Chromatin-remodelling factor Brg1 regulates myocardial proliferation and regeneration in zebrafish

Abstract

The zebrafish possesses a remarkable capacity of adult heart regeneration, but the underlying mechanisms are not well understood. Here we report that chromatin remodelling factor Brg1 is essential for adult heart regeneration. Brg1 mRNA and protein are induced during heart regeneration. Transgenic over-expression of dominant-negative Xenopus Brg1 inhibits the formation of BrdU⁺ /Mef2C⁺ and Tg(gata4:EGFP) cardiomyocytes, leading to severe cardiac fibrosis and compromised myocardial regeneration. RNA-seq and RNAscope analyses reveal that inhibition of Brg1 increases the expression of cyclin-dependent kinase inhibitors such as cdkn1a and cdkn1c in the myocardium after ventricular resection; and accordingly, myocardial-specific expression of dn-xBrg1 blunts myocardial proliferation and regeneration. Mechanistically, injury-induced Brg1, via its interaction with Dnmt3ab, suppresses the expression of cdkn1c by increasing the methylation level of CpG sites at the cdkn1c promoter. Taken together, our results suggest that Brg1 promotes heart regeneration by repressing cyclin-dependent kinase inhibitors partly through Dnmt3ab-dependent DNA methylation.

ARTICLE

Received 25 May 2016 | Accepted 1 Nov 2016 | Published 8 Dec 2016

Chenglu Xiao1,2,3,*, Lu Gao1,2,3,*, Yu Hou4,5, Congfei Xu6,7, Nannan Chang1,2,3, Fang Wang8, Keping Hu9,10,

Aibin He1,2, Ying Luo8, Jun Wang6,7, Jinrong Peng11, Fuchou Tang4,5, Xiaojun Zhu1,2,3 & Jing-Wei Xiong1,2,3

The zebrash possesses a remarkable capacity of adult heart regeneration, but the underlying mechanisms are not well understood. Here we report that chromatin remodelling factor Brg1 is essential for adult heart regeneration. Brg1 mRNA and protein are induced during heart regeneration. Transgenic over-expression of dominant-negative Xenopus Brg1 inhibits the formation of BrdU /Mef2C and Tg(gata4:EGFP) cardiomyocytes, leading to severe cardiac brosis and compromised myocardial regeneration. RNA-seq and RNAscope analyses reveal that inhibition of Brg1 increases the expression of cyclin-dependent kinase inhibitors such as cdkn1a and cdkn1c in the myocardium after ventricular resection; and accordingly, myocardial-specic expression of dn-xBrg1 blunts myocardial proliferation and regeneration. Mechanistically, injury-induced Brg1, via its interaction with Dnmt3ab, suppresses the expression of cdkn1c by increasing the methylation level of CpG sites at the cdkn1c promoter. Taken together, our results suggest that Brg1 promotes heart regeneration by repressing cyclin-dependent kinase inhibitors partly through Dnmt3ab-dependent DNA methylation.

1 Institute of Molecular Medicine, Peking University, Beijing 100871, China. 2 Beijing Key Laboratory of Cardiometabolic Molecular Medicine, Peking University, Beijing 100871, China. 3 State Key Laboratory of Natural and Biomimetic Drugs, Beijing 100871, China. 4 Biodynamic Optical Imaging Center, Peking University, Beijing 100871, China. 5 College of Life Sciences, Peking University, Beijing 100871, China. 6 School of Life Sciences, University of Science and Technology of China, Hefei 230026, China. 7 Hefei National Laboratory for Physical Sciences at the Microscale, University of Science and Technology of China, Hefei 230026, China. 8 Department of Biomedical Engineering, College of Engineering, Peking University, Beijing 100871, China. 9 Institute of Medicinal Plant Development, Chinese Academy of Medical Sciences, Beijing 100193, China. 10 Peking Union Medical College, Beijing 100730, China. 11 College of Animal Sciences, Zhejiang University, Hangzhou 310058, China. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to X.Z. (email: mailto:[email protected]

Web End [email protected] ) or to J.-W.X. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

DOI: 10.1038/ncomms13787 OPEN

Chromatin-remodelling factor Brg1 regulates myocardial proliferation and regenerationin zebrash

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787

The high mortality and morbidity following myocardial infarction is a public health problem worldwide. Myocar-dial infarction results in the loss of billions of cardiomyo

cytes in heart failure patients while myocardial regeneration is severely limited. Various cell-based and cell-free strategies are being explored for promoting heart regeneration in animal models and human patients13. However, the efcacy of cardiac cell-based therapy is still uncertain, with frequent occurrence of engraftment-induced arrhythmia, so the clinical implications remain unclear4. In contrast, lower vertebrates such as zebrash can perfectly regenerate the injured heart by cardiomyocyte dedifferentiation and proliferation58. Although cardiac regeneration after ventricular resection occurs in mouse neonatal heart at 1 day after birth, this regenerative capacity is lost within 7 days after birth9, suggesting that regenerative potential is gradually lost during mouse heart development and maturation. In spite of the very limited regenerative capacity, mammalian cardiomyocytes are able to divide and renew in adulthood1012. Therefore, harnessing the mechanisms underlying zebrash heart regeneration may provide insights into mammalian heart regeneration and have therapeutic applications.

ATP-dependent chromatin remodelling is involved in controlling chromatin structure that in turn regulates many physiological and pathological processes. Instead of covalently modifying DNA or histones, the SWI/SNF (SWI/sucrose non-fermentable)-like complex, a member of the family of ATP-dependent chromatinremodelling complexes, uses energy from ATP hydrolysis, and regulates gene transcription by rearranging nucleosome positions and histoneDNA interactions, and thus facilitates the transcriptional activation or repression of targeted genes13. The SWI/SNF complex contains 410 components, of which brahma-related gene 1 (BRG1, or SMARCA4) is one of the central ATPase catalytic subunits. This complex plays an important role in the development of the central nervous system, thymocytes, heart and other organs. Brg1 is essential for zygote genome activation14, erythropoiesis15, cardiac development16,17 and neuronal development18,19. Other members of the mammalian SWI/SNF complex are also required for heart morphogenesis, including Baf60c (ref. 20), Baf180 (ref. 21) and Baf250a (ref. 22). In particular, Brg1 controls cardiovascular development in a time-and tissue-specic manner. Brg1 deletion in mice results in embryonic lethality before implantation23. Endothelial and endocardial depletion of Brg1 results in embryonic death and failure of myocardial trabeculation around E10.5 in mice16. Mice with myocardial depletion of Brg1 die around E11.5 due to thin compact myocardium and the absence of the interventricular septum17. In embryos, Brg1 promotes cardiomyocyte proliferation by maintaining Bmp10 and suppressing p57kip2 (cdkn1c) expression17. Brg1 suppresses Ask1 and Cdkn1a to inhibit apoptosis and promote proliferation of neural crest cells24. Besides its effects on cardiomyocyte proliferation, Brg1 also controls a and b myosin heavy-chain switching in the embryonic and adult hearts under hypertrophic stimulations17. The function of Brg1 in heart development is evolutionarily conserved between zebrash and mammals. Mutation of brg1 in zebrash causes cardiac hypoplasia and severe arrhythmia with abnormal expression patterns of several heart-specic genes25. Besides its functions in organ development, Brg1 is also required for hair regeneration and epidermal repair. Brg1 knockdown impairs bulge cell proliferation partly through elevating the cyclin-dependent kinase inhibitor p27Kip1 (cdkn1b)26.

Although several subunits of the SWI/SNF complex are essential for cardiac development, little is known about how this complex orchestrates zebrash heart regeneration at the chromatin level. To address this important question, we set out to determine whether and how the disruption of Brg1 affects

zebrash heart regeneration. Here we nd that brg1 mRNA and protein are induced during the course of cardiac regeneration, and inhibition of Brg1 leads to severe cardiac brosis and compromised myocardial regeneration. Myocardial-specic expression of dn-xBrg1 blunts myocardial proliferation and regeneration by increasing cell-cycle-dependent inhibitors in the myocardium. Furthermore, injury-induced Brg1 interacts with Dnmt3ab to suppress the expression of cdkn1c by increasing the methylation level of CpG sites at the cdkn1c promoter. This study has gained molecular insights of Brg1 into zebrash heart regeneration and has shed light on potential intervention of this complex for promoting heart repair and regeneration in humans.

ResultsBrg1 is upregulated after ventricular apex amputation. In spite of great efforts in many laboratories, it remains challenging to induce mammalian cardiomyocytes to re-enter mitosis by either activating a single cyclin-dependent kinase or inactivating a single cyclin-dependent kinase inhibitor2729. We hypothesized that a global epigenetic change might occur during zebrash heart regeneration and so manipulating epigenetic programmes might be an efcient means of inducing mammalian cardiomyocytes to re-enter mitosis. To evaluate the functions of the SWI/SNF complex during zebrash cardiac regeneration, we performed in situ hybridization screens to identify expression patterns of the complex components after ventricular apex amputation. Interestingly, several members of this complex (brg1, baf60c and baf180) were induced during regeneration (Fig. 1 and Supplementary Fig. 1). brg1 transcripts were upregulated as early as 2 days post amputation (d.p.a.), peaked and concentrated proximal to the injury site at 3, 7 and 14 d.p.a., and became undetectable at 30 d.p.a., when regeneration was nearly complete (Fig. 1). These data support our hypothesis that the chromatinremodelling BAF complex is associated with zebrash heart regeneration.

