A long non-coding RNA protects the heart from pathological hypertrophy

Summary

The role of long noncoding RNA (lncRNA) in adult hearts is unknown; also unclear is how lncRNA modulates nucleosome remodeling. An estimated 70% of mouse genes undergo antisense transcription¹, including myosin heavy chain 7 (Myh7) that encodes molecular motor proteins for heart contraction². Here, we identify a cluster of lncRNA transcripts from Myh7 loci and show a new lncRNA–chromatin mechanism for heart failure. In mice, these transcripts, which we named Myosin Heavy Chain Associated RNA Transcripts (MyHEART or Mhrt), are cardiac-specific and abundant in adult hearts. Pathological stress activates the Brg1-Hdac-Parp chromatin repressor complex³ to inhibit Mhrt transcription in the heart. Such stress-induced Mhrt repression is essential for cardiomyopathy to develop: restoring Mhrt to the pre-stress level protects the heart from hypertrophy and failure. Mhrt antagonizes the function of Brg1, a chromatin-remodeling factor that is activated by stress to trigger aberrant gene expression and cardiac myopathy³. Mhrt prevents Brg1 from recognizing its genomic DNA targets, thus inhibiting chromatin targeting and gene regulation by Brg1. Mhrt binds to the helicase domain of Brg1, and this domain is crucial for tethering Brg1 to chromatinized DNA targets. Brg1 helicase has dual nucleic acid-binding specificities: it is capable of binding lncRNA (Mhrt) and chromatinized—but not naked—DNA. This dual-binding feature of helicase enables a competitive inhibition mechanism by which Mhrt sequesters Brg1 from its genomic DNA targets to prevent chromatin remodeling. A Mhrt-Brg1 feedback circuit is thus crucial for heart function. Human MHRT also originates from MYH7 loci and is repressed in various types of myopathic hearts, suggesting a conserved lncRNA mechanism in human cardiomyopathy. Our studies identify the first cardioprotective lncRNA, define a new targeting mechanism for ATP-dependent chromatin-remodeling factors, and establish a new paradigm for lncRNA–chromatin interaction.

By 5′- and 3′-rapid amplification of cDNA ends, we discovered an alternative splicing of Myh7 anti-sense transcription into a cluster of RNAs of 709 to 1147 nucleotides (Mhrts), containing partial sequences of Myh7 introns and exons (Fig. 1a, Supplementary information). Mhrts were cardiac-specific (Fig. 1b), present at low levels in fetal hearts, with increasing abundance as the hearts matured and Myh6/Myh7 increased (Fig. 1c). RNA in situ analysis showed that Mhrts resided in the myocardium but not endocardium or epicardium (Fig. 1d, Extended Data Fig. 1a). Quantitation of nuclear/cytoplasmic RNA in heart extracts revealed that Mhrts were primarily nuclear RNAs (Fig. 1e). Coding substitution frequencies^4,5 of Mhrts predicted a negative/low protein-coding potential, in vitro translation of Mhrts yielded no proteins, and ribosome profiling⁶ revealed no/minimal ribosomes on Mhrt (Fig. 1f, Extended Data Fig. 1b–f, Supplementary text). Consequently, Mhrts are non-coding RNAs in cardiomyocyte nuclei.

Figure 1. Profile of the non-coding RNA Mhrt.

a, Schematic illustration of Mhrts originating from the intergenic region between Myh6 and Myh7 and transcribed into Myh7. Myh7 exons and introns are indicated. F1 and R1, targeting 5' and 3′ Mhrt common sequences, are the primers used for subsequent PCR.

b, RT-qPCR of Mhrts using primers targeting common regions of Mhrts in tissues from 2-month-old mice. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

c, RT-qPCR of Mhrt, Myh6 and Myh7 in mouse hearts at different ages. Mhrt and Myh6/Myh7 ratio of E11 hearts are set as 1. Error bar: SEM.

d, RNA in situ analysis of Mhrt (blue) in adult hearts. The RNA probe targets all Mhrt species. Red: nuclear fast red. White arrowheads: myocardial nuclei. Black arrowheads: nuclei of endothelial, endocardial or epicardial cells. Dashed lines demarcate the myocardium from endocardium (endo) or from epicardium (epi). Scale= 50 μm.

e, RT-qPCR of nuclear/cytoplasmic RNA in adult hearts. TfIIb, Hprt1, and 28SrRNA are primarily cytoplasmic RNAs; Neat1, nuclear lncRNA. TfIIb ratio is set as 1. P-value: Student’s t-test. Error bar: SEM.

f, Ribosome profiling: ribosome density on coding RNAs and lncRNAs.

Mhrts were down-regulated by 46–68% in hearts pressure-overloaded by transaortic constriction (TAC)³, beginning by 2 days and lasting for ≥42 days after TAC (Fig. 2a). Such Mhrt reduction coincided with TAC-induced Myh6/7 isoform switch characteristic of cardiomyopathy^7-9 (Extended Data Fig. 2a). To define Mhrt function, we focused on Mhrt779, the most abundant Mhrt species with 779 nucleotides (Fig. 2b and 2c, Extended Data Fig. 2b–e). We generated a transgenic mouse line to restore Mhrt779 level in stressed hearts. This transgenic line, driven by tetracycline response element (Tre-Mhrt779), was crossed to a cardiac-specific driver line (Tnnt2-rtTA)³ that employs troponin promoter (Tnnt2) to direct expression of reverse tetracycline-dependent transactivator (rtTA). The resulting Tnnt2-rtTA;Tre-Mhrt779 line (abbreviated as Tg779) enabled the use of doxycycline to induce Mhrt779 expression in cardiomyocytes. Within 7–14 days of doxycycline (dox) treatment, Mhrt779 increased by ~1.5-fold in left ventricles of Tg779 mice; this offset Mhrt779 suppression in TAC-stressed hearts to maintain Mhrt779 at pre-stress level (Fig. 2d). Six weeks after TAC, dox-treated control mice (Tre-Mhrt779, Tnnt2-rtTA, or wild-type) developed severe cardiac hypertrophy and fibrosis with left ventricular (LV) dilatation and reduced fractional shortening (FS). Conversely, dox-treated Tg779 hearts—with Mhrt779 maintained at pre-stress level—developed much less pathology, with 45.7% reduction of ventricle/body-weight ratio (Fig. 2e), 61.3% reduction of cardiomyocyte size (Fig. 2f, Extended Data Fig. 3a), minimal/absent cardiac fibrosis (Fig. 2g), 45.5% improvement of FS (Fig. 2h, Extended Data Fig. 3b), normalized LV size (Fig. 2i), and reduced pathological changes of Anf, Bnp, Serca2, Tgfb1, and Opn expression^10-13 (Extended Data Fig. 3c and 6e). To further test cardioprotective effects of Mhrt, we induced Mhrt779 after 1–2 weeks of TAC when hypertrophy had begun. This approach reduced hypertrophy by 23% and improved FS by 33% in 8 weeks after TAC (Extended Data Fig. 3d–f). The efficacy of late Mhrt779 introduction suggests that a sustained repression of Mhrt in stressed hearts is essential for continued decline of cardiac function.

