GATA5 SUMOylation is indispensable for zebrafish cardiac development

BBA - General Subjects xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

BBA - General Subjects journal homepage: www.elsevier.com/locate/bbagen

GATA5 SUMOylation is indispensable for zebrafish cardiac development Bin Wena, Hao Yuana, Xiaohui Liua, Haihong Wanga, Saijuan Chena, Zhu Chena, Hugues de Thea,b, Jun Zhoua,⁎, Jun Zhua,b,⁎ a CNRS-LIA Hematology and Cancer, Sino-French Research Center for Life Sciences and Genomics, State Key Laboratory of Medical Genomics, Rui-Jin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China b Université de Paris 7/INSERM/CNRS UMR 944/7212, Equipe Labellisée No. 11 Ligue Nationale Contre le Cancer, Hôpital St. Louis, Paris, France

A R T I C L E I N F O

A B S T R A C T

Keywords: Protein modification SUMOylation Transcription factor GATA5 Heart development ubc9

Background: SUMOylation is a critical regulatory protein modification in eukaryotic cells and plays a pivotal role in cardiac development and disease. Several cardiac transcription factors are modified by SUMO, but little is known about the impact of SUMOylation on their function during cardiac development. Methods: We used a zebrafish model to address the impact of SUMOylation on GATA5, an essential transcription factor in zebrafish cardiac development. GATA5 SUMOylation was probed by western blot, the subcellular localization and transcriptional activity of GATA5 mutants were examined by immunostaining and luciferase reporter assay. The in vivo function of GATA5 SUMOylation was evaluated by gata5 mutants mRNA microinjection and in situ hybridization in gata5 morphants and ubc9 mutants. Results: Firstly, we identified GATA5 as a SUMO substrate, and lysine 324 (K324) and lysine 360 (K360) as two major modification sites. Conversion of lysine to arginine at these two sites did not affect subcellular localization, but did affect the transcriptional activity of GATA5. Secondly, in vivo experiments demonstrated that the wild type (WT) and K324R mutant of gata5 could rescue impaired cardiac precursor differentiation, while the K360R mutant of gata5 drastically lost this potency in gata5 morphant. Furthermore, in SUMOylationdeficient ubc9 mutants, the abnormal expression pattern displayed by the early markers of cardiac development (nkx2.5 and mef2cb) could be restored using a sumo-gata5 fusion, but not with a WT gata5. Conclusion: GATA5 SUMOylation is indispensable for early zebrafish cardiac development. General significance: Our studies highlight the potential importance of transcription factor SUMOylation in cardiac development.

1. Introduction

represents a hydrophobic amino acid and X represents any residue) [1,3,4]. In vertebrates, three SUMO paralogs exist, namely; SUMO1, SUMO2 and SUMO3. SUMO2/3 show a high degree of similarity to each other and are distinct from SUMO1. Whilst the different SUMO paralogs have a common conjugation function, they also have some specificities, such as subcellular distribution, SUMO chain formation properties and substrate preferences [2–4]. For example, the promyelocytic leukemia protein (PML) can be modified by all three SUMO paralogs, whilst RanGAP1 is preferentially modified by SUMO1, and topoisomerase II is SUMOylated by SUMO2/3 during mitosis [4–7].

SUMOylation is a crucial post-translational modification, conserved from yeast through to humans [1,2]. The multistep enzymatic pathway of SUMOylation begins with the activation of SUMO itself by a heterodimer of SUMO-activating enzyme subunits 1 and 2 (SAE1/ SAE2). Activated SUMO is then transferred to the unique E2 conjugating enzyme, Ubc9, and together with the specific E3 ligase (PIAS, RanBP2 and Pc2), SUMO is ultimately attached to a specific substrate lysine residue usually located in a consensus Ψ-K-x-D/E motif (ψ

Abbreviations: WT, wild-type; UBC9, ubiquitin-conjugating enzyme E2I; SAE1/SAE2, SUMO-activating enzyme subunit 1 and 2; PML, promyelocytic leukemia protein; SENP2, sentrinspecific protease 2; SENP5, sentrin-specific protease 5; ASDs, atrial septal defects; VSDs, ventricular septal defects; Nkx2, Nk2 homeobox; MEF2, myocyte enhancer factor 2; GATA, GATA binding protein; Tbx, T-box protein; Hand, heart and neural crest derivatives expressed; CHDs, congenital heart diseases; BAF, Brg1 associated factor; K282, lysine 281; K324, lysine 324; K360, lysine 360; RT-PCR, reverse-transcription polymerase chain reaction; RT, room temperature; TALEN, transcription activator-like effector nucleases; MO, morpholino; NBT, nitro blue tetrazolium; BCIP, X-phosphate; PIAS1, protein inhibitor of activated STAT 1; hpf, hours post fertilization; ALPM, anterior lateral plate mesoderm; cmlc2, cardiac myosin light chain 2; vmhc, ventricular myosin heavy chain; OC, outer curvature; GFP, green fluorescent protein; dpf, days post fertilization; PTM, post-translational modification; ZIC3, zic family member 3; ss, somites ⁎ Corresponding authors at: CNRS-LIA Hematology and Cancer, Sino-French Research Center for Life Sciences and Genomics, Rui-Jin Hospital, Shanghai Jiao Tong University School of Medicine, 197 Rui-Jin Road II, Shanghai, China. E-mail addresses: [email protected] (J. Zhou), [email protected], [email protected] (J. Zhu). http://dx.doi.org/10.1016/j.bbagen.2017.03.005 Received 30 October 2016; Received in revised form 7 January 2017; Accepted 7 March 2017 0304-4165/ © 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Wen, B., BBA - General Subjects (2017), http://dx.doi.org/10.1016/j.bbagen.2017.03.005

