Tbx20 Is an Essential Regulator of Embryonic Heart Growth in Zebrafish

The molecular mechanisms that regulate cardiomyocyte proliferation during embryonic heart growth are not completely deciphered yet. In a forward genetic N-ethyl-N-nitrosourea (ENU) mutagenesis screen, we identified the recessive embryonic-lethal zebrafish mutant line weiches herz (whz). Homozygous mutant whz embryos display impaired heart growth due to diminished embryonic cardiomyocyte proliferation resulting in cardiac hypoplasia and weak cardiac contraction. By positional cloning, we found in whz mutant zebrafish a missense mutation within the T-box 20 (Tbx20) transcription factor gene leading to destabilization of Tbx20 protein. Morpholino-mediated knock-down of Tbx20 in wild-type zebrafish embryos phenocopies whz, indicating that the whz phenotype is due to loss of Tbx20 function, thereby leading to significantly reduced cardiomyocyte numbers by impaired proliferation of heart muscle cells. Ectopic overexpression of wild-type Tbx20 in whz mutant embryos restored cardiomyocyte proliferation and heart growth. Interestingly, ectopic overexpression of Tbx20 in wild-type zebrafish embryos resulted, similar to the situation in the embryonic mouse heart, in significantly reduced proliferation rates of ventricular cardiomyocytes, suggesting that Tbx20 activity needs to be tightly fine-tuned to guarantee regular cardiomyocyte proliferation and embryonic heart growth in vivo.


Introduction
The adult mammalian heart largely lacks regenerative capacity, making it refractory to structural and functional recovery after acute or chronic damage [1][2][3]. Intriguingly, in the embryonic and neonatal mammalian heart, there seems to be still pronounced plasticity and adaptive cell proliferation, which however is gradually lost during early childhood, rendering cardiomyocytes mostly post-mitotic in adulthood [4,5]. By contrast and similar to the situation in the embryonic and neonatal mammalian heart, both, embryonic and adult zebrafish cardiomyocytes are not post-mitotic, warranting the efficient regenerative proliferation of pre-existing, spared cardiomyocytes after myocardial injury [6]. To which extent molecular programs used during cardiac development are re-deployed during heart regeneration is incompletely understood. Furthermore, the molecular mechanisms that regulate cardiomyocyte proliferation during embryonic heart development in vertebrates are not thoroughly understood. Recently, large-scale N-ethyl-N-nitrosourea (ENU)-mutagenesis screens for cardiac hypo-and hyperplasia mutants in zebrafish were shown to be suitable to identify novel regulators of embryonic cardiac growth. In this context, an activating mutation of the transcriptional repressor RuvB-like AAA ATPase 2 Reptin in the zebrafish mutant liebeskummer increases embryonic cardiomyocyte proliferation, at least in part via the modulation of β-catenin signaling [7]. Furthermore, the calcium channel voltage-dependent, L type, alpha 1C subunit (CAC-NA1C) in the zebrafish mutant island beat was also shown to regulate heart growth. [8]. Here, disruption of CACNA1C and thereby altered calcium signaling within embryonic cardiomyocytes led to hypoplastic hearts. Moreover, the zebrafish SWI/SNF related, matrix associated, actin dependent regulator of chromatin, subfamily a, member 4 (SMARCA, BRG1; zebrafish mutant brg1 s481 ) was found to be critical for cardiogenesis in zebrafish and to play an essential role in the regulation of embryonic cardiomyocyte proliferation in vivo [9].
The family of T-box (Tbx) transcription factors plays a crucial role during embryonic development in all metazoans. Especially, T-box factors such as Tbx1, Tbx2, Tbx3, Tbx5, Tbx18 and Tbx20 are described to be involved in virtually all steps of cardiogenesis and mutations are associated with developmental syndromes and cardiac abnormalities like DiGeorge syndrome (Tbx1) or Holt-Oram syndrome (Tbx5) [10][11][12]. Additionally, human Tbx20 mutations were associated with atrial septal defects and cardiomyopathy, whereas Tbx20 null mice die at midgestation due to the lack of proper heart chamber formation and cardiac hypoplasia [13,14]. In zebrafish, Morpholino-mediated knock-down of Tbx20 was found to cause defective heart chamber morphology and impaired development of the atrio-ventricular boundary in embryonic hearts [15].
Here, we show that embryonic cardiomyocytes in the ENU-induced recessive lethal zebrafish mutant weiches herz (whz m245 ) fail to proliferate, leading to severe cardiac hypoplasia and finally heart failure. By positional cloning, we demonstrated that a missense mutation in the zebrafish tbx20 gene, leading to the destabilization of Tbx20 protein, is responsible for the whz phenotype. Consistently, embryonic cardiac hypoplasia can also be induced by targeted knock-down of Tbx20, whereas cardiomyocyte proliferation can be restored in whz mutant hearts by ectopic overexpression of wild-type Tbx20. Remarkably, ectopic overexpression of Tbx20 in wild-type zebrafish embryos resulted in reduced proliferation rates of ventricular cardiomyocytes, suggesting that Tbx20 levels need to be tightly regulated during embryonic heart growth. Taken together, our findings demonstrate a pivotal role of Tbx20 in the regulation of cardiomyocyte proliferation in the embryonic zebrafish heart.

