1. Chromosomes and Gene Expression
  2. Genetics and Genomics

An atrial fibrillation-associated regulatory region modulates cardiac Tbx5 levels and arrhythmia susceptibility

  1. Fernanda M Bosada
  2. Karel van Duijvenboden
  3. Alexandra E Giovou
  4. Mathilde R Rivaud
  5. Jae-Sun Uhm
  6. Arie O Verkerk
  7. Bastiaan J Boukens
  8. Vincent M Christoffels  Is a corresponding author
  1. Department of Medical Biology, Amsterdam Cardiovascular Sciences, Amsterdam Reproduction and Development, Amsterdam University Medical Centers, University of Amsterdam, Netherlands
  2. Department of Experimental Cardiology, Amsterdam Cardiovascular Sciences, Amsterdam University Medical Centers, University of Amsterdam, Netherlands
  3. Department of Cardiology, Severance Hospital, College of Medicine, Yonsei University, Republic of Korea
  4. Department of Physiology, University of Maastricht, Cardiovascular Research Institute Maastricht, Maastricht University Medical Center, Netherlands
https://doi.org/10.7554/eLife.80317
Share this article
Cite this article
  1. Fernanda M Bosada
  2. Karel van Duijvenboden
  3. Alexandra E Giovou
  4. Mathilde R Rivaud
  5. Jae-Sun Uhm
  6. Arie O Verkerk
  7. Bastiaan J Boukens
  8. Vincent M Christoffels
(2023)
An atrial fibrillation-associated regulatory region modulates cardiac Tbx5 levels and arrhythmia susceptibility
eLife 12:e80317.
https://doi.org/10.7554/eLife.80317
  2. Download BibTeX
  3. Download .RIS

Abstract

Heart development and rhythm control are highly Tbx5 dosage-sensitive. TBX5 haploinsufficiency causes congenital conduction disorders, whereas increased expression levels of TBX5 in human heart samples has been associated with atrial fibrillation (AF). We deleted the conserved mouse orthologues of two independent AF-associated genomic regions in the Tbx5 locus, one intronic (RE(int)) and one downstream (RE(down)) of Tbx5. In both lines, we observed a modest (30%) increase of Tbx5 in the postnatal atria. To gain insight into the effects of slight dosage increase in vivo, we investigated the atrial transcriptional, epigenetic and electrophysiological properties of both lines. Increased atrial Tbx5 expression was associated with induction of genes involved in development, ion transport and conduction, with increased susceptibility to atrial arrhythmias, and increased action potential duration of atrial cardiomyocytes. We identified an AF-associated variant in the human RE(int) that increases its transcriptional activity. Expression of the AF-associated transcription factor Prrx1 was induced in Tbx5RE(int)KO cardiomyocytes. We found that some of the transcriptional and functional changes in the atria caused by increased Tbx5 expression were normalized when reducing cardiac Prrx1 expression in Tbx5RE(int)KO mice, indicating an interaction between these two AF genes. We conclude that modest increases in expression of dose-dependent transcription factors, caused by common regulatory variants, significantly impact on the cardiac gene regulatory network and disease susceptibility.

Editor's evaluation

Molecular mechanisms of atrial fibrillation, a highly prevalent arrhythmia, have been challenging to contextualize. The investigators, in a series of experiments using the deletion of two loci in TBX5 (one intronic and one downstream), achieved a modest increase in Tbx5 in the atria. In a series of elegant experiments they show that this increase in TbX5 in turn impacted the cardiac gene regulatory network that resulted in a higher susceptibility to AF. These results will provide useful insights into the mechanisms of disease especially the variable phenotypes observed with functional variants of a gene.

https://doi.org/10.7554/eLife.80317.sa0

Introduction

The lifetime risk of developing a common disease, such as cardiovascular or neurodegenerative conditions, is influenced by genetic predisposition resulting from large numbers of inherited common genetic variants (single-nucleotide polymorphisms, SNPs). Disease-associated variants are typically found in noncoding genomic regions and are thought to affect the functionality of regulatory elements (REs) such as enhancers or elements involved in chromatin conformation (Degtyareva et al., 2021; Deplancke et al., 2016; Hormozdiari et al., 2018; Maurano et al., 2012; Schaub et al., 2012). These regulatory variants are pleiotropic and their effects on target gene expression are often specific to particular cell-types, conditions or stages of development (GTEx Consortium, 2020; Sobreira et al., 2021; Strober et al., 2019; Watanabe et al., 2019). It remains challenging to identify the causal variants among the many associated variants in a disease-associated noncoding DNA region and the REs that are affected by such variants (Hormozdiari et al., 2017). Moreover, common variants typically have a small effect on phenotype, and the different functional variants may act additively, synergistically, or oppositely. As a consequence, very few biological mechanisms linking disease-associated variant(s) to phenotype have been uncovered (Lappalainen and MacArthur, 2021; Timpson et al., 2018; Visscher et al., 2017). Here, we set out to investigate how particular noncoding regions harboring clustered variants associated with a common disease modulate expression of a disease-associated transcription factor gene in a tissue-specific manner, and how this expression change affects phenotype in vivo.

Genome-wide association studies (GWAS) have identified many common variants in over 100 genetic loci associated with atrial fibrillation (AF) risk, the most prevalent arrhythmia associated with high comorbidity and increased mortality risk (Christophersen et al., 2017; Chugh et al., 2014; Low et al., 2017; Nielsen et al., 2018; Roselli et al., 2018; Staerk et al., 2017). The identification of functional variants, REs, and target genes underlying AF will provide important insights into the molecular mechanisms of disease (van Ouwerkerk et al., 2020). AF-associated variants have been identified in loci harboring transcription factor-encoding genes, including PITX2, TBX5, and PRRX1, suggesting that altered expression levels of such factors cause imbalances in gene regulatory networks that control heart rhythm and function (van Ouwerkerk et al., 2020). Indeed, using mouse models, insufficiency of these transcriptional regulators was shown to cause arrhythmia susceptibility (Bosada et al., 2021; Dai et al., 2019; Kirchhof et al., 2011; Laforest et al., 2019; Nadadur et al., 2016; Tao et al., 2014; Tucker et al., 2017; Wang et al., 2010; Zhang et al., 2019). Heterozygous loss- or gain of function variants in TBX5 can cause Holt-Oram syndrome in humans, characterized by congenital heart defects and cardiac conduction anomalies, as a result of profound changes in the gene regulatory networks controlling heart development and function (Basson et al., 1997; Bruneau et al., 1999; Kathiriya et al., 2021; Li et al., 1997; Mori et al., 2006; Moskowitz et al., 2004). Interestingly, duplications of TBX5 as well as intragenic duplications have been reported in families with (atypical) Holt-Oram syndrome including cardiac defects (Cenni et al., 2021; Kimura et al., 2015; Patel et al., 2012). Furthermore, a gain-of-function pathological missense variant in TBX5 causes paroxysmal AF (Postma et al., 2008; van Ouwerkerk et al., 2022). Moreover, a previous study uncovered a 30% increase, rather than a reduction, in cardiac TBX5 expression in human heart tissues has been associated with AF (Roselli et al., 2018). The effects of small but potentially physiologically relevant dosage increase in transcriptional regulators such as TBX5 are not well characterized.

We deleted the mouse orthologues of two AF variant-rich regions in the human TBX5 locus to investigate how variant regions associated with a common disease modulate phenotype in a tissue- and developmental stage-specific manner. Each deletion caused a modest 30% increase in Tbx5 expression in different heart compartments and at different stages of life. We report the relatively large effect of this modest increase in Tbx5 expression on atrial function including arrhythmia susceptibility, and on the gene regulatory network. Decreased expression of Prrx1 has been associated with AF in human and mouse models (Roselli et al., 2018; Bosada et al., 2021; Tucker et al., 2017). We observed a genetic interaction between Tbx5 and Prrx1, and found that some of the transcriptional and functional changes in the atria caused by increased Tbx5 expression were rescued by reducing cardiac Prrx1 expression.

Results

Identification of two AF-associated regulatory regions in the TBX5 locus

The topologically associated domain (Dixon et al., 2012; Nora et al., 2012) harboring Tbx5 shows very limited contact with the adjacent domains harboring Tbx3 and Rbm19, respectively (van Weerd et al., 2014). Because all AF-associated variants in this locus are found within the topologically associated domain of TBX5 (Figure 1A), we anticipate that REs affected by the risk variants modulate the expression of TBX5 only. Promoter capture Hi-C maps from iPSC-derived cardiomyocytes (Montefiori et al., 2018) show distinct contacts between the promoter of Tbx5 and distal AF-associated regions, including the region in the last intron of TBX5 (Figure 1A). To identify possible regulatory elements within the AF-associated regions, we analyzed epigenomic datasets in both human (Gilsbach et al., 2018; van Ouwerkerk et al., 2019) and mouse orthologous region (Figure 1A–B). We selected two regions: the first situated in the last intron (RE(int)), and the second immediately downstream (RE(down)) of Tbx5. Both fragments contain evolutionary conserved regions, and RE(int) harbors accessible chromatin sites in the left atria and ventricles, and EMERGE enhancer prediction signal (van Duijvenboden et al., 2019; van Ouwerkerk et al., 2019). Additionally, RE(int) and RE(down) contain regions associated with cardiac H3K4me1, and RE(int) with H3K27ac (Gilsbach et al., 2014; Figure 1B). The SNPs in the last intron and those downstream of the gene are clustered into two distinct haplotypes, suggesting that these two regions are independently inherited (Figure 1—figure supplement 1; Machiela and Chanock, 2015). Using CRISPR/Cas9 genome editing, we deleted these candidate REs from the mouse genome to test their function in vivo.

Figure 1 with 1 supplement see all
Download asset Open asset
AF-associated noncoding variants are found in the TBX5 locus.

(A) Hi-C heatmap from human lymphoblastoid line GM12878 shows AF-associated variants are found in the regulatory domain of TBX5. Zoom-in of the AF-associated region overlaid with promoter capture Hi-C (red arcs), regions of open chromatin in whole left atria (ATACseq LA), H3K27ac and H3K4me1 ChIPseq signatures in cardiomyocytes. (B) Mouse orthologue of the human region including ATACseq from left atrial and ventricular CMs, H3K27ac ChIPseq from whole juvenile hearts, EMERGE, and conservation tracks. CRISPR/Cas9-generated deletions in light gray (REint), and dark gray (REdown).

Tbx5 expression and arrhythmia predisposition are increased in atria of Tbx5RE(int)KO and Tbx5RE(down)KO mice

Before birth, expression of Tbx5 in the atria was not different between genotypes (Figure 2—figure supplement 1A). However, in juvenile atria of RE(int) KO mice (p=0.001), Tbx5 expression was slightly increased compared to controls (Figure 2A). Expression of Rbm19, Tbx3, and Med13l, which neighbor Tbx5, remained unchanged in atria or ventricles of both mutants (Figure 2—figure supplement 1). Both atria of Tbx5RE(int)KO adult mice expressed approximately 30% more Tbx5, whereas only the left atrium and lungs of Tbx5RE(down)KO adult mice expressed more Tbx5 (Figure 2B–C). While we were not able to assess quantitative differences in protein expression, we observed that Tbx5 protein was selectively present in PCM-1 +cardiomyocyte nuclei in left and right atria of both control and of Tbx5RE(int)KO mice (Figure 2—figure supplement 2). We confirmed absence of aberrant splicing caused by the intronic deletion in Tbx5RE(int)KO (Figure 2—figure supplement 3). Interestingly, eQTL analysis indicated that AF is associated with a statistically significant 30% increase in TBX5 expression in cardiac tissue (Roselli et al., 2018). This suggests that AF-associated variants in the corresponding human regions may mediate the increase in TBX5 expression observed in patients with risk variants.

Figure 2 with 3 supplements see all
Download asset Open asset
Deletion of AF-associated regions results in increased Tbx5 in adult atrial tissue.

(A) Tbx5 expression in atria, ventricles, and lungs from P21 control, Tbx5RE(int)KO, and Tbx5RE(down)KO mice determined by RT-qPCR. (B–C) Tbx5 expression levels in adult control and Tbx5RE(int)KO (B) or Tbx5RE(down)KO (C) left and right atria. (D) UCSC browser view of the human intronic region (gray) and tested fragments containing AF-associated SNPs (blue) overlaid with chromatin conformation, EMERGE, and conservation tracks. (E) Luciferase assay (n=6) shows enhancer activity differences between non-risk (dark gray) or risk (gray) alleles (Kruskal-Wallis p=0.0019). Statistical significance within each tissue type was determined with ANOVA followed by pairwise comparisons using Dunnett’s T3 multiple comparison test in A, unpaired t-tests in B, C, and Kruskal-Wallis tests followed by pairwise unpaired t-tests (p values shown in figure) in E.

To identify functional AF-associated variants within the human intronic RE region, we tested enhancer activity of five fragments containing an AF-associated SNPs (p<10–6) (Roselli et al., 2018) using luciferase assays in the atrial cardiomyocyte-like cell line HL-1 (Figure 2D). Of the tested fragments, the fragment containing rs7312625 showed increased enhancer activity, and the fragment containing rs4767237 showed minimal decreased activity (Figure 2E). Comparative TF motif analysis between the regions containing the major (A) or minor (G) allele of rs7312625 revealed that the minor (G) allele causes disruption of motifs for Arabidopsis SOL1/2 (TCX/TCX2) (Supplementary file 1), which are homologues of animal LIN54 (Marceau et al., 2016; Simmons et al., 2019), a DNA-binding component of the DREAM complex regulating cell cycle-dependent transcription (Sadasivam and DeCaprio, 2013). We also observed gain of a sine oculis (SIX) homeodomain transcription factor binding motif (Meurer et al., 2021; Supplementary file 1). Lin54 and both Six4 and Six5 are expressed in atrial tissue of mice (see Supplementary files 2 and 3).

