JunB-CBFβ signaling is essential to maintain sarcomeric Z-disc structure and when defective leads to heart failure

In muscle cells, a complex network of Z-disc proteins allows proper reception, transduction and transmission of mechanical and biochemical signals. Mutations in genes encoding different Z-disc proteins such as integrin-linked kinase (ILK) and nexilin have recently been shown to cause heart failure by distinct mechanisms such as disturbed mechanosensing, altered mechanotransduction or mechanical Z-disc destabilization. We identified core-binding factor β (CBFβ) as an essential component for maintaining sarcomeric Z-disc and myofilament organization in heart and skeletal muscle. In CBFβ-deficient cardiomyocytes and skeletal-muscle cells, myofilaments are thinned and Z-discs are misaligned, leading to progressive impairment of heart and skeletal-muscle function. Transcription of the gene encoding CBFβ mainly depends on JunB activity. In JunB-morphant zebrafish, which show a heart-failure phenotype similar to that of CBFβ-deficient zebrafish, transcript and protein levels of CBFβ are severely reduced. Accordingly, ectopic expression of CBFβ can reconstitute cardiomyocyte function and rescue heart failure in JunB morphants, demonstrating for the first time an essential role of JunB-CBFβ signaling for maintaining sarcomere architecture and function.


Introduction
The sarcomeric Z-disc is of great importance in the vertebrate heart to allow cardiomyocytes to adapt to continually changing hemodynamic needs (Hoshijima, 2006). In humans, mutations in Z-disc proteins, such as muscle Lim protein (MLP), integrin-linked kinase (ILK) and nexilin, cause dilated cardiomyopathy (DCM), the leading cause of premature death of the young in western countries (Bendig et al., 2006;Hassel et al., 2009;Knoll et al., 2002;Knoll et al., 2007). However, the exact in vivo function of most of the Z-disc constituents is largely unknown. Interestingly, a subset of Z-disc proteins such as calsarcin and MLP were recently found to localize at both the sarcomeric Z-disc and the cell nucleus, suggesting that, upon mechanical stress, these proteins might shuttle from the mechanical integration site of muscle cells, the sarcomeric Z-disc, to the nucleus of muscle cells to regulate transcriptional processes (Frey et al., 2004;Knoll et al., 2002).
The core-binding factor (CBF) transcription complex consists of a heterodimer of one and one -subunit (Kamachi et al., 1990;Ogawa et al., 1993;Wang et al., 1993). CBF contains a Runt domain, which is responsible for DNA-binding activity and heterodimer formation with CBF, thereby enhancing DNAbinding activity of CBF (Bae et al., 1993;Kagoshima et al., 1993;Levanon et al., 1998;Ogawa et al., 1993). In endothelial cells, CBF transcription mainly depends on the activating protein 1 (AP-1) factor JunB, which is known to be involved in various aspects of cellular adaptation and stress response (Alfonso-Jaume et al., 2006;Licht et al., 2006). Targeted deletion of CBF in mice revealed an essential role of CBF in several biological processes such as the induction of definitive hematopoiesis in the fetal liver and the regulation of osteogenesis (Erickson et al., 1992;Kundu et al., 2002;Kundu and Liu, 2003;Sasaki et al., 1996). However, the exact role of CBF in the living animal, especially in the heart and skeletal muscle, where it was found to be highly expressed at the sarcomeric Z-disc, is still not defined because CBF-deficient mice die very early in utero [by embryonic day (E)11.5-E13.5], most probably owing to hemorrhage in the central nervous system (Chiba et al., 1997;Sasaki et al., 1996). In contrast to mice and other mammals, zebrafish embryos do not require a functional cardiovascular system for survival during embryogenesis and early larval stages, allowing careful analysis of cardiac defects without the confounding context of a dying organism (Pelster and Burggren, 1996). Furthermore, external fertilization and transparency allows unrestricted in vivo analyses of cardiac morphology and function throughout development.
To elucidate for the first time the in vivo function of CBF in the vertebrate heart, we generated CBF-deficient zebrafish. CBF morphants show cardiac-chamber dilation and reduced systolic force, which are phenotypic hallmarks of human heart failure, and additionally show signs of skeletal-muscle myopathy. CBF deficiency results in structural abnormalities of sarcomeres and Zdiscs. Similar to the situation in rodents, we found CBF protein to be localized at both the sarcomeric Z-disc and the cell nucleus of heart and skeletal-muscle cells. CBF expression in heart and skeletal muscle is regulated by the AP-1 transcription factor JunB. Accordingly, overexpression of CBF rescues the heart-failure phenotype of JunB morphants. In summary, JunB-CBF signaling JunB-CBF signaling is essential to maintain sarcomeric Z-disc structure and when defective leads to heart failure is essential to maintain muscle sarcomere structure, and when defective leads to heart and skeletal-muscle failure.