Zebrafish strains and injection procedures
All procedures and experiments in this study were carried out after appropriate institutional approvals (Tierforschungszentrum (TFZ) Ulm University, No. 0183), which conform to the EU Directive 2010/63/EU.

Isolation of zebrafish protein lysates and Western blot analysis
For each of the four independent Western blot experiments, 100 embryonic hearts from wildtype and whz mutant zebrafish, respectively, expressing GFP exclusively in cardiomyocytes (Tg(cmlc2:rasEGFP)) were dissected out at 72 hpf and protein lysates were prepared. For Western blot analysis each protein lysate was boiled in 3x Laemmli Buffer and loaded on a precast 10% SDS gel (Bio-Rad). Proteins were separated by SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membrane. After blocking in 5% milk powder in TBST for 2h at RT, the membrane was incubated with the primary antibody over night at 4˚C. The secondary antibodies were incubated for 2h at RT [16].
The following primary antibodies were used: Tbx20 primary antibody (1:1000, abcam; ab93058). Signals were detected by chemiluminescence (anti-rabbit IgG HRP-linked, Cell Signaling), using a luminescent image analyzer (Image Quant Las4000 mini). In-vitro synthesized Tbx20 protein was generated following standard protocols (TnT Kit, Promega) and used as positive control for western blot analysis. Western blots were quantified by using Image Quant LAS4000 software.

In situ hybridization and Histology
A standard protocol was followed for in situ hybridization of embryos after fixation and dechorionation at the indicated stages. Probes were prepared by in vitro transcription. Wholemount in situ hybridization was performed using Digoxigenin labelled antisense RNA probe and visualized using anti-Digoxigenin Fab fragments conjugated with alkaline phosphatase (Roche Molecular Biochemicals) as described [17,18]. Embryos were processed and hybridized as described [19], except that 10 mg/ml of proteinase K in PBS/0.1% Tween-20 was used for 10 to 30 minutes depending on the age of the collected embryos.
Genetic mapping, positional cloning and mutation detection DNA from 48 whz mutants and 48 wild-type siblings was pooled and bulked segregation analysis was performed. The critical genomic interval for whz was defined by genotyping 1750 mutant embryos for polymorphic markers in the area. The whz locus was restricted to three overlapping bacterial artificial chromosomes (BAC) zC215K15, zC78C9 and zC281F5. Further recombination analyses using single nucleotide polymorphisms (SNPs) and simple sequence length polymorphisms (SSLP) derived from sequence of the overlapping BAC clones, the whz mutation interval could be restricted to a 89 kb region on BAC zC215K15 that contains 2 open reading frames encoding the proteins HERPUD family member 2 (herpud2; NP_956482) and the zebrafish T-box transcription factor Tbx20 (NP_571581). The stop codon mutation was detected by sequencing of the entire coding region of zebrafish herpud2 and tbx20 from wildtype and whz mutant cDNA.