Next, we investigated where Brg1 proteins were expressed in the cardiac cells of injured hearts using immunouorescence staining. Consistent with its mRNA expression pattern, Brg1 was almost undetectable in the mock-operated hearts but highly induced around the injured area at 3, 7 and 14 d.p.a.; it then declined from 21 to 30 d.p.a. (Fig. 2af). Co-staining with 4,6-diamidino-2-phenylindole showed that Brg1 was located in the nuclei (Fig. 2g), in accord with the fact that Brg1 is a nuclear ATPase of the SWI/SNF complex30. In addition, the specicity of this anti-human BRG1 antibody was conrmed by its recognition of zebrash/frog Brg1 proteins in brg1 morphants and dn-xBrg1-over-expressing embryos, as well as pull down of endogenous zebrash Brg1 protein by immunoprecipitation (Supplementary Figs 2a and 3b). To identify Brg1-expressing cells, we carried out co-immunostaining for Brg1 and myocardium-specic myosin heavy chain (MF20), endocardial/endothelial reporter Tg(k1:nucEGFP), macrophage/neutrophil reporter Tg (coronin1a:EGFP), epicardial reporter Tg(tcf21:DsRed) or myocardial reporter Tg(gata4:EGFP). Brg1 was detected within MF20-positive myocardial cells at 7 d.p.a. (Fig. 2h). Furthermore, Brg1 was co-localized in the injury site with Tg(k1:nucEGFP)-positive endocardium (Supplementary Fig. 2b), Tg (coronin1a:EGFP)-positive macrophages/neutrophils (Supplementary Fig. 2c), Tg(tcf21:DsRed)-positive epicardium (Supplementary Fig. 2d) and Tg(gata4:EGFP)-positive myocardium (Supplementary Fig. 2e). Additional analyses showed that about 2030% of Brg1 cells were co-localized with MF20-positive myocardium (Supplementary Fig. 2g,h) or with k1:nucEGFP-positive endocardium (Supplementary Fig. 2i,j) from 3 to 14 d.p.a.; about 2030% of Brg1 cells were co-

2 NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787 ARTICLE

a b

mock

1 d.p.a.

c

2 d.p.a.

d

3 d.p.a.

e

7 d.p.a.

f

14 d.p.a.

g

21 d.p.a.

h

30 d.p.a.

Figure 1 | brg1 is upregulated during cardiac regeneration in zebrash. In situ hybridization was performed on parafn sections of mock-operated zebrash (a) and those with amputated ventricular apexes (bh) at the indicated time points using a digoxigenin-labelled anti-sense brg1 RNA probe. Note induced expression of brg1 in the injured heart from 1 to 14 d.p.a. (bf). Dashed lines mark the resection sites; the right upper corner is high-magnication image of the framed area; similar results were conrmed by performing three independent experiments. Scale bar, 100 mm.

a b c d

MF20

Brg1

mock

3 d.p.a.

7 d.p.a. 14 d.p.a.

e f g h

DAPI

Brg1

7 d.p.a.

21 d.p.a.

30 d.p.a.

Brg1, MF20, DAPI, 7 d.p.a.

Figure 2 | Brg1 is activated in multiple types of cells during cardiac regeneration in zebrash. (af) Immunouorescence staining of Brg1 and cardiac sarcomere myosin heavy chain (MF20) was performed on parafn sections of mock-operated zebrash (a) and those with amputated ventricular apexes (bf) at the indicated time points. The right upper corners are high-magnication images of the frame area in af, showing Brg1 co-localization in MF20-positive myocytes. (g) Co-staining of Brg1 and 4,6-diamidino-2-phenylindole (DAPI) in parafn sections of amputated apexes at 7 d.p.a. The right upper corner is high-magnication image of the framed area. (h) Immunouorescence staining of Brg1 and MF20 of amputated heart at 7 d.p.a., showing the co-localization of Brg1 and MF20. The right upper corner is high-magnication image of the framed area. These data were conrmed by performing three independent experiments. Scale bars, 100 mm.

localized with tcf21:DsRed-positive epicardium from 7 to 21 d.p.a. (Supplementary Fig. 2k,l); and about 7% of Brg1 cells were co-localized with coronin1a:EGFP-positive leukocytes (Supplementary Fig. 2m,n) from 7 to 14 d.p.a. Together, our data suggested that Brg1 is induced in multiple types of cells in the heart after ventricular apex amputation.

Inhibition of Brg1 blocks heart regeneration in zebrash. Elevated expression of brg1 mRNA and protein following ventricular apex amputation suggested that brg1 might participate in regeneration. To determine the function of Brg1, we applied a dominant-negative Xenopus Brg1 (dn-xBrg1) that carries a

K770T771-to-A770A771 mutation in the ATP-binding pocket19. This mutant Brg1 protein can bind to the other components of the SWI/SNF complex but its ATPase domain is disrupted and thus plays a dominant-negative role. Brg1 is highly conserved with 84% identity in amino acids between zebrash and frog, and the ATP-binding pocket, which is mutated in the Xenopus dominant-negative Brg1, is identical between the two proteins. We generated a Tg(hsp70:dn-xbrg1) transgenic strain in which the dn-xBrg1 was driven by the zebrash heat shock promoter31. As expected, over-expression of Xenopus dn-xBrg1 inhibited Brg1 function and caused mutant heart to display stenosis shown by myocardial markers (cmlc2, vmhc, amhc and nppa) and had slight expanded expression domains of bmp4 and tbx2b while decreased

NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787

expression of notch1b in the atrioventricular canal as those in chemical-induced zebrash brg1 mutant embryos at 48 or 60 h.p.f. (ref. 25; Supplementary Fig. 3). We then performed ventricular amputation in Tg(hsp70:dn-xBrg1) zebrash and their wild-type siblings followed by heating at 37 C for 30 min daily from 5 to 30 d.p.a. Acid fuchsin orange G (AFOG) stain for extracellular matrix showed increased brosis and lack of sealing of the wound in dn-xBrg1 transgenic hearts (Fig. 3b and Supplementary Fig. 4c) compared with wild-type siblings (Fig. 3a and Supplementary Fig. 4a) at 30 d.p.a. Myocardial regeneration, visualized by MF20 immunostaining, was defective in dn-xBrg1 transgenic hearts (Fig. 3d and Supplementary Fig. 4d,e) compared with the fully regenerated myocardium in wild-type siblings (Fig. 3c and Supplementary Fig. 4b,e) at

30 d.p.a. After heat shock from 5 to 30 d.p.a., we stopped heat shock treatment and examined the hearts at 60 d.p.a. by AFOG and MF20 staining. Heat shock caused comparable lethality between dn-xBrg1 transgenic zebrash and their wild-type siblings (Supplementary Fig. 4f). We still found cardiac brosis and compromised myocardial regeneration in dn-xBrg1 transgenic hearts (Supplementary Fig. 5c,d) compared with wild-type sibling hearts (Supplementary Fig. 5a,b), suggesting that inhibiting Brg1 caused permanent defects in heart regeneration. On the other hand, conditional over-expression of wild-type brg1 had no effects on myocardial proliferation (Supplementary Fig. 6), suggesting that other members of SWI/ SNF complex are also required for this process. Taken together, these data demonstrated that inhibition of Brg1 caused a

a b

Sibling (+)HS 11/13

Tg (+) HS 13/14

AFOG

c d

MF20, 30 d.p.a.

Sibling (+) HS Tg (+) HS

15

10

5

0

1.5

1.0

0.5

0.0

e f g

BrdUMef2C 14 d.p.a.

20

BrdU+ myocytes (%)

Gata4:eGFP+ cells

***

Sibling

(+) HS Sibling

(+) HS

n=6

n=7

Tg(+) HS Tg

(+) HS

Sibling

(+) HS Tg

(+) HS

h i

Tg[gata4:EGFP];DAPI;14 d.p.a.

Sibling (+) HS

j

**

n=6

n=6

Tg (+) HS

Figure 3 | Inhibition of brg1 impairs cardiac regeneration. (ad) Representative sections from wild-type siblings (a,c) and Tg(hsp70:dn-xbrg1) (b,d) hearts at 30 d.p.a., evaluated by AFOG staining (a,b), and immunouorescence staining with anti-myosin heavy chain (MF20) (c,d). Note massive brosis (b) and compromised myocardial regeneration (d) in Tg(hsp70:dn-xBrg1) hearts (tg). Dashed lines mark the resection site. (eg) Parafn sections of 14 d.p.a. regenerating heart of wild-type sibling (e) and Tg(hsp70:dn-xBrg1) (f) hearts co-stained for BrdU (green), Mef2C (red) and 4,6-diamidino-2-phenylindole (DAPI; blue). Higher-magnication images of areas in squares are shown in the upper-right corners, and Mef2C /BrdU double-positive cardiomyocytes are indicated by arrowheads. (g) Percentages of Mef2C/BrdU cardiomyocytes in the injured area (***Po0.001; n 6 for siblings and 7 for transgenic

hearts; data are mean percentagess.e.m., paired Students t-test). (hj) Parafn sections of 14 d.p.a. wild-type Tg(gata4:EGFP) sibling (h) and Tg(hsp70:dn-xbrg1; gata4:EGFP) (i) hearts stained with anti-EGFP and DAPI. The average of uorescence intensity was calculated using Imaris software (j) (**Po0.01; n 6; data are mean percentagess.e.m.; paired Students t-test). Scale bars, 100 mm.

4 NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787 ARTICLE

regenerative defect in the zebrash heart and so Brg1 is required for cardiac regeneration.