Next, we recorded in vivo electrocardiograms (ECGs) to determine the functional consequences of such a modest increase in Tbx5. Both Tbx5RE(int)KO and Tbx5RE(down)KO mice had slower and more variable heart rates (R-R interval (RR); standard deviation of normal to normal R-R intervals (SDNN)) (Figure 3A and B). Additionally, we detected instances of sinus pauses and inverted P waves in mice with both deletions (Figure 3—figure supplement 1A). PR interval was significantly increased in Tbx5RE(int)KO mice, but remained unaffected in Tbx5RE(down)KO mice (Figure 3C). Other ECG parameters were not affected (Figure 3—figure supplement 1). Heart rate corrected sinus node recovery time (cSNRT) and Wenckebach cycle length (WBCL) measured during transesophageal pacing (Bosada et al., 2021; Verheule et al., 2004) were increased in both mouse models (Figure 3D and E), suggesting altered conduction system function. We next tested whether AA could be induced in both mouse lines using in vivo burst pacing (Figure 3G [typical AA traces]). The total duration of all AA episodes per mouse was greater in both deletion models when compared to control littermates (Figure 3F [top graph]). Yet, we were able to induce AAs more often in Tbx5RE(int)KO, but not Tbx5RE(down)KO mice compared to controls (Figure 3F [below graph]).

Figure 3 with 1 supplement see all
Download asset Open asset
Increased Tbx5 in adult atria results in altered in vivo electrophysiology.

(A–C) Graphs show individual and average ECG measurements for heart rate (RR) (A), heart rate variation (SDNN) (B), and PR-interval (C) of adult wild-type, Tbx5RE(int)KO, and Tbx5RE(down)KO mice. (D, E) Graphs show changes in heart rate-corrected sinus node recovery times at 120ms (cSNRT120) (D), and Wenckebach cycle length (WBCL) (E). (F) Bar graph depicts the total time each mouse spent in an atrial arrhythmia (AA) episode after two pacing passes, with the number of mice in which at least one episode lasting >1 s was observed below each bar. (G) Representative traces from wild-type (top) and two Tbx5RE(int)KO (bottom) individuals showing disappearance of p waves or the start of atrial arrhythmia with variability in ventricular response after pacing stimulus. Significance for in vivo parameters was determined with Kruskal-Wallis test followed by Dunn’s multiple comparison tests in A, B, D, and AA duration in F (top graph), and one-way ANOVA followed by Tukey’s multiple comparisons test in C and E. AA inducibility significance was determined with pairwise Fisher’s exact test (F bottom of graph).

To further characterize the electrophysiological phenotypes of Tbx5RE(int)KO mice we analyzed the properties of single isolated left atrial cardiomyocytes using patch-clamp. At 6 Hz stimulation, we observed APD prolongation in mutants at all measured repolarization stages, with no changes in AP upstroke velocity, resting membrane potential and maximal AP amplitude (Figure 4A–C). APD increase was present at all frequencies measured (Figure 4D). Because Tbx5 has important roles in intracellular Ca2+ handling (Dai et al., 2019; Laforest et al., 2019; Zhu et al., 2008), we measured intracellular Ca2+ concentration ([Ca2+]i) in isolated atrial CMs using fluorescent calcium indicator Indo-1. We did not observe any changes in systolic or diastolic [Ca2+]i concentration (Figure 4E). Together, our data show that a slight increase in Tbx5 disturbs atrial function and can predispose to arrhythmia.

Figure 4
Download asset Open asset
Atrial function abnormalities in Tbx5RE(int)KO mice.

(A) Typical action potentials (APs) and upstroke velocity measured at 6 Hz in single left atrial cardiomyocytes using the amphotericin-perforated patch clamp technique. (B) Average resting membrane potential (RMP), AP amplitude (APA), and AP upstroke velocity remain unchanged in mutants. (C) AP duration at 20, 50, and 90% of repolarization (APD20, APD50, APD90, respectively) is increased in mutants. (D) APD90 was prolonged at all tested frequencies in mutant left atrial cardiomyocytes. Error bars are SD. (E) Diastolic and systolic intracellular Ca2+ concentrations ([Ca2+]i) were not changed in Tbx5RE(int)KO compared to controls. Statistical significance in B was determined with unpaired t-tests with Welch’s correction, and unpaired t-tests were used in B. Experimental groups were compared using two-way repeated measures ANOVA (D, E), followed by pairwise unpaired t-tests (C). Error bars are SEM.

Sensitivity of the atrial gene regulatory network to modestly increased Tbx5 dosage

To gain insight into the mechanism underlying the changes in electrophysiological properties in mutants, we performed transcriptional profiling of whole left atria. We detected 13790 and 13564 different transcripts in Tbx5RE(int)KO and Tbx5RE(down)KO, respectively, of which 816 were significantly up- and 1235 were downregulated in Tbx5RE(int)KO (Figure 5A; adjusted p<0.05), and 112 were up- and 123 were downregulated in Tbx5RE(down)KO (Figure 5B; adjusted p<0.05). Both datasets shared a high proportion of deregulated genes, with fewer significantly deregulated genes observed in Tbx5RE(down)KO compared to Tbx5RE(int)KO (Figure 5C). These data indicate that transcriptomes in Tbx5RE(int)KO atria and in Tbx5RE(down)KO atria are comparably affected qualitatively, but not quantitatively, consistent with the slightly higher increase in Tbx5 expression in Tbx5RE(int)KO compared to Tbx5RE(down)KO. Gene Ontology analysis (Mi et al., 2019) revealed that processes involved in cellular compartment organization, ion transport, and cardiac conduction characterized the transcripts found in the upregulated genes, and extracellular matrix organization, vasculature development, and actin cytoskeleton organization were found in the downregulated set in both deletions (Figure 5D). Accordingly, several Tbx5 target genes known to affect APD or whose deregulation may result in electrophysiological changes, (Dai et al., 2019; Laforest et al., 2019; Nadadur et al., 2016; Zhu et al., 2008) were significantly deregulated in one or both deletion lines (Figure 5F, Figure 5—figure supplement 1).

Figure 5 with 4 supplements see all
Download asset Open asset
Transcriptomic analysis of Tbx5RE(int)KO left atria.

(A, B) Volcano plot showing differentially expressed transcripts in Tbx5RE(int)KO (n=3)(A) and Tbx5RE(down)KO (n=3) (B) from wild-type (n=4) left atria. Dark red and light red dots indicate significantly upregulated genes by raw p-value and p-adjusted for multiple testing, respectively. Correspondingly, dark blue and light blue dots indicate downregulated genes. p Values were adjusted for multiple testing using the false discovery rate (FDR) method of Benjamini-Hochberg. (C) X-Y plot of all transcripts in Tbx5RE(int)KO (x axis) and Tbx5RE(down)KO (y axis), with deregulated genes common to both deletion mutants in purple, Tbx5RE(int)KO deregulated genes in blue, and Tbx5RE(down)KO deregulated genes in red. (D) Gene ontology (GO) analysis of upregulated and downregulated genes in Tbx5RE(int)KO and Tbx5RE(down)KO mutants. (F) Graph depicts fold change expression in control and mutant samples of genes known to affect action potential duration. * denotes significantly deregulated in one or both mutant lines.

Next, we compared the transcriptional response to increased Tbx5 in left atria (Tbx5RE(int)KO) to that of a gain-of-function missense mutation (Tbx5G125R/+) (van Ouwerkerk et al., 2022), and noticed that it was unexpectedly divergent, with only 80 transcripts significantly deregulated in the same direction in both datasets (Figure 5—figure supplement 2A, Supplementary file 4). In contrast, comparison with left atria of inducible adult-specific Tbx5 deletion (Tbx5iKO) (Nadadur et al., 2016) revealed that the majority of significantly deregulated genes common to both datasets were changed in opposite direction (Figure 5—figure supplement 2B, Supplementary file 5), thus indicating the Tbx5-dependent regulatory network responds in opposite direction to increased- or decreased Tbx5 dosage, respectively. PITX2 is strongly associated with AF and functionally implicated in atrial rhythm control (Kirchhof et al., 2011; Tao et al., 2014; Wang et al., 2010; Zhang et al., 2019), and was previously found to be a Tbx5 target in the left atrium (Nadadur et al., 2016). However, we did not detect a change in Pitx2 expression in either the left or right atria of Tbx5RE(int)KO mice (Figure 5—figure supplement 3).

To determine whether a slight increase in Tbx5 expression in cardiomyocytes results in changes in chromatin accessibility, we performed ATAC-sequencing profiling (ATACseq) of Tbx5RE(int)KO left atrial cardiomyocytes (Buenrostro et al., 2015). After peak-calling, we found a total of 85,569 accessible sites common to both genotypes, and only 15 sites of increased accessibility and 32 of decreased accessibility in mutants (Figure 5—figure supplement 4A). Two of the sites with decreased accessibility were found in the locus of Tbx5 (one of which caused by the deletion of the intronic genomic fragment in Tbx5RE(int)KO mice), suggestive for increased direct transcriptional autoregulation in Tbx5RE(int)KO mice (Figure 5—figure supplement 4B). The very minor changes in chromatin accessibility indicate that a modest increase in Tbx5 expression does not significantly change transcription factor occupancy or epigenetic state.

An interaction between Tbx5 and Prrx1, two AF-associated genes

We considered whether the presence of two AF-risk alleles would exacerbate the phenotype(s) consistent with AF. PRRX1, encoding the transcription factor Paired Related Homeobox 1, has been linked to AF-predisposition (Bosada et al., 2021; Roselli et al., 2018; Tucker et al., 2017). Reduced PRRX1 expression in human cardiac tissues has been associated with AF (Roselli et al., 2018). We previously deleted the mouse orthologous variant region near PRRX1 (Prrx1(enh)) to investigate its role in gene regulation and rhythm control (Bosada et al., 2021). Prrx1(enh)KO mice express less Prrx1 specifically in cardiomyocytes compared to controls, and show atrial conduction slowing, lower AP upstroke velocity (indicative for lower Na+ current densities; Berecki et al., 2010), as well as increased systolic and diastolic [Ca2+]i concentration that culminate in increased susceptibility to atrial arrhythmia induction (Bosada et al., 2021). To explore whether and how these two AF-risk genes may interact, we intercrossed Tbx5RE(int)KO with Prrx1(enh)KO mice (decreased Prrx1 in cardiomyocytes), and investigated cardiac transcriptomes and phenotypes across genotypes. We isolated cardiomyocyte and non-cardiomyocyte nuclei from whole hearts using a PCM-1 antibody and interrogated Tbx5 and Prrx1 expression across all genotypes. Tbx5 was upregulated in Tbx5RE(int)KO cardiomyocytes, as expected, but not deregulated in Prrx1(enh)KO or double mutant cardiomyocytes (Figure 6A). Prrx1 levels were increased in Tbx5RE(int)KO cardiomyocytes and decreased in Prrx1(enh)KO, as expected, and also decreased in double mutant cardiomyocytes (Figure 6B). There were no statistically significant changes detected in the non-cardiomyocyte fractions (Figure 6A and B). These data indicate that in cardiomyocytes, Tbx5 regulates Prrx1, and that a regulatory feedback loop modulates Tbx5 levels when Prrx1 expression is reduced due to the Prrx1(enh) deletion.

Figure 6 with 1 supplement see all
Download asset Open asset
A genetic interaction between Tbx5 and Prrx1.

(A, B) Graph shows reference gene-corrected expression of Tbx5 (A), and Prrx1 (B) in PCM1-positive and -negative nuclei fractions of wild-type, Tbx5RE(int)KO, Prrx1(enh)KO and double mutant whole hearts determined by RT-qPCR. (C) Principal component analysis of transcriptomes of wild-type, Tbx5RE(int)KO and DoubleKO left atrial samples. (D–H) Graphs show individual and average ECG measurements for RR (D), heart rate variation (SDNN) (E), PR interval (F), cSNRT at 120ms pacing (G), and WBCL (H) of wild-type, Tbx5RE(int)KO, Prrx1(enh)KO and DoubleKO mice. (I) Bar graph depicts the total time each mouse spent in an AA episode after two pacing passes, with the number of mice in which at least one episode lasting >1 s was observed below each bar. Statistical significance was determined using one-way ANOVA followed by Tukey’s test for pairwise comparisons in A, B, F, H, Kruskal-Wallis test followed by Dunn’s test for pairwise comparisons in D, E, G, and AA duration in I (top of graph), Fisher’s exact test for pairwise comparisons of AA inducibility in I (bottom of graph).

We next performed transcriptome analysis and found that deregulated genes in left atria of Tbx5RE(int)KO;Prrx1(enh)KO double mutants were more similar to the transcript changes found in Tbx5RE(int)KO than the ones found in Prrx1(enh)KO, when comparing all lines to WT, suggesting a dominant contribution of Tbx5RE(int)KO to the transcriptional changes in double mutants (Figure 6—figure supplement 1A-C, Supplementary files 6-7). Using principal component analysis, we found that transcriptomes clustered according to genotype (Figure 6C). Compared to atrial transcriptomes of Tbx5RE(int)KO mice, those of Tbx5RE(int)KO;Prrx1(enh)KO double homozygous showed less variance with WT atrial transcriptomes (Figure 6C, Figure 6—figure supplement 1, Table 1). This suggested an interaction between Tbx5 and Prrx1, in which reduced Prrx1 expression partially normalizes increased Tbx5-induced transcriptomic changes.