RNA extraction and quantitative real-time PCR
For each the 4 biological replicates, a pool of 25 embryos were collected at 72 hpf. RNA extraction was carried out by RNeasy1 Mini Kit (Qiagen) according to the manufacturer's instructions. Total RNA (1 μg) was reverse transcribed to produce cDNA using Superscript III reverse transcriptase (Life Technologies).

Terminal deoxynucleotidyl transferase Dig-dUTP nick end-labeling (TUNEL)
To analyze cardiomyocyte apoptosis, TUNEL stainings were performed on zebrafish embryos at 72 hpf according to the manufacturer's instructions (Merck Millipore). For histology, the TUNEL stained zebrafish embryos were dehydrated prior to embedding in JB-4 (Polysciences). 5-μm sections were cut using a Leica RM2255 microtome (Leica, Wetzlar, Germany), dried, and stained with Hematoxylin and Eosin as described [21]. Images were recorded using a Zeiss Axioskop 2 plus.

Isolation of embryonic hearts and proliferation assay
Proliferating cardiomyocytes in zebrafish hearts at 72 hpf were counted upon the detection of DNA synthesis by the incorporation of 5-ethynyl-2'-deoxyuridine (EdU) using the Click-iT EdU Imaging Kit (Life Technologies, Carlsbad, CA, USA). Cardiomyocyte nuclei were counterstained using an antibody against Mef-2 (Santa Cruz Biotechnology C-21 HO310) at a dilution of 1:200. Anti-rabbit Alexafluor 555 antibodies were used to detect the fluorescence by an iMIC digital microscope (Till Photonics) or a confocal microscope (Leica TCS SP8).

Statistical analysis
All graphs and statistical analyses were prepared with the help of Prism4 (GraphPad software). If not further specified, results are expressed as mean ± S.D. Analyses were performed using nonparametric Mann-Whitney tests and a value of P<0.05 was accepted as statistically significant.

Embryonic cardiomyocyte proliferation is reduced in weiches herz embryos
In search for novel regulators of embryonic cardiomyocyte proliferation, we identified here by visual screening the "small heart" zebrafish mutant weiches herz (whz m245 ) [22] (Fig 1A and  1B). Homozygous whz mutant embryos display small and only weakly contracting hearts and die 5 days post-fertilization (dpf) (Fig 1A-1D).
By 72 hpf, myocardial and endocardial cell layers are clearly defined in whz mutant hearts but the ventricular chambers remain small and the myocardium single-layered (Fig 1E and  1F). Starting at 48 hpf, ventricular myocardium in zebrafish usually thickens by the addition of new cardiomyocytes [7,8]. To assess whether the observed "small heart" phenotype is due to reduced cardiomyocyte numbers in whz mutant hearts, we evaluated dissected wild-type and whz mutant hearts stained with antibodies against MEF-2 (Myocyte Enhancer Factor-2), known to mark nuclei of cardiomyocytes, and S46, exclusively labeling atrial cardiomyocytes, at 48 and 72 hpf. At 72 hpf, we found significantly less ventricular cardiomyocytes in whz mutants compared to wild-type littermates (sib: 144.2±10 SD and whz: 94.9±10 SD, n = 10 hearts, 3 independent experiments; p<0.0001) (Fig 1G-1I), whereas at 48 hpf cardiomyocyte numbers in whz mutant ventricles were not reduced compared to control embryos (wt: 93.4 ±10 SD and whz: 88.2±10 SD, n = 10 hearts, 3 independent experiments; p>0.05) (Fig 1I), suggesting that weiches herz interferes with cardiac growth after 48 hpf.
To investigate whether increased cardiomyocyte apoptosis is the cause for cardiac hypoplasia in whz at 72 hpf, we conducted TUNEL assays, performed serial histological sections and quantified the number of apoptotic cardiomyocytes. We found no increase in the number of apoptotic cardiomyocytes in whz mutant hearts compared to wild-type littermates (wt = 0.40 ± 0.16; whz = 0.30±0.15, n = 10 hearts, 3 independent experiments; p = 0.6601) (Fig  2A and 2B).
These findings indicate that rather impaired cardiomyocyte proliferative growth accounts for the small heart phenotype in whz mutant zebrafish embryos and not increased cardiomyocyte apoptosis.