Figure 2. Mhrt inhibits cardiac hypertrophy and failure.

a, Quantitation of cardiac Mhrts 2–42 days after TAC operation. P-value: Student’s t-test. Error bar: SEM.

b, RT-PCR of Mhrts in adult heart ventricles. Primers (F1 and R1, Fig. 1a) target Mhrt common regions. Size controls 779, 826, 709 are PCR products of recombinant Mhrt779, Mhrt826, and Mhrt709, respectively.

c, Northern blot of Mhrts in adult heart ventricles. The probe targets common regions of Mhrts. Negative: control RNA from 293T cells. Size control 826 is recombinant Mhrt826; 643 (not a distinct Mhrt species) contains the 5′ common region of Mhrt.

d Quantitation of Mhrt779 in control or Tg779 mice with/without doxycycline (Dox) or TAC operation. Mhrt779-specific PCR primers were used. Ctrl: control mice. Tg779: Tnnt2-rtTA;Tre-Mhrt779 mice. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

e, Ventricle-body weight ratio of hearts 6 weeks after TAC. P-value: Student’s t-test. Error bar: SEM. Scale=1 mm.

f, Quantitation of cardiomyocyte size in control and Tg779 mice 6 weeks after TAC by wheat-germ agglutinin staining. P-value: Student’s t-test. Error bar: SEM.

g, Trichrome staining in control and Tg779 hearts 6 weeks after TAC. Red: cardiomyocytes. Blue: fibrosis.

h, i, Echocardiographic measurement of left ventricular fractional shortening (h) and internal dimensions at end-diastole (LVIDd) and end-systole (LVIDs) (i) 6 weeks after TAC. P-value: Student’s t-test. Error bar: SEM.

To study Mhrt regulation, we examined the 5′ upstream region of Mhrts (−2329 to +143) (Extended Data Fig. 4a) for signatures of lncRNA promoter: RNA Polymerase II (PolII), histone 3 tri-methylated lysine 4 (H3K4me3), and histone 3 tri-methylated lysine 36 (H3K36me3)^4,14,15. By chromatin immunoprecipitation (ChIP) of left ventricles, we found that this putative promoter contained four evolutionarily conserved elements (a1 to a4)³ that were enriched with PolII (a1 to a4), H3K4me3 (a1 and a4), and H3K36me3^14,16-18 (a1 and a3/a4) (Extended Data Fig. 4a–d). Conversely, no PolII, H3K4me3, or H3K36me3 enrichment was found in control Shh and Vegfa promoters or in thymus and lungs that did not express Mhrts (Extended Data Fig. 4b–d). These results reveal an active, cardiac-specific lncRNA promoter controlling Mhrt expression.

We then asked how Mhrt was repressed in stressed hearts. We postulated that cardiac stress activated Brg1 to occupy a1–a4 promoter to repress Myh6³ and Mhrt in opposite transcription directions (Extended Data Fig. 4a). Indeed, Mhrt repression required Brg1: TAC suppressed Mhrts in control but not Brg1-null hearts (Tnnt2-rtTA;Tre-Cre;Brg1^f/f)³ (Extended Data Fig. 4e). To test Brg1 activity on Mhrt promoter, we cloned the a1-a4 promoter in Mhrt transcription direction (−2329 to +143) into an episomal luciferase reporter pRep4 that allows promoter chromatinization¹⁹. Brg1 was then transfected into Brg1-deficient SW13 cells²⁰ to reconstitute Brg1/BAF complex for reporter assays. Brg1 transfection caused ~50% reduction of Mhrt promoter activity (p<0.0001), and such Mhrt repression was virtually abolished by Hdac inhibition with trichostatin-A or Parp inhibition with PJ-34²¹ (Extended Data Fig. 4f), indicating a cooperative repressor function between Brg1, Hdac, and Parp. ChIP verified that Mhrt promoter (a1-a4) was occupied by Brg1, Hdac2/9 and Parp1 in stressed hearts³ and in pRep4 reporter episome (Extended Data Fig. 4g). These findings indicate that Mhrt is repressed by stress-induced Brg1–Hdac-Parp complex³ through the a1–a4 promoter.

Because Myh6 and Mhrt were both regulated by the a1–a4 promoter, we hypothesized that a1–a4 contained two elements to regulate Myh6 and Mhrt—with a1 element controlling Myh6 and a3/4 Mhrt (Extended Data Fig. 4a). On a1 and a3/4 (but not a2), we found cardiac-specific enrichment of Brg1³, H3K4me3 and H3K36me3 (Extended Data Fig. 4c-d), and DNaseI genomic footprints (Fig. 3a)²². To test a3/4 for Mhrt regulation, we conducted deletional analysis of a1-a4 promoter in Mhrt transcription direction. In reporter assays, a3/4 was necessary and sufficient for Mhrt promoter activity and for Brg1-dependent Mhrt repression, whereas a1 was not essential for such controls (Extended Data Fig. 4h). Conversely, a1 is necessary and sufficient for Brg1 to repress Myh6 promoter³, but a3/4 is not required³. Therefore, a1 and a3/4 are two functionally distinct elements for Brg1 to separately control Myh6 and Mhrt.

Figure 3. Mhrt complexes with Brg1 through the helicase domain.

a, DNaseI digital footprinting of Myh6/Mhrt promoter loci from ENCODE. Myh6 and Mhrt are transcribed in opposite directions as indicated by arrows. Bars represent DNA fragments protected from DNaseI digestion. Black boxes (a1-a4) refer to promoter regions with high sequences homology (Extended Data Fig. 4a).

b, Quantitation of Myh7/Myh6 ratio in control (Ctrl) and Tg779 (Tg) hearts 2 weeks after TAC. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

c, RNA immunoprecipitation (RNA-IP) of Mhrt-Brg1 in ventricles from control hearts (Ctrl) with Sham/TAC operation; Tg779 hearts after TAC; Brg1-null (Tnnt2-rtTA;Tre-Cre;Brg1^f/f) hearts after TAC; P1 hearts. P-value: Student’s t-test. Error bar: SEM.

d, ChIP analysis of Brg1 in control (Ctrl) and Tg779 hearts 2 weeks after TAC. Error bar: SEM.