Table 1
Gene ontology (GO) analysis of upregulated and downregulated genes in Tbx5RE(int)KO, Prrx1(enh)KO, and Double homozygous adult left atria.
Prrx1(enh)KOTbx5RE(int)KODouble
GO biological processUP/DOWNFold enrichP-valueFDRUP/DOWNFold enrichP-valueFDRUP/DOWNFold enrichP-valueFDR
 Regulation of biological quality (GO:0065008)DOWN1.641.7E-083E-06UP1.551.5E-108E-07UP1.563.4E-081.5E-05
 Localization (GO:0051179)DOWN2.111.5E-141.6E-11UP1.451.7E-096.5E-06UP1.442.9E-078E-05
 Developmental process (GO:0032502)UP2.336.1E-201.6E-16UP1.396.8E-092.1E-05UP1.611.8E-142.9E-10
 Cellular component organization (GO:0016043)DOWN1.854.3E-086.9E-06UP1.411.4E-083.1E-05UP1.541.3E-102.2E-07
 Regulation of ion transport (GO:0043269)DOWN2.688.7E-081.3E-05UP2.221.7E-070.00018UP2.367.6E-070.00018
 Regulation of cell communication (GO:0010646)DOWN2.182.8E-181.1E-14UP1.483.3E-070.00029UP1.573.4E-078.9E-05
 Positive regulation of cell projection organization (GO:0031346)DOWN2.070.000330.0158UP2.441.5E-060.001UP2.083.9E-070.0001
 Cardiac conduction (GO:0061337)DOWN7.10.000110.00597UP6.573.6E-060.00203UP7.361E-050.0015
 Cell differentiation (GO:0030154)UP2.252.8E-118.2E-09UP1.397.9E-060.00276UP1.654.8E-105.8E-07
 Metal ion transport (GO:0030001)DOWN3.434.5E-087.3E-06UP2.181E-050.00342UP2.250.000690.049
Regulation of cell motility (GO:2000145)UP2.962.1E-141.2E-11DOWN3.352.3E-389.1E-35DOWN2.282.5E-050.00564
Extracellular matrix organization (GO:0030198)UP7.572E-243.1E-20DOWN6.123.5E-344.2E-31DOWN7.873.2E-175.1E-13
Circulatory system development (GO:0072359)UP3.32.8E-173.7E-14DOWN3.23.3E-322.5E-29DOWN3.142.3E-109.1E-07
Vasculature development (GO:0001944)UP4.114.6E-174.8E-14DOWN3.971.6E-311E-28DOWN3.233.2E-070.00017
Regulation of response to stimulus (GO:0048583)UP1.911.8E-172.6E-14DOWN1.791.3E-285.8E-26DOWN1.612.2E-060.00072
Actin cytoskeleton organization (GO:0030036)DOWN3.174.6E-087.3E-06DOWN3.914.9E-281.9E-25DOWN3.267.9E-070.00034
Cell adhesion (GO:0007155)UP2.923.9E-121.5E-09DOWN3.11.8E-276.6E-25DOWN2.991.1E-081.9E-05
Biological adhesion (GO:0022610)UP2.942E-128.2E-10DOWN3.064.9E-271.6E-24DOWN2.951.4E-082E-05
Angiogenesis (GO:0001525)UP4.54.9E-121.8E-09DOWN4.874.9E-271.6E-24DOWN4.173.7E-070.00019
Animal organ development (GO:0048513)UP2.038.4E-177.4E-14DOWN1.841.3E-233.3E-21DOWN1.783.4E-070.00018

Next, we considered whether the adult electrophysiological changes observed in Tbx5RE(int)KO would also be rescued by the presence of this second AF-risk allele. In vivo ECGs and burst pacing experiments revealed that RR, PR interval, and WBCL were normalized in double homozygous mice, whereas SDNN, cSNRT, and AA inducibility remained unchanged (Figure 6D–I).

Taken together, our findings suggest that an interaction between Tbx5 and Prrx1 exists; in which Prrx1 expression is induced by Tbx5, where Prrx1 is required to induce Tbx5 expression in Prrx1(enh)KO mice, and increased Tbx5 imposes changes in expression and electrophysiological properties, in part through increasing Prrx1 expression in cardiomyocytes.

Discussion

Our study reveals that mouse orthologues of two independent variant regions in the TBX5 locus modulate Tbx5 expression in tissue- and stage-specific manners, causing distinct specific phenotypes. While the impact of 50–100% dose reduction of Tbx5 in humans and model systems has been well-investigated (Bamshad et al., 1997; Bruneau et al., 2001; Kathiriya et al., 2021Li et al., 1997; Luna-Zurita et al., 2016; Mori et al., 2006; Moskowitz et al., 2004), we here show the effect of physiologically relevant increases in expression of Tbx5. This is significant as both decreased and increased expression of genes has been associated with increased AF risk.(Roselli et al., 2018). Our study reveals that atrial- and postnatal-specific increase in Tbx5 levels of only 30% affects postnatal atrial gene regulation, function and arrhythmia propensity. Furthermore, we provide an example of the interaction between the effect of two independent AF-associated variant regions (TBX5 and PRRX1) on phenotype. Our models provide insight into the mechanisms underlying the pleiotropic effects and interactions of disease-associated regulatory variants, which typically cause small changes in target gene expression in particular cell-types, conditions or stages of development (GTEx Consortium, 2020; Sobreira et al., 2021; Strober et al., 2019; Watanabe et al., 2019).

We found that the mouse orthologues of two noncoding regions studied here containing AF-associated variants in the TBX5 locus harbor REs that independently controlled Tbx5 expression in the heart in vivo. Deletion of either of the two RE-containing regions only affected Tbx5 expression, consistent with the notion that Tbx5 and surrounding RE regions are largely confined to one topologically associated domain not shared with adjacent genes (van Weerd et al., 2014). The mechanism of the repressive action of the REs and cross-talk with the other REs in the locus (Smemo et al., 2012) remain to be established. For RE(int), which showed the largest effect size in atria upon deletion, we identified two AF-associated SNPs that caused differential transcriptional activity of the RE sub fragment in HL1 atrial cardiomyocyte-like cells; one variant caused ~sixfold increase and the other minimally decreased activity. The risk allele of the first variant, rs7312625, disrupts a potential Lin54 motif, implying loss of interaction with the DREAM complex (Sadasivam and DeCaprio, 2013), and thus suggesting possible cell cycle-dependent regulation that remains to be elucidated. We speculate that in individuals carrying the variant that increases the activity of RE(int) in HL-1 cells, the repressive activity of RE(int) is reduced in the context of the entire regulatory system, thus causing increased TBX5 expression. The opposite could be true for the second variant, yet the small effect would likely be experimentally undetectable. On the other hand, the AF-associated variants in RE(down) may increase Tbx5 expression in both the left atrium and pulmonary vein myocardium. The pulmonary veins have been strongly implicated in AF as they are the most common source of triggered activity (Haïssaguerre et al., 1998; Jai¨s et al., 1997). TBX5 levels in pulmonary vein myocardium may influence gene regulation independently from the atrial TBX5 levels (Steimle et al., 2020). The variants in RE(down) segregate independently from those in RE(int), indicating these variant RE regions act through distinct tissue-specific transcriptional mechanisms that are influenced by AF-predisposing common variants. A further implication is they may cumulatively increase AF predisposition in a manner dependent on risk variant dose, in which homozygous carriers of risk haplotypes in RE(int) and RE(down) have the largest relative predisposition. These relations could be addressed in future genotype-phenotype analyses of human atrial and pulmonary vein samples.

Heart rates of the Tbx5RE(int)KO mice were slower and more variable than of controls. In addition, PR interval was prolonged in Tbx5RE(int)KO mice combination with increased Wenckebach cycle lengths. The latter is in line with the important role of Tbx5 in sinus node and AV node development and function (Mori et al., 2006; Moskowitz et al., 2007; Moskowitz et al., 2004; van Eif et al., 2018). Moreover, atrial arrhythmias could be more easily induced in Tbx5RE(int)KO mice suggesting electrophysiological remodeling of the atria. It has been shown that mice haploinsufficient for Tbx5 show atrial downregulation of genes encoding proteins involved in cardiac conduction (e.g. Gja1/Cx43 and Scn5a/Nav1.5) and Ca2+ handling (e.g Atp2a2/Serca2a and Ryr2/Ryr2) (Nadadur et al., 2016; Zhu et al., 2008). This leads to slower conduction, reduced Ca2+ concentration in the sarcoplasmic reticulum [Ca2+]SR, increased incidence of early and delayed afterdepolarization, and accordingly, increased AF propensity (Laforest et al., 2019; Nadadur et al., 2016). In the atria of our Tbx5RE(int)KO mice, where expression of Tbx5 was slightly elevated, transcriptomic analysis indicated upregulation of genes involved in cardiac conduction including Scn5a and Cx43. The Ca2+ handling genes Atp2a2/Serca2a and Ryr2/Ryr2 were not differently expressed between Tbx5RE(int)KO mice and controls. Accordingly, in our mice, we did not observe any changes in Ca2+ transients and long decay constants as seen in Tbx5 haploinsufficient mice (Laforest et al., 2019). Expression of Cacna1c and Cacna2d2/Cacna2d3, encoding subunits of the L-type Ca2+ channel, was upregulated in the atrium of Tbx5RE(int)KO mice. This is consistent with the observed longer atrial APD in Tbx5RE(int)KO mice compared to controls, which may predispose to increased arrhythmia inducibility, as seen in mice harboring the pathogenic TBX5-G125R variant or Prrx1 enhancer deletion (Bosada et al., 2021; van Ouwerkerk et al., 2022). Although shortened APD is typically found in AF patients with sustained arrhythmia (Kim et al., 2002; Wu et al., 2001), prolonged APD has been previously associated with increased risk of developing AF (Bosada et al., 2021; Lee et al., 2016; Olson et al., 2006; Simpson et al., 1988; van Ouwerkerk et al., 2022).

The molecular mechanisms underlying electrophysiological remodeling have remained unclear. For example, atrial cardiomyocytes of Tbx5 iKO (Tbx5 deletion induced in adult mice), of Tbx5+/G125R (heterozygous for a gain of function missense mutation) and of Tbx5RE(int)KO mice show APD prolongation (Nadadur et al., 2016; van Ouwerkerk et al., 2022). When comparing the transcriptional changes in left atria of these mouse models, we observed that it was unexpectedly divergent between the Tbx5RE(int)KO and Tbx5+/G125R models, whereas significantly differentially expressed genes common to both Tbx5RE(int)KO and Tbx5 iKO models were changed in opposite direction (Figure 5—figure supplement 2, Supplementary files 4 and 5). This suggests that the degree and direction of change of expression of, for example, genes implicated in APD (including Cacna2d3, Camk2a, Kcnh2, and Ryr2) do not necessarily predict changes in APD. Furthermore, Pitx2 was downregulated in left atria of Tbx5 iKO mice and upregulated in right atria of Tbx5+/G125R mice (van Ouwerkerk et al., 2022), and Tbx5 and Pitx2 were found to work antagonistically in the left atrium to tightly regulate the expression of genes impacting on cardiac electrophysiology (Nadadur et al., 2016). Yet, we did not detect a change in Pitx2 expression in the atria of Tbx5RE(int)KO mice. Together, these findings indicate that decreasing or increasing Tbx5 dose or changing Tbx5 function induces a large number of divergent transcriptional responses that disturb the balance in the genetic networks underlying functional (electrophysiological) properties of the atria, such as APD, leading to similar phenotypic outcomes.

The risk of developing complex diseases such as AF is strongly influenced by a large number of pleiotropic variants that each confers a small change in overall risk. We found that Tbx5 drives Prrx1 expression, and that Tbx5 levels are Prrx1-dependent (or Prrx1 target-dependent) in specific contexts. We then asked whether combining two alleles modeling the changes in gene expression conferred by AF-associated variant regions would cause a greater effect on heart phenotype, and by extension, arrhythmia predisposition, than each allele alone. Here, we found that introduction of this AF-susceptibility allele Tbx5RE(int)KO mice rescued heart rate, PR interval, and atrioventricular node phenotypes. This example shows that two variant regions that are both independently associated with increased risk for a particular disease, do not necessarily cumulatively increase disease predisposing phenotype, but may neutralize each other’s effect. Previously, the atrial electrophysiological phenotype of Tbx5 haploinsufficient mice was observed to be rescued by haploinsufficiency of either Pitx2 or Gata4, both of which have been associated with AF as well (Laforest et al., 2019; Nadadur et al., 2016). Although the direction of change of expression of the transcription factors in these models does not necessarily correspond to the direction of change in AF patients, this highlights the inherent robustness of the underlying gene regulatory networks, which in general remain stable when individual quantitative parameters such as transcription factor dose or binding sites affinity change (Albergante et al., 2014). This also implies that evaluating specific interactions between AF risk loci will be necessary for ascertaining individual risk from genetic association data.

In conclusion, our data provide unique mechanistic insights into the biological effects of variant region-driven modest physiologically relevant changes in expression of crucial transcriptional regulators, and into the impact of interactions between transcriptional output-modulating risk loci on atrial biology.

Materials and methods

Usage of epigenomic datasets

The following publicly available epigenomic datasets were used: Hi-C (Rao et al., 2014), AF variants (Roselli et al., 2018), promoter capture Hi-C maps from iPSC-derived cardiomyocytes (Montefiori et al., 2018), accessible chromatin in human and mouse left atria (van Ouwerkerk et al., 2019), cardiac H3K27ac and H3K4me1 ChIP-seq signatures in human (Gilsbach et al., 2014), and EMERGE enhancer prediction signal (van Duijvenboden et al., 2019; van Ouwerkerk et al., 2019).