BBA - General Subjects xxx (xxxx) xxx–xxx

B. Wen et al.

2. Materials and methods

Over a thousand SUMO substrates have since been identified since the discovery of SUMO in the 1990s [8]. Protein SUMOylation is involved in a wide variety of biological processes, including transcriptional regulation, DNA replication, nuclear transport, maintenance of genome integrity and signal transduction [1]. The exact roles of SUMOylation within these diverse biological processes remain an open question. Accumulating evidence has shown that SUMOylation plays essential roles in developmental processes and diseases including cardiogenesis, hematopoiesis and gametogenesis [9]. Cardiogenesis is a complex process that involves cardiac progenitor specification, differentiation and migration, followed by intricate tissue morphogenesis and remodeling [10]. Perturbation of any of these steps results in abnormal heart development and malformation. Mutations or transgenic alterations of components of the SUMOylation pathway have also been shown to lead to cardiac developmental defects. For instance, both SUMO1 knockout and cardiac-specific SUMO protease 2 (SENP2) transgenic mice, which both manifest reduced levels of SUMOylation in the heart, show atrial and ventricular septal defects (ASDs/VSDs) [11,12]. On the other hand, hyperSUMOylation in SENP2 null mutants result in accumulation of SUMOylated polycomb group protein, leading to hypocellular endocardial cushions and myocardium hypoplasia [13]. However, little is known about exactly how the SUMOylation pathway affects cardiac development. Heart development is controlled by an evolutionarily conserved network of transcription factors that connect signaling pathways to the genes for muscle growth, patterning, and contractility [14]. The core cardiac transcription factor families, such as Nkx2, MEF2, GATA, Tbx, and Hand, control myocardial gene expression, cardiac cell fate, and morphogenesis [14]. To control the complex development of the heart, the activity of transcription factors and their interactions require intricate regulation. Several cardiac transcription factors have been reported as SUMO substrates. For instance, SUMOylation of Nkx2.5 can stabilize Nkx2.5 containing complexes that boost transcriptional activity and underlie the development of human congenital heart diseases (CHDs) [15]. SUMOylation can also play an important role in controlling MEF2A and MEF2C transcriptional activity [16,17]. GATA4 SUMOylation leads to enhanced transcriptional activity and altered nuclear localization [18]. Although the Nkx2.5 SUMOylation mutant K51R was tested in a mouse model [19], the exact role of in vivo SUMOylation of transcription factors during cardiac development remains largely unknown. The GATA5 transcription factor is expressed in cardiac progenitor cells and the endocardium of both embryo and adult [20]. GATA5's role in cardiac development and congenital heart disease is well established. In humans, GATA5 mutation is relevant to CHD, including ASD and VSD [21,22]. In mice, loss of gata5 leads to bicuspid aortic valves [23]. The role of GATA5 is essentially conserved in zebrafish cardiac development. Gata5 mutants show a decrease in numbers of cardiac progenitor cells and gata5 overexpression is sufficient to produce ectopic beating tissue [24]. GATA5 cooperates with the Brg1 associated factor (BAF) chromatin remodeling complex to promote cardiac specification [25]. Recent work suggesting that GATA5 directs the efficient generation of cardiomyocytes in ESC derivatives underscores the upstream role of GATA5 in directing cardiac fate during development [26]. Herein, we show that GATA5 is a novel SUMOylated substrate. Lysine 324 (K324) and lysine (K360) are identified as major SUMOylation sites. K360 is identified as the critical GATA5 functional SUMOylation site in vivo, as a GATA5 K360R mutant was severely compromised in its ability to rescue the cardiac defects of a GATA5 zebrafish morphant. Furthermore, a sumo1-gata5 fusion could partially (but effectively) restore the normal heart phenotype, whilst WT gata5 was unable to rescue the severe cardiac defects observed in ubc9deficient zebrafish embryos. Taken together, our data highlight the significance of GATA5 SUMOylation in development of the normal heart.

2.1. Zebrafish maintenance and breeding All experimental procedures followed the rules of the Committee on Animal Care of Shanghai, China. All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of Shanghai Jiao Tong University. Zebrafish were raised and maintained at 28.5 °C in a circulating system that continuously filters, UV treats and aerates the circulating water. The embryos were collected in dishes and reared in an incubator at 28.5 °C. 2.2. Generation of constructs Gata5 and pias1 were amplified by RT-PCR with the indicated primers from the cDNA of WT zebrafish embryos (Tuebingen strain). The gata5 PCR fragment was ligated into the EcoR I and Xho I sites, and pias1 PCR fragment was ligated into the Xho I and Xba I sites of the pCS2 + vector, respectively. Sumo1 and sumo2 expression vectors have been described previously [27]. To generate the anf luciferase reporter, the 0.9-kb upstream fragment of the mouse anf gene was amplified by PCR from the genomic DNA of a WT mouse (C57BL/6 J strain), and ligated into the Kpn I and Xho I sites of the PGL3 basic vector (Promega, Maddison, WI, USA). The zebrafish gata5 serial mutants were generated using a Quick Change Site-Directed Mutagenesis Kit (Agilent, Cedar Creek, TX, USA) with the indicated primers. For the gata5 morphant rescue experiments, we generated MO-resistant gata5 mRNA by changing 10 nucleotides at the binding site of translation blocking gata5 MO to prevent direct interaction with the gata5 MO, while keeping the GATA5 amino acids unchanged. To generate sumo1-gata5 fusion, flag tagged sumo1 fragment without last 6 amino acids GGCRND was amplified from sumo1 expression vector by PCR, then ligated in-frame into the BamH I and EcoR I sites of gata5 expression vector. To generate sumo2-gata5 fusion, the sumo2cebpa vector, which have been described previously [28], was digested with EcoR I and Xho I to remove cebpa. Gata5 was then ligated into the EcoR I and Xho I sites of this vector. The primers used were as following: gata5, forward (5′-ATGTATTCGAGCCTGGCTTTATCTTC-3′) and reverse (5′-GTCTCGGATCACGCTTGAGACAG-3′); pias1, forward (5′- AATCTCGAGATGGACTACAAAGACGATGACGACAAGCACAAGATGGCGGAGAG TGC-3′) and reverse (5′-AATTCTAGAAAAGAGCCAGGAGTTCGTCA-3′); mouse anf promoter, forward (5′-AAGGTACCGGTGGGACCACC ACATATTTC-3′) and reverse (5′- CACTCGAGGGGGCACGATCTGA TGTTTG-3′); gata5 K281R, forward (5′-CAAGGCCATTAGCTATGAGAA AAGAAAGCATTCAGAC-3′) and reverse (5′-GTCTGAATGCTTTCTTTTC TCATAGCTAATGGCCTTG-3′); gata5 K324R, forward (5′- GAAAACGC CTCTACAATAAGAAGTGAACCTAGTATCG-3′) and reverse (5′- CGGACACAGGCAGTCTTCTTATTGTAGAGGCGTTTTC-3′); gata5 K360R, forward (5′-CCATGTGGACATCAGATATGAAGACTACACATACAC3′) and reverse (5′-GTGTATGTGTAGTCTTCATATCTGATGTCC ACATGG-3′); gata5 rescue, forward (5′- CCGAATTCATGTACAGCTCA CTCGCATTGTCTTCCAAC -3′) and reverse (5′-GTCTCGGATCACGCTTGAGACAG-3′). All constructs were verified by sequencing. 2.3. Cell culture and transfection Human cardiomyocyte AC16 cells and human embryonic kidney 293T cells were cultured in DMEM/F12 and DMEM medium (Life technologies, Grand Island, NY, USA) supplemented with 10% fetal calf serum (Life technologies, Grand Island, NY, USA) respectively, and subcultured every two days to maintain cells in a proliferative state. For Western blot, 1 μg plasmid was transfected into 293T cells which were 80% confluent in each well of a 6-well plate using 3 μl polyjet reagent (SignaGen Laboratories, Ljamsville, MD, USA) according to the manufacturer's instructions. Lysates were generated by homogenization in 2× laemmli buffer (Sigma-Aldrich, St. Louis, MO, USA). For the 2

