic profiles of Shox2 and Nkx2-5 ChIP-seq assays on the right atrial tissues of E12.5 mouse embryonic heart, which displayed a substantial genome-wide co-occupancy of Shox2 and Nkx2-5 (Ye et al., 2015). Among the co-binding peaks, two sites (termed Gja5-S1 and Gja5-S2) co-occupied by Shox2 and Nkx2-5 downstream of Gja5 raised our particular interest (Fig. 1A), because Gja5 expression has been proven to be regulated by Shox2 and Nkx2-5 during SAN development (Li et al., 2019; Ye et al., 2015). Gja5-S1 locates at around 9-Kb downstream of the Gja5 coding region, whereas Gja5-S2 sits downstream of Gja5-S1 separated by a 12-Kb non-coding sequence. Additional analysis integrating accessible public data showed that Gja5-S1 and Gja5-S2 are located in the same topological associated domain (TAD) as Gja5 (Dixon et al., 2012), and are both marked as putative open chromatin in embryonic hearts but not in any other tissues at the same stage (ENCODE Project Consortium, 2012) (Fig. S1), further suggesting that these two elements may have cardiac-specific enhancer activity.
To assess the enhancer activity of Gja5-S1 and Gja5-S2 in vivo, we initially cloned each of them into a reporter construct containing the Hsp68 minimal promoter (Hsp68mp) and β-galactosidase coding sequence (LacZ) (Kothary et al., 1989). However, neither element exhibited enhancer activity in transient transgenic reporter assays (Fig. 1C). We wondered whether the two elements are required to function together, and therefore subsequently conjugated both elements together as a fusion fragment, termed Gja5-eh, and generated a reporter construct named Gja5-eh-LacZ to test its enhancer activity in transgenic mouse embryos. Strikingly, all the Gja5-eh-LacZ transgene-positive embryos (15/15) showed beta-galactosidase (β-gal) activity in the developing heart (Fig. 1B, C). Section immunostaining and X-gal staining further revealed the specific expression of LacZ in the myocardium but its absence in the SAN of the transgenic animals, similar to the pattern of endogenous Gja5 expression (Fig. 1B, D). These observations indicate that both Gja5-S1 and Gja5-S2 are required to act together in cis to drive cardiac-specific gene transcription. We subsequently generated Gja5-eh-LacZ permanent transgenic reporter mouse lines to facilitate further studies.
Gja5-eh recapitulates endogenous Gja5 expression in Nkx2-5+ myocardium
Next, we examined whether the enhancer activity of Gja5-eh could faithfully recapitulate the endogenous Gja5 expression in the developing heart. Since Nkx2-5 binds to Gja5-S1 and Gja5-S2 and is essential for Gja5 expression in the heart (Dupays et al., 2005), we conducted co-immunostaining on Nkx2-5, Cx40, and β-gal in the hearts of Gja5-eh-LacZ reporter mice. The results showed co-expression of β-gal and Cx40 in the Nkx2-5+ myocardium of the ventricle and atrium of Gja5-eh-LacZ hearts, not only at the embryonic stage (E12.5) but also at adulthood (P60) (Fig. 2A-D, single-channel images in Fig. S2). Similar to the results seen in the transient transgenic reporter assay, β-gal was not detectable in the Shox2+ SAN region in contrast to its presence in the adjacent Nkx2-5+/Cx40+ atrial myocardium (Fig. 2A, Fig. S2M-P). Notably, we also observed the absence of β-gal in the Nkx2-5−/Cx40+ coronary arteries (CA) in contrast to its intense expression in the Nkx2-5+/Cx40+ ventricular trabeculae (VT) (Fig. 2D, Fig. S2I-L). Similar observations were found in other developmental stages (data not shown). These results demonstrate that Gja5-eh acts as a functional distal enhancer and recapitulates the endogenous Gja5 expression. However, its enhancer activity is restricted only to the Nkx2-5+ myocardium, suggesting a positive role of Nkx2-5 in the activation of the enhancer.