Previous studies suggest that regenerated cardiomyocytes originate from the de-differentiation and proliferation of cardiomyocytes near injury site7,32. Thus, we evaluated whether inhibition of Brg1 had any effect on the de-differentiation of cardiomyocytes after injury. We found comparable de-differentiation of cardiomyocytes in the injury site, as measured disassembly of cardiac sarcomeres by transmission electron microscopy (Supplementary Fig. 7ad), or disrupted Z-disks labelled by cypher-EGFP fusion protein33 (Supplementary Fig. 7el) in both wild-type sibling and Tg(hsp70:dn-xbrg1) transgenic hearts at 14 d.p.a. We next compared the index of cardiomyocyte proliferation in Tg(hsp70:dn-xBrg1) and wild-type sibling hearts at 14 d.p.a. after heat shock by quantifying the percentage of BrdU and myocardial marker Mef2C double-positive nuclei near the injury area. Compared with B14%

BrdU /Mef2C cardiomyocytes in wild-type sibling hearts at 14 d.p.a. (Fig. 3e,g), we found only 2.4% proliferating cardiomyocytes in Tg(hsp70:dn-xBrg1) hearts (Fig. 3f,g), which was further conrmed by measuring PCNA /Mef2C proliferating cardiomyocytes (Supplementary Fig. 8df). In addition, both Tg(hsp70:dn-xBrg1) transgenic and wild-type sibling hearts, without heat shock treatments, had similar BrdU /Mef2C proliferating cardiomyocytes (Supplementary

Fig. 8ac), suggesting the minimal effects of heat shock on the index of proliferating cardiomyocytes. As previously reported7,34, Gata4-positive cardiomyocytes appeared near the injured area in wild-type Tg(gata4:EGFP) hearts at 14 d.p.a. (Fig. 3h). However, there were markedly fewer Gata4:EGFP-positive cardiomyocytes in Tg(hsp70:dn-xBrg1; gata4:EGFP) hearts at 14 d.p.a. (Fig. 3i,j). These data suggested that Brg1 is essential for cardiomyocyte proliferation and regeneration after ventricular apex amputation in zebrash.

The endocardium and epicardium are also known to be activated after ventricular amputation35. Immunostaining revealed that epicardial and endocardial marker Raldh2 (Supplementary Fig. 9ac) and epicardial reporter Tg(tcf21:DsRed) (Supplementary Fig. 9df) were similarly expressed in wild-type sibling and Tg(hsp70:dn-xBrg1) hearts, suggesting that the organ-wide activation of endocardium and epicardium was not affected after inhibition of Brg1. In addition, coronary vessel regeneration accompanied myocardial regeneration after ventricular resection36. By labelling coronary vessels and endocardium with the Tg(k1:EGFP) transgene, we found fewer k1 endothelium and endocardium in the injured area of Tg(hsp70:dn-xBrg1)

hearts than in wild-type sibling hearts (Supplementary Fig. 9gi), suggesting that Brg1 is also critical for the formation of newly regenerated coronary vessels since the Raldh2 endocardium is less affected during heart regeneration.

Brg1 represses cyclin-dependent kinase inhibitor genes. To decipher the molecular mechanisms underlying the regulation of myocardial proliferation by Brg1, we performed RNA-seq to compare the transcriptomes in Tg(hsp70:dn-xBrg1) and wild-type sibling hearts after heat shock daily from 5 to 14 d.p.a. We obtained B4 million 100 bp pair-end reads for each sample.

Further bioinformatics analyses revealed that 1,204 genes were upregulated while 1,092 genes were downregulated after the inhibition of Brg1 (Supplementary Data Sets 1 and 2). The genes with a twofold difference were selected for further analysis. cdkn1a was one of the upregulated genes in the Tg(hsp70:dnxBrg1) hearts (Fig. 4a). Since Brg1 inhibition had a pronounced effect on myocardial proliferation and brosis in Tg(hsp70:dnxBrg1) hearts, we focused on the genes regulating cell-cycle

progression and proliferation as well as cardiac brosis. Using reverse transcriptionPCR (RTPCR), we found that other CDK inhibitors cdkn1a, cdkn1ba, cdkn1bb, cdkn1c and cdkn1d were also increased in the dn-xBrg1 transgenic heart at 14 d.p.a. (Fig. 4b). A previous study reported that Meis1 is required for the transcriptional activation of CDK inhibitors (cdkn2a, cdkn2b and cdkn1a) in mice37. Consistently, meis1a, meis2a and meis2b were induced in the dn-xBrg1 transgenic heart (Fig. 4b). In addition, several brotic genes col1a1a, col1a2, tgfb1a, tgfb3 and vimentin were also upregulated in the dn-xBrg1 transgenic heart (Fig. 4b). Importantly, brg1 was induced while cdkn1c was reciprocally repressed in the early phases of regeneration (Fig. 4c). Taken together, our data suggested that Brg1 normally represses the expression of CDK inhibitors for priming heart regeneration in zebrash.

a c

1.51 0.5

1

1.5

0.50

cdkn1c/GAPDH brg1/GAPDH

1.5

1.0

0.5

0.0

**

***

**

nkx2.5 eif2ak2 col1a2 col1a1a cdkn1a mcl1a bmp4 tgfb1a

4

3

2

1

***

(+) HS sib2

(+) HS Tg1

(+) HS Tg2

(+) HS

sib1

0 mock 3 d.p.a. 5 d.p.a.

b

Sibling (+) HS

Tg(hsp70:dn-xBrg1) (+) HS

***

***

**

***

***

***

Fold change

6

4

2

***

*** ***

* *

*** ***

***

cdkn1a

0 cdkn1ba

cdkn1bb

cdkn1c

cdkn1d

meis1a

meis1b

meis2a

meis2b

meis3

col1a1a

col1a2

tgfb1a

tgfb2

tgfb3 vimentin

Figure 4 | Transgenic inhibition of Brg1 induces expression of cyclin-dependent kinase inhibitors. (a) Heat map of Z-score values showing genes differentially expressed between Tg(hsp70:dn-xBrg1)(tg1 and tg2) and wild-type sibling (sib1 and sib2) hearts. The FPKM (fragments per kilobase of exon per million fragments mapped) value of each gene was normalized using Z-scores. Genes were ranked by the mean Z-scores in the highest-expression group. (b) Tg(hsp70:dn-xBrg1) and wild-type sibling zebrash were heat-shocked daily from 5 to 14 d.p.a., and total RNA was isolated from their hearts at 14 d.p.a. Quantitative PCR showed that the cdkn and meis genes, as well as brotic markers (col1a1a, col1a2, tgfb1a, tgfb2, tgfb3 and vimentin) were upregulated in transgenic hearts (*Po0.05, ***Po0.001; data are mean fold changes after normalized to

GAPDH and expressed as means.e.m.; paired Students t-test).(c) Quantitative PCR showed higher expression of brg1 but lower expression of cdkn1c in wild-type hearts at 3 and 5 d.p.a. than mock hearts, suggesting a repressive role of brg1 in regulating cdkn1c. GAPDH was used to normalize the RNA level (**Po0.01, ***Po0.001; data are means.e.m.; one-way analysis of variance followed by Dunnetts multiple comparison test, mock served as control).

NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787

Since Brg1 is induced in injured cardiomyocyte (Fig. 2h) and is essential for cardiomyocyte proliferation (Fig. 3eg and Supplementary Fig. 8df), we tested whether the Brg1-cdkn axis acted in cardiomyocytes during heart regeneration. RNAscope in situ hybridization analysis revealed that both cdkn1a (Fig. 5a,b,i,j) and cdkn1c (Fig. 5e,f,m,n) were very lowly expressed in wild-type heart with or without injury, but both cdkn1a and cdkn1c was markedly induced in Tg(hsp70:dn-xBrg1) transgenic hearts (Fig. 5d,h,l,p) compared with wild-type sibling hearts (Fig. 5c,g,k,o). Interestingly, co-staining with MF20 showed that both cdkn1a and cdkn1c were enriched in the myocardium (Fig. 5l,p). Furthermore, by generating Tg(myl7:CreER;ubi: loxP-DsRed-STOP-loxP-dn-xBrg1) transgenic zebrash, tamoxifen-induced myocardial-specic inhibition of Brg1 resulted in decreased PCNA /Mef2C proliferating cardiomyocytes at 7 d.p.a. (Fig. 6ac), as well as increased cardiac brosis (Fig. 6d,e)

and compromised myocardial regeneration at 30 d.p.a. (Fig. 6f,g). Consistently, we also found that both cdkn1a and cdkn1c were upregulated in the MF20 myocardium in Tg(myl7:CreER;ubi:-loxP-DsRed-STOP-loxP-dn-xBrg1) transgenic hearts (Supplementary Fig. 10j,l,n,p) compared with Tg(ubi:loxP-DsRed-STOP-loxP-dn-xBrg1) control hearts (Supplementary Fig. 10i,k,m,o) after 4-HT induction by RNAscope. Together, these data support our hypothesis that Brg1 acts to suppress cdkn1a and cdkn1c in the myocardium to regulate heart regeneration in zebrash.

We then asked how Brg1 regulates the transcriptional activation of CDK inhibitors such as cdkn1c during regeneration. Since DNA methylation is an important mechanism for regulating gene expression, we determined the DNA methylation

pattern of the promoters of cdkn1c, meis1a and tgfb1a by performing bisulfate sequencing of 810 individual CpG sites. Bisulfate sequencing showed less methylation in these promoters of Tg(hsp70:dn-xBrg1) transgenic hearts than that in wild-type sibling hearts (Fig. 7a,b and Supplementary Fig. 11). Furthermore, methylation of the cdkn1c promoter increased in injured hearts at 3 and 5 d.p.a. while brg1 was reciprocally induced compared with mock controls (Fig. 7c,d), conrming a potential endogenous role of Brg1 in repressing the expression of cdkn1c. To investigate the mechanism underlying the effects of Brg1 inhibition on cdkn1c promoter methylation, we performed chromatin immunoprecipitation (ChIP) and quantitative ChIP assays, and found that Brg1 bound with the cdkn1c promoter region, which was demethylated after inhibition of Brg1 during cardiac regeneration (Fig. 7a,e). Consistent with Brg1-cdkn1a/1c function in the myocardium, we found that cdkn1c promoter was hypomethylated in Tg(myl7:CreER;ubi:loxP-DsRed-STOP-loxP-dn-xBrg1) transgenic hearts compared with Tg(ubi:loxP-DsRed-STOP-loxP-dn-xBrg1) control hearts after 4-HT induction by bisulte sequencing (Supplementary Fig. 12a); and that myc-tagged dn-Brg1 directly bound to the cdkn1c promoter in Tg(myl7:CreER;-ubi:loxP-DsRed-STOP-loxP-dn-xBrg1) transgenic hearts by ChIP assay (Supplementary Fig. 12d).