e, Luciferase reporter assay of Myh6 and Myh7 promoters in neonatal rat cardiomyocytes. Vector: pAdd2 empty vector. Mhrt: pAdd2-Mhrt779. P-value: Student’s t-test. Error bar: SEM.

f, RNA-IP and EMSA of recombinant Brg1 proteins and in vitro transcribed Mhrt779. Biotin-labeled Mhrt779: 50 nM; unlabeled Mhrt779: 500 nM. P-value: Student’s t-test. Error bar: SEM.

g, Schematics of mouse Brg1 protein. The helicase core includes DExx-c and HELIC-c domain.

h, EMSA of Mhrt779 and Brg1 helicase. MBP: maltose binding protein. MBP-D1: MBP fused to Brg1 D1 (aa 774–913). MBP-D2: MBP fused to Brg1 D2 (aa 1086–1310). MBP-D1D2: MBP fused to Brg1 D1D2 (aa 774–1310).

i, Binding affinity of Mhrt779 for MBP-tagged D1D2 determined by EMSA. Error bars represent the standard error from multiple independent measurements. Nonlinear regression curves were generated by GraphPad Prism.

In stressed hearts Brg1 represses Myh6 and activates Myh7³, causing a pathological switch of Myh6/7 expression, characteristic of cardiomyopathy²³. Such stress/Brg1-dependent Myh switch was largely eliminated by Mhrt779 (Fig. 3b), and the inhibition of Myh switch by Mhrt did not involve RNA-RNA sequence interference between Mhrt and Myh (Extended Data Fig. 5a-j, Supplementary text). Instead, it required physical interaction between Mhrt and Brg1. RNA immunoprecipitation of TAC-stressed adult hearts or Brg1-expressing neonatal hearts showed that Brg1 co-immunoprecipitated with Mhrt779 but not control RNAs, and Mhrt779 complexed with Brg1 but not polycomb Ezh2 or Suz2 (Fig. 3c, Extended Data Fig. 6a and 6b). The Brg1-Mhrt complex was minimal in unstressed adult hearts with low Brg1³ or in stressed Brg1-null hearts (Tnnt-rtTA;Tre-Cre;Brg1^f/f)³ (Fig. 3c, Supplementary text). These results suggest that Mhrt binds to Brg1 to influence its gene regulation.

We then tested how Mhrt regulated Brg1 activity on its in vivo target genes, including Myh6³, Myh7³, and Opn (Osteopontin, critical for cardiac fibrosis¹²)(Extended Data Fig. 6c–e, Supplementary text). In dox-treated, TAC-stressed Tg799 hearts, Mhrt779—without affecting Brg1 mRNA/protein level (Extended Data Fig. 7a-f)—reduced Brg1 occupancy on Myh6, Myh7 and Opn promoters by 60–90% (Fig. 3d), causing a 56–76% loss of Brg1-controlled Myh switch and Opn activation (Fig. 3b, Extended Data Fig. 6e and 7g). We then used primary rat ventricular cardiomyocytes to conduct reporter assays. In these cells, as observed in vivo, Brg1 repressed Myh6 and activated Myh7 and Opn promoters; Mhrt779 reduced Brg1 activity on these promoters by 54–80% (Fig. 3e). Accordingly, Mhrt prevents Brg1 from binding to its genomic targets to control gene expression.

How Brg1 or ATP-dependent chromatin remodelers recognize their target promoters is an important but not fully understood issue in chromatin biology. Biochemically, recombinant Brg1 proteins and in vitro transcribed Mhrt779 could directly co-immunoprecipitate without involving other factors (Fig. 3f). Electrical mobility shift assay (EMSA) showed that Brg1 shifted biotin-labeled Mhrt779 to form a low mobility protein–RNA complex that was competitively disrupted by unlabeled Mhrt779 (Fig. 3f). Brg1, which belongs to SWI/SNF family of chromatin-remodeling factors, contains a helicase/ATPase core that is split by an insertion into two RecA-like domains: DExx-c (DEAD-like helicase superfamily c-terminal domain, D1) and HELIC-c (helicase superfamily c-terminal domain, D2)^24,25 with signature motifs of DEAD-box, superfamily 2 RNA helicase^25,26 (Fig. 3g, Extended Data Fig. 8). SWI/SNF proteinş although conserved with RNA helicases, were observed to bind DNA²⁷ and mediate DNA structural changes and repair¹⁹. The binding properties of Brg1 remained undefined. To test if Mhrt could bind to Brg1 helicase, we generated MBP-tagged recombinant proteins that contained Brg1 DExx-c domain (MBP-D1, amino acid 774–913), HELIC-c domain with C-terminus extension (MBP-D2, 1086–1310), or entire helicase (MBP-D1D2, 774–1310)(Extended Data Fig. 9a). D1D2 showed the highest Mhrt binding affinity (K_d = 0.76 μM); D1 moderate (K_d = 1.8 μM); D2 modest (K_d > 150 μM); MBP no binding (Fig. 3h and 3i). Therefore, Brg1 helicase binds Mhrt with high affinity.

Contrary to its potent RNA binding, Brg1 helicase had no detectable binding to the naked DNA of Myh6 promoter (596bps, −426 to +170, critical for Brg1’s control of Myh6³) (Extended Data Fig. 9b). To test if Brg1 helicase could bind chromatinized DNA, we generated nucleosomal DNA in vitro by assembling histone octamers (histone 2A, 2B, 3, and 4)²⁸ on Myh6 promoter DNA, as well as on control Neo and 5SrDNA. We achieved 50–65% efficiency of nucleosome assembly, comparable among Myh6, Neo, and 5SrDNA (Fig. 4a). Because the large nucleosome size precluded a clear EMSA resolution, we used amylose to pull down MBP-tagged D1D2 proteins. We found that D1D2 pulled down nucleosomal Myh6 promoter DNA but not the naked one (Fig. 4b). The pull-down efficiency of nucleosomal Myh6 was ~3–6-fold that of Neo or 5SrDNA (Fig. 4c), and Mhrt779 was capable of disrupting D1D2-Myh6 pull-down (Fig. 4d). Although D1D2 bound to histone 3 (Fig. 4e), histone binding was insufficient to anchor D1D2 to nucleosomal DNA, as D1D2 bound poorly to nucleosomal Neo and 5SrDNA that also contained histones (Fig. 4c). Therefore, chromatinized DNA targets are biochemically recognized by Brg1 helicase, and this process is inhibited by Mhrt.