Gja5-eh enhancer activity is dependent on Nkx2-5 but is inhibited by Shox2
To determine whether Nkx2-5 regulates Gja5-eh enhancer activity in vivo, we compounded Gja5-eh-LacZ allele onto an Nkx2-5-null background using the Nkx2-5Cre knock-in allele (Moses et al., 2001; Ye et al., 2015). As we expected, the enhancer activity of Gja5-eh was abolished in the Nkx2-5Cre/Cre;Gja5-eh-LacZ embryonic hearts as compared to controls (Fig. 2E-J), indicating an indispensable role of Nkx2-5 in the activation of Gja5-eh. Since Shox2 also binds to Gja5-S1 and Gja5-S2 while represses Gja5 expression in the heart (Espinoza-Lewis et al., 2009; Ye et al., 2015), we next examined whether Shox2 inhibits Gja5-eh enhancer activity in vivo. To do this, we compounded Gja5-eh-LacZ allele with ROSA26mTmG and Shox2Cre alleles to generate Shox2Cre/Cre;Gja5-eh-LacZ;ROSA26mTmG mice, which enabled monitoring of Gja5-eh enhancer activity and tracing of Shox2 lineage in Shox2-null background (Sun et al., 2013). In contrast to the absence of β-gal in the Nkx2-5+ cells derived from the Shox2 lineage cells that were labeled by mGFP in controls (Shox2+/Cre;Gja5-eh-LacZ;ROSA26mTmG (Fig. 2K-N), we observed the presence of Nkx2-5+/mGFP+/β-gal+ cells in the SAN junction, but not the SAN head domain, of Shox2Cre/Cre;Gja5-eh-LacZ;ROSA26mTmG mice (Fig. 2O-R), indicating the cell-autonomous ectopic activation of Gja5-eh enhancer in the absence of Shox2. This result is consistent with the ectopic expression of Gja5 in the cardiac structures in Shox2-null embryos (Espinoza-Lewis et al., 2009; Sun et al., 2015), suggesting that Shox2 suppresses Gja5 expression by binding to Gja5-S1 and Gja5-S2.
Gja5-S1 and Gja5-S2 are essential for myocardial Gja5 expression
To investigate whether Gja5-S1 and Gja5-S2 serve as a bona fide distal enhancer that drives endogenous Gja5 expression, we generated a knockout allele termed Gja5Δ14, which carries a deletion of a 14-Kb region that includes both Gja5-S1, Gja5-S2, and the sequence between these two sites (Fig. 3A). As a control, we also generated a pseudo-knockout allele termed Gja5Δ12, in which the 12-Kb sequence between Gja5-S1 and Gja5-S2 was deleted (Fig. 3B). We subsequently examined and compared the endogenous Gja5 expression in the hearts of Gja5Δ14/Δ14, Gja5Δ12/Δ12, as well as wild type mice. Strikingly, we observed that in the perinatal Gja5Δ14/Δ14 hearts, Cx40 expression in the atrial and ventricular trabecular myocardium was almost completely abolished but remained unaltered in the coronary arteries (Fig. 3D, G). We also verified by Western blotting the dramatically decreased levels of Cx40 proteins in the atria and ventricles of P0 Gja5Δ14/Δ14 mice as compared to Gja5+/+ controls (Fig. 3I). Further quantification via real-time qPCR (RT-qPCR) showed that the Gja5 transcription levels remained only 5%±0.2% in the atria and 31%±3.7% in the ventricles of Gja5Δ14/Δ14 mice as compared to their counterparts in Gja5+/+ controls (Fig. 3J). These residual Gja5 expressions in Gja5Δ14/Δ14 groups may largely be contributed by coronary arteries where Gja5 expression is unaffected. In contrast, immunohistochemical staining revealed indiscernible change in Cx40 expression in both myocardium and coronary arteries in Gja5Δ12/Δ12 perinatal hearts as compared to that of Gja5+/+ mice (Fig. 3E, H). These results demonstrate that Gja5-S1 and Gja5-S2 are explicitly required in the myocardium to drive Gja5 