Methyltransferases are known to maintain the patterns of methylated cytosine residues in the mammalian genome and are the key molecules in regulating the level of DNA methylation. Seven DNA methyltransferases were annotated in the zebrash genome website (Zv9.0). By RTPCR, we found that dnmt1, dnmt3aa and dnmt3ab were expressed in adult zebrash heart,

a b c d

cdkn1a

i j k l

WT mock

WT mock

WT 3 d.p.a.

Sibling(+) HS 14 d.p.a. Tg (+) HS 14 d.p.a.

Tg (+) HS 14 d.p.a.

MF20 cdkn1a

WT mock WT 3 d.p.a.

Sibling (+) HS 14 d.p.a.

Tg (+) HS

14 d.p.a.

e f g h

cdkn1c

m n o p

MF20 cdkn1c

WT 3 d.p.a.

Sibling(+) HS 14 d.p.a.

WT mock WT 3 d.p.a.

Sibling (+) HS 14 d.p.a.

Tg (+) HS

14 d.p.a.

Figure 5 | cdkn1a and cdkn1c are induced and enriched in the myocardium of Tg(hsp70:dn-xBrg1) transgenic hearts. (ah) RNAscope in situ hybridization analysis with cdkn1a (ad) and cdkn1c (eh) probes on frozen sections of uninjured wild-type (WT) hearts (a,e), injured WT hearts at 3 d.p.a. (b,f), injured WT sibling hearts at 14 d.p.a. (c,g) and injured Tg(hsp70:dn-xBrg1) transgenic hearts at 14 d.p.a. (d,h). Note the robust induction of cdkn1a (d) and cdkn1c (h) induction in Tg(hsp70:dn-xBrg1) transgenic hearts compared with WT sibling hearts at 14 d.p.a. after heat shock. Black arrowheads indicate the cdkn1a or cdkn1c signals. The panels below ah are higher-magnication images of areas in squares of ah. (ip) Bright-eld images of cdkn1a (il) and cdkn1c (mp) expression by RNAscope merged with immunostaining signal images of MF20 on frozen sections of uninjured WT hearts (i,m), injured WT hearts at 3 d.p.a. (j,n), injured WT sibling hearts at 14 d.p.a. (k,o) and injured Tg(hsp70:dn-xbrg1) transgenic hearts at 14 d.p.a. (i,p). Higher-magnication images of squared areas of ip are shown below their respective panels. Note that both cdkn1a (i) and cdkn1c (m) are normally expressed in cardiomyocytes in uninjured WT hearts, and that they are highly induced in MF20-positive cardiomyocytes of Tg(hsp70:dn-xBrg1) transgenic hearts (l,p) compared with WT sibling hearts (k,o) at 14 d.p.a. White arrowheads show cdkn1a or cdkn1c signals in cardiomyocytes. Tg, Tg(hsp70:dn-xBrg1); ( ) HS,

heat shock; Scale bars, 100 mm.

6 NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787 ARTICLE

ubi:DsRed-dn-xBrg1

ubi:DsRed-dn-xBrg1

myl7:CreER; ubi:DsRed-dn-xBrg1

ubi:DsRed-dn-xBrg1 mvyl7:CreER; ubi:DsRed-dn-xBrg1

AFOG, 30 d.p.a., 4-HT

ubi:DsRed-dn-xBrg1 myl7:CreER; ubi:DsRed-dn-xBrg1

MF20, 30 d.p.a., 4-HT 4/4

a b c

Mef2C 7 d.p.a.

PCNA 4-HT

myl7:CreER; ubi:DsRed-dn-xBrg1

CM proliferation index (%)

25

20

15

10

5

*

0 n=7 n=4

d e

4/4

4/4

f g

4/4

Figure 6 | Myocardial-specic inhibition of Brg1 interferes heart regeneration. (ac) PCNA /Mef2C proliferating cardiomyocytes decreased in Tg(myl7:CreER; ubi:DsRed-dn-xBrg1) transgenic hearts (b) compared with control Tg(ubi:DsRed-dn-xBrg1) transgenic hearts (a) at 7 d.p.a. Statistics of cardiomyocyte proliferation index is shown (*Po0.05; data presented are means.e.m.; paired Students t-test) (c). White arrowheads point to

PCNA/Mef2C proliferating cardiomyocytes; n, the number of hearts analysed; ubi:DsRed-dn-xBrg1 stands for Tg(ubi:loxP-DsRed-STOP-loxP-dn-xBrg1); tamoxifen (4-HT) was applied at 3 days before injury. (d,e) AFOG staining revealed accumulated brin and brosis in Tg(myl7:CreER; ubi:DsRed-dn-xBrg1)

transgenic hearts (e) compared with control Tg(ubi:DsRed-dn-xBrg1) transgenic hearts (d) at 30 d.p.a. (f,g) MF20 staining showed compromised myocardial regeneration in Tg(myl7:CreER; ubi:DsRed-dn-xBrg1) transgenic hearts (g) compared with control Tg(ubi:DsRed-dn-xBrg1) transgenic hearts (f) at 30 d.p.a. 4/4, all 4 hearts analysed showed the same phenotype. Scale bars, 100 mm.

suggesting that both de novo and maintenance DNA methylation might occur during heart regeneration. By over-expressing Brg1 and Dnmt3ab in 293T cells, or H9C2 cardiac cells, we demonstrated that either Brg1 or dn-xBrg1 and Dnmt3ab form a protein complex by co-immunoprecipitation experiments (Fig. 7f,g). These data are consistent with previous studies in cancer cells38 and hypertrophic hearts39. An additional reporter system was further used to examine whether this interaction affected the cdkn1c promoter activity. Indeed, over-expression of either brg1 or dnmt3ab inhibited the cdkn1c promoter activity, which was synergistically enhanced by co-expression of brg1 and dnmt3ab in 293T cells (Fig. 7h) or P4 rat neonatal cardiomyocytes (Supplementary Fig. 12c). Furthermore, we also found that baf60c increased while dnmt3ab decreased in dnxBrg1 transgenic hearts compared with their sibling hearts at 14 d.p.a. (Supplementary Fig. 13). These data suggest a feedback effect of SWI/SNF complex on baf60c, as well as synergistic interaction of Brg1 and Dnmt3ab in regulating DNA methylation

of cdkn inhibitors on heart regeneration by over-expression of dn-xBrg1. Importantly, dnmt3ab was induced from 3 to 14 d.p.a. and peaked at 7 d.p.a. compared with that at mock hearts (Supplementary Fig. 14a), and nanoparticle-mediated dnmt3ab siRNA decreased BrdU /Mef2C proliferating cardiomyocytes (Supplementary Fig. 14bd), supporting the role of dnmt3ab during heart regeneration. We have previously reported methodology that nanoparticle-delivered siRNA efciently inhibits targeted gene expression in adult zebrash hearts40. Together, our data support the notion that Brg1 and Dnmt3ab form a protein complex in 293T cells and H9C2 cardiomyocytes, and this complex might be utilized to increase DNA methylation of cdkn1c promoter and so repressing its transcription for promoting zebrash heart regeneration.

Cdkn1a/1c mediate effects on myocardial proliferation. We then asked whether CDK inhibitors functionally act downstream

NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787

a b c d

TSS cdkn1c

Set A

mock

3 d.p.a. 5 d.p.a.

cdkn1c

Set A

TSS

Methylated Unmethylated

Sibling (+) HS

Tg (+) HS

Methylated Unmethylated

Sibling

(+) HS

72.8%

27.2%

47.2%

33.9%

66.1%

84.4%

88.3%

5 d.p.a.

mock

3 d.p.a.

Tg

(+) HS

52.8%

15.6%

11.7%

e

cdkn1c promoter

cdkn1c

100 bpladder Input brg1 IgG

Brg1 IgG

400 bp 300 bp

Fold enrichment

0.5

0.4

0.3

0.2

0.1

0.0

mock 5 d.p.a.