Generation of mutant mice

Mutant mice were generated using CRISPR/Cas9. Guide RNA (sgRNA) constructs were designed with ZiFiT Targeter (Sander et al., 2010). The sgRNA constructs were transcribed in vitro using MEGAshortscript T7 (Invitrogen AM1354) and mMessage Machine T7 transcription kit (Invitrogen AM1344) according to manufacturer instructions. One-cell FVB/NRj zygotes were microinjected with 10 ng/µL each sgRNA and 25 ng/µL Cas9 mRNA to generate mouse founders. Deletions were validated by PCR and Sanger sequencing. The sgRNA target sequences are the following: RE(int) guide 1 (GGGAAATCGCCTTACCTTTC), guide 2 (GGACTGTTGGGTCACCTTGT), and RE(down) guide 1 (GGCTCCTTCGTCAGTAAATA), guide 2 (AAACAAGGGCTCTCTGGCGTTT). The deleted coordinates are the following in mm10: RE(int) chr5:119,874,013–119,880,148, and RE(down) chr5:119,891,017–119,915,744. Founders were backcrossed with wildtype FVB/NJ mice to obtain stable lines. Downstream experiments were performed on F3-F7 mice, backcrossed with wild-type FVB mice. All transgenic mice were maintained on a FVB/NJ background commercially obtained from Jackson laboratory (stock number 100800).

For tissue harvest, animals were euthanized by 20% CO2 inhalation followed by cervical dislocation.

EdU cardiomyocyte proliferation assay

Timed pregnant female mice received intraperitoneal injection of 100 mg/kg EdU one hour before sacrifice by isoflurane and cervical dislocation. E14.5 fetuses were isolated in 1xPBS on ice, the head was removed and the body fixed in 4% PFA for 24 hr. 7 μm sections were stained with mouse-anti-actin (1:400; Sigma A9357), goat-anti-nkx2.5 (1:150; Santa Cruz sc-14033), DAPI (1:1000) and EdU Click-IT (ThermoFisher, C10340) before imaging and counting. Nkx2.5 positive nuclei were designated cardiomyocytes, Nkx2.5 and EdU double positive nuclei were designated proliferating cardiomyocytes. Ratios were normalized within each litter.

qPCR

Total RNA was isolated from atria, ventricle and lungs from P21 and left and right atria from adult male and female mice using ReliaPrep RNA Tissue Miniprep System (Promega, Z6112) according to the manufacturer’s protocol. cDNA was reverse transcribed with oligo dT primers from 500 ng of total RNA, or random hexamers from 500 pg of CM nuclear RNA, according to the manufacturer’s protocol of the Superscript II Reverse Transcriptase system (Thermo Fisher Scientific, 18064014). Expression levels of candidate target genes were determined by quantitative real-time PCR using a LightCycler 480 Instrument II (Roche Life Science, 05015243001). Expression levels were measured using LightCycler 480 SYBR Green I Master (Roche, 04887352001) and the primers had a concentration of 1 pmol/L. The amplification protocol consisted of 5 min 95 °C followed by 45 cycles of 10 s 95 °C, 20 s 60 °C and 20 s 72 °C. Relative start concentration (N0) was calculated using LinRegPCR (Ruijter et al., 2009). Values were normalized to the geometric mean of two reference genes per experiment (Hprt, Ppia, or Rpl32) (Ruiz-Villalba et al., 2017). The primer sequences are as follows: (all 3’ to 5’) Hprt: TGTTGGATATGCCCTTGACT, GATTCAACTTGCGCTCATCT; Ppia: GGGTGGTGACTTTACACGCC, CTTGCCATCCAGCCATTCAG; Rpl32: GCCTCTGGTGAAGCCCAAG, TTGTTGCTCCCATAACCGATGT; Tbx5: CCCGGAGACAGCTTTTATCG, TGGTTGGAGGTGACTTTGTG; Prrx1: CACAAGCAGACGAAAGTGTGG, GTTGTCCTGTTTCTCCGCTG; Tbx3: CGCCGTTACTGCCTATCAGAA; GCCATTGCCAGTGTCTCGAA.

Cell culture and transfection luciferase assays

RE sub fragments were cloned into a modified pGL2-Basic plasmid containing an SV40 minimal promoter and an adjusted multiple cloning site for in vitro analysis by a transfection luciferase assay. HL1 cells (RRID:CVCL_0303) were grown in 24-well plates in Claycomb medium (Sigma-Aldrich, 51,800 C) supplemented with chemically defined HL-1 FBS substitute (Lonza, 77227), Glutamax (ThermoFisher Scientific, 35050–061) and Pen/Strep (ThermoFisher Scientific, 15070–063). Cells were transfected using polyethylenimine 25 kDa (PEI, Brunschwig, 23966–2) at a 1:3 ratio (DNA:PEI). Standard transfections were carried out using 200 ng of reporter construct per well. 24 hr after transfection, cells were lysed using Renilla luciferase assay lysis buffer (Promega, E291A-C) and luciferase activity measured. Luciferase measurements were performed using a GloMax Explorer (Promega, GM3500). During the measurement, 100 µL D-Luciferin (p.j.k, 102111) was injected (150 µL/s) followed by a 1-s delay and 5 s of measurement. Transfections were carried out at least three times and measured in duplicates. HL-1 cell cultures were routinely tested negative for mycoplasma contamination. HL-1 cell lines were easily distinguished based on cellular morphology and contractility. HL-1 is not found in the database of commonly misidentified cell lines that is maintained by the International Cell Line Authentication Committee.

In vivo electrophysiology

12–20 week old male mice were anesthetized with 5% Isoflurane (Pharmachemie B.V. 061756) and placed on thermostated mat (36 °C) with a steady flow of 1.5% isoflurane during all experiments. Electrodes were inserted subcutaneously in the limbs and connected to an ECG amplifier (Powerlab 26T, AD Instruments). The electrocardiogram (ECG) was measured for 5 min. ECG parameters were determined in Lead II (L-R) based on the last 60 s of the recording. For atrial stimulation, an octapolar CIB’er electrode (NuMED) was advanced through the esophagus to achieve atrial capture. Atrial capture thresholds were determined for each mouse, and all pacing protocols were performed at 2 x threshold. For sinus node recovery time (SNRT) measurements, a 4-s pacing train with a cycle length of 120 or 100ms was used. SNRT was defined as the interval between the last pacing stimulus and onset of the first P wave. To control for differences in sinus rate, SNRT was normalized to resting heart rate (cSNRT = SNRT – RR interval). To determine the Wenckebach cycle length (WBCL) we applied a 4s pacing train starting at a cycle length of 100ms, and decreasing by 2ms until atrioventricular (AV) block was first observed. Atrial arrhythmia (AA) was induced by 1 or 2 s bursts starting with a cycle length of 60ms, decreasing successively with a 2 ms decrement, down to a cycle length of 10ms. AA duration was the sum of time each mouse spent under and AA episode after completion of the two passes. Atrial arrhythmia (AA) inducibility was scored as the number of mice in which at least one episode lasting >1 s of arrhythmia was induced after pacing. Mice of both sexes were used for atrial arrhythmia induction experiments.

Cellular electrophysiology

Single cells were isolated from left atria of adult male mice by enzymatic dissociation. Therefore, excised hearts were perfused for 5 min in a Langendorff system with a modified Tyrode’s solution containing (in mmol/L): NaCl 140, KCl 5.4, CaCl2 1.8, MgCl2 1.0, glucose 5.5, HEPES 5.0; pH 7.4 (set with NaOH). Subsequently, the hearts were perfused with Tyrode’s solution containing a low Ca2+-concentration (10 μmol/L) for 10 min, after which Liberase TM research grade (Roche Diagnostics, GmbH, Mannheim, Germany, 5401119001) and Elastase from porcine pancreas (Bio-Connect B.V., Huissen, Netherlands, W59168R) were added for 12 min at a concentration of 0.038 mg/mL and 0.01 mg/mL, respectively. All solutions were saturated with 100% O2 and the temperature was maintained at 37 °C. To obtain single cells, the digested left atria was cut into small pieces which were triturated for 4 min through a pipette (tip diameter: 0.8 mm) in the low Ca2+ Tyrode’s solution, supplemented with 10 mg/ml Bovine Serum Albumin (Roche Diagnostics, essential fatty free, fraction V). Single cells were stored at room temperature for at least 45 min before they were used. Quiescent single rod-shaped cells with smooth surfaces were selected for electrophysiological measurements.

Action potentials (APs) were recorded with the amphotericin-B perforated patch-clamp technique using an Axopatch 200B amplifier (Molecular Devices Corporation, Sunnyvale, CA). Voltage control and data acquisition were performed as described previously (Kamel et al., 2021), and APs were low-pass filtered at 5 kHz and digitized at 40 kHz. Potentials were corrected for the estimated liquid junction potentials (Barry and Lynch, 1991).

APs were recorded at 36 ± 0.2°C using the modified Tyrode’s solution. Pipettes (borosilicate glass (Harvard Apparatus, UK)) were filled with solution containing (in mmol/l): K-gluc 125, KCl 20, NaCl 5, amphotericin-B 0.44, HEPES 10, pH 7.2 (KOH). APs were elicited at 2–8 Hz by 3ms,~1.2 × threshold current pulses through the patch pipette. We analyzed resting membrane potential (RMP), maximal AP amplitude (APA), maximum AP upstroke velocity, and AP duration at 20, 50, and 90% repolarization (APD20, APD50, and APD90, respectively). Parameters from 10 consecutive APs were averaged.

Intracellular Ca2+ measurements

Intracellular calcium concentrations: [Ca2+]i were measured at 25 °C in HEPES solution ((mmol/l): Na+ 156, K+ 4.7, Ca2+ 1.3, Mg2+ 2.0, Cl− 150.6, HCO3 4.3, HPO2− 1.4, HEPES 17, Glucose 11 and 1% fatty acid free albumin, pH 7.3) using the fluorescent probe Indo-1 as described previously (Verkerk et al., 2003). In brief, isolated myocytes were exposed to 5 µmol/l of the acetoxymethyl esters of indo-1 during 30 min at 37 °C. Myocytes were attached to a poly-D-lysine (0.1 g/l) treated cover slip placed on a temperature-controlled microscope stage of an inverted fluorescence microscope (Nikon Diaphot) with quartz optics. A temperature-controlled perfusion chamber (height 0.4 mm, diameter 10 mm, volume 30 µL, temperature 37 °C), with two needles at opposite sides for perfusion purposes, was tightly positioned over the cover slip. The contents of the chamber could be replaced within 100ms. Bipolar square pulses for field stimulation (40 V/cm) were applied through two thin parallel platinum electrodes at a distance of 8 mm. One quiescent single myocyte was selected (myocytes with more than one spontaneous oscillation per 10 s were excluded) and the measuring area was adjusted to the cell surface with a rectangular diaphragm. The wavelength of excitation of Indo-1 was 340 nm, applied with a stabilized xenon-arc lamp (100 W). Fluorescence was measured in dual emission mode at 410 and 516 nm. Emitted light passed a barrier filter of 400 nm, a dichroic mirror (450 nm) and respective narrow band interference filters in front of two photomultipliers (Hamamatsu R-2949). Signals were digitized at 1 kHz and corrected for background signals recorded from Indo-1 free myocytes. Ten subsequent Ca2+ transients were averaged from which apparent [Ca2+]i was calculated according to the ratio equation (Grynkiewicz et al., 1985).

Isolation of CM nuclei

Nuclei isolation was performed as follows: snap frozen adult left and right atria from adult male and female mice were trimmed and homogenized in lysis buffer containing RNase inhibitor using an Ultra-Turrax homogenizer. Samples were further homogenized with a loose pestle douncer (10 strokes). After a 10 min incubation in the lysis buffer, an additional 10 strokes were performed with a tight pestle. The lysis procedure was monitored by light microscopy to ensure complete tissue and cell lysis and efficient nuclear extraction. The crude lysate was successively passed through 100  and 30 µm mesh filters. The final lysate was spun at 1000 × g for 5  min and the resulting pellet was resuspended in 500  µl staining buffer (5% BSA in PBS) supplemented with RNAse inhibitor. Isolated nuclei were incubated with rabbit polyclonal antibodies specific for pericentriolar material 1 (PCM1) (Sigma-Aldrich; HPA023370) at a dilution of 1:400 for 1  hr rotating at 4  °C. Next, Alexa Fluor 647-conjugated donkey-anti-rabbit 647 antibodies (ThermoFisher Scientific A-31573; 1:500 dilution), and DAPI (1:1000 dilution) were added and the incubation was continued for an additional hour. Samples were spun at 1000 × g for 10  min and washed with 500  µl staining buffer before resuspension in 500  µl staining buffer supplemented with RNAse inhibitor. Intact CM nuclei were sorted on a BD Influx FACS on the basis of DAPI and Alexa Fluor 647 positivity into cold BL +TG buffer from the ReliaPrep RNA Tissue Miniprep System (Promega, Z6112) for RNA isolation, and into resuspension buffer for ATACseq (Buenrostro et al., 2015). RNA was isolated following a gDNA depletion step according to the manufacturer’s instructions. RNA yield and purity was assessed using an Agilent 2100 Bioanalyzer in combination with the RNA Pico chips.

Library preparation and sequencing

For RNA isolated from left atria, 500 ng was used for library generation with the KAPA mRNA HyperPrep kit (Roche) and sequenced on the HiSeq4000 system (Illumina) with 50 bp single-end reads. RNAseq sample sizes are as follows: Whole adult left atria: 4 WT, 3 Tbx5RE(int)KO, 3 Tbx5RE(down)KO, 4 Prrx1(enh)KO, and 3 Double homozygous.