Figure 4. Mhrt inhibits chromatin targeting and gene regulation by Brg1.

a, Gel electrophoresis and quantitation of nucleosomal 5SrDNA, Myh6 promoter and Neo DNA. Arrowheads: DNA-histone complex. Arrows: naked DNA. Nucleosome assembly efficiency is defined as the fraction of DNA bound to histones (arrowheads). P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

b-d, Quantification of amylose pull-down of MBP-D1D2 (D1D2) with nucleosomal and naked Myh6 promoter DNA (b), with nucleosomal Myh6 promoter, Neo, and 5SrDNA (c), or with nucleosomal Myh6 promoter in the presence of Mhrt779 (d). P-value: Student’s t-test. Error bar: SEM.

e, Amylose pull-down of MBP-D1D2 and histone 3. Anti-histone 3 and anti-MBP antibodies were used for western blot analysis.

f, ChIP analysis of Brg1 on chromatinized and naked Myh6 promoter in rat ventricular cardiomyocytes. GFP: green fluorescence protein control. P-value: Student’s t-test. Error bar: SEM.

g, h, Luciferase reporter activity of Brg1 on naked Myh6 promoter (g) or of helicase-deficient Brg1 on chromatinized Myh6 promoter (h) in rat ventricular cardiomyocytes. ΔD1: Brg1 lacking amino acid 774–913; ΔD2: Brg1 lacking 1086–1246. GFP: green fluorescence protein control. ChIP: H-10 antibody recognizing N-terminus, non-disrupted region of Brg1. P-value: Student’s t-test. Error bar: SEM.

i, j, ChIP analysis in SW13 cells of chromatinized Myh6 promoter in the presence of Mhrt779 (i) or helicase-deficient Brg1 (j). Vector: pAdd2 empty vector. Mhrt: pAdd2-Mhrt779. P-value: Student’s t-test. Error bar: SEM.

k, Schematic illustration and PCR of human MHRT. MHRT originates from MYH7 and is transcribed into MYH7. MYH7 exons and introns are indicated. R1 and R2 are strand-specific PCR primers; F1 and R1 target MHRT and MYH7; F2 and R2 are specific for MHRT.

l, Quantification of MHRT in human heart tissues. Ctrl: control. LVH: left ventricular hypertrophy. ICM: ischemic cardiomyopathy. IDCM: idiopathic dilated cardiomyopathy. P-value: Student’s t-test. Error bar: SEM.

m, Working model of a Brg1-Mhrt negative feedback circuit in the heart. Brg1 represses Mhrt transcription, whereas Mhrt prevents Brg1 from recognizing its chromatin targets. Brg1 functions through two distinct promoter elements to bidirectionally repress Myh6 and Mhrt expression.

n, Molecular model of how Brg1 binds to its genomic DNA targets. Brg1 helicase (D1D2) binds chromatinized DNA, C-terminal extension (CTE) binds histone 3 (H3), and bromodomain binds acetylated (Ac) histone 3 or 4 (H4).

To test the ability of Brg1 to distinguish chromatinized from naked DNA promoters in cells, we cloned Myh6 promoter into the luciferase reporter plasmid pREP4 (allowing promoter chromatinization¹⁹) and pGL3 (containing naked, non-chromatinized promoter). In rat ventricular cardiomyocytes and SW13 cells, ChIP and luciferase analyses showed that Brg1 bound and repressed chromatinized but not naked Myh6 promoter (Fig. 4f and 4g, Extended Data Fig. 9c and 9d). However, without D1/D2 domain or in the presence of Mhrt, Brg1 was unable to bind or repress chromatinized Myh6 promoter (Fig. 4h–j, Extended Data Fig. 9e), indicating the necessity of D1D2 for the interaction between Brg1, chromatin, and Mhrt. Consistently, all our genetic, biochemical, and cellular studies show that Brg1 requires the helicase domain to bind to chromatinized DNA targets, and Mhrt seizes the helicase to disrupt Brg1–chromatin binding.

We then asked how Brg1 surpassed its basal suppression by Mhrt to control Myh, Mhrt, Opn, or other genes to trigger cardiomyopathy (Supplementary text). Amylose pull-down experiments showed that Brg1 dose-dependently escaped from Mhrt inhibition to occupy Mhrt promoter (Extended Data Fig. 10). Brg1 protein, which increases under stress conditions³, could therefore outrun Mhrt and gain control over Mhrt promoter to repress Mhrt expression and tip the balance toward Brg1. Contrary to the endogenous Mhrt that was repressible by Brg1, the Mhrt transgene (Tg779)—driven by Tnnt2/Tre promoters—was not subject to Brg1’s repression and thus able to keep Mhrt at pre-stress level to inhibit Brg1 and reduce hypertrophy. This further demonstrates the necessity of Mhrt repression for myopathy to develop.

Human MYH7 loci encoded RNA that resembled Mhrt in primary sequence and secondary structure predicted by minimal free energy²⁹ (Fig. 4k, Extended Data Fig. 11a and 11b). The human MHRT was also repressed in stressed hearts, with 82.8%, 72.8%, and 65.9% reduction of MHRT in hypertrophic, ischemic or idiopathic cardiomyopathy tissues (Fig. 4l, Extended Data Fig. 11c). This suggests a conserved Mhrt mechanism of human cardiomyopathy.

Discussion

Mhrt is the first example of lncRNA that inhibits myopathy and chromatin remodelers. Reciprocal Mhrt-Brg1 inhibition constitutes a negative feedback circuit critical for maintaining cardiac function (Fig. 4m). The helicase core of Brg1, combined with histone-binding domains of Brg1/BAF complex, adds a new layer of specificity control to Brg1/BAF targeting and chromatin remodeling (Fig. 4n). The Mhrt–helicase interaction also exemplifies a new mechanism by which lncRNA controls chromatin structure. To further elucidate chromatin regulation, it will be essential to define helicase domain function in all ATP-dependent chromatin-remodeling factors and to identify new members of lncRNA that act through this domain to control chromatin. The cardioprotective Mhrt may have translational value, given that RNA can be chemically modified and delivered as a therapeutic drug. This aspect of lncRNA–chromatin regulation may also inspire new therapies for human disease.

METHODS

Mice, animal sample size, and randomization

For the generation of Tg779 mice, Mhrt779 was cloned into pTRE2 backbone (Clonetech, CA). DNA fragment containing the Tre promoter, and Mhrt779 was injected into the pronucleus of fertilized oocytes (B6C3H/F1). Embryos were implanted into a pseudopregnant CD-1 mouse. Tre-Mhrt779 transgene was identified by PCR genotyping using primers (CGCCTGGAGACGCCATCCAC; TGTCTTCAAAGCTGACTCCCT). Tre-Mhrt779 mice with ~3 copies of the transgene were backcrossed with Tnnt2-rtTA mice as described previously^3,30 to generate Tnnt2-rtTA; Tre-Mhrt779 (Tg779) mice. The number of animals used (N) was denoted in each test in the figures, including technical replicates when applicable. We routinely used mouse littermates to control and perform our experiments. Each subgroup of experiments had N = 3 to 14 biological replicates, many of which had technical replicates of three. The assignment to each experimental subgroup was based on the genotypes. Littermate mice with the same genotypes regardless of gender were randomly selected from the cage and assigned to different control and experimental subgroups. Major procedures were blinded. The use of mice for studies was in compliance with the regulations of Indiana University, Stanford University, and National Institute of Health.