Weiches herz (whz m245 ) encodes zebrafish t-box transcription factor 20 (tbx20)
To reveal the ENU-induced genetic defect in whz m245 mutant embryos, we performed a genome-wide study of microsatellite marker segregation by bulked segregant analysis and located whz in-between the microsatellite marker Z1215 and Z6240 on chromosome 16.
To finally identify the ENU-induced mutation in whz, we sequenced the entire coding sequences of zebrafish herpud2 and tbx20 from wild-type and whz mutant cDNA. We found the whz mutation to be a transversion of nucleotide Adenine to Cytosine at codon 1225 of zebrafish tbx20 (Fig 3B), whereas in herpud2 no mutation was detected by cDNA sequencing. Interestingly, the identified mutation is predicted to change the tbx20 stop codon into the amino acid cysteine (Fig 3B) thereby resulting in prolongation of the whz mutant Tbx20 protein by 29 amino acids (Fig 3C).
To test whether the identified loss-of-stop mutation in whz results in degradation of whz mutant tbx20 mRNA, we performed whole-mount antisense RNA in situ hybridizations on wild-type and whz mutant embryos at 72 hpf using tbx20-specific riboprobes. We found no alterations in tbx20 mRNA levels and location between whz mutants and wild-type littermates (Fig 3D and 3E). To confirm these findings from whole-mount antisense in situ hybridizations, we next performed quantitative RT-PCR analyses of tbx20 mRNA levels in wild-type and whz mutant embryos at 72 hpf. We found that tbx20 mRNA levels were not affected by the whz mutation (n = 4; p = 0.5957), suggesting that whz mutant tbx20 mRNA is not degraded by nonsense-mediated RNA decay ( Fig 3F).
Next, to assess whether the predicted Tbx20 protein elongation in whz mutant embryos destabilizes Tbx20 proteins, we conducted Western blot analyses using a Tbx20-specific antibody and found that Tbx20 protein levels were significantly reduced (n = 4 experiments; p = 0.0286) (Fig 3G and 3H). These findings suggest that the whz loss-of-stop mutation and the addition of 29 amino acids at the C-terminus of Tbx20 leads to a destabilization and degradation of the protein.
Targeted knock-down of Tbx20 in wild-type zebrafish leads to inhibition of embryonic cardiomyocyte proliferation Our Tbx20 Western blot analyses imply that rather loss of Tbx20 function than a dominantnegative effect of elongated Tbx20 appears to cause the whz ventricular hypoplasia phenotype. To further validate the finding that ventricular hypoplasia in whz mutants is caused by loss of Tbx20 function, we injected Morpholino-modified antisense oligonucleotides (MOs), either directed against the translational start site (MO1-tbx20) or the splice donor site of exon 3 (MO2-tbx20) of zebrafish tbx20 into wild-type zebrafish embryos at the one-cell stage. When injected with 2.5 ng of MO1-tbx20 or 3.1 ng of MO2-tbx20, 78% and 76% of the injected embryos (n = 150, respectively) displayed the whz mutant phenotype (Fig 4B, 4D and 4G) indicating a mono-layered, small ventricle (Fig 4F). Injection of the same amount of the corresponding five base pair mismatch MOs as control did not result in any pathology (Fig 4A, 4C, 4E and 4G).
Ectopic expression of wild-type tbx20 mRNA restores cardiomyocyte proliferation in whz mutants To finally prove that impaired cardiomyocyte proliferation in whz mutant embryos is indeed caused by loss of Tbx20 function, we next performed a genetic rescue experiment by ectopically expressing wild-type tbx20 (tbx20 wt ) mRNA in homozygous mutant whz embryo. To do so, we injected 8 ng of tbx20 wt into one-cell stage whz mutant embryos derived from intercrossing heterozygous whz zebrafish (n = 150, 3 independent experiments). As shown in Fig 5, in 73% ± 3.4% of homozygous mutant whz embryos, cardiac growth was completely restored by injection of wild-type tbx20 mRNA (Fig 5A-5C). Cardiomyocyte numbers of tbx20 mRNA injected whz mutant ventricles were comparable to wild-type littermates and significantly increased compared to control-injected whz mutants at 72 hpf (whz + tbx20 mRNA: 142.5±10 SD, whz + KCl: 87.82±10 SD, n = 10 hearts, 3 independent experiments; p = 0.0001) (Fig 5F). Furthermore, to validate that the reconstitution of cardiac growth in tbx20 mRNA injected whz mutant was due to re-initiated cardiomyocyte proliferation, we evaluated cell proliferation by EdU stainings in tbx20 mRNA and control injected whz mutant embryos. In fact, we found significantly increased numbers of EdU-positive cardiomyocytes at 72 hpf (whz + tbx20 wt : 7 ±3% SD, whz + KCl: 1±2% SD, n = 10 hearts, 3 independent experiments; p = 0.0001) (Fig 5D,  5E and 5G), indicating that ectopic expression of wild-type tbx20 mRNA in whz mutant embryos is able to restore cardiomyocyte proliferation.