Differential expression analysis

Reads were mapped to the mm10 build of the mouse transcriptome using STAR (Dobin et al., 2013). Differential expression analysis was performed using the DESeq2 package based on a negative binomial distribution model (Love et al., 2014). p-Values were corrected for multiple testing using the false discovery rate (FDR) method of Benjamini-Hochberg. We have used 0.05 as FDR control level. Unsupervised hierarchical clustering was performed on differentially expressed genes using the R package pheatmap version 1.0.8. (http://cran.rproject.org/web/packages/pheatmap/index.html). PANTHER (Mi et al., 2019) was used for gene ontology (GO) biological process analysis. Benjamini–Hochberg correction was performed for multiple testing-controlled p values. Statistically significant enriched terms were functionally grouped and visualized.

ATACseq on CM nuclei

ATACseq on FACS-sorted PCM1 +nuclei was performed and analyzed as described in Buenrostro et al., 2015. Approximately 50 k nuclei were used as input. The library was sequenced (paired-end 125 bp) and data was collected on a HiSeq4000.

Peak-calling and motif analysis of ATACseq

Reads from ATAC-seq data were mapped to mm10 build of the mouse genome using BWA (Li and Durbin, 2009), the default settings were used. The BEDTools suite was used to distribute the genome wide signal into bins of 500 bp (Quinlan and Hall, 2010). Bins with less than 89 cumulative tags across all 10 samples were discarded as noise. Differential accessibility was assessed using the DESeq2 package based on a model using the negative binomial distribution (Love et al., 2014). p-Values were corrected for multiple testing using the false discovery rate (FDR) method of Benjamini-Hochberg. We have used 0.05 as FDR control level. Continuous bins with differential signal were subsequently merged together using the BEDTools suite.

200 bp summits were determined of ATAC-seq performed in atrial cells (Fernandez-Perez et al., 2019). In total, HOMER (Heinz et al., 2010) was performed on sequences of 2000 neutral called peaks (randomly sampled out of 85,570 peaks), 2000 peaks up (randomly sampled out of 3551 peaks), and 1776 peaks down in Tbx5RE(int)KO, all with random genome as background in HOMER.

Unsupervised hierarchical clustering was performed on differentially detected peaks and genes in RNA-seq and ATAC-seq using the R package pheatmap, version 1.0.8. (http://cran.r-project.org/web/packages/pheatmap/index.html).

Usage of EMERGE

The genome-wide heart enhancer prediction track generated by EMERGE (van Duijvenboden et al., 2016) was used as a proxy for the presence of putative heart enhancers. The particular prediction used aimed at identifying robustly active heart enhancers. This was accomplished by further annotating the available true positives (Visel et al., 2007) to train the algorithm with on the basis of the consistency in heart activity patterns shown. Full details are outlined in van Ouwerkerk et al., 2019.

Statistics

The experimenters were blind to mouse genotype during all measurements and outcome assessment. Datasets were tested for normality using Shapiro-Wilk test unless specified otherwise. Whole tissue RT-qPCR fold change vs WT in fetal ventricles and adult whole left atria corrected expression changes were analyzed with unpaired t-tests and reference gene-corrected Tbx5 in juvenile whole tissue was analyzed with Welch’s ANOVA followed by Dunnet’s T3 multiple comparison tests within each tissue type. Luciferase assays on HL1 cells were analyzed using Kruskal-Wallis test. In vivo electrophysiology was analyzed with Kruskal-Wallis followed by Dunn’s multiple comparison tests or one-way ANOVA followed by Tukey’s multiple comparison tests. Significance of atrial arrhythmia (AA) duration was determined with Mann Whitney Wilcoxon test and differences in AA inducibility were tested using Fisher’s exact test. Conduction velocity was analyzed with unpaired t-tests with Welch’s correction. For single cell and whole tissue action potential (AP) duration (APD) normality and equal variance assumptions were tested with the Kolmogorov-Smirnov and the Levene median test, respectively. Two groups were compared with unpaired t-test or repeated measures ANOVA followed by pairwise comparison using the Student-Newman-Keuls test. Differences in Ca2+ transient amplitude as well as changes in diastolic and systolic Ca2+ concentrations in atria were tested using two-way ANOVA. Multiple testing corrections were performed independently within each hypothesis. Data are presented as individual data points and mean or mean ± standard error of the mean (SEM) or standard deviation (SD), as indicated, and <0.05 defines statistical significance. Statistical analysis was performed using GraphPad Prism 9.

Data availability

Adult left atrial RNAseq and ATACseq have been deposited under GEO accession numbers GSE189342 and GSE189498.

The following data sets were generated
    1. Bosada FM
    (2022) NCBI Gene Expression Omnibus
    ID GSE189342. Tbx5 regulatory element controls atrial function and heart size.
    https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE189342
    1. Bosada FM
    (2022) NCBI Gene Expression Omnibus
    ID GSE189498. Tbx5 regulatory element controls atrial function and heart size.
    https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE189498

References

    1. Christophersen IE
    2. Rienstra M
    3. Roselli C
    4. Yin X
    5. Geelhoed B
    6. Barnard J
    7. Lin H
    8. Arking DE
    9. Smith AV
    10. Albert CM
    11. Chaffin M
    12. Tucker NR
    13. Li M
    14. Klarin D
    15. Bihlmeyer NA
    16. Low SK
    17. Weeke PE
    18. Müller-Nurasyid M
    19. Smith JG
    20. Brody JA
    21. Niemeijer MN
    22. Dörr M
    23. Trompet S
    24. Huffman J
    25. Gustafsson S
    26. Schurmann C
    27. Kleber ME
    28. Lyytikäinen LP
    29. Seppälä I
    30. Malik R
    31. Horimoto A
    32. Perez M
    33. Sinisalo J
    34. Aeschbacher S
    35. Thériault S
    36. Yao J
    37. Radmanesh F
    38. Weiss S
    39. Teumer A
    40. Choi SH
    41. Weng LC
    42. Clauss S
    43. Deo R
    44. Rader DJ
    45. Shah SH
    46. Sun A
    47. Hopewell JC
    48. Debette S
    49. Chauhan G
    50. Yang Q
    51. Worrall BB
    52. Paré G
    53. Kamatani Y
    54. Hagemeijer YP
    55. Verweij N
    56. Siland JE
    57. Kubo M
    58. Smith JD
    59. Van Wagoner DR
    60. Bis JC
    61. Perz S
    62. Psaty BM
    63. Ridker PM
    64. Magnani JW
    65. Harris TB
    66. Launer LJ
    67. Shoemaker MB
    68. Padmanabhan S
    69. Haessler J
    70. Bartz TM
    71. Waldenberger M
    72. Lichtner P
    73. Arendt M
    74. Krieger JE
    75. Kähönen M
    76. Risch L
    77. Mansur AJ
    78. Peters A
    79. Smith BH
    80. Lind L
    81. Scott SA
    82. Lu Y
    83. Bottinger EB
    84. Hernesniemi J
    85. Lindgren CM
    86. Wong JA
    87. Huang J
    88. Eskola M
    89. Morris AP
    90. Ford I
    91. Reiner AP
    92. Delgado G
    93. Chen LY
    94. Chen YDI
    95. Sandhu RK
    96. Li M
    97. Boerwinkle E
    98. Eisele L
    99. Lannfelt L
    100. Rost N
    101. Anderson CD
    102. Taylor KD
    103. Campbell A
    104. Magnusson PK
    105. Porteous D
    106. Hocking LJ
    107. Vlachopoulou E
    108. Pedersen NL
    109. Nikus K
    110. Orho-Melander M
    111. Hamsten A
    112. Heeringa J
    113. Denny JC
    114. Kriebel J
    115. Darbar D
    116. Newton-Cheh C
    117. Shaffer C
    118. Macfarlane PW
    119. Heilmann-Heimbach S
    120. Almgren P
    121. Huang PL
    122. Sotoodehnia N
    123. Soliman EZ
    124. Uitterlinden AG
    125. Hofman A
    126. Franco OH
    127. Völker U
    128. Jöckel KH
    129. Sinner MF
    130. Lin HJ
    131. Guo X
    132. Dichgans M
    133. Ingelsson E
    134. Kooperberg C
    135. Melander O
    136. Loos RJF
    137. Laurikka J
    138. Conen D
    139. Rosand J
    140. van der Harst P
    141. Lokki ML
    142. Kathiresan S
    143. Pereira A
    144. Jukema JW
    145. Hayward C
    146. Rotter JI
    147. März W
    148. Lehtimäki T
    149. Stricker BH
    150. Chung MK
    151. Felix SB
    152. Gudnason V
    153. Alonso A
    154. Roden DM
    155. Kääb S
    156. Chasman DI
    157. Heckbert SR
    158. Benjamin EJ
    159. Tanaka T
    160. Lunetta KL
    161. Lubitz SA
    162. Ellinor PT
    163. METASTROKE Consortium of the ISGC
    164. Neurology Working Group of the CHARGE Consortium
    165. AFGen Consortium
    (2017) Large-scale analyses of common and rare variants identify 12 new loci associated with atrial fibrillation
    Nature Genetics 49:946–952.
    https://doi.org/10.1038/ng.3843
    1. Roselli C
    2. Chaffin MD
    3. Weng L-C
    4. Aeschbacher S
    5. Ahlberg G
    6. Albert CM
    7. Almgren P
    8. Alonso A
    9. Anderson CD
    10. Aragam KG
    11. Arking DE
    12. Barnard J
    13. Bartz TM
    14. Benjamin EJ
    15. Bihlmeyer NA
    16. Bis JC
    17. Bloom HL
    18. Boerwinkle E
    19. Bottinger EB
    20. Brody JA
    21. Calkins H
    22. Campbell A
    23. Cappola TP
    24. Carlquist J
    25. Chasman DI
    26. Chen LY
    27. Chen Y-DI
    28. Choi E-K
    29. Choi SH
    30. Christophersen IE
    31. Chung MK
    32. Cole JW
    33. Conen D
    34. Cook J
    35. Crijns HJ
    36. Cutler MJ
    37. Damrauer SM
    38. Daniels BR
    39. Darbar D
    40. Delgado G
    41. Denny JC
    42. Dichgans M
    43. Dörr M
    44. Dudink EA
    45. Dudley SC
    46. Esa N
    47. Esko T
    48. Eskola M
    49. Fatkin D
    50. Felix SB
    51. Ford I
    52. Franco OH
    53. Geelhoed B
    54. Grewal RP
    55. Gudnason V
    56. Guo X
    57. Gupta N
    58. Gustafsson S
    59. Gutmann R
    60. Hamsten A
    61. Harris TB
    62. Hayward C
    63. Heckbert SR
    64. Hernesniemi J
    65. Hocking LJ
    66. Hofman A
    67. Horimoto ARVR
    68. Huang J
    69. Huang PL
    70. Huffman J
    71. Ingelsson E
    72. Ipek EG
    73. Ito K
    74. Jimenez-Conde J
    75. Johnson R
    76. Jukema JW
    77. Kääb S
    78. Kähönen M
    79. Kamatani Y
    80. Kane JP
    81. Kastrati A
    82. Kathiresan S
    83. Katschnig-Winter P
    84. Kavousi M
    85. Kessler T
    86. Kietselaer BL
    87. Kirchhof P
    88. Kleber ME
    89. Knight S
    90. Krieger JE
    91. Kubo M
    92. Launer LJ
    93. Laurikka J
    94. Lehtimäki T
    95. Leineweber K
    96. Lemaitre RN
    97. Li M
    98. Lim HE
    99. Lin HJ
    100. Lin H
    101. Lind L
    102. Lindgren CM
    103. Lokki M-L
    104. London B
    105. Loos RJF
    106. Low S-K
    107. Lu Y
    108. Lyytikäinen L-P
    109. Macfarlane PW
    110. Magnusson PK
    111. Mahajan A
    112. Malik R
    113. Mansur AJ
    114. Marcus GM
    115. Margolin L
    116. Margulies KB
    117. März W
    118. McManus DD
    119. Melander O
    120. Mohanty S
    121. Montgomery JA
    122. Morley MP
    123. Morris AP
    124. Müller-Nurasyid M
    125. Natale A
    126. Nazarian S
    127. Neumann B
    128. Newton-Cheh C
    129. Niemeijer MN
    130. Nikus K
    131. Nilsson P
    132. Noordam R
    133. Oellers H
    134. Olesen MS
    135. Orho-Melander M
    136. Padmanabhan S
    137. Pak H-N
    138. Paré G
    139. Pedersen NL
    140. Pera J
    141. Pereira A
    142. Porteous D
    143. Psaty BM
    144. Pulit SL
    145. Pullinger CR
    146. Rader DJ
    147. Refsgaard L
    148. Ribasés M
    149. Ridker PM
    150. Rienstra M
    151. Risch L
    152. Roden DM
    153. Rosand J
    154. Rosenberg MA
    155. Rost N
    156. Rotter JI
    157. Saba S
    158. Sandhu RK
    159. Schnabel RB
    160. Schramm K
    161. Schunkert H
    162. Schurman C
    163. Scott SA
    164. Seppälä I
    165. Shaffer C
    166. Shah S
    167. Shalaby AA
    168. Shim J
    169. Shoemaker MB
    170. Siland JE
    171. Sinisalo J
    172. Sinner MF
    173. Slowik A
    174. Smith AV
    175. Smith BH
    176. Smith JG
    177. Smith JD
    178. Smith NL
    179. Soliman EZ
    180. Sotoodehnia N
    181. Stricker BH
    182. Sun A
    183. Sun H
    184. Svendsen JH
    185. Tanaka T
    186. Tanriverdi K
    187. Taylor KD
    188. Teder-Laving M
    189. Teumer A
    190. Thériault S
    191. Trompet S
    192. Tucker NR
    193. Tveit A
    194. Uitterlinden AG
    195. Van Der Harst P
    196. Van Gelder IC
    197. Van Wagoner DR
    198. Verweij N
    199. Vlachopoulou E
    200. Völker U
    201. Wang B
    202. Weeke PE
    203. Weijs B
    204. Weiss R
    205. Weiss S
    206. Wells QS
    207. Wiggins KL
    208. Wong JA
    209. Woo D
    210. Worrall BB
    211. Yang P-S
    212. Yao J
    213. Yoneda ZT
    214. Zeller T
    215. Zeng L
    216. Lubitz SA
    217. Lunetta KL
    218. Ellinor PT
    (2018) Multi-Ethnic genome-wide association study for atrial fibrillation
    Nature Genetics 50:1225–1233.
    https://doi.org/10.1038/s41588-018-0133-9

Decision letter

  1. Kalyanam Shivkumar
    Reviewing Editor; UCLA Health, United States
  2. Matthias Barton
    Senior Editor; University of Zurich, Switzerland
  3. Kalyanam Shivkumar
    Reviewer; UCLA Health, United States
  4. Benoit Bruneau
    Reviewer; University of California, San Francisco, United States

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "An atrial fibrillation-associated regulatory region modulates cardiac Tbx5 levels and arrhythmia susceptibility" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, including Kalyanam Shivkumar as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by a Senior Editor. The following individual involved in the review of your submission has agreed to reveal their identity: Benoit Bruneau (Reviewer #2).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

Please address the discussion points raised by reviewers # 1 and 2.