3 d.p.a.

f g h

Luciferase reporter assay for cdkn1c

IgG

Myc Input Marker

1.5

1.0

0.5

0.0

Brg1

*** *

*

170 kDa

*

IgG Myc Input Marker

Myc Dnmt3ab

DN-xBrg1 Flag

Myc Dnmt3ab

*

100 kDa

Brg1

180 kDa

Myc Dnmt3ab

100 kDa

IgG Myc Input Marker

170 kDa

100 kDa

Fold difference

Ad-lacZ Ad-lacZ Ad-brg1 Ad-dnmt3ab

+ +

+

+

+

+

+ +

Figure 7 | Brg1 represses cdkn1c expression by increasing the level of DNA methylation in its promoter region. (a) Methylation patterns of 10 individual CpG sites in the cdkn1c promoter of Tg(hsp70:dn-xBrg1) and wild-type sibling hearts after daily heat shock from 5 to 14 d.p.a. Upper panel, schematic of10 CpG island sites (set A) of the cdkn1c promoter region and transcription start site (TSS); lower panels, cdkn1c methylation patterns of wild-type sibling (sibling) and dn-xBrg1 transgenic (tg) hearts, with open circles for unmethylated and lled circles for methylated CpG islands. Methylated DNA sequences were obtained by bisulte sequencing. Note decreased methylation of ckkn1c promoter in dn-xBrg1 transgenic hearts (b). (c) cdkn1c promoter methylation of 10 individual CpG sites (set A) of mock, 3 d.p.a. and 5 d.p.a. wild-type hearts. The percentages of unmethylated (white) and methylated (black) DNA from a and b are shown in b and d. (e) Left panel, ChIP assays with anti-Brg1 antibody. Right panel, quantitation of Brg1 immunoprecipitated cdkn1c promoter in wild-type mock, 3 d.p.a. and 5 d.p.a. hearts. Data are presented as Brg1 enrichment relative to control IgG. The 335 bp DNA fragment within the cdkn1c promoter region ( 1,625 to 1,290 bp) was amplied from immunoprecipitated DNA of mock, 3 d.p.a. and 5 d.p.a. hearts by anti-Brg1

antibody or control IgG. (f) Immunoprecipitation by anti-Myc antibody in 293T cells over-expressing Brg1 and Myc-tagged Dnmt3ab. (g) Upper panel, immunoprecipitation of Brg1 and Myc-Dnmt3ab by Myc antibody or control IgG antibody in H9C2 cells over-expressing Brg1 and Myc-tagged Dnmt3ab. Lower panel, immunoprecipitation of dn-xBrg1-Flag and Myc-Dnmt3ab by Myc antibody or IgG antibody in H9C2 cells over-expressing dn-xBrg1-Flag and Myc-Dnmt3ab. (h) Luciferase reporter assays showed that over-expression of zebrash brg1 and dnmt3ab synergistically suppressed the transcriptionof cdkn1c in 293T cells. 293T cells were transfected/infected with the indicated adenoviral constructs and luciferase reporter constructs, and those cells were then collected and measured for luciferase activity at 24 h after transfection/infection. Equal amounts of adenovirus were used for each group. Firey luciferase activity was normalized by Renilla luciferase activity (*Po0.05, ***Po0.001; data are means.e.m.; one-way analysis of variance followed by Bonferronis multiple comparison test).

from Brg1 during heart regeneration. Using nanoparticle-mediated siRNA knockdown method40, we were able to decrease the expression of cdkn1a and cdkn1c in 2 d.p.a. hearts at 24 h after siRNA injection (Fig. 8a,b). We found comparable BrdU /

Mef2C proliferating cardiomyocytes in wild-type sibling hearts without (Fig. 8c) or with control siRNA injection (Fig. 8d), showing that siRNA injections daily from 5 to 14 d.p.a. had no effect on injury-induced cardiomyocyte proliferation (Fig. 8h).

8 NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787 ARTICLE

a b

cdkn1a/GAPDH

1.5

1.0

0.5

0.0

cdkn1c/GAPDH

1.5

1.0

0.5

0.0

*

***

Control

siRNA siRNA

Control

c

BrdU;Mef2C;14 d.p.a. sibling (+) HS

d

Sibling+control siRNA (+) HS

e

Tg (+) HS

f

Tg+cdkn1a siRNA (+) HS

Tg+cdkn1c siRNA (+) HS

g

h

BrdU+ myocytes (%)

20

15

10

5

***

NS

** **

n=8 n=8 n=8 n=8

n=4

0 Control

cdkn1c

cdkn1a

Sibling

(+) HS

Tg

(+) HS

Figure 8 | siRNA knockdown of either cdkn1a or cdkn1c partially rescues proliferating cardiomyocytes in the Tg(hsp70: dn-xBrg1) heart.(a,b) Quantitative PCR showed that nanoparticle-encapsulated siRNA efciently decreased the RNA levels of cdkn1a and cdkn1c in wild-type hearts at2 d.p.a., into which control and cdkn1a (a) or cdkn1c (b) siRNA were injected at 1 d.p.a. The RNA level was normalized to GAPDH (*Po0.05, ***Po0.001;

data presented are means.e.m.; paired Students t-test). (ch) Ventricular apex amputation was performed in wild-type siblings and Tg(hsp70:dn-xBrg1) zebrash, followed by heat shock treatment for 30 min daily from 5 to 14 d.p.a. The Mef2C /BrdU double-positive cardiomyocytes were comparable in control siRNA-injected (d) and uninjected (c) hearts. Either encapsulated cdkn1a (f) or cdkn1c (g) siRNA partially rescued the ratio of Mef2C/BrdU double-positive cardiomyocytes in Tg(hsp70:dn-xBrg1) hearts compared with those in uninjected control transgenic hearts (e). Scale bar, 100 mm.(h) Statistics of cg (**Po0.01, ***Po0.001; data are means.e.m.; one-way analysis of variance followed by Bonferronis multiple comparison test).

The number (n) of hearts analysed in each group is indicated in each bar.

Importantly, we found more BrdU /Mef2C proliferating cardiomyocytes in Tg(hsp70:dn-xBrg1) hearts with either cdkn1a (Fig. 8f) or cdkn1c siRNA (Fig. 8g) than in control dn-xBrg1 transgenic hearts (Fig. 8e,h). Another independent siRNA for either cdkn1a or cdkn1c was also able to rescue the proliferation index in Tg(hsp70:dn-xBrg1) transgenic hearts (Supplementary Fig. 15af). However, siRNA knockdown of either cdkn1a or cdkn1c had little or no effect on BrdU /

Mef2C proliferating cardiomyocytes in wild-type sibling hearts at 14 d.p.a. (Supplementary Fig. 16), consistent with very low levels of these genes in wild-type hearts after ventricle resection (Fig. 5). These data further support the notion that Brg1 promotes heart regeneration by repressing CDK inhibitors such as cdkn1a and cdkn1c.

DiscussionBrg1 plays an essential role in embryonic development such as in zygote genome activation14, erythropoiesis15 and the development of T cells41,42, heart16,17 and neurons18,19. However, its function in adult heart regeneration has not been addressed. Here we showed that Brg1 was activated during zebrash heart regeneration; either global or myocardial-specic over-expression of dn-xBrg1 interfered with myocardial proliferation and

regeneration; and mechanistically, Brg1 promoted regeneration by suppressing the CDK inhibitors cdkn1c. Although Brg1 was broadly expressed in cardiomyocytes, endothelial/endocardial cells, epicardium and inammatory cells (macrophages and neutrophils), myocardial-specic enrichment and function of Brg1 and cdkn1c suggest that the Brg1-cdkn1c axis acts in the myocardium to regulate cardiomyocyte proliferation and regeneration. Future studies are warranted to parse Brg1 function in the endocardium, epicardium and inammatory leukocytes in adult heart regeneration.

Previous studies have shown that the components of the SWI/SNF complex (also called the BAF complex) are differentially expressed in embryonic stem cells, neuronal progenitors and differentiated neurons43, suggesting the existence of cell- or tissue-specic SWI/SNF complexes. Embryonic stem cell BAF contains Brg1, BAF250a, BAF60a/b, BAF155, BAF57, BAF47, BAF53a, BAF45a/d and SS18 (ref. 13). Neuronal progenitor BAF consists of Brg1 or Brm, BAF250a/b, two BAF155 homodimers or BAF155/170 heterodimers, BAF53a, BAF45a, BAF60a/c, BAF47, BAF45a/d and SS18, which are essential for maintaining the stem cell state44. After neuronal progenitor differentiation into neurons, the BAF complex changes from BAF53a to 53b, SS18 to CREST, and BAF45a/d to 45b/c to form neuronal BAF45. It remains unclear whether there is a cardiac regenerative BAF

NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787

complex. During zebrash heart regeneration, we found that brg1, baf60c and baf180 were induced on heart sections. They had no or little expression in control hearts and were induced from 1 d.p.a. with peak expression around 714 d.p.a., and then declined from 14 to 30 d.p.a. for brg1 and baf60c while declined from 21 to 30 d.p.a. for baf180. Overall, brg1 and baf60c mRNA are more abundant while baf180 is less expressed during heart regeneration. Therefore, we propose that a similar cardiac-regenerative BAF complex might contain at least brg1, baf60c and baf180 during heart regeneration in zebrash. However, this hypothesis remains to be tested in future studies.

Previous studies by Field and colleagues have shown that either deletion of a CDK inhibitor or activation of a CDK has limited effects on promoting mammalian cardiomyocyte proliferation2,12,2729. Our data support the notion that an increase of CDK inhibitors and related meis genes occurs after inhibition of Brg1 during heart regeneration in zebrash; this was partly supported by RNA-seq analysis of Tg(hsp70:dn-xBrg1) and wild-type sibling hearts after daily heat shock from 5 to 14 d.p.a., as well as being further conrmed by quantitative RTPCR and its myocardial enrichment by RNAscope in situ hybridization analysis. During regeneration, cdkn1c was downregulated in injured hearts at 3 and 5 d.p.a., while brg1 was reciprocally upregulated, suggesting that Brg1 normally represses cdkn1c expression during this process, consistent with previous reports that cell-cycle-dependent kinase inhibitors are downstream of Brg1 in cardiac development17, mammalian neural crest cell development24, bulge stem cells during tissue regeneration26 and adult neural stem cells maintenance46, as well as Brg1 directly binds to the cdkn1c promoter as predicted by ChIP-seq analysis47. Furthermore, our data showed that Brg1 directly bound to the promoter of cdkn1c by ChIP assay, repressed the cdkn1cluciferase reporter, and siRNA knockdown of either cdkn1a or cdkn1c partially rescued the blunted myocardial proliferation with transgenic over-expression of dn-xBrg1. Taken together, we have shown, for the rst time, that Brg1 promotes adult cardiomyocyte proliferation by repressing CDK inhibitors, specically by direct repression of cdkn1c during heart regeneration.