CBF localizes to sarcomeric Z-discs and nuclei in cardiomyocytes
To study for the first time the function of CBF in the vertebrate heart, we identified the zebrafish orthologous sequence by searching the zebrafish expressed sequence tag (EST) database with the National Center for Biotechnology Information (NCBI) tblastn program. Zebrafish cbf cDNA sequence NM_199209 showed the highest matching score with an amino acid identity between human and zebrafish of 88% and between mouse and zebrafish of 91% (Fig. 1).
Next, we evaluated temporal and spatial mRNA distribution of cbf in the zebrafish embryo by whole-mount RNA antisense in situ hybridization. By 36 hours post-fertilization (hpf), we found significant levels of cbf mRNA in the zebrafish heart and skeletal muscle. By 72 hpf, besides its strong expression in muscle tissue, cbf mRNA was detectable in the retina, brain, branchial arches, fin buds and gastrointestinal tract ( Fig. 2A,B). To evaluate in detail the expression of CBF within heart and skeletal-muscle cells, we performed immunostainings with a polyclonal antibody directed against human CBF ( Fig. 2C-F). As revealed by coimmunofluorescence imaging with anti--actinin and anti-Nexilin antibodies, zebrafish CBF protein specifically localized at sarcomeric Z-discs of zebrafish cardiomyocytes and skeletal-muscle cells ( Fig. 2D-F), whereas other sarcomeric structures were devoid of CBF expression. Furthermore, in accordance with its known function as a transcriptional cofactor, we found CBF protein in nuclei of heart and skeletal-muscle cells (Fig. 2C).

Inactivation of CBF leads to heart failure and skeletalmuscle myopathy in zebrafish embryos
To investigate the structural and functional role of CBF in the vertebrate heart in vivo, we inactivated zebrafish CBF by injecting morpholino-modified antisense oligonucleotides directed against either the translational start site (MO1-cbf) or the splice-acceptor site of exon 2 (MO2-cbf) of zebrafish cbf into one-cell-stage zebrafish embryos (for phenotypes of MO-control and MO2-cbf embryos see Fig. 3A,B,D,E). First, to evaluate the efficacy of the CBF knockdown, we analyzed whole-embryo cbf-mRNA abundance in CBF morphants at 48 and 72 hpf. MO2-cbf is predicted to block splicing at the intron-1-exon-2 boundary, leading either to skipping of exon 2 (band at 167 bp) or integration of intron 1 (band at 861 bp) and consecutively to premature termination of translation of zebrafish CBF. As shown in Fig. 3G (inset), in MO2-cbf-injected embryos, two aberrant splice products that either lack exon 2 or show integration of intron 1 can be detected, and the amount of the wild-type form of cbf is strongly reduced. Accordingly, as shown in Fig. 3I,J and discussed later in the manuscript, we found that, in knockdown embryos, CBF protein levels were severely downregulated, demonstrating high efficacy of the MO1-cbf start-site morpholino. As a result of cbf knockdown, 86% of the MO1-cbf (n100; 8 ng/embryo)-injected and 95% of the MO2-cbf (n100; 8 ng/embryo)-injected zebrafish embryos developed severe heart failure, whereas embryos injected with a control morpholino (MO-control; n100, 8 ng/embryo) Fig. 1. CBF is highly conserved across species. Amino acid sequence alignment of zebrafish (dr), human (hs) and mouse (mm) CBF, demonstrating the high cross-species homology. Black boxes indicate amino acid identity; gray boxes indicate amino acids with similar chemical properties.