1. Please provide data for this comment by reviewer # 2-Does the increase in Tbx5 gene expression result in an increase in protein expression?

2. Please address the extra data sought by reviewer # 3 in the detailed comments.

Reviewer #1 (Recommendations for the authors):

1. Why is the assumption that the regulatory elements "modulate the expression of TBX5 only" as opposed to have trans effects on distant gene encoding regions? For example, in reference 41 on the topic of promoter capture Hi-C maps, "Remarkably, more than 90% of SNP-target gene interactions did not involve the nearest gene, while 40% of SNPs interacted with at least two genes, demonstrating the importance of considering long-range chromatin interactions when interpreting functional targets of disease loci." Aside from the two regions for selected to study, were additional regions identified?

2. As "13790 and 13564 different transcripts in RE(int)-/- and RE (down)-/-" were identified, how is it known that the transcriptional changes are from increased TBX5 levels as opposed to a direct, trans effects due to the deletion of the two loci at TBX5? This seems at odds with the claim that "[d]eletion of either of the two RE-containing regions only affected Tbx5 expression".

3. As the SNP of interest rs7312625 has been identified as a variant associated with the electrocardiographic P wave (https://pubmed.ncbi.nlm.nih.gov/28794112/) and the transgenic mice in this study had slower heart rates, sinus pauses, inverted P waves (presumably ectopic atrial rhythm), PR prolongation, corrected sinus node recovery time, and Wenckebach cycle length, could the findings be more explanatory for nodal (sinus and atrioventricular) function and the cardiac conduction system more generally rather than atrial fibrillation in particular? The study seems as though it could have been re-framed as an evaluation of nodal function rather than atrial fibrillation. Of note, these findings were of note in the RE(down)-/- mice, which did not have increased atrial arrhythmia inducibility.

4. The authors cite reference 19 as showing a "slight increase…in cardiac TBX5 expression in human heart tissues [as being] associated with AF". What was the magnitude of the increase? Challenging to gather this information on this reviewer's review of reference 19.

5. Of all the genes that are modulated by changes in TBX5 expression, why was Prrx1 chosen? How do the authors explain that lack of change noted for Pitx2?

Reviewer #2 (Recommendations for the authors):

The rationale for focusing on RE(down) is not clearly appreciated based on Figure 1. This is in contrast to the apparent evidence in support of RE(int). Promoter Hi-C data shows DNA interactions in human that appears to span beyond the presumptive syntenic region of RE(down) in mouse, as well as a lack of chromatin accessibility, histone marks or EMERGE prediction for RE(down). Perhaps highlighting relevant features for RE(down) more explicitly would address this minor concern.

For supplemental figure 1, please define LD as "linkage disequilibrium" prior to abbreviation, as appropriate, and list the abbreviation at the beginning of the manuscript. As well, a scale bar is not apparent.

Does the increase in Tbx5 gene expression result in an increase in protein expression?

It would be interesting to compare these mouse models to the recently published TBX5 gain of function model.

For Figure 3, please define RR as "R-R interval" in legend and SDNN as "standard deviation of normal to normal R-R intervals", as appropriate, and add to abbreviation list.

In Figure 5, where is Prrx1 highlighted in the transcriptomic analysis?

For Figure 6C, how does Prrx1(enh)-/- compare in the PCA analysis? Please display it.

Reviewer #3 (Recommendations for the authors):

1. TBX5 mutations in humans and animal models are known to generate congenital heart defects and upper limb/radial malformations. Perhaps of relevance to this manuscript, human patients with TBX5 duplications also show many of these same phenotypes (Heinritz et al. 2005 Heart; Patel et al. 2012 Eur J Hum Genet; Kimura et al. 2015 Pedatr Cardiol; Cenni et al. 2021 Eur J Hum Genet). Do RE(int) and RE(down) homozygous mutations result in congenital heart or limb malformations? Structural heart defects can certainly impact cardiac electrophysiology and reporting structural defects, or lack of them can help put the phenotype into context.

2. The RE(int) deletion appears to be more severe than the RE(down) deletion. Does the RE(int) deletion interfere with the splicing of the last two exons? Depending on library construction, perhaps the RNA-seq can address this.

3. In Figure 2E, the authors report that rs7312625 A>G appears to be the functional SNP as it alone results in increased Tbx5 expression in vitro. As this is a predicted regulatory element, what transcription factor binding motif(s) is this SNP predicted to interfere or generate? Does this motif change make sense with gained Tbx5 expression in cardiomyocytes/HL1 cells?

4. In Figure 3C, the authors report RE(int) mutants have an increased PR interval and in Figure 4C, the authors report RE(int) have an increased APD20, 50, and 90. Single-copy loss of Tbx5 in mice has been reported to cause these same phenotypes (e.g. Bruneau et al. 2001 Cell; Nadadur et al. 2016 Sci Trans Med). Would you not expect the opposite phenotype, i.e., decreased PR interval and APD20/50/90? What is the explanation for these paradoxical results?

5. Several transcriptional profiling datasets have been generated for various Tbx5 mutations in mouse and human including a few by the authors (by cursory search of GEO for "TBX5"). How do the gene expression changes seen in the current datasets compare with previously generated datasets? Are the changes largely consistent with expectations of increased Tbx5 levels in RE(int) and RE(down)? It appears most public datasets are in Tbx5/TBX5 loss-of-function contexts, so we are expecting the opposite results. Does this data support that hypothesis or is there something else going on?

6. Figures 2A and 2C report that RE(down) demonstrated increased expression of Tbx5 in the adult left atria and P21 lungs, specifically, but there is little to no mention of this in the text or discussion. Given the importance of the pulmonary veins in the development of atrial fibrillation and prior publication history from the authors and colleagues on the topic of left atrial Tbx5, can the authors expand their discussion on the RE(down) related SNPs? The SNPs associated with RE(down) are independently segregating and the differences observed in this manuscript may hint at an independent, tissue-dependent function for that regulatory element.

https://doi.org/10.7554/eLife.80317.sa1

Author response

Essential revisions:

Please address the discussion points raised by reviewers # 1 and 2.

1. Please provide data for this comment by reviewer # 2-Does the increase in Tbx5 gene expression result in an increase in protein expression?

2. Please address the extra data sought by reviewer # 3 in the detailed comments.

We would like to thank the reviewers for their constructive criticism and suggestions. In our revised manuscript, we have addressed all comments and questions raised as detailed below. We believe that the incorporated changes have greatly improved the manuscript.

Reviewer #1 (Recommendations for the authors):

1. Why is the assumption that the regulatory elements "modulate the expression of TBX5 only" as opposed to have trans effects on distant gene encoding regions? For example, in reference 41 on the topic of promoter capture Hi-C maps, "Remarkably, more than 90% of SNP-target gene interactions did not involve the nearest gene, while 40% of SNPs interacted with at least two genes, demonstrating the importance of considering long-range chromatin interactions when interpreting functional targets of disease loci." Aside from the two regions for selected to study, were additional regions identified?

Although a regulatory element, potentially containing SNP(s), usually shares the topologically associating domain (TAD) with its target genes, it does not necessarily interact with the nearest gene. In the referred study, promoter capture Hi-C identified only a fraction of the interactions, and showed a bias against detecting proximal interactions because of the high background signal associated with genomic locations close to each other. Moreover, the interactions reflect conformation (3D structure) of the chromatin and do not necessarily imply functional interactions. Nevertheless, our group has previously shown that the Tbx5 regulatory region is confined to its own TAD (Weerd et al., 2014). Additionally, we examined the expression of genes adjacent to the Tbx5 TAD, Rbm19, Tbx3 and Med13l, in atria and ventricles of mice with deletions by RT-qPCR, and found no changes in their expression (Figure 2—figure supplement 1). These data confirm that the deleted regulatory regions target Tbx5 and not genes in adjacent TADs.

We identified at least one additional AF-associated variant region with regulatory potential in the TBX5/Tbx5 locus, namely, in the second to last intron, which can be seen in Figure 1. However, we chose not to study this region because of the decreased accessibility in atrial tissue (ATACseq) and the comparably lower degree of conservation across species (GERP).

2. As "13790 and 13564 different transcripts in RE(int)-/- and RE (down)-/-" were identified, how is it known that the transcriptional changes are from increased TBX5 levels as opposed to a direct, trans effects due to the deletion of the two loci at TBX5? This seems at odds with the claim that "[d]eletion of either of the two RE-containing regions only affected Tbx5 expression".

Although we cannot fully rule out that the deleted regulatory regions target other genes in addition to Tbx5, we anticipate that the transcriptional changes are a result of increased Tbx5 levels rather than a consequence of the deletion of the variant regions. As mentioned above in response to comment 1 of the reviewer, previous research indicated that the regulatory elements of Tbx5 are restricted to the Tbx5 TAD (Weerd et al., 2014), and we confirmed that nearby genes in adjacent TADs were not influenced by the deletions (see Figure 2—figure supplement 1 in the revised manuscript). Regulatory element-target promoter interactions are generally assumed to be confined to the same TAD (Dixon et al., 2012; Nora et al., 2012; Sexton et al., 2012), and it is considered very unlikely that these interactions span multiple TADs or chromosomes. Furthermore, we observed the upregulation of several well established TBX5 target genes (e.g. Gja5, Ryr2), suggesting that increased Tbx5 dose is causal. Finally, we compared the transcriptomes of left atria of RE(int)-/- mice (deletion in the last intron of Tbx5; increased Tbx5 expression in atria) to the reported transcriptome of left atria of mice in which Tbx5 was inactivated (Nadadur et al., 2016). We observed a strikingly opposite transcriptional response (see new Figure 5—figure supplement 2), indicating that the Tbx5-dependent regulatory network responds in opposite direction to increased and decreased Tbx5 dose, respectively.

3. As the SNP of interest rs7312625 has been identified as a variant associated with the electrocardiographic P wave (https://pubmed.ncbi.nlm.nih.gov/28794112/) and the transgenic mice in this study had slower heart rates, sinus pauses, inverted P waves (presumably ectopic atrial rhythm), PR prolongation, corrected sinus node recovery time, and Wenckebach cycle length, could the findings be more explanatory for nodal (sinus and atrioventricular) function and the cardiac conduction system more generally rather than atrial fibrillation in particular? The study seems as though it could have been re-framed as an evaluation of nodal function rather than atrial fibrillation. Of note, these findings were of note in the RE(down)-/- mice, which did not have increased atrial arrhythmia inducibility.

We thank the reviewer for this comment. The study the reviewer refers to states: “The P wave on an ECG is a measure of atrial electric function, and its characteristics may serve as predictors for atrial arrhythmias.” Therefore, the SNPs identified in this study may impact on expression of genes, including TBX5 and SCN5A, involved in atrial cardiomyocyte electrophysiology. We addressed this aspect here, and found that atrial cardiomyocytes showed altered electrophysiological properties facilitating arrhythmia. The observed inverted P waves, changed SAN recovery times and changed WCL indeed suggest that SAN and AVN function may be affected in addition. Therefore, it is possible that the atrial arrhythmia is a culmination of multiple triggers, including changes in nodal and working cardiomyocytes. While we certainly want to further investigate a role of Tbx5 in SAN function, we feel that further characterization of nodal function remains outside the scope of this study.

4. The authors cite reference 19 as showing a "slight increase…in cardiac TBX5 expression in human heart tissues [as being] associated with AF". What was the magnitude of the increase? Challenging to gather this information on this reviewer's review of reference 19.

We thank the reviewer for pointing this out. We have amended the text in the revised manuscript to clarify a 30% increase in cardiac TBX5 expression in lines 96-98 and 144-145.