RACE and cloning of full length of Mhrt transcripts

The 3′ and 5′ RACE were performed using the FirstChoice RLM-RACE Kit (Ambion) following the manufacture’s instruction. RNA was extracted from adult heart ventricles. Primers used for 3′ and 5′ RACE were designed based on the known sequence information: TCATTGGCACGGACAGCATC (First-round Mhrt 3′-prime specific) and GAGCATTTGGGGATGGTATAC (Second-round Mhrt 3′-prime specific); CAACACTTTTCATTTTCCTCTTT (First-round Mhrt 5′-prime specific) and TCTGCTTCATTGCCTCTGTTT (Second-round Mhrt 5′-prime specific). Once we reached the 5′- and 3′-cDNA ends, we used primers F1 (Fig. 1a, AAGAGCCCTACAGTCTGATGAACA) and R1 Fig. 1a, CCTTCACACAAACATTTTATTT) to amply the full-length Mhrt transcripts and cloned into pDrive TA cloning vector (Qiagen) and send for sequencing. Mhrts were also further cloned into shuttle vector pAdd2^31,32 for expression in cells.

Northern blot and in situ Hybridization

We obtained 5 μg of total RNA using Quick-RNA Mini Kit (Zymo Research). RNA blot was performed using NorthernMax Kit (Ambion) following the manufacturer’s protocol. Single stranded RNA probe was generated by in vitro transcription with MaxIscript SP6/T7 kit (Ambion) with ATP [α-³²P] (PerkinElmer) using full-length Mhrt779, Myh6 and Myh7 as the template and followed by digestion with DNase I (Ambion). Hybridization was performed at 65°C. The blot was washed and imaged by Phosphor storage scanning by Typhoon 8600 Imager (GE Healthcare). In situ hybridization experiments were performed as previously described^3,33.

RNA fractionation

To isolate cytosolic and nuclear RNAs from adult heart tissues, we used PARIS kit (Ambion) and followed the manufacturer’s instruction. 10 mg of tissues were homogenized in Cell Fractionation Buffer thoroughly before centrifuge for 5min at 500g. Supernatant was collected as cytosolic fraction, while nuclear pellet was washed and lysed by Cell Disruption Buffer. Such samples were further mixed with 2X Lysis/Binding Solution before extracting RNA using the manufacturer’s protocol.

Codon substitution frequency predication

To measure the coding potential of Mhrt, we used the previously described codon substitution frequencies (CSF) method^4,5 to evaluate the evolutionary characteristics in their alignments with orthologous regions in 6 other sequenced mammalian genomes (Rat, Human, Hamster, Rhesus Monkey, Cat and Dog). CSF generates a likelihood score for a given sequence considering all codon substitutions observed within its alignment across multiple species, which was based on the relative frequency of similar substitutions occurred in known coding and noncoding regions. CSF compares two empirical codon models; one generated from alignments of known coding regions and the other according to noncoding regions, producing a likelihood ratio. The ratio reflects whether the protein-coding model better explains the alignment.

Ribosome profiling and RNA deep sequencing

For ribosome profiling⁶, over-expression of predominant specie of Mhrt (Mhrt779) along with HOTAIR were achieved through co-transfecting pAdd2-779 and pAdd2-HOTAIR into SW13 cells. The cells were then lysed to extract ribosome-associated RNA fragments using ARTseq™ Ribosome Profiling Kit (Epicenter, Illumina). The RNA fragments were further converted into a DNA library through end repair, adaptor ligation, reverses transcription circularization, and PCR amplification. A conventional RNA-seq library was also prepared with total RNA extracted from those cells with miRNeasy Mini Kit (Qiagen #217004). The libraries were further processed according to MiSeq Sample Prep sheet, and MiSeq 50 cycle kit was used for sequencing. 1.25 picomoles PCR products were used for sequencing. 600K~700K reads were properly paired and used for further analysis. The resulting reads were aligned to the human hg19 or mouse mm10 genome using Bowtie2 v2.0.0.6³⁴. Mapped reads were visualized on the UCSC browser as bigwig files generated using samtools v0.1.18³⁵, bedtools v2.16.1³⁶, bedClip and bedGraphToBigWig. For quantification of FPKM values (fragments per kilobase of exon per million fragments mapped), cuffdiff as part of the tophat suite V2.0.8b³⁷ was run on a merged bam file containing the human and the Mhrt reads using a custom gtf file comprising of the human hg19 iGenome and the Mhrt transcripts. To generate scatter plot of the genes, cuffdiff files were used for visualization with cummerbund v2.3.1³⁷. The data were uploaded to GEO (Gene Expression Omnibus) with accession number: GSE49716.

In vitro translation and biotin labeling

TNT Quick Coupled Transcription/Translation System (Promega) was used for in vitro translation. Briefly, 1 ug plasmids of control (Luciferase) and various Mhrt species inserted into pDrive vector were added to 40 ul rabbit reticulocyte lysates containing S³⁵-methionine. After 1hr of incubation, the reactions were analyzed on 10–20% Tris-Tricine gel. The gel was dried and visualized by the Typhoon 8600 Imager (GE Healthcare). Biotin-NTP was added to in vitro translation reaction. Total RNAs were extracted and the biotin-labeled RNAs were detected subsequently by IRDye 680 Streptavidin (Li-COR, 926-68079) using Odyssey Infrared Imaging System.

Transaortic constriction (TAC)

The TAC surgery was performed as described³ on adult mice of 8–10 weeks of age and between 20 and 25 grams of weight. Mice were fed with doxycycline food pellets (6 mg doxycycline/kg of food, Bioserv, Frenchtown, NJ) 7–14 days prior to the TAC operation. Mice were anesthetized with isoflurane (2–3%, inhalation) in an induction chamber and then intubated with a 20-gauge intravenous catheter and ventilated with a mouse ventilator (Minivent, Harvard Apparatus, Inc). Anesthesia was maintained with inhaled isoflurane (1–2%). A longitudinal 5-mm incision of the skin was made with scissors at midline of sternum. The chest cavity was opened by a small incision at the level of the second intercostal space 2–3 mm from the left sternal border. While opening the chest wall, the chest retractor was gently inserted to spread the wound 4–5 mm in width. The transverse portion of the aorta was bluntly dissected with curved forceps. Then, 6–0 silk was brought underneath the transverse aorta between the left common carotid artery and the brachiocephalic trunk. One 27-gauge needle was placed directly above and parallel to the aorta. The loop was then tied around the aorta and needle, and secured with a second knot. The needle was immediately removed to create a lumen with a fixed stenotic diameter. The chest cavity was closed by 6–0 silk suture. Sham-operated mice underwent similar surgical procedures, including isolation of the aorta, looping of aorta, but without tying of the suture. The pressure load caused by TAC was verified by the pressure gradient across the aortic constriction measured by echocardiography. Only mice with a pressure gradient >30 mmHg were analyzed for cardiac hypertrophy, echocardiography and other purposes.