BBA - General Subjects xxx (xxxx) xxx–xxx

B. Wen et al.

Fig. 1. GATA5 is modified by SUMO.A) Schematic representation of the structure of zebrafish GATA5 and alignment of the SUMOylation sites of GATA5 predicted by bioinformatics across human, mouse and zebrafish. TAD indicates transactivation domain; ZF1 and 2, zinc-finger 1 and 2; NLS, nuclear localization signal; H.s, Homo sapiens; M.s, Mus musculus; D.r, Danio rerio. Arrowheads and numbers indicate the evolutionarily conserved lysine residue within the consensus and its location in the protein, respectively.B) Western blot of GATA5 SUMOylation was performed on 293T cells transfected with blank pCS2+, gata5, or gata5 together with sumo1 or sumo2 vector. GATA5 arrow indicates intact GATA5 band, and SUMOGATA5 arrow indicates two possible SUMOylated bands. ACTIN was used as loading control.C) Western blot of GATA5 SUMOylation was performed on 293T cells transfected with different combinations of gata5, sumo1, sumo2 and pias1 vectors. GATA5 arrow indicates intact GATA5 band, and SUMO-GATA5 arrow indicates two strikingly enhanced SUMOylated bands. ACTIN was used as loading control.

luciferase reporter assay, 0.5 μg plasmid was transfected into 293T cells in each well of a 24-well plate using 1.5 μl polyjet reagent. After 48 h, cells were collected and homogenized in 1× passive lysis buffer (Promega, Maddison, WI, USA). For immunoprecipitations, 5 μg plasmid was transfected into 293T cells in 10 mm dishes using 15 μl polyjet. After 48 h, cells were collected and homogenized in weak RIPA lysis buffer (Beyotime Biotechnology, Beijing, China).

revealed with the appropriate secondary antibody at RT for 2 h. The blots were detected by SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, Rochford, IL, USA). For immunoprecipitations, cell lysates were purified by anti-Flag affinity gel (Biotool, Houston, TX, USA), followed by Western blot with GFP or FLAG antibody. The antibodies used were as follows: mouse anti-FLAG and anti-HA (Sigma-Aldrich, St. Louis, MO, USA), rabbit anti-SUMO1 (Sigma-Aldrich, St. Louis, MO, USA), rabbit anti-SUMO2/3 and antiUBC9 (Zymed laboratories, South San Francisco, CA, USA), mouse antiTUBULIN (Sigma-Aldrich, St. Louis, MO, USA), mouse anti-GFP (Roche Diagnostics, Indianapolis, IN, USA), rabbit anti-GATA5 (GeneTex, Irvine, CA,USA).

2.4. Western blot and immunoprecipitation Zebrafish embryos were harvested at the indicated time points and de-yolked as described previously [22]. Total proteins from 293T cells or zebrafish embryos were separated by 10% SDS-PAGE (SigmaAldrich, St. Louis, MO, USA) and transferred onto nitrocellulose membranes (GE Healthcare Life sciences, Pittsburgh, PA, USA). After blocking the membranes with 5% non-fat milk at room temperature (RT) for 1 h, primary antibodies in 5% non-fat milk were incubated with the membranes at 4 °C overnight. The primary antibodies were

2.5. Luciferase reporter assay For luciferase reporter assays, human cardiomyocyte AC16 cells were seeded in each well of a 24-well plate at a density of 5 × 104 for 24 h, and transfected with 100 ng anf promoter, 400 ng gata5 serial 3

BBA - General Subjects xxx (xxxx) xxx–xxx

B. Wen et al.

Fig. 2. K324 and K360 are major SUMOylation sites in GATA5.A) Western blot of GATA5 SUMOylation was performed on 293T cells transfected with WT or mutants of gata5, together with sumo1 vectors. SUMO1-GATA5 was SUMO1 fused GATA5 loaded as control of molecular weight. GATA5 arrow indicates intact GATA5 band, and SUMO-GATA5 arrow indicates possible SUMOylated bands. ACTIN was used as loading control.B) Western blot of SUMOylation of GATA5 mutants which harbor distinct double mutations was performed on 293T cells transfected with WT or mutants of gata5, together with sumo1 or sumo2 vectors. ACTIN was used as loading control.C) Immunoprecipitation was conducted on 293T cells transfected with gfp, gata5 or K324/K360R mutant vectors, together with gfp-sumo1 vector, respectively. Extracts were subjected to immunoprecipitation with anti-FLAG antibody and probed by anti-GFP or anti-FLAG antibodies. Upper panel, precipitates probed by anti-GFP antibodies; middle panel, precipitates probed by anti-FLAG antibodies; lower panel, input probed by anti-FLAG antibody.D) Western blot was performed on 293T cells transfected with gata5 or K324/K360R mutant vector, together with sumo1 and pias1 vectors as indicated. ACTIN was used as loading control.

2.6. Immunostaining

mutants, and 10 ng PRL-TK plasmids. After 48 h, cells were collected and analyzed by Dual-Luciferase Reporter Assay Kit (Promega, Maddison, WI, USA) according to the manufacturer's instructions. Data were represented as means ± se from seven independent assays, each carried out in triplicate.

293T cells transfected with gata5 or gata5 3 K mut were cultured for 48 h in dishes containing 10% FBS DMEM medium. The cells were fixed with 4% PFA for 25 min at RT and permeabilized with 0.4% Triton X100 (Sigma-Aldrich, St. Louis, MO, USA) in PBS for 10 min at RT. Cells were incubated with FLAG antibody (1:200, Sigma-Aldrich, St. Louis, 4

BBA - General Subjects xxx (xxxx) xxx–xxx

B. Wen et al.

MO, USA) in 10% goat serum at 37 °C for 1 h. After washing the cells 3 times with PBS, primary antibodies were revealed using Alexa fluor 488 goat anti-mouse antibody (1: 400, Life Technologies, Eugene, OR, USA) with 10% goat serum in PBS for 30 min at RT. The cells were observed and photographed with a fluorescence microscope (Zeiss).

values of < 0.05 were considered to indicate statistical significance. Significant differences between two groups were noted by * to ****.

2.7. TALEN construction

3.1. GATA5 can be modified by SUMO

The pair of TALENs recognizing the exon 3 of zebrafish ubc9.1 were designed using TALE-NT [29] and the TAL effector repeats were constructed using the FastTALE TALEN kit (SIDANSAI biotechnology, Shanghai, China). TALEN mRNA was synthesized by in vitro transcription using the mMESSAGE mMACHINE SP6 Transcription Kit (Life Technologies-Ambion, Austin, TX, USA). The fragments containing the TALEN target site were amplified by PCR and digested with Kpn I for genotyping. Approximately 150 pg mRNAs encoding each of the two TALEN arms were injected into the cytoplasm of one-cell stage zebrafish embryos. The genomic DNA was isolated from a mixture of ten embryos at 48 hpf, and the fragments containing the TALEN target site were amplified by PCR and digested with Kpn I to examine the efficiency of TALENs. The PCR primers for genotyping are: Forward 5′GTGAGTAACACTTCTGTTATGTTAAC-3′; Reverse 5′-CTGCAGCAGATCGTCTTGACCATGTC-3′.