5. Of all the genes that are modulated by changes in TBX5 expression, why was Prrx1 chosen?

In GWAS studies, the PRRX1 locus has been strongly associated with AF. PRRX1 encodes a transcription factor that we and others have shown to be causally linked to atrial fibrillation susceptibility. Therefore, we chose to intercross RE(int)-/- and Prrx1(enh)-/- mice to test whether the two established AF-risk alleles would aggravated the arrhythmogenic phenotype. Strikingly, we found a partial rescue of the phenotypes instead. We later determined that Prrx1 was upregulated in RE(int)-/- specifically in cardiomyocytes, but not in bulk RNAseq, likely because Prrx1 is most highly expressed in non-cardiomyocyte cell populations.

We have changed the revised manuscript to clarify this in 229-2243, as follows:

“We considered whether the presence of two AF-risk alleles would exacerbate the phenotype(s) consistent with AF. PRRX1, encoding the transcription factor Paired Related Homeobox 1, has been linked to AF-predisposition (19, 24, 25). Reduced PRRX1 expression in human cardiac tissues has been associated with AF (19). We previously deleted the mouse orthologous variant region near PRRX1 (Prrx1(enh)) to investigate its role in gene regulation and rhythm control (25). Prrx1(enh)-/- mice express less Prrx1 specifically in cardiomyocytes compared to controls, and show atrial conduction slowing, lower AP upstroke velocity (indicative for lower Na+ current densities) (58), as well as increased systolic and diastolic [ca2+]i concentration that culminate in increased susceptibility to atrial arrhythmia induction (25). To explore whether and how these two AF-risk genes may interact, we intercrossed RE(int)-/- with Prrx1(enh)-/- mice (decreased Prrx1 in cardiomyocytes), and investigated cardiac transcriptomes and phenotypes across genotypes.”

How do the authors explain that lack of change noted for Pitx2?

We did not observe a change in Pitx2 expression in the left atria of our mutant mice. We surmise that this is because a 30-35% increase of Tbx5 expression is not sufficient to induce a proportional increase in Pitx2. We recently characterized a Tbx5 gain-of-function mouse model (van Ouwerkerk et al., 2022) in which Pitx2 was increased in the right atrium, while the left atrium remained unaffected. We therefore examined whether Pitx2 expression was changed in the right atria of RE(int)-/- by RT-qPCR and found that it is indeed not affected, thus suggesting that the modest increase in Tbx5 level does not influence Pitx2 expression.

We included this result in lines 215-216, new Figure 5—figure supplement 3, and discussion in lines 348-356.

Reviewer #2 (Recommendations for the authors):

The rationale for focusing on RE(down) is not clearly appreciated based on Figure 1. This is in contrast to the apparent evidence in support of RE(int). Promoter Hi-C data shows DNA interactions in human that appears to span beyond the presumptive syntenic region of RE(down) in mouse, as well as a lack of chromatin accessibility, histone marks or EMERGE prediction for RE(down). Perhaps highlighting relevant features for RE(down) more explicitly would address this minor concern.

We would like to thank the reviewer for this suggestion. While the absence of interactions as assayed in Promoter Capture Hi-C does not necessarily indicate a lack of interaction, the interactions beyond the RE(down) region indeed indicate possible distal downstream regulatory elements. Consistently, Smemo et al. (Smemo et al., 2012) identified multiple enhancer candidates downstream of the REdown region, one of which was found to harbor a SNP associated with CHD (non-HOS). We chose to include the RE(down) region in our study because it harbors AF-associated variants, and H3K4me1 modifications in cardiomyocytes. We have clarified this in lines 125-126 of the revised manuscript.

For supplemental figure 1, please define LD as "linkage disequilibrium" prior to abbreviation, as appropriate, and list the abbreviation at the beginning of the manuscript. As well, a scale bar is not apparent.

We thank the reviewer for pointing out these oversights. We have changed the text in the supplement as indicated.

Does the increase in Tbx5 gene expression result in an increase in protein expression?

We attempted to quantify protein expression in adult atria of RE(int)-/- and WT mice by Western Blot. However, we were unable to detect Tbx5 protein. We performed immunofluorescence staining on adult atria to show that indeed Tbx5 protein is detected in the atria of mutant and control mice, even though a 30% increase is not quantifiable using this method. However, the analysis shows that Tbx5 remains selectively expressed in PCM-1+ cardiomyocyte nuclei in RE(int)-/- atria (see new Figure 2—figure supplement 2).

It would be interesting to compare these mouse models to the recently published TBX5 gain of function model.

We thank the reviewer for this valuable suggestion. We now include an additional table and a scatter plot showing a comparison between adult left atria from RE(int)-/- and Tbx5G125R/+ as Supplementary file 4 and in Figure 5—figure supplement 2, respectively. Surprisingly, we found very few differentially expressed genes (DEGs) common to both models, while their electrophysiological phenotypes were more alike- inverted P-waves, slower and more variable heart rates, increased atrial arrhythmia induction, and prolonged APD. Conspicuously, an inducible loss-of-function Tbx5 mouse model (Nadadur et al., 2016) also displays irregular RR, prolonged APD and arrhythmia. We have also compared transcriptomes of left atria of our RE(int) model to those of this induced adult Tbx5 deletion model (Supplementary file 5 and in Figure 5—figure supplement 2) and found opposite response of many of the same DEGs, as expected. Together, this suggests that perturbation of a Tbx5–dependent network independently of direction of change of many of the network genes results in overlapping phenotypic changes culminating in atrial dysfunction.

We have included this comparison in lines 204-208.

For Figure 3, please define RR as "R-R interval" in legend and SDNN as "standard deviation of normal to normal R-R intervals", as appropriate, and add to abbreviation list.

We apologize for the oversight. The definitions have been added to the revised manuscript as indicated.

In Figure 5, where is Prrx1 highlighted in the transcriptomic analysis?

Prrx1 was, in fact, not found to be differentially expressed in left atrial RNAseq data. We found that the induction of Prrx1 occurs only in atrial cardiomyocytes of RE(int)-/- mice, consistent with cardiomyocyte-specific expression-, and thereby induction in RE(int)-/-, of Tbx5 (see also immunofluorescence analysis mentioned above). We surmise that this change was not detectable in bulk RNAseq because Prrx1 is mostly expressed in non-cardiomyocytes, which is likely not affected in our mutants.

We understand that the rationale for studying the interaction between Tbx5 and Prrx1 is muddled without this observation. Therefore, we have changed the text in lines 229-238, as specified in reviewer 1 point 5.

For Figure 6C, how does Prrx1(enh)-/- compare in the PCA analysis? Please display it.

Unfortunately, we are unable to directly compare the Prrx1(enh)-/- RNAseq dataset in the PCA because it was obtained, along with its own WT control group, in a different session (months in between experiments; all experiments performed with alleles on same FVB background though). It appeared that this difference in session greatly influences the PCs. However, we compared the L2FC (and p(adj)) values in each transcriptomic comparison (WT vs RE(int)-/-, WT vs Prrx1(enh)-/-, WT vs RE(int)-/-;Prrx1(enh)-/- (Double mutant)), as these parameters are relatively unaffected by inter-session variation. We observed that (1) the number of genes differentially expressed between mutant and WT decreased going from RE(int)-/-, Double mutant, Prrx1(enh)-/-, (2) the direction of change of DEGs was similar between RE(int)-/- and Double mutant, but different from Prrx1(enh)-/-, (3) the distribution of the L2FCs was slightly smaller in Double mutant than RE(int)-/-. This is shown in the revised Supplemental figure 10. Together, these analyses suggest that RE(int)-/- has a much larger effect on the atrial transcriptome than Prrx1(enh)-/-, and that the effect of RE(int)-/- is slightly normalized by the inclusion of Prrx1(enh)-/- in RE(int)-/-.

Reviewer #3 (Recommendations for the authors):

1. TBX5 mutations in humans and animal models are known to generate congenital heart defects and upper limb/radial malformations. Perhaps of relevance to this manuscript, human patients with TBX5 duplications also show many of these same phenotypes (Heinritz et al. 2005 Heart; Patel et al. 2012 Eur J Hum Genet; Kimura et al. 2015 Pedatr Cardiol; Cenni et al. 2021 Eur J Hum Genet). Do RE(int) and RE(down) homozygous mutations result in congenital heart or limb malformations? Structural heart defects can certainly impact cardiac electrophysiology and reporting structural defects, or lack of them can help put the phenotype into context.

We thank the reviewer for this suggestion. We have included the notion of CHD in patients with duplications and these references in the Introduction section of the revised manuscript in lines 85-86. We did not observe any structural defects of limbs or heart. Tbx5 expression in RE(down)-/- mice is selectively induced in atrial and lung tissue after birth, and therefore limb defects are not expected. Similarly, atrial induction in RE(int)-/- mice is observed after birth. However, we noted a mild induction of Tbx5 expression in prenatal ventricles of RE(int)-/- mice (not in RE(down)-/- mice). While this does not lead to structural defects, the ventricles may be slightly larger. We are currently investigating this possible phenotype.

2. The RE(int) deletion appears to be more severe than the RE(down) deletion. Does the RE(int) deletion interfere with the splicing of the last two exons? Depending on library construction, perhaps the RNA-seq can address this.

This is a very insightful suggestion. Inspection of the RNAseq data sets revealed that the exon read coverage of exons relative to each other, including the ones flanking the intra-intronic deletion, is highly similar between WT and RE(int)-/- (see new Figure 2—figure supplement 3). Moreover, when zooming in on the intronic region, we clearly see absence of reads mapped to the deleted region, validating the absence of this genomic sequence. Furthermore, when we performed RT-qPCR of WT and RE(int)-/- atrial samples using intron/exon-spanning primers in exons 6-8, 7-9 and 8-9, respectively, we obtained the same result for both groups, and the obtained amplicons did not include fragments in which exon 8 was skipped (new Figure 2—figure supplement 3).

3. In Figure 2E, the authors report that rs7312625 A>G appears to be the functional SNP as it alone results in increased Tbx5 expression in vitro. As this is a predicted regulatory element, what transcription factor binding motif(s) is this SNP predicted to interfere or generate? Does this motif change make sense with gained Tbx5 expression in cardiomyocytes/HL1 cells?

We analyzed whether the rs7312625 substitution interfered/changed predicted binding motifs and found that the risk allele disrupts motifs for SOL1/2 (TCX/TCX2/CHC), which are homologues of the animal CHC protein LIN54 that binds DNA in a sequence-specific manner and is a component of DREAM complexes regulating cell cycle-dependent transcription. Furthermore, we observed gain of a sine oculis (SIX) homeodomain transcription factor binding motif with the risk allele (new Supplementary file 1). SIX transcription factors can act as transcriptional activators or repressors depending on interactions with other highly conserved regulators including Paired-box (Pax), Eyes absent (Eya), Dachshund (Dach), and Groucho (Grg) proteins (Meurer et al., 2021). Both Six4 and Six5 are expressed in atrial tissue of mice (see Supplementary table 1) and human (see (Meurer et al., 2021)). We speculate the SNP could create a SIX binding site, causing loss of repressive activity of the RE. We have added this information to the revised manuscript in lines 152-159 of the Results section and 295-298 in the discussion.

4. In Figure 3C, the authors report RE(int) mutants have an increased PR interval and in Figure 4C, the authors report RE(int) have an increased APD20, 50, and 90. Single-copy loss of Tbx5 in mice has been reported to cause these same phenotypes (e.g. Bruneau et al. 2001 Cell; Nadadur et al. 2016 Sci Trans Med). Would you not expect the opposite phenotype, i.e., decreased PR interval and APD20/50/90? What is the explanation for these paradoxical results?

This observation is puzzling to us as well, and may reflect the complexity of responses to either decreased or increased levels of Tbx5. When comparing the transcriptional response of increased Tbx5 in left atria (RE(int)-/-) to that of Tbx5G125R/+ (gain of function missense mutation; (van Ouwerkerk et al., 2022) (new Supplementary file 4, Figure 5—figure supplement 2)) we observed that it was unexpectedly divergent. When comparing the published left atrial transcriptomes of Tbx5 iKO ((Nadadur et al., 2016); Tbx5 deletion induced in adult mice) with that of RE(int)-/-, we found, as expected, that most significantly differentially expressed genes (DEGs) common to both datasets were changed in opposite direction (new Supplementary file 5, Figure 5—figure supplement 2). Remarkably, all three Tbx5 mouse models display APD prolongation. This suggests that a (probably) large number of transcriptional changes occur in RE(int)-/- that disturb the balance between AP duration-increasing and -decreasing gene expression, leading to phenotypic outcome similar to that of both Tbx5 loss of function and Tbx5 G125R missense variant. This further indicates that the response to Tbx5 dose is complex. Moreover, as indicated by the reviewer, loss- and gain of TBX5 gene copy results in CHD phenotypes, consistent with the notion that both reduced and increased Tbx5 levels can result in similar phenotypic outcomes.

5. Several transcriptional profiling datasets have been generated for various Tbx5 mutations in mouse and human including a few by the authors (by cursory search of GEO for "TBX5"). How do the gene expression changes seen in the current datasets compare with previously generated datasets? Are the changes largely consistent with expectations of increased Tbx5 levels in RE(int) and RE(down)? It appears most public datasets are in Tbx5/TBX5 loss-of-function contexts, so we are expecting the opposite results. Does this data support that hypothesis or is there something else going on?

We have included a new supplemental figure and accompanying tables in which we compare transcriptional profiles of left atria of RE(int)-/- mice to those of a Tbx5 gain-of-function variant mouse model (van Ouwerkerk et al., 2022) and an adult-specific knockout mouse model (Nadadur et al., 2016). Please see reviewer 2 point 4, and the response this reviewer above.