expression, consistent with the specific enhancer activity of Gja5-eh in the Nkx2-5+ myocardium, but not in the Nkx2-5− coronary arteries (Fig. S2I-L). Despite the diminished myocardial Gja5 expression in Gja5Δ14/Δ14 hearts, these mice are viable and fertile with no overt phenotype. We subsequently performed surface ECG on adult Gja5Δ14/Δ14 mice along with controls but did not find obvious differences between mutants and controls (N = 10 for each genotype, with 5 males and 5 females, data not shown). Indeed, Gja5 null mutant mice were viable and fertile, with only minor cardiac physiological abnormalities (prolonged P-wave manifested by ECG) (Bagwe et al., 2005). However, simultaneous loss of Cx37 and Gja5(Cx40) results in perinatal lethality with severe vascular defect, indicating that Gja5 functions importantly but redundantly in the endothelial lineage (Simon and McWhorter, 2002). Therefore, the normal cardiac function of Gja5Δ14/Δ14 mice, revealed by ECG, can be explained by the unaffected expression of Gja5 in the endothelial lineage. Additionally, although a number of mutations within GJA5 coding regions have been associated with atrial fibrillation, none was reported occurring in the non-coding regulatory element associated with GJA5 so far. We speculate that the pathogenicity of mutations in GJA5 may arise from alterations of biochemical properties of the CX40 protein rather than changes in overall CX40 levels. Admittedly, we acknowledge that the physiological consequences of the 14Kb deletion could be tested in more detail by additional methods, such as echocardiography and optical mapping.
An Nkx2-5-dependent interaction between Gja5-S1/Gja5-S2 and the Gja5 promoter
Recent studies have reported prevalent long-range enhancer-promoter interactions in cardiomyocytes (Montefiori et al., 2018; Rosa-Garrido et al., 2017) and association of disrupted enhancer-promoter looping with cardiac malfunctions (Man et al., 2019). Nkx2-5 was reported to bind to the Gja5 promoter and upregulate its expression in vitro (Linhares et al., 2004), but its functional mechanisms remain elusive. To determine whether Gja5-S1 and Gja5-S2 have direct interaction with the Gja5 promoter and if such interaction requires the presence of Nkx2-5, we performed Chromatin Conformation Capture (3C) (Dekker et al., 2002) assays on E10.5 wild type and Nkx2-5Cre/Cre hearts. The results showed that in wild type groups, fragments embracing primer pairs F2+R1 and F4+R1 bore significantly higher relative interaction frequency (RIF) than other primer pairs (Fig. 4A, B), indicating that both Gja5-S1 and Gja5-S2 form a strong loop with the Gja5 promoter. In comparison, the RIF between Gja5-S1/Gja5-S2 and the Gja5 promoter of Nkx2-5Cre/Cre (Nkx2-5 null) hearts was significantly reduced (Fig. 4A, B). These results demonstrate an Nkx2-5-dependent loop formation between both Gja5-S1/Gja5-S2 and the Gja5 promoter, illustrating a critical enhancer-pairing and promoter-docking function of Nkx2-5 in the conjugated activation of Gja5-S1 and Gja5-S2. This bridging function of Nkx2-5 may result from its ability to form homodimers or heterodimerize with other Nkx2 family transcription factors (Kasahara et al., 2001). Interestingly, we did not observe the binding peaks of Nkx2-5 to the Gja5 promoter from our ChIP-seq data (Fig. 1A, Fig. S1), raising the possibility that Nkx2-5 may have different binding preferences at a context-specific manner.