Discussion
Due to its optical translucency during embryogenesis, the zebrafish is a highly suitable model system to investigate heart development in vivo. In particular, the cellular and molecular processes that are involved in embryonic cardiac growth can be easily assessed in zebrafish in vivo. In search for novel drivers of embryonic heart growth, we isolated here in a large scale ENU mutagenesis screen the "small heart" zebrafish mutant weiches herz (whz) and found that proliferation of ventricular cardiomyocytes is severely reduced in whz mutants. By positional cloning, as well as loss-and gain-of function experiments we demonstrated here for the first time that zebrafish Tbx20 is essentially involved in the regulation of embryonic heart growth by orchestrating cardiomyocyte proliferation.
Loss of Tbx20 in weiches herz mutant zebrafish embryos results in pathological cardiac hypoplasia. In contrast to zebrafish embryos deficient for the PAF1 complex component Ctr9 or the cardiac sodium channel Scn5Lab [23], we find that cardiac hypoplasia in whz is not caused by a reduced amount of cardiac precursor cells in the anterior lateral plate mesoderm. These findings are consistent with data derived from Tbx20 morphant embryos demonstrating that the expression of nkx2.5 in the anterior lateral plate mesoderm at the 12 somite stage is unaltered by the targeted Morpholino-mediated loss of Tbx20 [15]. In fact, by EdU-staining that specifically mark cells in the S-phase of the cell cycle, we found here that embryonic cardiomyocytes in Tbx20-deficient whz mutant embryos fail to proliferate, suggesting that Tbx20 might be crucial to control embryonic cardiomyocyte proliferation in zebrafish.
The whz mutation in zebrafish Tbx20 is a transversion of the nucleotide Adenine to Cytosine at the stop codon of zebrafish tbx20, resulting in a change of the tbx20 stop codon to the amino acid Cysteine ("nonstop" mutation) and thereby an extension of the Tbx20 protein by 29 amino acids. As shown by Western Blot analyses, this putative protein prolongation results in the destabilization of Tbx20. Tbx20 mRNA levels are unaffected by the nonstop mutation. Similar effects of stop codon mutations on protein stability were observed in other species such as zebrafish, mice and humans [24][25][26][27]. As shown in mammalian cells, expression levels of nonstop mRNAs generally appeared unaltered, whereas ribosome stalling at the 3-prime end of the elongated nonstop mRNA interfered with regular translation and led to premature termination of protein synthesis [28][29][30].
We found here that the whz mutation interferes with Tbx20 protein stability resulting in complete loss of Tbx20 function. By Morpholino-mediated knock-down of zebrafish tbx20 and the subsequent evaluation of cardiomyocyte proliferation we confirmed the loss of function effect of the whz mutation. In a previous study, where zebrafish Tbx20 was also partially inactivated by a Morpholino-based antisense oligonucleotide approach, Szeto and co-workers found defective heart chamber morphology and impaired development of the boundaries between both cardiac chambers [15], phenotypic characteristics that were also present in our whz mutant embryos, whereas, the impact of loss of zebrafish Tbx20 function on embryonic cardiomyocyte proliferation was not evaluated in this study.