Transcription factor GATA5, together with GATA4 and 6, expresses in cardiac tissue and involves in heart development [20,33]. GATA5 contains two conserved GATA-type zinc fingers and three activation domains (Fig. 1A). The SUMO-targeted lysine usually lies in a canonical sequence Ψ-K-x-D/E. Using a predictive bioinformatics approach, we retrieved three putative SUMO consensus motifs within zebrafish GATA5: MKKE (lysine 281), IKSE (lysine 324) and IKYE (lysine 360) (Fig. 1A). K281 is located within a nuclear location signal, while K324 and K360 are located in the transactivation domain. Among them, K281 and K360 are highly conserved among human, mouse and zebrafish. We set out to study whether GATA5 could be SUMOylated and to determine its possible impact on GATA5 function. Firstly, Gata5 was transfected either alone or together with sumo1 or sumo 2/3 in 293T cells, and Western blot analysis then performed to detect SUMOylation adducts. As shown in Fig. 1B, a slower migrating band was initially observed in the presence of GATA5 alone. Co-expression of sumo1or sumo2 resulted in an increased intensity of the retarded bands. Consistent with our in vitro results, we observed a putative endogenous SUMOylated GATA5 band in WT zebrafish embryos (Fig. S1). One of the SUMO E3 ligases, PIAS1, was shown to efficiently boost SUMOylation in cardiac transcriptional factors [15,18]. We co-transfected gata5, pias1, together with sumo1 or sumo2 and observed that gata5, pias1 and sumo1 coexpression resulted in more obviously retarded bands, while gata5, pias1 and sumo2 co-expression did not show such effects. The result suggests that PIAS1 might be a SUMO1 specific E3 ligase of GATA5 (Fig. 1C).

3. Results

2.8. Morpholinos and mRNA microinjection The microinjection of morpholinos and mRNAs was performed at the one-cell stage. The morpholinos (MO) specific to ubc9.2 and gata5 were described previously [30,31]. The morpholino (Gene Tools, Philomath, OR, USA) was dissolved in nuclease free water to make a stock solution and used at the following working concentration: ubc9.2 MO 0.25 mM, splice blocking gata5 MO (ssG5 MO) 1.5 mM and translation blocking gata5 MO (tbG5 MO) 1 mM, respectively. Fulllength capped zebrafish gata5 and sumo1-gata5 fusion mRNA was prepared using the mMessage mMachine SP6 kit (Life TechnologiesAmbion, Austin, TX, USA). For the gata5 morphant rescue experiment, G5 MO alone or together with 70 ng/μl WT or mutant gata5 mRNA were injected into the WT embryos. For the ubc9-deficient embryo rescue experiment, ubc9.2 MO alone or together with 40 ng/μl gata5 or sumo1-gata5 mRNA was injected into the ubc9.1 −/− embryos.

3.2. Lysine 324 and lysine 360 are the two major SUMOylation sites within zebrafish GATA5 To directly determine the SUMO targeting sites, we mutated the three lysines into arginines individually or in different combinations via mutagenesis. Both the K281/K324/K360R triple mutant (3 K mut) and the K324/K360R double mutant abrogated SUMO1 and SUMO2 modification of GATA5, whilst the K281R single mutant did not change the GATA5 SUMOylation pattern, suggesting that K324 and K360 were the major SUMO1 and SUMO2 modified sites (Fig. 2A,B). To further confirm that this band was indeed a SUMO conjugate, an immunoprecipitation assay was performed in 293T cells co-transfected with GFP tagged SUMO1 (gfp-sumo1) and FLAG tagged GATA5 (flag-gata5) expression vectors. GATA5 was precipitated using a FLAG antibody and the lysate was then analyzed by Western blot using a GFP or FLAG antibody. We observed a clear GFP-reactive and FLAG-reactive band in the FLAG-precipitated GATA5 (Fig. 2C). We also performed an immunoprecipitation experiment using GATA5 K324/360R and found that the SUMO1 modified bands disappeared (Fig. 2C). Moreover, PIAS1 lost its ability to augment SUMOylation on this double mutant (Fig. 2D). The K324R and K360R single mutants showed one narrow SUMOylation band with a distinct migration speed (Fig. 2A), suggesting that both K324 and K360 could be mono-SUMOylated. In an attempt to test the preference of these two sites for SUMO1 or SUMO2 modification, it appears that these two sites are preferentially modified more by SUMO1 than SUMO2 in vitro (Fig. 2B). Within GATA5, these results clearly demonstrate that K324 and K360 represent two major SUMO modification sites.

2.9. Whole-mount mRNA in situ hybridization Whole mount in situ hybridization was performed according to a published procedure [27]. Briefly, embryos at the various embryonic times indicated in the results were fixed in 4% paraformaldehyde overnight at 4 °C, and then dehydrated with gradient methanol. After proteinase K treatment, embryos were placed into pre-hybridization buffers and incubated for 2.5 h at 68 °C. Probes were then added into the hybridization buffer and incubated at 68 °C overnight. After washing, the embryos were incubated with phosphatase-conjugated antibody against digoxigenin (Roche Diagnostics, Indianapolis, IN, USA) at 4 °C overnight. The next day, embryos were stained with NBT (Nitro Blue Tetrazolium)-BCIP (X-phosphate) (Vector laboratories, Burlingame, CA, USA). The stained embryos were then photographed using a stereomicroscope (Nikon) equipped with a digital camera. The quantification of signal intensity in ALPM was performed using the Image J Software according to a published procedure [32]. Data were represented as means ± sd from at least two independent experiments. 2.10. Statistical analysis The values of luciferase reporter assay were represented as means ± se, and the values of quantification of whole-mount mRNA in situ hybridization were represented as means ± sd. Comparisons between groups were performed using an unpaired Student's t-test. P 5

BBA - General Subjects xxx (xxxx) xxx–xxx

B. Wen et al.

Fig. 3. Mutation of GATA5 SUMOylation sites does not alter subcellular localization, but does affect transcriptional activity.A–F) Immunostaining of 293T cells transfected with FLAG tagged GATA5 (A–C) or GATA5 3 K mutant (D–F) vectors, respectively. FLAG antibody was used to verify GATA5 subcellular localization and DAPI was used to stain the nucleus. 3 K mut indicates GATA5 mutant in which all three putative SUMOylation sites K281/K324/K360 were mutated to arginine in GATA5.G) Luciferase activities were determined in AC16 cells transfected with blank PCS2 +, gata5 or gata5 mutants as indicated, together with the anf promoter reporter vector. * and **** denote the p value < 0.05 or 0.0001 between two indicated groups, respectively.