It has been shown that Brg1 interacts directly or indirectly with other transcription factors or epigenetic components17,2426,38,48, and SWI/SNF chromatin-remodelling factors can induce changes in DNA methylation to regulate gene expression49. We then hypothesized that Brg1 might interact with other transcription and/or epigenetic factors to repress the transcription of cdkn1c. DNA methyltransferases are known to catalyse the reaction of transferring the methyl group to DNA from S-adenosyl methionine. Dnmt3a and Dnmt3b are de novo DNA methyltransferases that normally act as transcriptional repressors by DNA methylation or transcriptional co-repressors38,5056. Here we showed that inhibition of Brg1 led to a decreased level of DNA methylation in the promoters of cdkn1c, meis1a and tgfb1a, and increased the expression of cdkn1c accordingly. Importantly, the level of DNA methylation in the cdkn1c promoter increased after ventricular resection at 3 and 5 d.p.a., which is consistent with the repression of cdkn1c during normal regeneration. Indeed, RTPCR showed that zebrash dnmt3ab was induced during heart regeneration and it is required for myocardial proliferation, and co-immunoprecipitation analysis showed that it directly interacted with Brg1 in 293T cells and H9C2 cardiomyocytes, consistent with the previous report that Brg1 and Dnmt3a interact in cancer cells38 and in hypertrophic cardiomyocytes39. Our data reveal that dnmt1, dnmt3aa and dnmt3ab are expressed in adult zebrash heart, and their respective role in heart regeneration need to be addressed in the future. Together, Brg1 suppresses expression of cdkn1c and possible other CDK inhibitors, at least, partly through its

interacting with Dnmt3ab to increase the level of DNA methylation in the cdkn1c promoter, leading to an automatic regenerative capacity in the heart of adult zebrash. Therefore, conditional activation of the BAF complex and related signalling pathways might shed light on improving mammalian myocardial regeneration.

Methods

Zebrash lines. Zebrash were raised and handled according to a zebrash protocol (IMM-XiongJW-3) approved by the Institutional Animal Care and Use Committee at Peking University, which is fully accredited by AAALAC International. Tg(hsp70:dn-xBrg1), Tg(myl7:cypher-EGFP), Tg(myl7:CreER), Tg(ubi:loxPDsRed-STOP-loxP-Brg1) and Tg(ubi:loxP-DsRed-STOP-loxP-dn-xBrg1) zebrash lines were generated by using Tol2-based transgenesis57. The dominant-negative Xenopus Brg1 (dn-xBrg1) plasmid clone was kindly provided by Dr Kristen L. Kroll (Washington University at St Louis)19,58, and the Tg(myl7:CreER) and Tg(ubi:loxP-DsRed-STOP-loxP-EGFP) plasmid clones were kindly providedby C Geoffrey Burns (Massachusetts General Hospital, Boston, MA, USA)59. Tg(coronin1a:EGFP)60 and Tg(k1:nucEGFP)61 lines were provided by Dr Zilong Wen (Hong Kong University of Science and Technology, Hong Kong, China) and Dr Feng Liu (Institute of Zoology, Chinese Academy of Sciences, Beijing, China); Tg(gata4:EGFP) line62 was provided by Dr Todd Evans (Weill Cornell Medical College, New York, USA); and Tg(tcf21:DsRed) line63 was provided by C. Geoffrey Burns (Massachusetts General Hospital). Heterozygous transgenic zebrash and their wild-type siblings were used for all experiments.

For heat shock experiments, we crossed heterozygous Tg(hsp70:dn-xBrg1) with wild-type TL zebrash, and so expected to have 50% heterozygous transgenic sh and 50% wild-type siblings. Heterozygous Tg(hsp70:dn-xBrg1) transgenic and wild-type sibling adult zebrash received a daily heat shock in 37 C water for30 min from 5 to either 14 or 30 d.p.a. Each cycle of heat shock was carried out by transferring zebrash to system water at 31 C, accompanying with gradually increasing temperature from 31 to 37 C for about 10 min and then remaining at 37 C for another 20 min. To induce the Cre recombination in adult zebrash, Tg(myl7:CreER;ubi:loxP-DsRed-STOP-loxP-dn-xBrg1) or control Tg(ubi:loxPDsRed-STOP-loxP-dn-xBrg1) transgenic zebrash were bathed in 5 mM tamoxifen (Sigma, St Louis, MO) for 24 h, which was made from a 10 mM stock solution dissolved in 100% ethanol at room temperature. Zebrash were treated with tamoxifen at a density of 34 per 150 ml of water, and then returned to circulating zebrash system water. Ventricular resections were performed at 3 days after tamoxifen treatment. Zebrash were conrmed for their genotyping and randomly picked for all experiments.

Adult zebrash heart resection. The ventricular resection was performed according to a well-established procedure5,34. Briey, adult zebrash were anaesthetized with tricaine and the pericardial sac was exposed by removing surface scales and a small piece of skin. The apex of the ventricle was gently pulled up and removed with Vannas scissors. The zebrash was then placed back into a water tank, and water was puffed over the gills with a plastic pipette until it breathed and swam regularly. The surface opening sealed automatically within a few days. In particular, we crossed heterozygous Tg(hsp70:dn-xBrg1) with wild-type TL zebrash, and so expected to have 50% heterozygous transgenic sh and 50% wild-type siblings for performing ventricular resections.

Construction and sequencing of high-throughput RNA-seq libraries. Total RNA was isolated using an RNeasy Mini kit (QIAGEN). After conrming the quality and integrity of RNA on agarose gels, we used B1 mg of total RNA to construct the RNA-seq libraries by applying a TruSeq RNA Sample Prep kit (Illumina, San Diego, CA). We carried out the RNA-seq using Illumina HiSeq 2500 to generate B4 million 100 bp pair-end reads for each sample. Low-quality reads and sequencing adapters were removed from the raw sequencing data, and the clean reads were mapped onto the zebrash transcriptome (danRer7) using Tophat64. The expression level of each gene was calculated using Cufinks, and the genes differentially expressed between samples were calculated using Cuffdiff64. Genes with 42-fold difference between the two groups were selected for further analyses.

siRNA delivery into adult zebrash heart. siRNAs were encapsulated in polyethylene glycolpolylactic acid nanoparticles using a double emulsion-solvent evaporation technique and then injected into the pericardial sac40,65,66. Briey, zebrash were allowed to recover for 1 day after ventricular resection. To evaluate the effect of siRNA on its target gene expression, the hearts were collected at2 d.p.a., and total RNA was isolated to assess the expression of the respective genes by quantitative RTPCR. To evaluate the effect of genes on cardiomyocyte proliferation, 50 ml polyethylene glycolpolylactic acid nanoparticle-encapsulated siRNAs was injected rst, and B1 h later, 50 ml 2.5 mg ml 1 BrdU (B5002; Sigma)

was injected into the thoracic cavity daily from 7 to 14 d.p.a. The hearts at 14 d.p.a.

10 NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787 ARTICLE

were collected for subsequent experiments. siRNA sequences for cdkn1a, cdkn1c and dnmt3ab are shown in Supplementary Table 2.

mRNA or protein detection assays and AFOG staining. In situ hybridization and AFOG staining were performed on parafn sections34. Adult zebrash hearts were xed in 4% paraformaldehyde at room temperature for 2 h, dehydrated and then embedded in parafn and sectioned at 5 mm. Zebrash brg1 cDNA was cloned from an embryonic cDNA library, of which primer sequences are shown as brg1-F and brg1-R in Supplementary Table 1. Digoxigenin-labelled brg1 probes were synthesized using T7 RNA polymerase (Roche).

For immunouorescence staining, adult zebrash hearts were xed in 4% paraformaldehyde at room temperature for 2 h, dehydrated and then embedded in parafn and sectioned at 5 mm. The sections were dewaxed in xylene, rehydrated with a series of ethanol and then washed in PBS. To repair the antigen, the citric acid buffer (CW0128S; CWBIO) and the microwave treatment were used. After washing in water and PBS, the sections were blocked in 10% FBS in PBT (1% tween 20 in PBS), and then incubated with primary antibodies (1:50 diluted in PBT containing 10% FBS) overnight at 4 C. The primary antibodies used for immunouorescence were anti-BrdU (B8434; Sigma), anti-Mef2c (sc-313; Santa Cruz), anti-GFP (A-11122; Invitrogen), anti-PCNA (18-0110; Invitrogen), anti-GFP (BE2001; EASYBIO), anti-RFP (BE2023; EASYBIO), anti-myosin heavy-chain monoclonal antibody (hybridoma product MF20; Developmental Studies Hybridoma Bank, Iowa City, IA) and The Brg1 JI antibody, which was raised against a glutathione S-transferaseBRG1 fusion protein (human BRG1 amino acids 1,0861,307)30,67. The primary antibodies were then washed and sections were incubated with secondary antibodies for 2 h at room temperature. Secondary antibodies (1:100 diluted in PBT) were Alexa Fluor 488 goat anti-mouse IgG (A21121; Invitrogen), Alexa Fluor 488 goat anti-rabbit IgG (A11034; Invitrogen), Alexa Fluor 555 goat anti-mouse IgG (A21424; Invitrogen) and Alexa Fluor 555 goat anti-rabbit IgG (A21428; Invitrogen).