Similar to Tbx20-deficient whz mutant embryos, targeted ablation of Tbx20 in mice was recently shown to lead to hypoplastic hearts due to impaired cardiomyocyte proliferation, and thereby embryonic lethality at mid-gestation [13,14,[31][32][33][34]. Cardiomyocyte-specific ablation of Tbx20 in adult mice leads to cardiomyopathies accompanied by arrhythmias, heart failure and ultimately death [31]. Remarkably, cardiomyocyte-specific overexpression of Tbx20 can enhance cardiomyocyte proliferation in fetal ventricular cardiomyocytes leading to increased Tbx20 Regulates Heart Growth thickening of the myocardium [35], whereas overexpression of Tbx20 in the embryonic mouse heart at E9.5 resulted in reduced cardiomyocyte proliferation rates, suggesting an opposite impact of Tbx20 overexpression on cardiomyocyte proliferation in embryonic and fetal cardiomyocytes of mice [35]. We found here that, similar to the situation in embryonic mice, overexpression of Tbx20 in the zebrafish embryo resulted in significantly reduced cardiomyocyte numbers due to impaired proliferation, suggesting conserved roles of Tbx20 during embryonic heart growth in zebrafish and mice. Whether Tbx20 overexpression in larval or adult zebrafish leads to reduced or increased cardiomyocyte proliferation rates is currently not known. Dissected wild-type hearts injected with KCl (A) and tbx20 mRNA (B) at 72 hpf, stained against MEF-2 (red) after incorporation of 5-ethynyl-2'-deoxyuridine (EdU; green) to visualize the effect of tbx20 overexpression on cardiomyocyte proliferation. (C) Wild-type zebrafish hearts injected with wild-type tbx20 mRNA show significantly reduced numbers of ventricular cardiomyocytes at 72 hpf compared to control injected hearts (wt+KCl: 148.6±2.9; wt + tbx20 mRNA: 64.0±2.017, n = 10, p< 0.0002). (D) Ventricular cardiomyocyte proliferation in wild-type embryos injected with tbx20 mRNA is significantly reduced compared to control injected embryos at 72 hpf (wt+KCl: 8.9 ± 0.38; wt+tbx20 mRNA: 1.1±0.28, n = 10, p = 0.0001). Interestingly, in adult murine hearts, Tbx20 overexpression beginning in the fetal period promotes the activation of cardiac proliferative pathways such as BMP2/pSMAD1/5/8 and PI3K/ AKT/GSK3β/β-catenin signaling, thereby maintaining adult cardiomyocyte in an immature state in vivo [35]. Very recently, Xiang et al. were even able to show that Tbx20 overexpression significantly increased cardiomyocyte proliferation and cardiac repair after induced myocardial infarction (MI) [36].
Tbx20 was recently shown to physically bind and thereby repress the cell cycle inhibitory proteins p21 and Meis1 as well as the antiproliferative protein Btg2, thereby activating cardiomyocyte proliferation [36]. Transcriptional profiling of Tbx20-deficient weiches herz mutant zebrafish hearts as well as hearts with stable, cardiomyocyte-specific overexpression of Tbx20 will help us to identify the Tbx20-dependent regulatory transcriptional network that controls cardiomyocyte proliferation.