We also looked for consensus binding sites of Nkx2-5 and Shox2 within both Gja5-S1 and Gja5-S2 by searching JASPAR-2020, a recently updated database of in vitro DNA binding motifs (Fornes et al., 2019). We found numerous predicted transcription factor binding sites within Gja5-S1 (Fig. S4) and Gja5-S2 (Fig. S5). However, there is only one Nkx2-5 binding site (-AACCACTCCAG-) inside Gja5-S1 but not Gja5-S2, while no Shox2 binding site was found. Meanwhile, binding sites of two other NK family transcription factors, Nkx2-3 and Nkx2-8, were found in both Gja5-S1 (-GCACTTGAG-) and Gja5-S2 (-CCACTTGAC-). All these three sites share motif -(C/G)CACT(T/C)-, consistent with our Nkx2-5 ChIP-seq results (Ye et al., 2015) and results from another study (He et al., 2011). Our previous studies showed that Shox2 has distinct binding sites among different tissues during organogenesis (Wang et al., 2020; Xu et al., 2019; Ye et al., 2016, 2015), but a common motif is yet to be characterized. The binding preferences of Shox2 may be partially dependent upon other co-factors, such as Hox family transcription factors, as reported previously (Ye et al., 2016).
Based on the work presented here, we propose that Gja5-S1 and Gja5-S2 are a pair of Nkx2-5-Shox2 co-responsive elements that act together as a myocardial-specific Gja5 distal enhancer by interacting with the Gja5 promoter (Fig. 4C). We acknowledge that as the ChIP-seq and 3C experiments were performed on bulk tissues consisting of both myocardium and SAN, we can not conclude that Gja5-S1 and Gja5-S2 are co-bound by both Nkx2-5 and Shox2 simultaneously. One interpretation of the model could be that in the SAN junction, Shox2 competes with Nkx2-5 for binding to these two elements and thus disrupts the interaction of the enhancers with the Gja5 promoter by keeping the enhancers at a distal status. Another interpretation is that Shox2 and Nkx2-5 form non-functional protein complexes that can not bind to DNA or have diminished DNA binding specificities, since Shox2 and Nkx2-5 have physical interaction in vitro (Ye et al., 2015).
While we report that Gja5-S1 alone does not have enhancer activity (Fig. 1C), a 705-bp Gja5 enhancer that embraces Gja5-S1 was reported (Hashimoto et al., 2019). Gja5-S1 (553-bp) locates in the center of this 705-bp enhancer, flanked by an additional 152-bp sequence (90 bp upstream/62 bp downstream). This additional 152-bp sequence, which contains Hand2 binding sites and is enriched in H3K27ac, appears to contribute to the discrepant enhancer activity via H3K27ac deposition recruited by Hand2 (Creyghton et al., 2010; Hashimoto et al., 2019). In contrast, we found that Gja5-S2 is marked by H3K4me2, H3K4me3, and occupied by CCCTC-binding factor (CTCF) (Fig. S3) (ENCODE Project Consortium, 2012). Both H3K4me2 and H3K4me3 are associated with active transcription (Heintzman et al., 2007; Kim and Buratowski, 2009; Mikkelsen et al., 2007; Santos-Rosa et al., 2002), and CTCF is a critical regulator of enhancer-promoter interactions and chromatin topology (Barrington et al., 2019; Nora et al., 2017; Shin, 2019; Wutz et al., 2017). These distinct histone modifications and binding of chromatin modifiers between Gja5-S1 and Gja5-S2 may display complementary actions that contribute to their conjugated activation. Based on the proposed model, we also speculate that a smaller deletion of either Gja5-S1 or Gja5-S2 should abolish Gja5 expression in the myocardium and phenocopy Gja5Δ14/Δ14 mice.