(MO), with the resultant morphants exhibiting similar cardiac development defects as those reported within gata5 mutants. We utilized a series of GATA5 SUMOylation site mutants to test their ability to ‘rescue’ cardiac defects in the gata5 morphants. Since a major defect in the gata5 mutants was an obvious impairment of myocardial precursor differentiation [24], we examined cmlc2 and vmhc expression which marks the differentiation of pan myocardial and ventricular myocardial precursors in anterior lateral plate mesoderm (ALPM), respectively. Compared with WT embryos (Fig. 4A,F), cmlc2 and vmhc expression in gata5 morphants was remarkably reduced (Fig. 4B,G). Both WT and K324R mRNA could efficiently rescue cmlc2 and vmhc expression (Fig. 4C,D,H,I), while the K360R mutant was unable to rescue in gata5 morphants (Fig. 4E,J). Nkx2.5 represents a target gene regulated by GATA5 and is expressed in part of the ALPM during early development. Similar to cmlc2 and vmhc, the expression of nkx2.5 was markedly diminished in gata5 morphants (Fig. 4L). Again, WT and K324R mRNA could efficiently rescue nkx2.5 expression (Fig. 4M,N), but K360R failed to do so (Fig. 4O). Furthermore, we also examined the heart morphology at 48 hpf when cardiac morphogenesis is complete. Anf expression is normally confined to the outer curvature (OC) of the atrium and ventricle in WT embryos at 48 hpf, but was dramatically increased and ectopically expressed in the entire malformed heart of gata5 morphants (Fig. 4Q). Using the same constructs, similar rescue effects were

3.3. SUMO modification is dispensable for GATA5 nuclear localization, but does affect transcriptional activity Protein SUMOylation often results in altered subcellular localization and/or activity. We firstly asked whether the SUMOylation of GATA5 could influence protein nuclear localization. FLAG-tagged WT or 3 K mut GATA5 expression plasmids were transfected into 293T cells. Immunostaining showed that even the 3 K mut was still localized within the nucleus, suggesting that SUMOylation was dispensable for GATA5 nuclear localization (Fig. 3A–F). We next asked whether GATA5 SUMOylation was correlated with transcriptional activity. Luciferase reporter assays were performed in AC16 cells using the promoter of anf, a direct target gene of GATA5. Compared with control, GATA5 could potently activate the anf promoter, while the K360R mutant showed impaired transactivation ability. There was no significant difference of transactivation ability between WT and K324R (Fig. 3G). These results suggest that the two SUMOylation sites exert distinct effects on GATA5 transcriptional activity. 3.4. K360 is the major functional SUMOylation site of GATA5 in vivo To investigate the impact of SUMOylation on GATA5 in vivo, we generated a gata5-deficient zebrafish model using gata5 morpholino 6

BBA - General Subjects xxx (xxxx) xxx–xxx

B. Wen et al.

Fig. 4. K360R significantly loses its ability to rescue the cardiac developmental defects of gata5 morphants.A–E) In situ hybridization depicts expression of cmlc2 in WT (A) and gata5 morphant embryos which had been injected with gata5 or gata5 mutant mRNA (B–E). Dorsal views, anterior to the top, at 17 ss.F–J) In situ hybridization depicts expression of vmhc in WT (F) and gata5 morphant embryos which had been injected with gata5 or gata5 mutant mRNA (G–J). Dorsal views, anterior to the top, at 18 ss.K–O) In situ hybridization depicts expression of nkx2.5 in WT (K) and gata5 morphant embryos which had been injected with gata5 or gata5 mutant mRNA (L–O). Dorsal views, anterior to the top, at 12 ss.P–T) In situ hybridization depicts expression of anf in WT (P) and gata5 morphant embryos which had been injected with gata5 or gata5 mutant mRNA (Q–T). Ventral views, anterior to the top, at 48 hpf. ss indicates somites.U–V) Quantification analyses of cmlc2 (U) and vmhc (V) positive fields in ALPM. **** denotes the p value < 0.0001 between two indicated groups. NS denotes no significant difference.

Hand [14]. Some of these transcription factors are SUMO targets, thus, we sought to investigate the potential role of transcription factor SUMOylation during cardiac development. Ubc9 is the unique E2 conjugating enzyme which catalyzes isopeptide bond formation between SUMO and its substrates [1–3]. Thus, loss of UBC9 function demonstrates the impact of SUMOylation loss throughout the entire organism. Ubc9 homozygous knockout mice are embryonic lethal due to the extensive apoptosis which occurs within the blastocyst inner cell mass, hindering studies of cardiac development [34]. To circumvent this hurdle, we took advantage of the zebrafish model because zebrafish larvae have the unique capacity to oxygenate through diffusion alone for the first 7 dpf. In zebrafish, two paralogs of ubc9 exist, ubc9.1 and ubc9.2, with ubc9.1 being the major contributor to UBC9 function [30]. Thus, we generated a ubc9.1 mutant zebrafish line by using a transcription activator-like effector nucleases (TALEN) approach (Fig. S3A,B), and used ubc9.2 MO to knockdown ubc9.2 expression in

observed on anf expression (Fig. 4R–T). It has previously been reported that gata5 overexpression is sufficient to induce the expression of nkx2.5 in zebrafish, hence we asked whether mutation of the GATA5 SUMOylation sites could affect this induction. Compared with uninjected nkx2.5:GFP zebrafish embryos (Fig. S2A), WT gata5 and K324R mRNA injection induced enhanced GFP expression in ALPM (Fig. S2B,C), while K360R mRNA failed to do so (Fig. S2D). These results suggest that the SUMOylation of K360 is indispensable for GATA5's ability to regulate the differentiation of myocardial precursors. 3.5. A sumo1-gata5 fusion can partially rescue heart defects in ubc9deficient embryos Cardiac development is regulated by an evolutionarily conserved network of transcription factors including Nkx2, GATA, MEF2, Tbx, and 7

BBA - General Subjects xxx (xxxx) xxx–xxx

B. Wen et al.

Fig. 5. The expression of cardiac regulators gata4, gata5, mef2cb and nkx2.5 is impaired in the ALPM of ubc9-deficient embryos.In situ hybridization depicts expression of gata4 (A–C), gata5 (D–F), gata6 (G–I), nkx2.5 (J–L), and mef2cb (M–O) in WT (A, D, G, J, M) and ubc9-deficient (B–C, E–F, H–I, K–L, N–O) embryos. Dorsal views, anterior to the top, at 12 ss. In comparison to WT embryos (A, D, J, M), ubc9-deficient embryos display decreased expression of gata4 (B, C), gata5 (E, F), nkx2.5 (K, L) and mef2cb (N, O), but unaltered gata6 expression (H, I) in ALPM. ss indicates somites. The red bar indicates the width of gene expression in ALPM.P–Q) Quantification analyses of nkx2.5 (P) and mef2cb (Q) positive fields in ALPM. **** denotes the p value < 0.0001between two indicated groups.

ubc9.1 −/−. The resulting loss of UBC9 resulted in severe cardiac defects, including abnormal heart chamber morphology and looping due to deficient SUMOylation (Fig. S3C–H,I). During cardiac development, both myocardial and endocardial precursors arise from the ALPM, and the patterning and differentiation of ALPM are crucial for heart tube formation. In light of this, we examined GATA family (gata4, gata5, and gata6), nkx2.5 and mef2cb gene expression in ALPM. While gata4 and gata5 expression was reduced (Fig. 5A–F), gata6 expression was not evidently altered,

suggesting that ALPM specification and maintenance were not perturbed (Fig. 5G–I). The expression of nkx2.5 and mef2cb was also decreased, and their expression fields were expanded (Fig. 5J–O). It has been reported previously that GATA5 regulates the expression of nkx2.5 in zebrafish, and furthermore that the GATA family limits the boundary of the nkx2.5 expression field during embryonic development [24,35]. Our data also showed that the expression of nkx2.5 and mef2cb was decreased in gata5 morphants (Fig. S4A–D).Therefore, we tested whether GATA5 could restore the expression of nkx2.5 and mef2cb in 8