Echocardiography

The echocardiographer was blinded to the genotypes and surgical procedure. Transthoracic ultrasonography with a GE Vivid 7 ultrasound platform (GE Health Care, Milwaukee, WI) and a 13 MHz transducer was used to measure aortic pressure gradient and left ventricular function. Echocardiography was performed on control and Tnnt2-rtTA;Tre-Mhrt779 (Tg779) mice designated time points after the TAC procedure. To minimize the confounding influence of different heart rates on aortic pressure gradient and left ventricular function, the flow of isoflurane (inhalational) was adjusted to anesthetize the mice while maintaining their heart rates at 450–550 beats per minute. The peak aortic pressure gradient was measured by continuous wave Doppler across the aortic constriction. The left ventricular function was assessed by the M-mode scanning of the left ventricular chamber, standardized by two-dimensional, short-axis views of the left ventricle at the mid papillary muscle level. Left ventricular (LV) chamber size and wall thickness were measured in at least three beats from each projection and averaged. LV internal dimensions at diastole and systole (LVIDd and LVIDs, respectively) were measured. The fractional shortening (FS) of the left ventricle was defined as 100% × (1 - LVIDs/LVIDd), representing the relative change of left ventricular diameters during the cardiac cycle. The mean FS of the left ventricle was determined by the average of FS measurement of the left ventricular contraction over 5 beats. P-values were calculated by the Student-t test. Error bars indicate standard error of mean.

Histology, trichrome staining, and morphometric analysis of cardiomyocytes

Histology and trichrome staining were performed as described^38,39. Trichrome Stain (Masson) kit (Sigma) was used and manufacture’s protocol was followed. For morphometric analysis of cardiomyocytes, paraffin sections of the heart were immunostained with a fluoresecin isothiocyanate-conjugated Wheat Germ Agglutinin (WGA) antibody (F49, Biomeda, Foster City, CA) that highlighted the cell membrane of cardiomyocytes. Cellular areas outlined by WGA were determined by the number of pixels enclosed using Image J software (NCBI). Approximately 250 cardiomyocytes of the papillary muscle at the mid left ventricular cavity were measured to determine the size distribution. P-values were calculated by the Student-t test. Error bars indicate standard error of mean.

Reverse transcription–quantitative PCR analysis (RT–qPCR) and strand specific reverse transcription PCR analysis

RT–qPCR analyses were performed as described^3,38. The following primer sequences (listed below) were used. RT-qPCR reactions were performed using SYBR green master mix (BioRad) with an Eppendorf realplex, and the primer sets were tested to be quantitative. Threshold cycles and melting curve measurements were performed with software. P-values were calculated by the Student-t test. Error bars indicate standard error of mean. To conduct strand specific RT PCR analysis, human total RNA and Superscript III First-Strand Synthesis System (Invitrogen) was used. Primers R1 (Fig. 4k, CTACAGAATGAGATCGAGGACT) and R2 (Fig. 4k, GGGGCTGAAGAGTGAGCCTT) were designed based on known sequence and used for individual RT respectively. To detect MHRT, primers F1 (Fig. 4k, CTGGAGCTGGGACAGGTCAGCA) and R1 were used. Those primers could also amplify endogenous MYH7 and thus serve as controls. Primers F2 (Fig. 4k, TGGGGAACACGGCGTTCTTGA) and R2 were used to specifically amplify MHRT and used in RT-qPCR analysis.

PCR primers for RT-qPCR of mRNA:

Mouse TfIIb-F (CTCTGTGGCGGCAGCAGCTATTT),

Mouse TfIIb-R (CGAGGGTAGATCAGTCTGTAGGA),

Mouse Hprt1-F (GCTGGTGAAAAGGACCTCT),

Mouse Hprt1-R (CACAGGACTAGAACACCTGC),

Mouse Anf-F (GACTAGGCTGCAACAGCTTCCG),

Mouse Anf-R (GCCACAGTGGCAATGTGACCAA),

Mouse Serca2a-F (CATTTGCATTGCAGTCTGGAT),

Mouse Serca2a-R (CTTTGCCATCCTACGAGTTCC),

Mouse Tnnt2-F (TACAGACTCTGATCGAGGCTCACTTC),

Mouse Tnnt2-R (TCATTGCGAATACGCTGCTGCTC),

Mouse Mhrt-F (common) (GAGCATTTGGGGATGGTATAC),

Mouse Mhrt-R (common) (TCTGCTTCATTGCCTCTGTTT),

Mouse Mhrt779-F (TCTGGCCACAGCCCGCAGCTTC),

Mouse Mhrt779-R (AGTCATGTATACCATCCCCAA),

Mouse Neat1-F (TCTCCTGGAGCCACATCTCT)

Mouse Neat1-R (GCTTTTCCTTAGGCCCAAAC)

Mouse 28S-rRNA-F (GGTAGCCAAATGCCTCGTCAT)

Mouse 28S-rRNA-R (CCCTTGGCTGTGGTTTCG)

Human TFIIb-F (ACCACCCCAATGGATGCAGACAG),

Human TFIIb-F (ACGGGCTAAGCGTCTGGCAC),

Human MHRT-F (F2, TGGGGAACACGGCGTTCTTGA),

Human MHRT-R (R2, GGGGCTGAAGAGTGAGCCTT).