RNAscope (Advanced Cell Diagnostics, Hayward, CA) was performed on10 mm sections from freshly frozen hearts embedded in O.C.T. Compound (Embedding Medium for Frozen Tissue Specimens to ensure Optimal Cutting Temperature; SAKURA; 4583). Tissues were xed in pre-chilled 10% neutral buffered formalin, followed by dehydration, then treated with Pretreat 1 for 10 min at room temperature. After Pretreat 1, slides were washed with water and incubated for 30 min at room temperature with Pretreat 4. Following Preteat 4, the RNAscope2.0 HD Detection Kit Brown was applied for visualizing hybridization signals. Three injured and mock hearts were used for each RNAscope experiment. Immunostaining was performed with primary antibodies (1:50 diluted in PBT containing 10% FBS) incubated overnight at 4 C.

RNA in situ hybridization, RNAscope in situ hybridization and AFOG staining were analysed and documented under a uorescence microscope (DM5000B; Leica, Germany). Immunouorescence images were captured on a confocal microscope (LSM510; Carl Zeiss, Germany). A Zeiss 700 confocal microscope was usedfor RNAscope with Immunostaining images. The BrdU /Mef2C , PCNA /

Mef2C , Brg1 , MF20 /Brg1 , Flk1 /Brg1 and Coronin1a /Brg1 were counted manually. Fluorescence intensity was quantitated using MBF Image J.

RTPCR analysis. Total RNA was isolated and puried using an RNeasy mini kit (74106; Qiagen). About 1 mg RNA was used for reverse transcription with a Prime Script RT Reagent kit (RR037A; TakaRa), and quantitative RTPCR was performed using a SYBR Premix DimerEraser kit (RR091A; TakaRa). Primer sequences are listed in Supplementary Table 1.

Chromatin immunoprecipitation. Chromatin was isolated from zebrash hearts using Chromatin Prep Module (Catalogue# 26158; Thermo Scientic Pierce). ChIP assays were performed using Agarose ChIP Kit (Catalogue# 26156; Thermo Scientic Pierce)34. Chromatin was immunoprecipitated using anti-Brg1 (J1) antibody30, and Brg1-bound sequences were amplied with the respective gene primers. The primer sequences are listed in Supplementary Table 1. The original uncropped images of gels are shown in Supplementary Fig. 17.

DNA methylation and immunoprecipitation. Zebrash genomic DNA was extracted, puried and resuspended in 1 TE buffer. For each sample, a total

of 500 ng genomic DNA was treated using MethylCode bisulte conversion kit (MECOV-50; Invitrogen). The CpG islands of the cdkn1c promoter region were primarily located between 1,586 and 1,354 bp. The bisulte-treated DNA was

subjected to PCR to amplify the cdkn1c promoter region. The primers were designed according to the website (http://www.urogene.org/methprimer/

Web End =http://www.urogene.org/methprimer/) and are listed in Supplementary Table 1. Alterations of promoter methylation were conrmed by Sanger sequencing. In addition, zebrash dnmt3ab was isolated from an embryonic cDNA library and then subcloned into the pCDNA3.1 vector. Myc-tagged dnmt3ab and brg1 were co-transfected into 293T cells (CRL-1573, American Type Culture Collection; ATCC), which were then collected for immunoprecipitation after 24 h (ref. 34). For immunoprecipitation, H9C2 cells (CRL-1446, ATCC) were infected with either Ad-brg1/Ad-Myc-dnmt3ab or Ad-dn-xbrg1-Flag/Ad-Myc dnmt3ab, and infected cells were then collected for immunoprecipitation after 24 h. The recombinant adenovirus was constructed

and amplied by SinoGenoMax Co., Ltd, Beijing, China. The antibody for immunoprecipitation were anti-Myc (2276S; Cell Signaling Technology), anti-Flag (AP1013a, ABGENT) and anti-Brg1 (J1) as described previously30. The original uncropped images of blots are shown in Supplementary Fig. 17.

Luciferase assays. The cdkn1c promoter (from 1,624 to 1192 bp) was cloned

into the chromatinized pREP4 vector to form pREP4-cdkn1c-Luc. 293T cells (CRL-1573, ATCC) or primary cultured cardiomyocytes from neonatal P4 rats were transfected with pREP4-cdkn1c-Luc (495 ng) and pREP4-renilla (5 ng) by Lipofectamine 3000 (Invitrogen). At 20 h after transfection, the 293T cells were co-infected with Ad-lacZ, Ad-brg1, Ad-dnmt3ab or Ad-brg1/Ad-dnmt3ab, respectively. Luciferase assays were carried out 48 h after adenovirus infection68. Firey luciferase activity was normalized by Renilla luciferase activity.

Transmission electron microscopy. Tg(hsp70:dn-xBrg1) and wild-type sibling hearts were collected at 14 d.p.a. and xed in 2% glutaraldehyde, 2% paraformaldehyde and 0.1 M PBS overnight at 4 C. Subsequent embedding, ultra-thin section preparation and staining were performed by the Electron Microscopy Core Facility of Peking University. A TecnaiT20 (LaB6, 200KV) transmission electron microscope (FEI, Hillsboro, OR, USA) was used to image stained sections33. Three transgenic and wild-type sibling hearts were used for transmission electron microscopy.

Statistical analysis. All statistics were calculated using Prism 5 Graphpad Software. The statistical signicance between two groups was determined using paired Students t-test, with tow-tailed P value, and the data were reported as means.e.m. Among three or more groups, one-way analysis of variance followed by Bonferronis multiple comparison test or Dunnetts multiple comparisontest was used for comparisons.

Data availability. Data that support the ndings of this study have been deposited in Gene Expression Omnibus with the accession code GSE81627. All other relevant data are available from the corresponding authors on reasonable request.

References

1. Sahara, M., Santoro, F. & Chien, K. R. Programming and reprogramming a human heart cell. EMBO J. 34, 710738 (2015).

2. Lin, Z. & Pu, W. T. Strategies for cardiac regeneration and repair. Sci. Transl. Med. 6, 239rv231 (2014).

3. Garbern, J. C. & Lee, R. T. Cardiac stem cell therapy and the promise of heart regeneration. Cell Stem Cell 12, 689698 (2013).

4. Chong, J. J. et al. Human embryonic-stem-cell-derived cardiomyocytes regenerate non-human primate hearts. Nature 510, 273277 (2014).

5. Poss, K. D., Wilson, L. G. & Keating, M. T. Heart regeneration in zebrash. Science 298, 21882190 (2002).

6. Jopling, C., Boue, S. & Izpisua Belmonte, J. C. Dedifferentiation, transdifferentiation and reprogramming: three routes to regeneration. Nat. Rev. Mol. Cell Biol. 12, 7989 (2011).

7. Kikuchi, K. et al. Primary contribution to zebrash heart regeneration by Gata4 cardiomyocytes. Nature 464, 601605 (2010).

8. Gupta, V. & Poss, K. D. Clonally dominant cardiomyocytes direct heart morphogenesis. Nature 484, 479484 (2012).

9. Porrello, E. R. et al. Transient regenerative potential of the neonatal mouse heart. Science 331, 10781080 (2011).

10. Bergmann, O. et al. Evidence for cardiomyocyte renewal in humans. Science 324, 98102 (2009).

11. Mollova, M. et al. Cardiomyocyte proliferation contributes to heart growth in young humans. Proc. Natl Acad. Sci. USA 110, 14461451 (2013).

12. Senyo, S. E., Lee, R. T. & Kuhn, B. Cardiac regeneration based on mechanisms of cardiomyocyte proliferation and differentiation. Stem Cell Res. 13, 532541 (2014).

13. Ho, L. & Crabtree, G. R. Chromatin remodelling during development. Nature 463, 474484 (2010).

14. Bultman, S. J. et al. Maternal BRG1 regulates zygotic genome activation in the mouse. Genes Dev. 20, 17441754 (2006).

15. Bultman, S. J., Gebuhr, T. C. & Magnuson, T. A Brg1 mutation that uncouples ATPase activity from chromatin remodeling reveals an essential role for SWI/ SNF-related complexes in beta-globin expression and erythroid development. Genes Dev. 19, 28492861 (2005).

16. Stankunas, K. et al. Endocardial Brg1 represses ADAMTS1 to maintain the microenvironment for myocardial morphogenesis. Dev. Cell 14, 298311 (2008).

17. Hang, C. T. et al. Chromatin regulation by Brg1 underlies heart muscle development and disease. Nature 466, 6267 (2010).

18. Eroglu, B., Wang, G., Tu, N., Sun, X. & Mivechi, N. F. Critical role of Brg1 member of the SWI/SNF chromatin remodeling complex during neurogenesis and neural crest induction in zebrash. Dev. Dyn. 235, 27222735 (2006).

NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 11

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787

19. Seo, S., Richardson, G. A. & Kroll, K. L. The SWI/SNF chromatin remodeling protein Brg1 is required for vertebrate neurogenesis and mediates transactivation of Ngn and NeuroD. Development 132, 105115 (2005).

20. Lickert, H. et al. Baf60c is essential for function of BAF chromatin remodelling complexes in heart development. Nature 432, 107112 (2004).

21. Wang, Z. et al. Polybromo protein BAF180 functions in mammalian cardiac chamber maturation. Genes Dev. 18, 31063116 (2004).

22. Lei, I., Gao, X., Sham, M. H. & Wang, Z. SWI/SNF protein component BAF250a regulates cardiac progenitor cell differentiation by modulating chromatin accessibility during second heart eld development. J. Biol. Chem. 287, 2425524262 (2012).

23. Bultman, S. et al. A Brg1 null mutation in the mouse reveals functional differences among mammalian SWI/SNF complexes. Mol. Cell 6, 12871295 (2000).