BBA - General Subjects xxx (xxxx) xxx–xxx

B. Wen et al.

Fig. 6. GATA5 cannot restore the expression pattern of cardiac regulators and normal heart morphology, while SUMO1-GATA5 partially achieves restoration in ubc9-deficient embryos.A–H) In situ hybridization depicts expression of nkx2.5 (A–D), mef2cb (E–H) in WT embryos (A, E), ubc9-deficient embryos (B, F), ubc9-deficient embryos injected with gata5 (C, G), and ubc9-deficient embryos injected with sumo1-gata5 (D, H). Dorsal views, anterior to the top, at 12 ss. Compared with WT embryos (A, E), ubc9-deficient embryos show decreased expression of nkx2.5 and mef2cb (B, F). Gata5 mRNA injection augments the expression of nkx2.5 in ALPM (C) but fails to restore the normal expression pattern observed in WT, while sumo1-gata5 restores the normal expression pattern of nkx2.5 in ALPM (D). As for mef2cb, gata5 mRNA injection is unable to restore its normal expression in ALPM (G), while sumo1-gata5 does (H).I–P) In situ hybridization of cmlc2 (I–L) and anf (M–P) was conducted to outline the heart morphology of WT embryos (I, M), ubc9-deficientembryos (J, N), ubc9-deficient embryos injected with gata5 (K, O), and ubc9-deficient injected with sumo1-gata5 (L, P). Ventral views, anterior to the top, at 48 hpf. Compared with WT embryos (I, M), ubc9-deficient embryos show overtly unlooped and shrunken hearts (J, N). Gata5 mRNA injection is unable to ameliorate the heart morphology of ubc9-deficient embryos (K, O), while sumo1-gata5 mRNA partially restores it (L, P). ss indicates somites.Q–R) Quantification analyses of nkx2.5 (Q) and mef2cb (R) positive fields in ALPM. *, *** and **** denotes the p value < 0.05, 0.001 and 0.0001 between two indicated groups, respectively.

examined the impact of WT gata5 and sumo1-gata5 on heart morphology. We observed that sumo1-gata5 mRNA injected into ubc9-deficient embryos partially restored heart looping and chamber shape (Fig. 6L,P). In contrast, we did not observe any improvement in ubc9-deficient embryos injected with gata5 mRNA (Fig. 6K,O). Jogging and looping are two key stages during heart morphogenesis, and perturbation of either of these leads to dysmorphic heart

ubc9-deficient embryos. WT gata5 mRNA injection showed minimal effects on expression pattern restoration for nkx2.5 and mef2cb (Fig. 6C,G). We then asked whether GATA5 SUMOylation was required for this process. Sumo1-gata5 fusion mRNA was injected into ubc9-deficient embryos, and indeed was able to efficiently recover the expression patterning of nkx2.5 and mef2cb in ALPM (Fig. 6D,H). We also 9

BBA - General Subjects xxx (xxxx) xxx–xxx

B. Wen et al.

necessary for the maintenance of GATA5 transcriptional activity. This is supported by the fact that K360 resides within the C terminal transactivation domain GATA5, and that K360R showed attenuated transactivation of GATA5 direct target genes (anf and nkx2.5). It appears that the reduced rescue ability of the K360R mutant in vivo experiment is more obvious than that observed in the luciferase reporter assay in vitro. Since GATA5 is able to cooperatively activate the expression of downstream myocardial genes with NKX2.5, we speculate that both the reduced activity of K360R and the decreased expression of nkx2.5 impair their synergistic gene activating ability, potentially accounting for the more severe in vivo outcome of the K360R mutant. Evidently, tight control of SUMOylation is essential for cardiac development, as both hypo- or hyper-SUMOylation lead to cardiac defects [39]. However, the molecular mechanism(s) underlying abnormal cardiac development in the context of dysfunctional SUMOylation are largely unknown. Kang et al. reported that SENP2 targets Pc2/ CBX4, a polycomb repressive complex 1 subunit, to suppress the expression of gata4 and gata6 [13]. Thus, perturbed SUMOylation of epigenetic regulators could be accountable for abnormal cardiac development. However, it cannot be excluded that SUMOylation of transcription factors is implicated in aberrant cardiac development. To test this idea, we generated ubc9-deficient embryos, representing a SUMOylation deficient model. These ubc9-deficient embryos displayed malformation of the heart chambers and atrial-ventricular canals, looping failure and severe pericardial edema. This phenotype is consistent with that reported in other SUMOylation deficient models. Furthermore, we found abnormal expression patterns of cardiac regulators including gata4, gata5, nkx2.5 and mef2cb in ALPM within ubc9-deficient embryos. Based on the fact that gata5 is required for nkx2.5 and mef2cb expression, and that SUMOylation is required for its function, we assessed the ability of gata5 and sumo1-gata5 to rescue nkx2.5 and mef2cb expression in ubc9-deficient embryos. Sumo1-gata5 fusion can partially (but effectively) restore the normal expression pattern of nkx2.5 and mef2cb and later heart morphology, whilst gata5 mRNA cannot. These results indicate that transcription factor SUMOylation is required for normal cardiac development. Congenital heart disease (CHD) is a consequence of aberrant cardiac development. Several lines of evidence support decreased levels of SUMOylation in CHD-associated mutants of cardiac transcription factors [19,40]. ZIC3, a zinc finger transcription factor is also implicated in CHD [41], and the CHD-associated ZIC3 mutant W255G has shown decreased SUMOylation and impaired cellular nuclear localization. Enhanced W255G SUMOylation has subsequently been shown to restore nuclear protein occupancy [40]. In a similar fashion to ZIC3, the CHD-associated nkx2.5 K51R mutant displays diminished SUMOylation and cardiomyocyte-specific expression of this mutant in nkx2.5 haploinsufficient mice leads to CHD [19]. Although the exact mechanism by which K51R causes CHD is unclear, this study causally links SUMOylation deficiency in a single site of NKX2.5 to CHD. As for GATA5, a number of mutations have been reported in CHD patients [22]. It would be intriguing to investigate the SUMOylation status and function of these GATA5 mutants. In summary, our studies indicate that GATA5 is a novel SUMO target, and K360 SUMOylation is indispensable for its function during zebrafish cardiac development. These findings highlight the importance of transcription factor SUMOylation in cardiac development and imply that restoring the expression of SUMO-targeted genes in a SUMOylation-deficient environment may not be sufficient for functional recovery. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.bbagen.2017.03.005.