Human HOTAIR-F (GGTAGAAAAAGCAACCACGAAGC)

Human HOTAIR-R (ACATAAACCTCTGTCTGTGAGTGCC)

Human GAPDH-F (CCGGGAAACTGTGGCGTGATGG)

Human GAPDH-R (AGGTGGAGGAGTGGGTGTCGCTGTT)

Chromatin immunoprecipitation–quantitative PCR (ChIP–qPCR)

ChIP assay was performed as described³ with modifications. Briefly, chromatin from hearts or SW13 cells was sonicated to generate average fragment sizes of 200–600 bp, and immunoprecipitated using anti-BRG1 J1 antibody^3,40, anti-Brg1 H-10 antibody (Santa Cruz Biotechnology, against 115–149 amino acids of N-terminus Brg1), anti-RNA polymerase II (Pol II) antibody (ab24759, Abcam), anti-H3K4me3 antibody (07–473, Millipore), anti-H3K36me3 antibody (17–10032, Millipore) or normal control IgG. Isolation and purification of immunoprecipitated and input DNA were done according to the manufacturer’s protocol (Magna ChIP Protein G Magnetic Beads, Millipore), and qPCR analysis of immunoprecipitated DNA were performed. ChIP–qPCR signal of individual ChIP reaction was standardized to its own input qPCR signal or IgG ChIP signal. PCR primers (listed below) were designed to amplify the promoter regions of mouse Myh6 (−426, −320), mouse Myh7 (−102, +58), mouse Shh (−7142, −6911), mouse Vegfa (+1, +150) human GAPDH (−45, +121). The DNA positions are denoted relative to the transcriptional start site (+1).

PCR primers for ChIP-qPCR:

Mouse ChIP-Myh6 promoter-F (GCAGATAGCCAGGGTTGAAA),

Mouse ChIP-Myh6 promoter-R (TGGGTAAGGGTCACCTTCTC),

Mouse ChIP-Myh7 promoter-F (GTGACAACAGCCCTTTCTAAAT),

Mouse ChIP-Myh7 promoter-R (CTCCAGCTCCCACTCCTACC),

Mouse ChIP-Shh promoter-F (GAGAACATTACAGGGTAGGAA),

Mouse ChIP-Shh promoter-R (GAAGCAGTGAGGTTGGTGG),

Mouse ChIP-Vegfa promoter-F (CAAATCCCAGAGCACAGACTC),

Mouse ChIP-Vegfa promoter-R (AGCGCAGAGGCTTGGGGCAGC),

Human ChIP-GAPDH promoter-F (TACTAGCGGTTTTACGGGCG),

Human ChIP-GAPDH promoter-R (TCGAACAGGAGGAGCAGAGAGCGA).

RNA Immunoprecipitation

RNA Immunoprecipitation (RNA-IP, RIP) was conducted as described⁴ with some modifications. Briefly, P1 hearts, sham/TAC’ed hearts, or SW13 cells were cross-linked and lysed with Lysis Buffer (10 mM HEPES pH 7.5, 85 mM KCl, 0.5% NP-40, 1 mM DTT, 1 × protease inhibitor) for tissue and Lysis Buffer (10 mM Tris-HCl pH 8.1, 10 mM NaCl, 1.5 mM MgCl_2, 0.5% NP-40, 1 mM DTT, 1X protease inhibitor) for cell. Nuclei were isolated and sonicated using Bioruptor (Diagenode) (30 s on, 30 s off, power setting H, 5 minute for twice) in Nuclear Lysis buffer (50 mM Tris-HCl pH 8.1, 150 mM NaCl, 0.1% NP-40, 1 mM DTT, protease inhibitor, Ribonuclease inhibitor). The nuclear extract was collected and incubated with primary antibodies at 4°C for overnight together with Manga ChIP Protein G Magnetic Beads (Millipore). The beads were washed by Wash Buffer I (20 mM Tris-Hcl pH 8.1, 150 mM NaCl, 1% Triton X-100 and 0.1% SDS) for three times, and Wash Buffer II (20 mM Tris-Hcl pH 8.1, 500 mM NaCl, 1% Triton X-100 and 0.1% SDS) for three times. Beads were then resuspended in 150 μl 150 mM RIPA (50 mM Tris pH 7.5, 150 mM NaCl, 1 mM EDTA, 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate) with 5 μl Proteinase K and incubated for 1 h at 65°C. We added 1 ml of TRIzol to the sample, and RNA was extracted using the Quick-RNA Mini Kit with the on-column DNAse I digest (ZymoResearch). RT and qPCR were then conducted with the purified RNA. The antibodies used for the immunoprecipitation are anti-BRG1 J1 antibody^3,40, Ezh2⁴¹ (Active Motif), Suz12^41,42 (Bethyl Laboratories) and normal IgG control.

Reporter assay and truncation of Mhrt promoter

For Mhrt promoter reporter assay, plasmid was constructed by inserting ~2.5Kbp mouse Mhrt promoter into the episomal pREP4-Luc plasmid^3,19,38,43 through cloning PCR amplified region of the promoter by using primers ACCGGCCTGAACCCCACTTCC and ATGTCGAGACAGGGAACAGAA. Mouse Myh6 (−426 to +170, based on new genome annotation) and Myh7 (−3561 to +222) reporter constructs were described previously³. These vectors were transfected into rat neonatal cardiomyocytes or SW13 cells using lipofectamine 2000 (Invitrogen) along with plasmids expressing mouse Brg1 (Actin-mBrg1-IRES-EGFP) or a matching empty vector plasmid (Gifts from Dr. G. Crabtree) as well as an episomal renilla-luciferase plasmid (pREP7-RL) to normalize transfection efficiency. The transfected cells were cultured for 48hrs and harvested for luciferase assay using the dual luciferase assay kit (Promega). For naked DNA reporter, mouse Myh6 promoter (−426 to +170) was inserted in pGL3 vector (Promega), and renilla-luciferase plasmid phRL-SV40 (Gifts from Dr. J. Chen) was used as a normalizer. Dual luciferase assay was performed according to the manufacture’s instruction 48hrs after transfection. For deletional analysis of Mhrt promoter, various regions of the promoter were deleted from the full-length pREP4-Mhrt. The constructs were further analysed by transfecting into SW13 cells.

RNA-EMSA and Kd calculation

Biotin-labeled RNA probe was generated by in vitro transcription with MAXIscript SP6/T7 kit (Ambion) with biotin labeling NTP mixture (Roche) using linearized pDrive-Mhrt779 construct as the template and followed by digestion with DNase I (Ambion). The EMSA was performed by using the LightShift Chemiluminescent RNA EMSA Kit (Thermo Scientific). The labeled probe was incubated with appropriate amount of recombinant proteins in 10 μl in the 1 × binding buffer (10 mM HEPES-KOH, pH 7.3, 10 mM NaCl, 1 mM MgCl₂, 1 mM DTT) with 5 μg tRNA carrier at room temperature for 30 min. The reactions were then loaded onto 1% 0.5 × TBE agarose gel and transferred to BrightStar-Plus positive charged membrane. The biotin-labeled probes were detected and quantified subsequently by IRDye 680 Streptavidin (Li-COR, 926–32231) using Odyssey Infrared Imaging System. The shifted signals were quantified and plotted against amount of the MBP, MBP-D1, MBP-D2 and MBP-D1D2 proteins using previously described method²⁶ with GraphPad Prism (GraphPad). The software facilitates the fitting of non-linear regression model and calculation of Kds (Dissociate Constant) based on the fitting curve. The errors and r-square (r²) were also generated from the fitting curve.