24. Li, W. et al. Brg1 governs distinct pathways to direct multiple aspects of mammalian neural crest cell development. Proc. Natl Acad. Sci. USA 110, 17381743 (2013).

25. Takeuchi, J. K. et al. Chromatin remodelling complex dosage modulates transcription factor function in heart development. Nat. Commun. 2, 187 (2011).

26. Xiong, Y. et al. Brg1 governs a positive feedback circuit in the hair follicle for tissue regeneration and repair. Dev. Cell 25, 169181 (2013).

27. Rubart, M. & Field, L. J. Cardiac regeneration: repopulating the heart. Annu. Rev. Physiol. 68, 2949 (2006).

28. Rubart, M. & Field, L. J. Cardiac repair by embryonic stem-derived cells. Handb. Exp. Pharmacol. 174, 73100 (2006).

29. Rubart, M. & Field, L. J. Cell-based approaches for cardiac repair. Ann. N Y Acad. Sci. 1080, 3448 (2006).

30. Wang, W et al. Purication and biochemical heterogeneity of the mammalian SWI-SNF complex. EMBO J. 15, 53705382 (1996).

31. Halloran, M. C. et al. Laser-induced gene expression in specic cells of transgenic zebrash. Development 127, 19531960 (2000).

32. Jopling, C. et al. Zebrash heart regeneration occurs by cardiomyocyte dedifferentiation and proliferation. Nature 464, 606609 (2010).33. Wu, Q. et al. Talin1 is required for cardiac Z-disk stabilization and endothelial integrity in zebrash. FASEB J. 29, 49895005 (2015).

34. Bu, Y. et al. Protein tyrosine phosphatase PTPN9 regulates erythroid cell development through STAT3 dephosphorylation in zebrash. J. Cell Sci. 127, 27612770 (2014).

35. Kikuchi, K. et al. Retinoic acid production by endocardium and epicardium is an injury response essential for zebrash heart regeneration. Dev. Cell 20, 397404 (2011).

36. Lepilina, A. et al. A dynamic epicardial injury response supports progenitor cell activity during zebrash heart regeneration. Cell 127, 607619 (2006).

37. Mahmoud, A. I. et al. Meis1 regulates postnatal cardiomyocyte cell cycle arrest. Nature 497, 249253 (2013).

38. Datta, J. Physical and functional interaction of DNA methyltransferase 3A with Mbd3 and Brg1 in mouse lymphosarcoma cells. Cancer Res. 65, 1089110900 (2005).

39. Han, P. et al. Epigenetic response to environmental stress: assembly of BRG1-G9a/GLP-DNMT3 repressive chromatin complex on Myh6 promoter in pathologically stressed hearts. Biochim. Biophys. Acta. 1863, 17721781 2016.

40. Diao, J. et al. PEG-PLA nanoparticles facilitate siRNA knockdown in adult zebrash heart. Dev. Biol. 406, 196202 (2015).

41. Chi, T. H. et al. Reciprocal regulation of CD4/CD8 expression by SWI/SNF-like BAF complexes. Nature 418, 195199 (2002).

42. Chi, Y., Senyuk, V., Chakraborty, S. & Nucifora, G. EVI1 promotes cell proliferation by interacting with BRG1 and blocking the repression of BRG1 on E2F1 activity. J. Biol. Chem. 278, 4980649811 (2003).

43. Son, E. Y. & Crabtree, G. R. The role of BAF (mSWI/SNF) complexes in mammalian neural development. Am. J. Med. Genet. C Semin. Med. Genet. 166C, 333349 (2014).

44. Lessard, J. et al. An essential switch in subunit composition of a chromatin remodeling complex during neural development. Neuron 55, 201215 (2007).

45. Yoo, A. S., Staahl, B. T., Chen, L. & Crabtree, G. R. MicroRNA-mediated switching of chromatin-remodelling complexes in neural development. Nature 460, 642646 (2009).

46. Petrik, D. et al. Chromatin remodeling factor Brg1 supports the early maintenance and late responsiveness of nestin-lineage adult neural stem and progenitor cells. Stem Cells 33, 36553665 (2015).

47. Attanasio, C. et al. Tissue-specic SMARCA4 binding at active and repressed regulatory elements during embryogenesis. Genome Res. 24, 920929 (2014).

48. Nagl, Jr N. G., Wang, X., Patsialou, A., Van Scoy, M. & Moran, E. Distinct mammalian SWI/SNF chromatin remodeling complexes with opposing roles in cell-cycle control. EMBO J. 26, 752763 (2007).

49. Banine, F. et al. SWI/SNF chromatin-remodeling factors induce changes in DNA methylation to promote transcriptional activation. Cancer Res. 65, 35423547 (2005).

50. Okano, M., Bell, D. W., Haber, D. A. & Li, E. DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99, 247257 (1999).

51. Rhee, I. et al. DNMT1 and DNMT3b cooperate to silence genes in human cancer cells. Nature 416, 552556 (2002).

52. Aasland, R., Gibson, T. J. & Stewart, A. F. The PHD nger: implications for chromatin-mediated transcriptional regulation. Trends Biochem. Sci. 20, 5659 (1995).

53. Aasland, R. & Stewart, A. F. The chromo shadow domain, a second chromo domain in heterochromatin-binding protein 1, HP1. Nucleic acids Res. 23, 31683173 (1995).

54. Fuks, F., Burgers, W. A., Brehm, A., Hughes-Davies, L. & Kouzarides, T. DNA methyltransferase Dnmt1 associates with histone deacetylase activity. Nat. Genet. 24, 8891 (2000).

55. Robertson, K. D. et al. DNMT1 forms a complex with Rb, E2F1 and HDAC1 and represses transcription from E2F-responsive promoters. Nat. Genet. 25, 338342 (2000).

56. Bachman, K. E., Rountree, M. R. & Baylin, S. B. Dnmt3a and Dnmt3b are transcriptional repressors that exhibit unique localization properties to heterochromatin. J. Biol. Chem. 276, 3228232287 (2001).

57. Kawakami, K. et al. A transposon-mediated gene trap approach identies developmentally regulated genes in zebrash. Dev. Cell 7, 133144 (2004).

58. Ochi, H., Hans, S. & Westereld, M. Smarcd3 regulates the timing of zebrash myogenesis onset. J. Biol. Chem. 283, 35293536 (2008).

59. Mosimann, C. et al. Ubiquitous transgene expression and Cre-based recombination driven by the ubiquitin promoter in zebrash. Development 138, 169177 (2011).

60. Li, Y. J. & Hu, B. Establishment of multi-site infection model in zebrash larvae for studying Staphylococcus aureus infectious disease. J. Genet. Genomics 39, 521534 (2012).

61. Roman, B. L. et al. Disruption of acvrl1 increases endothelial cell number in zebrash cranial vessels. Development 129, 30093019 (2002).

62. Heicklen-Klein, A., McReynolds, L. J. & Evans, T. Using the zebrash model to study GATA transcription factors. Semin. Cell Dev. Biol. 16, 95106 (2005).

63. Kikuchi, K. et al. tcf21 epicardial cells adopt non-myocardial fates during zebrash heart development and regeneration. Development 138, 28952902 (2011).

64. Trapnell, C., Pachter, L. & Salzberg, S. L. TopHat: discovering splice junctions with RNA-Seq. Bioinformatics 25, 11051111 (2009).

65. Yang, X. Z. et al. Systemic delivery of siRNA with cationic lipid assisted PEG-PLA nanoparticles for cancer therapy. J. Controlled Release 156, 203211 (2011).

66. Liu, J. et al. Functionalized dendrimer-based delivery of angiotensin type 1 receptor siRNA for preserving cardiac function following infarction. Biomaterials 34, 37293736 (2013).

67. Khavari, P. A., Peterson, C. L., Tamkun, J. W., Mendel, D. B. & Crabtree, G. R. BRG1 contains a conserved domain of the SWI2/SNF2 family necessary for normal mitotic growth and transcription. Nature 366, 170174 (1993).

68. Zhang, J. et al. Rad GTPase inhibits cardiac brosis through connective tissue growth factor. Cardiovasc. Res. 91, 9098 (2011).

Acknowledgements

We thank Dr Iain C Bruce (Zhejiang University) for reading the manuscript, the Pathology Core of Institute of Molecular Medicine for histology, Dr Weidong Wang (National Institute on Aging/NIH) for providing Brg1 anti-body and members ofDr J-W.X. laboratory for helpful discussion and technical assistance.

This work was supported by grants from the National Basic Research Program of China (2012CB944501 and 2010CB529503), the National Natural Science Foundation of China (31430059, 81470399, 31521062, 31271549 and 81270164) and a sponsored research programme from the AstraZeneca Innovation Center China.

Author contributions

C.X. and L.G. performed most of the experiments, analysed data and wrote the manuscript; Y.H. and F.T. designed and performed RNA-seq experiments, analysed data and contributed manuscript writing; C.X., J.W., F.W. and Y.L. made encapsulated siRNAs and analysed the data; N.C., K.H. and J.P. contributed essential reagents and transgenic zebrash lines; A.H. contributed to the design and analysis of DNA methylation experiments; and X.Z. and J.-W.X. conceived and designed this work, analysed data and wrote the manuscript.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interests.

12 NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13787 ARTICLE

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Xiao, C. et al. Chromatin-remodelling factor Brg1 regulates myocardial proliferation and regeneration in zebrash. Nat. Commun. 7, 13787doi: 10.1038/ncomms13787 (2016).

Publishers note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations.

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

r The Author(s) 2016

NATURE COMMUNICATIONS | 7:13787 | DOI: 10.1038/ncomms13787 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 13