Fig. 2. Expression and localization of CBF in zebrafish embryos.
(A,B)RNA antisense in situ hybridization demonstrates abundant cbf-mRNA expression, including in heart cells and skeletal muscle (skm). Shown is a lateral view of an embryo at the 72-hpf stage. wt, wild type. (C-F)In zebrafish heart (C,D) and skeletal muscle (E,F), CBF protein localizes to the cell nucleus (cell nuclei are co-stained with DAPI; blue) and the sarcomere. Localization of CBF at the sarcomeric Z-disc of heart and skeletal muscle cells is demonstrated by co-immunostaining with Nexilin (D,F) and -actinin were devoid of any abnormalities ( Fig. 3G and data not shown). By 24 hpf, similar to wild-type hearts, CBF-morphant hearts showed regular peristaltic contraction waves. By contrast, by 36 hpf, atrial and ventricular chambers of CBF morphants started to dilate and the contractile force of both heart chambers began to decline. From 48 to 72 hpf, heart failure in CBF morphants worsened even further owing to an additional drop in systolic function ( Fig. 3H; supplementary material Movie 1). Finally, impairment of atrial contractility and consecutive dilation of the atrium in morphant embryos promoted pronounced blood regurgitation from the atrium to the sinus venosus (Fig. 3D,E). Besides the loss of cardiac contractility, skeletal-muscle function was also impaired in CBF-knockdown embryos. Upon tactile stimulation, MO-cbf-injected embryos showed only weak skeletalmuscle movements resembling a quiver or rigid shudder (supplementary material Movie 2) -hindering them to swim or hatch -or even showed complete paralysis.

Heart failure in CBF morphants is due to disturbed sarcomerogenesis
In order to evaluate how CBF deficiency leads to heart failure in zebrafish, we next analyzed cardiac structure and ultrastructure in CBF morphants. Histological analysis of CBF-deficient hearts revealed defined cardiac chambers with each atrium and corresponding ventricle separated by the atrio-ventricular ring. Endocardial and myocardial layers of both heart chambers were developed properly with a multilayered ventricular myocardium (Fig. 4A,B). Atrial and ventricular cardiomyocytes of CBF morphants expressed myosin heavy chain isoenzymes in the typical heart-chamber-specific pattern, indicating normal molecular chamber specification (Fig. 3C,F).
To further elucidate whether, similar to the situation observed in the Cardiac myosin light chain 2 (cMLC2)-deficient zebrafish mutant tell tale heart (myl7) or the Titin-deficient zebrafish mutant pickwick (ttna), defective sarcomere assembly is responsible for impaired heart function in CBF morphants, we analyzed the ultrastructure of cardiomyocytes by transmission electron microscopy Xu et al., 2002). At 72 hpf, sarcomeric units of MO-control-injected zebrafish cardiomyocytes consisted of highly organized well-aligned thick and thin myofilaments, flanked by distinct Z-discs, where adjacent thinfilament bundles interconnect (Fig. 4C,E). By contrast, in CBFdeficient hearts, although primary organization of thick and thin filaments in hexagonal lattices, as well as organization of these bundles in higher ordered sarcomeric units, is preserved, these sarcomeres are thin and composed of only a few bundles of thick and thin filaments (Fig. 4D,F,G). In addition to the sarcomeric alteration, we found abnormally thick and blurry Z-discs and intercalated discs in CBF-morphant cardiomyocytes. A similarly altered ultrastructure of sarcomere architecture was observed in skeletal-muscle cells of CBF morphants ( Fig. 4E-G).