The behaviors of enhancers are multifaceted. During organogenesis, tissue-specific enhancers of the same gene could function redundantly to achieve phenotypic robustness by buffering the loss-of-function mutations of individual enhancers (Frankel et al., 2010; Osterwalder et al., 2018). Strong enhancers may compete for contacting promoters, and enhancers of weak or intermediate strength by themselves may function additively to enhance the expression of the same gene (Bothma et al., 2015). One enhancer may contain multiple binding sites for different transcription factors and is synergistically activated through proximal protein-protein interactions (Ambrosetti et al., 1997; Anderson et al., 2017; Grieder et al., 1997; Hashimoto et al., 2019). Here, our studies show that the nature of enhancers could be even more complicated. This conjugated activation of myocardial-specific activation of Gja5 by two regulatory elements may represent a novel mechanism for accuracy assurance of tissue-specific gene expression. Although Gja5-eh is a recombinant fusion of Gja5-S1 and Gja5-S2, we indeed observed a striking consistency between Gja5-eh-LacZ enhancer activity and molecular phenotype of Gja5Δ14/Δ14 mice. We reason that the fusion of Gja5-S1 and Gja5-S2 has mimicked their actual close contact at the Gja5 promoter loci in vivo, which makes Gja5-eh represent the spatiotemporal behaviors of Gja5-S1 and Gja5-S2 precisely. The Gja5Δ12 allele actually also simulates the fusion of Gja5-S1 and Gja5-S2 in situ and represents their conjugated activation (Fig. 3B, E, H). Collectively, these observations point out a unique pattern of conjugated enhancer activation and provide novel insights into characterization, design, and optimization of tissue/lineage-specific enhancers that may benefit biomedical research or therapeutic applications.
Materials and Methods
Cloning and plasmids
The 553-bp Gja5-S1 and 594-bp Gja5-S2 fragments were amplified by PCR, respectively, using primers Gja5-S1-F (5’ -AGTCTGATGACAACTTGTGAGAATCG-3’), Gja5-S1-R (5’-GGGTGACAGTAAGAAATGTCAGGTG-3’), and Gja5-S2-F (5’ -AATGAACAGGAAAGTGGGAG G-3’), Gja5-S2-R (5’-CAGGGCGGTCAGGCAG-3’). The 1147 bp fusion fragment Gja5-eh was synthesized commercially (Qinglanbiotech.com). Each of the fragments was subsequently cloned into plasmid hsp68-LacZ (Kothary et al., 1989) to generate Gja5-S1-LacZ, Gja5-S2-LacZ, and Gja5-eh-LacZ constructs for transient transgenic reporter assays. Gja5-eh-LacZ was also used for generating permanent transgenic reporter mouse lines.
Mouse models
The generation and genotyping methods of Shox2Cre, Nkx2-5Cre and ROSA26mTmG mice have been described previously (Moses et al., 2001; Muzumdar et al., 2007; Sun et al., 2013). Generation of transgenic embryos and mice were performed as described previously (Ye et al., 2016). The Gja5Δ14 and Gja5Δ12 alleles were generated by CRISPR/Cas9-mediated genome editing (Wang et al., 2013). Additional details of CRISPR/Cas9-mediated genome editing and transgenesis methods are described in the Supplementary Materials and Methods. All animal work in this study was approved by The Tulane University Institutional Animal Care and Use Committee (IACUC). Sample sizes were empirically determined based on previous experimental procedures (Ye et al., 2016). Mouse embryos were excluded from further analysis only if they did not carry alleles of interest.
Histology, immunohistochemistry, and X-gal staining
Hearts were harvested from properly euthanized mice or staged embryos, fixed in ice-cold 4% paraformaldehyde (PFA) overnight at 4◻, dehydrated through gradient ethanol, cleared in xylene, embedded in paraffin, and sectioned at 5μm for immunostaining as described previously (Li et al., 2019). The primary antibodies used in this study were: anti-Hcn4 (ab32675, Abcam; 1:200), anti-Nkx2-5 (AF2444, Novus Biologicals; 1:200), anti-Cx40 (Cx40-A; Alpha Diagnostic International; 1:200), anti-GFP (sc-9996, Santa Cruz Biotechnology; 1:200), anti-β-galactosidase (ab9361, Abcam; 1:200). The secondary antibodies were used at 1:1000 and all from Jackson ImmunoResearch: donkey anti-goat (705-545-147), donkey anti-mouse (705-585-151), donkey anti-rabbit (711-585-152, 711-545-152), donkey anti-rat (712-585-153), donkey anti-chicken (703-606-155).