development. To further ascertain the rescue effects using SUMO-fused GATA5 protein, we examined its effects on heart jogging and looping in ubc9-deficient embryos. By 24 hpf, the primary heart tube of WT embryos was fully elongated (Fig. S5A), but the one in ubc9.1 −/− ubc9.2 MO embryos was stunted (Fig. S5B), suggesting that defective heart tube assembly had occurred. By 36 hpf, while the heart tube of WT embryos underwent normal dextral looping and displayed an “Sshape”(Fig. S5E), that of ubc9.1 −/− ubc9.2 MO embryos showed a straight tube, and failed to form the expected “S-shape” (Fig. S5F), indicating a disrupted cardiac looping process. Interestingly, while Sumo1-gata5 mRNA can partially rescue both jogging and looping processes (Fig. S5D,H), gata5 mRNA failed to do so (Fig. S5C,G). We also tested sumo2-gata5 fusion mRNA injection and obtained the similar results. These findings strongly suggest that SUMOylation is indispensable for GATA5 to exert its full function. 4. Discussion Within the last decade, genetic and cell biological experiments have demonstrated a key role for SUMOylation in cardiac development [11–13,36], and several SUMO targeted transcription factors that modulate cardiac gene expression and cardiogenesis have been identified [15–18,37]. Nevertheless, the impact of SUMOylation on transcription factor function in cardiac development remains less well understood. Our studies herein have uncovered SUMOylation of GATA5 itself and have identified potential roles for this modification in zebrafish cardiac development. GATA5 is modified by SUMO, with K324 and K360 being the major SUMOylation sites. K360 appears to be the major functional GATA5 SUMOylation site in vivo, namely because conversion of lysine 360 to arginine abrogates GATA5's function in regulating the differentiation of cardiac precursors. Furthermore, we show that SUMOylation of GATA5 is required for the normal expression patterning of cardiac regulators nkx2.5 and mef2cb in ALPM during early cardiac development. Thus, our findings reveal that SUMOylation is indispensable for GATA5 to execute its critical function and indeed, underscore the potential importance of transcription factor SUMOylation during cardiac development. GATA5 plays an essential role in zebrafish cardiac development. Faust mutants carrying gata5 mutations manifest with cardia bifida, hypoplastic ventricles and reduced expression of several cardiac markers including nkx2.5 (24). Embryos depleted of gata5 and gata6 are heartless, indicating that gata5 and gata6 are necessary for the specification of cardiomyocytes [31]. In view of the essential role of GATA5 in cardiac development, it is therefore important to unravel the intricacies of GATA5 regulation. Previous studies have shown that gata5 expression is regulated by Bmp2b and Oep in myocardial precursors, and that Bmp2b and Oep function through gata5 to regulate nkx2.5 expression and promote myocardial differentiation [38]. Our studies have revealed another type of GATA5 regulation, post-translational modification (PTM) with SUMO. GATA5 was modified by both SUMO1 and SUMO2, and K324 and K360 were the major GATA5 SUMOylation sites. GATA4, another GATA family member, has been reported to be targeted by SUMO1 [18]. Mutation of the major GATA4 SUMOylation site impairs both its nucleocytoplasmic translocation and transcriptional activity. In contrast, combined mutation of K281, K324 and K360 did not affect the nuclear localization of GATA5. Analogous to GATA4 however, SUMOylation of GATA5 did exert effects on transactivation. K360R mutation effectively dampened the activity of GATA5. Although several cardiac transcription factors including GATA4, NKX2.5 and TBX5 were identified as SUMO targets in vitro, the in vivo functional effects of SUMOylation have not yet been interrogated. To evaluate the in vivo function of GATA5 SUMOylation, we tested the potency of GATA5 mutants in rescuing the cardiac defects of gata5 morphants. We found that both WT and K324R could restore cardiac defects, while K360R lost substantial rescue capacity in gata5 morphants. A plausible explanation is that the SUMOylation of K360 is

Author contributions B. Wen designed, performed experiments, analyzed data, and wrote 10

BBA - General Subjects xxx (xxxx) xxx–xxx

B. Wen et al.

[17] C. Riquelme, K.K. Barthel, X. Liu, SUMO-1 modification of MEF2A regulates its transcriptional activity, J. Cell. Mol. Med. 10 (2006) 132–144. [18] J. Wang, X.H. Feng, R.J. Schwartz, SUMO-1 modification activated GATA4dependent cardiogenic gene activity, J. Biol. Chem. 279 (2004) 49091–49098. [19] E.Y. Kim, L. Chen, Y. Ma, W. Yu, J. Chang, I.P. Moskowitz, J. Wang, Expression of sumoylation deficient Nkx2.5 mutant in Nkx2.5 haploinsufficient mice leads to congenital heart defects, PLoS One 6 (2011) e20803. [20] F. Charron, M. Nemer, GATA transcription factors and cardiac development, Semin. Cell Dev. Biol. 10 (1999) 85–91. [21] D. Wei, H. Bao, N. Zhou, G.F. Zheng, X.Y. Liu, Y.Q. Yang, GATA5 loss-of-function mutation responsible for the congenital ventriculoseptal defect, Pediatr. Cardiol. 34 (2013) 504–511. [22] J.Q. Jiang, R.G. Li, J. Wang, X.Y. Liu, Y.J. Xu, W.Y. Fang, X.Z. Chen, W. Zhang, X.Z. Wang, Y.Q. Yang, Prevalence and spectrum of GATA5 mutations associated with congenital heart disease, Int. J. Cardiol. 165 (2013) 570–573. [23] B. Laforest, G. Andelfinger, M. Nemer, Loss of Gata5 in mice leads to bicuspid aortic valve, J. Clin. Invest. 121 (2011) 2876–2887. [24] J.F. Reiter, J. Alexander, A. Rodaway, D. Yelon, R. Patient, N. Holder, D.Y. Stainier, Gata5 is required for the development of the heart and endoderm in zebrafish, Genes Dev. 13 (1999) 2983–2995. [25] X. Lou, A.R. Deshwar, J.G. Crump, I.C. Scott, Smarcd3b and Gata5 promote a cardiac progenitor fate in the zebrafish embryo, Development 138 (2011) 3113–3123. [26] H.K. Turbendian, M. Gordillo, S.Y. Tsai, J. Lu, G. Kang, T.C. Liu, A. Tang, S. Liu, G.I. Fishman, T. Evans, GATA factors efficiently direct cardiac fate from embryonic stem cells, Development 140 (2013) 1639–1644. [27] H. Yuan, J. Zhou, M. Deng, X. Liu, M. Le Bras, H. de The, S.J. Chen, Z. Chen, T.X. Liu, J. Zhu, Small ubiquitin-related modifier paralogs are indispensable but functionally redundant during early development of zebrafish, Cell Res. 20 (2010) 185–196. [28] H. Yuan, J. Zhou, M. Deng, Y. Zhang, Y. Chen, Y. Jin, J. Zhu, S.J. Chen, H. de The, Z. Chen, T.X. Liu, J. Zhu, Sumoylation of CCAAT/enhancer-binding protein alpha promotes the biased primitive hematopoiesis of zebrafish, Blood 117 (2011) 7014–7020. [29] E.L. Doyle, N.J. Booher, D.S. Standage, D.F. Voytas, V.P. Brendel, J.K. Vandyk, A.J. Bogdanove, TAL Effector-Nucleotide Targeter (TALE-NT) 2.0: tools for TAL effector design and target prediction, Nucleic Acids Res. 40 (2012) W117–W122. [30] M. Nowak, M. Hammerschmidt, Ubc9 regulates mitosis and cell survival during zebrafish development, Mol. Biol. Cell 17 (2006) 5324–5336. [31] A. Holtzinger, T. Evans, Gata5 and Gata6 are functionally redundant in zebrafish for specification of cardiomyocytes, Dev. Biol. 312 (2007) 613–622. [32] Y. Fan, S. Richelme, E. Avazeri, S. Audebert, F. Helmbacher, R. Dono, F. Maina, Tissue-specific gain of RTK signalling uncovers selective cell vulnerability during embryogenesis, PLoS Genet. 11 (2015) e1005533. [33] T. Peterkin, A. Gibson, M. Loose, R. Patient, The roles of GATA-4, -5 and -6 in vertebrate heart development, Semin. Cell Dev. Biol. 16 (2005) 83–94. [34] K. Nacerddine, F. Lehembre, M. Bhaumik, J. Artus, M. Cohen-Tannoudji, C. Babinet, P.P. Pandolfi, A. Dejean, The SUMO pathway is essential for nuclear integrity and chromosome segregation in mice, Dev. Cell 9 (2005) 769–779. [35] Y. Jiang, T.A. Drysdale, T. Evans, A role for GATA-4/5/6 in the regulation of Nkx2.5 expression with implications for patterning of the precardiac field, Dev. Biol. 216 (1999) 57–71. [36] E.Y. Kim, Y. Zhang, I. Beketaev, A.M. Segura, W. Yu, Y. Xi, J. Chang, J. Wang, SENP5, a SUMO isopeptidase, induces apoptosis and cardiomyopathy, J. Mol. Cell. Cardiol. 78 (2015) 154–164. [37] I. Beketaev, E.Y. Kim, Y. Zhang, W. Yu, L. Qian, J. Wang, Potentiation of Tbx5mediated transactivation by SUMO conjugation and protein inhibitor of activated STAT 1 (PIAS1), Int. J. Biochem. Cell Biol. 50 (2014) 82–92. [38] J.F. Reiter, H. Verkade, D.Y. Stainier, Bmp2b and Oep promote early myocardial differentiation through their regulation of gata5, Dev. Biol. 234 (2001) 330–338. [39] L. Mendler, T. Braun, S. Muller, The ubiquitin-like SUMO system and heart function: from development to disease, Circ. Res. 118 (2016) 132–144. [40] L. Chen, Y. Ma, L. Qian, J. Wang, Sumoylation regulates nuclear localization and function of zinc finger transcription factor ZIC3, Biochim. Biophys. Acta 1833 (2013) 2725–2733. [41] J. Cowan, M. Tariq, S.M. Ware, Genetic and functional analyses of ZIC3 variants in congenital heart disease, Hum. Mutat. 35 (2014) 66–75.