CBF deficiency does not interfere with ANF and VEGF-A expression
Proteins of the sarcomeric Z-disc, such as ILK, were recently found to be essential for cardiac mechanosensing and mechanotransduction in the zebrafish (Bendig et al., 2006). In the hearts of the ILK-deficient zebrafish mutant main squeeze (ilk), expression of the stretch-responsive factors Atrial natriuretic factor (ANF) and Vascular endothelial growth factor-A (VEGF-A) was severely reduced. Hence, to evaluate whether CBF is also essential for ILK-dependent cardiac mechanosensing and, when defective, leads to impairment of cardiac contractility via defective ILK signaling, we assayed levels of anf and vegf-a mRNA in CBFmorphant hearts. As shown in Fig. 5A-C, in contrast to ILKmutant zebrafish, levels of neither anf nor vegf-a mRNA were reduced in CBF-morphant hearts, suggesting that impairment of contractile function in CBF morphants is not caused by interference with ILK-ANF signaling.

JunB acts upstream of CBF in the zebrafish heart
Recently, it was shown that CBF expression in mouse endothelial cells is dependent on JunB signaling (Licht et al., 2006). To evaluate whether CBF expression in the vertebrate heart is also regulated by JunB, we inactivated JunB by morpholino-modified antisenseoligonucleotide injection in zebrafish. To do so, we first identified the zebrafish ortholog by Blast search (Fig. 6A). On the basis of this sequence information, we generated an in situ hybridization probe to evaluate temporal and spatial mRNA distribution of junb in the zebrafish embryo. By 48 and 72 hpf, we found significant levels of junb mRNA in the zebrafish heart and skeletal muscle (Fig. 6B-E). Furthermore, we found strong expression of junb in the retina, brain, branchial arches, fin buds and gastrointestinal tract.
Next, we designed a morpholino directed against the translational start site of JunB (MO-junb) and injected it into wild-type zebrafish embryos at the one-cell stage. Similar to CBF morphants, JunB morphants developed heart failure with cardiac-chamber dilation, reduced systolic function and precardial blood congestion (Fig.  7A,B,D; supplementary material Movie 3). Ventricular chambers of MO-junb-injected zebrafish embryos become hypo-contractile by 36 hpf. By 72 hpf, fractional shortening of the ventricular chamber in JunB morphants was reduced to 18±9% ( Fig. 7C; supplementary material Movie 2). Furthermore, similar to the situation in mouse endothelial cells, in JunB morphants levels of both cbf mRNA and CBF protein were severely downregulated, demonstrating the dependency of zebrafish cbf transcription on the transcriptional regulator JunB (Fig. 8A,B).
Next, to evaluate whether heart failure in JunB morphants is indeed mainly due to loss of CBF expression, we injected either capped cbf mRNA or control solution (200 mM KCl) into wildtype or JunB-morphant embryos. Overexpression of cbf mRNA in wild-type zebrafish embryos did not affect cardiac contractility (Fig. 8C). By contrast, JunB morphants injected with 4 ng of cbf mRNA showed much stronger heart contractility than control-injected JunB morphants. At 48 hpf, when signs of heart failure accompanied by progressive precardial congestion and pericardial edema usually proceed further in JunB-deficient embryos, cbf-mRNA-injected JunB morphants showed almost normal cardiac contractility (fractional shortening: control solution + MO-junb  19.1% vs cbf mRNA + MO-junb  37.8%; P<0.005; n10). Together, these findings indicate that CBF acts downstream of JunB to control cardiac contractility in the vertebrate heart.