For whole-mount X-gal staining, staged embryos were fixed with freshly-prepared, ice-cold 2% PFA and 0.2% glutaraldehyde in PBS for 1 hour at 4◻, rinsed with staining solution (5mM potassium ferricyanide, 5mM potassium ferrocyanide, 2mM MgCl2, 0.01% sodium deoxycholate, and 0.02% IGEPAL CA-630), followed by addition of X-gal stock solution (40mg/ml in dimethylformamide) at 1:40. For section X-gal staining, fixed embryos were dehydrated through gradient sucrose/OCT, embedded in OCT, cryo-sectioned at 10μm, and stained for X-gal as described above.
Western Blot and real-time quantitative PCR (RT-qPCR)
P0 pups of wild type and Gja5Δ14/Δ14 mice were euthanized by decapitation. Hearts were then isolated, and atria were separated from ventricles. Samples from about 20 pups of each genotype were pooled, homogenized, and subjected to protein extraction or RNA extraction followed by cDNA preparation, as previously described (Li et al., 2019; Ye et al., 2015). For Western Blot, the following primary antibodies were used: anti-Cx40 (Cx40-A; Alpha Diagnostic International; 1:2000), anti-GAPDH (2118S; Cell Signaling Technology; 1:2000). The secondary antibody used was: IRDye® 800CW Donkey anti-Rabbit (926-32213; LI-COR; 1:10000). For RT-qPCR, the following primers were used: Gja5 (F, 5’ -GGTCCACAAGCACTCCACAG-3’; R, 5′-CTGAATGGTATCGCACCGGAA-3′), Gapdh (F, 5’ -ATCAAGAAGGTGGTGAAGCAG-3’; R, 5′ -GAGTGGGAGTTGCTGTTGAAGT-3′).
Chromatin Conformation Capture (3C) assays
Briefly, embryonic hearts were dissected under a microscope, minced into small pieces and digested with Accutase (Invitrogen 00-4555-56) at 37◻ for about 20 minutes, passed through a 70μm cell strainer, followed by formaldehyde fixation and quenched with glycine. The cells were then cryoprotected at −80◻. 1×107 cells of each genotype (about 30 wild type embryos or 50 Nkx2-5Cre/Cre embryos) were pooled for each 3C library preparation following standard procedures (Cope and Fraser, 2009). The relative interaction frequency (RIF) of each fragment of interest was measured by RT-qPCR, as described previously (Li et al., 2019). Primer details are as follows: F1 (5’ -GCCAAGGCCCTCAAGGTGA-3’), F2 (5’ -GGAGGGATTTGATATG AATGTAAGCACTG-3’), F3 (5’ -ATTCATGTAAGAGGGTCTGATCTCCAG-3’), F4 (5’ -ACAACCTTATCTCCAAACCTTTGCTC-3’), R1 (5’ -CCATTCCCTTTAGGACGGTTACCTTC-3’), R2 (5’ -CATGGACTGACCTCATTGGAGTG-3’), R3 (5’ -GGGAGGGATTCAGGACATGTTG-3’).
Statistical analysis
All experiments were repeated at least three times to ensure scientific reproducibility. Quantification results are presented as mean±s.e.m., and statistical analysis was conducted using Student’s t-test in a GraphPad Prism 6 software. P<0.05 was considered significant.
Competing interests
The authors declare no competing or financial interests.
Author contributions
Conceptualization: Y.C., T.Y.; Methodology: Y.C., T.Y., Z.H.; Validation: Y.C., T.Y.; Formal analysis: T.Y., Z.H., H.L.; Investigation: T.Y., Z.H., H.L., L.W.; Resources: Y.C.; Data curation: Y.C., T.Y.; Writing-original draft: T.Y.; Writing – review & editing: Y.C.; Visualization: T.Y.; Supervision: Y.C.; Project administration: Y.C.; Funding acquisition: Y.C.
Funding
We acknowledge financial support by the National Institutes of Health (R01HL136326 to Y.C.) and an American Heart Association (AHA) Predoctoral Fellowship (20PRE35040002 to T.Y.). H.L. and Z.H. were supported in part by fellowships from Fujian Normal University, and L.W. received a fellowship from the China Scholarship Council.
Acknowledgments
We thank Mrs. Ann Mullin at the Tulane Transgenic Animal Core Facility for assistance with the generation of genetically modified mice.