the draft; H. Yuan, X.H. Liu and H.H. Wang performed experiments and collected data of experiments; S.J. Chen, Z. Chen, and H. de The analyzed data; J. Zhu and J. Zhou designed experiments, analyzed data and wrote the paper. Transparency document The http://dx.doi.org/10.1016/j.bbagen.2017.03.005 associated with this article can be found in online version. Acknowledgements We appreciate Dr. Gavine for proofreading this manuscript. The authors would also like to thank Dr. Bo Zhang for kindly providing several WISH plasmids. We are grateful for Mei Dong, Ming Deng, Yi Chen and Yi Jin for their excellent technical support. This work was supported by grants from the National Basic Research Program of China (973 Program) (2012CB910300), the Natural Science Foundation of Shanghai City (13ZR1425800) and the Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning. References [1] S. Creton, S. Jentsch, SnapShot: the SUMO system, Cell 143 (848–848) (2010) e841. [2] E.S. Johnson, Protein modification by SUMO, Annu. Rev. Biochem. 73 (2004) 355–382. [3] A. Flotho, F. Melchior, Sumoylation: a regulatory protein modification in health and disease, Annu. Rev. Biochem. 82 (2013) 357–385. [4] A.G. van der Veen, H.L. Ploegh, Ubiquitin-like proteins, Annu. Rev. Biochem. 81 (2012) 323–357. [5] M.J. Matunis, E. Coutavas, G. Blobel, A novel ubiquitin-like modification modulates the partitioning of the Ran-GTPase-activating protein RanGAP1 between the cytosol and the nuclear pore complex, J. Cell Biol. 135 (1996) 1457–1470. [6] R. Mahajan, C. Delphin, T. Guan, L. Gerace, F. Melchior, A small ubiquitin-related polypeptide involved in targeting RanGAP1 to nuclear pore complex protein RanBP2, Cell 88 (1997) 97–107. [7] Y. Azuma, A. Arnaoutov, M. Dasso, SUMO-2/3 regulates topoisomerase II in mitosis, J. Cell Biol. 163 (2003) 477–487. [8] B. Short, SUMO enters the ring, J. Cell Biol. 211 (2015) 490. [9] H. Lomeli, M. Vazquez, Emerging roles of the SUMO pathway in development, Cell. Mol. Life Sci. 68 (2011) 4045–4064. [10] D. Staudt, D. Stainier, Uncovering the molecular and cellular mechanisms of heart development using the zebrafish, Annu. Rev. Genet. 46 (2012) 397–418. [11] J. Wang, L. Chen, S. Wen, H. Zhu, W. Yu, I.P. Moskowitz, G.M. Shaw, R.H. Finnell, R.J. Schwartz, Defective sumoylation pathway directs congenital heart disease, Birth Defects Res. A Clin. Mol. Teratol. 91 (2011) 468–476. [12] E.Y. Kim, L. Chen, Y. Ma, W. Yu, J. Chang, I.P. Moskowitz, J. Wang, Enhanced desumoylation in murine hearts by overexpressed SENP2 leads to congenital heart defects and cardiac dysfunction, J. Mol. Cell. Cardiol. 52 (2012) 638–649. [13] X. Kang, Y. Qi, Y. Zuo, Q. Wang, Y. Zou, R.J. Schwartz, J. Cheng, E.T. Yeh, SUMOspecific protease 2 is essential for suppression of polycomb group protein-mediated gene silencing during embryonic development, Mol. Cell 38 (2010) 191–201. [14] E.N. Olson, Gene regulatory networks in the evolution and development of the heart, Science 313 (2006) 1922–1927. [15] J. Wang, H. Zhang, D. Iyer, X.H. Feng, R.J. Schwartz, Regulation of cardiac specific nkx2.5 gene activity by small ubiquitin-like modifier, J. Biol. Chem. 283 (2008) 23235–23243. [16] J. Kang, C.B. Gocke, H. Yu, Phosphorylation-facilitated sumoylation of MEF2C negatively regulates its transcriptional activity, BMC Biochem. 7 (2006) 5.

11