Discussion
Z-discs are the mechanical integration sites of both heart and skeletal-muscle cells, linking anchorage of myofilaments to force reception and processing. However, the Z-disc-associated signaling pathways that mediate these functions are largely unknown. We identified here JunB-CBF signaling to be essential for Z-disc stability and sarcomerogenesis; when this signaling cascade is defective, heart failure and skeletal myopathy result.
We and others have recently shown that proteins located at the sarcomeric Z-disc are of great importance for mechanical Z-disc stabilization, mechanosensing and mechanotransduction in the heart (Bendig et al., 2006;Hassel et al., 2009;Knoll et al., 2002). Whereas, for example, disruption of ILK does not interfere with Z-disc ultrastructure, it affects the expression of stretch-responsive genes such as those encoding ANF and VEGF-A in the myocardium (Bendig et al., 2006). By contrast, in CBF morphants, myocardial ANF and VEGF-A expression is unaffected, which suggests that CBF does not affect ILK-ANF-VEGF signaling. Protein shuttling between the sarcomere and the nucleus seems to be a prerequisite for cardiac mechanosensing and mechanotransduction (Knoll et al., 2002). Similar to other Zdisc proteins such as cardiac MLP or calsarcin, CBF is a good candidate to sense and transduce muscle stress signals, because it can reside at the mechanical integration site of muscle cells, the Z-disc, and also in the nuclei of these cells, where it is part of the CBF transcription complex (Frey et al., 2004). Recently, it was shown that CBF interacts with the filamin network via a filaminbinding domain (Yoshida et al., 2005). Filamins are integrators of cell mechanics and stress signaling, and are known to interact with both transcription factors and distinct structural proteins of the cytoskeleton, such as -spectrin, dystrophin and -actinin (Stossel et al., 2001;Stossel and Hartwig, 2003). The association of CBF with the filamin network and the cardiac Z-disc might trigger important messenger events, which are transduced via CBF-dependent transcriptional activity. However, the precise nuclear function of CBF in cardiac and skeletal muscle remains speculative, although our findings implicate a crucial role in the transcriptional regulation of myofilament and Z-disc proteins, because myofilament and Z-disc structures are severely altered upon CBF deprivation.
In CBF morphants, Z-discs are misaligned and myofilaments are thinned, leading to reduced force generation of both cardiacand skeletal-muscle cells. Z-disc abnormalities, as observed in CBF morphants, are found in various disease models of genetic heart failure. For instance, cardiomyopathies associated with mutations in the genes encoding desmin, metavinculin, cardiac muscle LIM protein and nexilin are accompanied by structural alterations of sarcomeric Z-discs Vogel et al., 2009). Besides its structural role at the Z-disc, CBF also regulates cardiac and skeletal sarcomerogenesis, because sarcomeres are significantly thinner in CBF morphants. Similar changes in myofilament structure can be observed in other models of heart failure and range from almost completely disrupted sarcomeric units, for example in the Troponin-T-deficient zebrafish mutant  silent heart (tnnt2a), the Titin-deficient mutant pickwick and the cMLC2-deficient zebrafish mutant tell tale heart, to minor changes in the content, size or organization of sarcomeres in Tropomyosin-4-or connexin Cx36.7-deficient zebrafish Sehnert et al., 2002;Sultana et al., 2008;Xu et al., 2002;Zhao et al., 2008).
Tight transcriptional control of cardiac proteins is essential for appropriate adaptation to varying mechanical needs. For instance, expression of stretch-responsive VEGF in the vertebrate heart is regulated by upstream effectors including ILK and Protein kinase B (PKB), thereby regulating cardiomyocyte contractility (Rottbauer et al., 2005). To elucidate the transcriptional control of CBF, we inactivated JunB and found, similar to the situation in endothelial cells (Licht et al., 2006), that JunB acts upstream of CBF to control CBF transcription, because levels of cbf transcript are severely reduced in JunB morphants. Concordantly, knockdown of JunB also leads to heart failure in zebrafish, which can be rescued by ectopic expression of zebrafish CBF. Hence, heart failure in JunB morphants seems to be caused by reduced expression levels of CBF.
JunB is part of the activator protein-1 (AP-1) transcriptionfactor complex, which has been previously implicated in cellular stress responses such as the adaptation to ischemia and cardiac hypertrophy (Alfonso-Jaume et al., 2006;Licht et al., 2006;Ricci et al., 2005;Shaulian and Karin, 2002). Knockout of Jun, also a member of the AP-1 family of transcription factors, leads to embryonic lethality around E13 owing to abnormalities in liver hematopoiesis, in cardiac outflow-tract development and in myocardial texturing of the right ventricle (Eferl et al., 1999). By contrast, cardiac function of JunB could not be thoroughly investigated in vivo owing to an even earlier lethality of JunBknockout mice (around E8.5-E10.0) (Schorpp-Kistner et al., 1999). However, overexpression studies have suggested that JunB is sufficient to compensate for loss of Jun in the vertebrate heart and does so by driving the expression of crucial target genes of Jun (Eferl et al., 1999;Passegue et al., 2002). Our experiments highlight an important role of JunB in the vertebrate heart and skeletal muscle in vivo. JunB deficiency results in reduced CBF expression and consequently in severe heart failure and skeletal-muscle myopathy caused by rarefication of myofilaments and alterations of Z-disc structure.
Identification of novel disease genes and dissection of the molecular pathways by which mutations in these genes lead to heart failure is one of the big challenges in current research. The zebrafish model facilitates easy and rapid in vivo characterization of newly identified candidate genes by reverse genetic approaches. We used here a reverse genetic strategy in zebrafish to demonstrate for the first time an essential role of JunB-CBF signaling in the vertebrate heart and skeletal muscle. Because we found that CBF is not only expressed at sarcomeric Z-discs, the mechanical integration site of contractile units, but also in musclecell nuclei, it will be interesting to evaluate in future studies the transcriptional targets of JunB-CBF signaling in the vertebrate heart and skeletal muscle. Furthermore, careful analyses of mutations of JunB-CBF signaling components will help to elucidate their potential role in the pathogenesis of human heart and skeletal-muscle diseases.