6. Figures 2A and 2C report that RE(down) demonstrated increased expression of Tbx5 in the adult left atria and P21 lungs, specifically, but there is little to no mention of this in the text or discussion. Given the importance of the pulmonary veins in the development of atrial fibrillation and prior publication history from the authors and colleagues on the topic of left atrial Tbx5, can the authors expand their discussion on the RE(down) related SNPs? The SNPs associated with RE(down) are independently segregating and the differences observed in this manuscript may hint at an independent, tissue-dependent function for that regulatory element.

This is an excellent suggestion. We have addressed this point in the Discussion lines 302-313 as follows:

“…the AF-associated variants in RE(down) may increase Tbx5 expression in both the left atrium and pulmonary vein myocardium. The pulmonary veins have been strongly implicated in AF as they are the most common source of triggered activity (64, 65). TBX5 levels in pulmonary vein myocardium may influence gene regulation independently from the atrial TBX5 levels (66). The variants in RE(down) segregate independently from those in RE(int), indicating these variant RE regions act through distinct tissue-specific transcriptional mechanisms that are influenced by AF-predisposing common variants. A further implication is they may cumulatively increase AF predisposition in a manner dependent on risk variant dose, in which homozygous carriers of risk haplotypes in RE(int) and RE(down) have the largest relative predisposition. These relations could be addressed in future genotype-phenotype analyses of human atrial and pulmonary vein samples.”

References

Dixon, J.R., Selvaraj, S., Yue, F., Kim, A., Li, Y., Shen, Y., Hu, M., Liu, J.S., Ren, B., 2012. Topological domains in mammalian genomes identified by analysis of chromatin interactions. Nature 485, 376–80. https://doi.org/10.1038/nature11082

Meurer, L., Ferdman, L., Belcher, B., Camarata, T., 2021. The SIX Family of Transcription Factors: Common Themes Integrating Developmental and Cancer Biology. Front. Cell Dev. Biol. 9.

Nadadur, R.D., Broman, M.T., Boukens, B., Mazurek, S.R., Yang, X., Van Den Boogaard, M., Bekeny, J., Gadek, M., Ward, T., Zhang, M., Qiao, Y., Martin, J.F., Seidman, C.E., Seidman, J., Christoffels, V., Efimov, I.R., McNally, E.M., Weber, C.R., Moskowitz, I.P., 2016. Pitx2 modulates a Tbx5-dependent gene regulatory network to maintain atrial rhythm. Sci. Transl. Med. 8. https://doi.org/10.1126/scitranslmed.aaf4891

Nora, E.P., Lajoie, B.R., Schulz, E.G., Giorgetti, L., Okamoto, I., Servant, N., Piolot, T., Van Berkum, N.L., Meisig, J., Sedat, J., Gribnau, J., Barillot, E., Blüthgen, N., Dekker, J., Heard, E., 2012. Spatial partitioning of the regulatory landscape of the X-inactivation centre. Nature 485, 381–385. https://doi.org/10.1038/nature11049

Sexton, T., Yaffe, E., Kenigsberg, E., Bantignies, F., Leblanc, B., Hoichman, M., Parrinello, H., Tanay, A., Cavalli, G., 2012. Three-dimensional folding and functional organization principles of the Drosophila genome. Cell 148, 458–472. https://doi.org/10.1016/j.cell.2012.01.010

Smemo, S., Campos, L.C., Moskowitz, I.P., Krieger, J.E., Pereira, A.C., Nobrega, M.A., 2012. Regulatory variation in a TBX5 enhancer leads to isolated congenital heart disease. Hum. Mol. Genet. 21, 3255–63. https://doi.org/10.1093/hmg/dds165

van Ouwerkerk, A.F., Bosada, F.M., van Duijvenboden, K., Houweling, A.C., Scholman, K.T., Wakker, V., Allaart, C.P., Uhm, J.-S., Mathijssen, I.B., Baartscheer, T., Postma, A.V., Barnett, P., Verkerk, A.O., Boukens, B.J., Christoffels, V.M., 2022. Patient-Specific TBX5-G125R Variant Induces Profound Transcriptional Deregulation and Atrial Dysfunction. Circulation 145, 606–619. https://doi.org/10.1161/CIRCULATIONAHA.121.054347

Weerd, J.H. van, Badi, I., Boogaard, M. van den, Stefanovic, S., Werken, H.J.G. van de, Gomez-Velazquez, M., Badia-Careaga, C., Manzanares, M., Laat, W. de, Barnett, P., Christoffels, V.M., 2014. A Large Permissive Regulatory Domain Exclusively Controls Tbx3 Expression in the Cardiac Conduction System. Circ. Res. 115, 432–441. https://doi.org/10.1161/CIRCRESAHA.115.303591

https://doi.org/10.7554/eLife.80317.sa2

Article and author information

Author details

  1. Fernanda M Bosada

    1. Department of Medical Biology, Amsterdam Cardiovascular Sciences, Amsterdam Reproduction and Development, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, Netherlands
    2. Department of Experimental Cardiology, Amsterdam Cardiovascular Sciences, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, Netherlands
    Contribution
    Conceptualization, Data curation, Formal analysis, Investigation, Visualization, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9817-7370

  2. Karel van Duijvenboden

    Department of Medical Biology, Amsterdam Cardiovascular Sciences, Amsterdam Reproduction and Development, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, Netherlands
    Contribution
    Data curation, Visualization, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5691-8097

  3. Alexandra E Giovou

    Department of Medical Biology, Amsterdam Cardiovascular Sciences, Amsterdam Reproduction and Development, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, Netherlands
    Contribution
    Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared

  4. Mathilde R Rivaud

    1. Department of Medical Biology, Amsterdam Cardiovascular Sciences, Amsterdam Reproduction and Development, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, Netherlands
    2. Department of Experimental Cardiology, Amsterdam Cardiovascular Sciences, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, Netherlands
    Contribution
    Formal analysis, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0680-3483

  5. Jae-Sun Uhm

    1. Department of Medical Biology, Amsterdam Cardiovascular Sciences, Amsterdam Reproduction and Development, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, Netherlands
    2. Department of Cardiology, Severance Hospital, College of Medicine, Yonsei University, Seoul, Republic of Korea
    Contribution
    Data curation, Formal analysis, Writing – review and editing
    Competing interests
    No competing interests declared

  6. Arie O Verkerk

    1. Department of Medical Biology, Amsterdam Cardiovascular Sciences, Amsterdam Reproduction and Development, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, Netherlands
    2. Department of Experimental Cardiology, Amsterdam Cardiovascular Sciences, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, Netherlands
    Contribution
    Data curation, Formal analysis, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-2140-834X

  7. Bastiaan J Boukens

    1. Department of Medical Biology, Amsterdam Cardiovascular Sciences, Amsterdam Reproduction and Development, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, Netherlands
    2. Department of Physiology, University of Maastricht, Cardiovascular Research Institute Maastricht, Maastricht University Medical Center, Maastricht, Netherlands
    Contribution
    Data curation, Formal analysis, Supervision, Investigation, Methodology, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6449-145X

  8. Vincent M Christoffels

    Department of Medical Biology, Amsterdam Cardiovascular Sciences, Amsterdam Reproduction and Development, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, Netherlands
    Contribution
    Conceptualization, Supervision, Funding acquisition, Writing – original draft, Writing – review and editing
    For correspondence
    v.m.christoffels@amsterdamumc.nl
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-4131-2636

Funding

CardioVasculair Onderzoek Nederland (Project 2014-18 CONCOR-genes Young Talent Program)

  • Fernanda M Bosada

Fondation Leducq (14CVD01)

  • Vincent M Christoffels

Dutch Cardiovascular Alliance (OUTREACH)

  • Vincent M Christoffels

ZonMw (TOP 91217061)

  • Vincent M Christoffels

Dutch Research Council (OCENW.GROOT.2019.029)

  • Vincent M Christoffels

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Berend de Jonge for technical assistance. This work was supported by CardioVasculair Onderzoek Nederland (CVON) project 2014–18 CONCOR-genes Young Talent Program (to FMB), CVON project 2014–18 CONCOR-genes (to VMC), Leducq Foundation 14CVD01 (to VMC), Dutch CardioVascular Alliance OUTREACH (to VMC) and ZonMW TOP 91217061 (to VMC).

Ethics

Housing, husbandry, all animal care and experimental protocols were in accordance with guidelines 310 from the Directive 2010/63/EU of the European Parliament and Dutch government. Protocols were 311 approved by the Animal Experimental Committee of the Amsterdam University Medical Centers. 312 Animal group sizes were determined based on previous experience.

Senior Editor

  1. Matthias Barton, University of Zurich, Switzerland

Reviewing Editor

  1. Kalyanam Shivkumar, UCLA Health, United States

Reviewers

  1. Kalyanam Shivkumar, UCLA Health, United States
  2. Benoit Bruneau, University of California, San Francisco, United States

Publication history

  1. Preprint posted: May 14, 2022 (view preprint)
  2. Received: May 27, 2022
  3. Accepted: January 29, 2023
  4. Accepted Manuscript published: January 30, 2023 (version 1)
  5. Version of Record published: February 14, 2023 (version 2)

Copyright

© 2023, Bosada et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 460
    Page views
  • 86
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Fernanda M Bosada
  2. Karel van Duijvenboden
  3. Alexandra E Giovou
  4. Mathilde R Rivaud
  5. Jae-Sun Uhm
  6. Arie O Verkerk
  7. Bastiaan J Boukens
  8. Vincent M Christoffels
(2023)
An atrial fibrillation-associated regulatory region modulates cardiac Tbx5 levels and arrhythmia susceptibility
eLife 12:e80317.
https://doi.org/10.7554/eLife.80317

Categories and tags

Research organism

Further reading

    1. Chromosomes and Gene Expression

    RNA Polymerase II transcription independent of TBP

    James ZJ Kwan, Thomas F Nguyen ... Sheila S Teves
    Research Article

    Transcription by RNA Polymerase II (Pol II) is initiated by the hierarchical assembly of the Pre-Initiation Complex onto promoter DNA. Decades of research have shown that the TATA-box binding protein (TBP) is essential for Pol II loading and initiation. Here, we report instead that acute depletion of TBP in mouse embryonic stem cells has no global effect on ongoing Pol II transcription. In contrast, acute TBP depletion severely impairs RNA Polymerase III initiation. Furthermore, Pol II transcriptional induction occurs normally upon TBP depletion. This TBP-independent transcription mechanism is not due to a functional redundancy with the TBP paralog TRF2, though TRF2 also binds to promoters of transcribed genes. Rather, we show that the TFIID complex can form and, despite having reduced TAF4 and TFIIA binding when TBP is depleted, the Pol II machinery is sufficiently robust in sustaining TBP-independent transcription.

    1. Chromosomes and Gene Expression
    2. Plant Biology

    Vernalization-triggered expression of the antisense transcript COOLAIR is mediated by CBF genes

    Myeongjune Jeon, Goowon Jeong ... Ilha Lee
    Research Article Updated

    To synchronize flowering time with spring, many plants undergo vernalization, a floral-promotion process triggered by exposure to long-term winter cold. In Arabidopsis thaliana, this is achieved through cold-mediated epigenetic silencing of the floral repressor, FLOWERING LOCUS C (FLC). COOLAIR, a cold-induced antisense RNA transcribed from the FLC locus, has been proposed to facilitate FLC silencing. Here, we show that C-repeat (CRT)/dehydration-responsive elements (DREs) at the 3′-end of FLC and CRT/DRE-binding factors (CBFs) are required for cold-mediated expression of COOLAIR. CBFs bind to CRT/DREs at the 3′-end of FLC, both in vitro and in vivo, and CBF levels increase gradually during vernalization. Cold-induced COOLAIR expression is severely impaired in cbfs mutants in which all CBF genes are knocked-out. Conversely, CBF-overexpressing plants show increased COOLAIR levels even at warm temperatures. We show that COOLAIR is induced by CBFs during early stages of vernalization but COOLAIR levels decrease in later phases as FLC chromatin transitions to an inactive state to which CBFs can no longer bind. We also demonstrate that cbfs and FLCΔCOOLAIR mutants exhibit a normal vernalization response despite their inability to activate COOLAIR expression during cold, revealing that COOLAIR is not required for the vernalization process.

    1. Chromosomes and Gene Expression

    X-chromosome target specificity diverged between dosage compensation mechanisms of two closely related Caenorhabditis species

    Qiming Yang, Te-Wen Lo ... Barbara J Meyer
    Research Article

    An evolutionary perspective enhances our understanding of biological mechanisms. Comparison of sex determination and X-chromosome dosage compensation mechanisms between the closely related nematode species C. briggsae (Cbr) and C. elegans (Cel) revealed that the genetic regulatory hierarchy controlling both processes is conserved, but the X-chromosome target specificity and mode of binding for the specialized condensin dosage compensation complex (DCC) controlling X expression have diverged. We identified two motifs within Cbr DCC recruitment sites that are highly enriched on X: 13-bp MEX and 30-bp MEX II. Mutating either MEX or MEX II in an endogenous recruitment site with multiple copies of one or both motifs reduced binding, but only removing all motifs eliminated binding in vivo. Hence, DCC binding to Cbr recruitment sites appears additive. In contrast, DCC binding to Cel recruitment sites is synergistic: mutating even one motif in vivo eliminated binding. Although all X-chromosome motifs share the sequence CAGGG, they have otherwise diverged so that a motif from one species cannot function in the other. Functional divergence was demonstrated in vivo and in vitro. A single nucleotide position in Cbr MEX can determine whether Cel DCC binds. This rapid divergence of DCC target specificity could have been an important factor in establishing reproductive isolation between nematode species and contrasts dramatically with conservation of target specificity for X-chromosome dosage compensation across Drosophila species and for transcription factors controlling developmental processes such as body-plan specification from fruit flies to mice.