Zebrafish strains
Care and breeding of zebrafish, Danio rerio, were as described (Rottbauer et al., 2005). The present study was performed after securing appropriate institutional approvals. It conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institute of Health (NIH Publication No. 85-23, revised 1996). For all mRNA and morpholino injection procedures, the TüAB wildtype strain was used.

Histology, transmission electron microscopy, immunostaining and immunoblotting
Histological analysis and electron micrographs of fish embryos were performed as described . For whole-mount stainings, sectioned hematoxylin and eosin stainings, and immunostainings, embryos were fixed in 4% paraformaldehyde. For immunoblotting, polyclonal rabbit anti-human CBF antibody (Abcam, MA) was used. Proteins were separated by SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membrane. Blots were probed with the aforementioned antibody as primary antibody and signals detected by chemiluminescence (monoclonal anti-rabbit-Hrp). In-vitro-translated zebrafish CBF was generated following standard protocols (TnT Kit, Promega) and probed with either rabbit-anti-CBF or, for biotinylated CBF, with hrp-streptavidine. Both antibodies detected a signal with the same molecular weight, indicating the specificity of the rabbit-anti-CBF against zebrafish CBF

RNA in situ hybridization and quantitative real-time PCR
RNA whole-mount in situ hybridization was used to detect expression of cbf and junb transcripts essentially as described (Rottbauer et al., 2001;Rottbauer et al., 2002). Quantitative real-time PCR was carried out according to standard protocols with the SYBR-Green method (Thermo Scientific) using an ABI 7000 system (ABI). cDNA was generated from 2 g total RNA using oligo(dT) primer and SuperScript II reverse transcriptase (Invitrogen). To do so, embryos were transferred to TRIzol reagent (Invitrogen) and the RNA was isolated according to the manufacturer's instructions. RNA integrity was assessed by gel electrophoresis and concentration determined by spectroscopy.

Functional assessment and statistical analysis
Still images and video films were recorded and digitized with a Zeiss microscope/MCU II. The functional assessment of cardiac contractility was carried out as described before Rottbauer et al., 2005). Fractional shortening, and atrial and ventricular diameters were measured with the help of the zebraFS software (http://www.benegfx.de). If not further specified, results are expressed as mean ± s.d. Analyses were performed using unpaired Student's t-test and a value of P<0.05 was accepted as statistically significant.