Protein expression and purification of Brg1 helicase domains

To generate MBP fusion proteins of mouse Brg1 helicase domains, the DExx-box domain (D1) (Amino acids 774–913 of Brg1), Helicase-C domain (D2) together with carboxyl-terminal extension (CTE) (Amino acids 1086–1310 of Brg1), as well as the entire helicase region (D1D2) (774–1310) were amplified by PCR and cloned into pMAL vector. MBP fusion proteins were induced by IPTG and purified by Amylose resin (E8021S, NEB).

Nucleosome assembly and amylose pull-down

Nucleosome assembly was performed by using EpiMark Nucleosome Assembly Kit (E5350S, NEB) following the manufacture’s instruction²⁸. In brief, recombinant human core histone octamer, which consist of the 2:1 mix of histone H2A/H2B dimer and histone H3.1/H4 tetramer, were mixed with purified 5SrDNA (208bp, N1202S, NEB), Neo (512bp, amplified from pST18-Neo, 1175025, Roche), Myh6 core promoter (596bp, −426 to +170) and Mhrt core promoter (a3a4, 596bp, −2290 to −1775) DNA at 2 M NaCl. PCR primers to amplify Neo are CGATGCGCTGCGAATCGGGA and CACTGAAGCGGGAAGGGACT. The salt concentration was gradually lowered by dilution to allow the formation of nucleosomes. The EMSA assay was used to assess the efficiency of nucleosome assembly. For amylose pull-down assay, the amylose resin (E8021S, NEB) was washed thoroughly and equilibrated with binding buffer (10 mM Tris-HCl, pH=7.5, 150 mM NaCl) before incubation with purified MBP, MBP-D1D2 proteins for 2hr. Nucleosome mixture or naked DNA mixture of 5S rDNA, Neo and Myh6 promoter DNA were added for incubation at 4°C for overnight. The resin was then washed excessively by washing buffer (20 mM Tris-HCl, pH=8.1, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.1% SDS) before decrosslinking and extraction of the DNA with Phenol: Chloroform: Isoamyl Alcohol. For competition assay, in vitro transcribed Mhrt779 was incubated with MBP-D1D2 in binding buffer (10 mM HEPES-KOH, pH 7.3, 10 mM NaCl, 1 mM MgCl₂, 1 mM DTT) with Ribonuclease inhibitor at room temperature for 30 minutes before adding nucleosomal DNA. The subsequent incubation, wash and DNA purification were performed as regular amylose pull-down assays. qPCR signal of individual pull-down reaction was standardized to its own input RT-qPCR signal. qPCR primers were designed to amplify the 5SrDNA (CAAGCAAGAGCCTACGACCA; ATTCGTTGGAATTCCTCGGG), Neo (TAAAGCACGAGGAAGCGGTC; TCGACCACCAAGCGAAACAT), Myh6 promoter (GCAGATAGCCAGGGTTGAAA; TGGGTAAGGGTCACCTTCTC) and Mhrt promoter (ATGCCAAATGGTTGCTCTTT; GAGCTTGAGAACCAGGCAGT).

Cloning of Brg1 truncation constructs

For cloning of the truncated Brg1 with deletion of amino acids 774–913 (ΔD1) or 1086–1246 (ΔD2), the primers with NheI restriction digestion site, which complement with the downstream and upstream of the truncated region (Δ D1: CCCGGGGCTAGCCTGCAGAACAAGCTACCGGAGCT and CCCGGGGCTAGCCAGGTTGTTGTTGTACAGGGACA; Δ D2: CCCGGGGCTAGCATCAAGAAGTTCAAATTTCCC and CCCGGGGCTAGCCTGCAGGCCATCCTGGAGCACGAGCAG) were used to amplify from pActin-Brg1-IRES-EGFP by KOD Xtreme Hot Start DNA Polymerase (Novagen). After digestion with NheI, the linearized fragment was subject to ligation and transformation. The truncation constructs were sequenced to confirm the fidelity of the cloning. Western Blot was further performed to assess the expression of the constructs. Monoclonal H-10 antibody (Santa Cruz Biotech, sc-374197), which were raised against Brg1 N-terminal amino acids, were used in the experiments involving the truncated Brg1.

Protein sequence analysis

Brg1 core helicase domain (774–1202) was applied for secondary structure prediction using the Fold & Function Assignment System (FFAS) server (http://ffas.burnham.org/ffas-cgi/cgi/ffas.pl). The output revealed that Brg1 core helicase domain are structural homologs of SF2 helicases: Vasa⁴⁴ (Fruit fly, PDB# 2DB3), Rad54^27,45 (Zebrafish PDB# 1Z3I, Sulfolobus solfataricus PDB# 1Z63) and Chd1⁴⁶ (Yeast, PDB# 3MWY). Those proteins together with Brg1 were further employed for multiple sequence alignment with T-Coffee, which is a program allowing combination of the results obtained with several alignment methods (http://www.ebi.ac.uk/Tools/msa/tcoffee/).

RNA secondary structural prediction

To predict the secondary structure for mouse Mhrt and human MHRT, the single stranded sequence of Mhrt779 and human MHRT were analyzed on the Vienna RNAfold web server (http://rna.tbi.univie.ac.at/cgi-bin/RNAfold.cgi) with calculation of minimum free energy^29,47-49.

Human heart tissue analysis

The human tissues were processed for RT-qPCR and strand-specific RT-PCR. The use of human tissues is in compliance with the regulation of Sanford/Burnham Medical Research Institute, Intermountain Medical Center, Stanford University, and Indiana University.

Primary cardiomyocyte culture

For functional studies in cardiomyocytes, neonatal rat ventricular cardiomyocytes were cultured as previously described^50,51. Briefly, P0 or P1 Sprague-Dawley rats were used. The ventricles were excised and trypsinized for 15 minutes for 4–5 times. Cells were then collected and resuspended in DMEM supplements with 10% FBS. The cells were plated for 1 h to allow the attachment of non-cardiomyocytes cells. The remaining cardiomyocytes were plated at a density of 2 × 10⁵/ml. The cells were transfected with Lipofectamine 2000 (Invitrogen) after 48 hr.

Supplementary Material

7

Extended Data Figure 6. RNA-IP controls; Opn is another target gene of Brg1 in stressed hearts

a, Immunostaining of Brg1 in P1 heart. Red: Brg1. Green: WGA. Blue: DAPI. Ctrl: control. Tg779: Tnnt2-rtTA;Tre-Mhrt779. Scale = 50 μm.

b, RNA immunoprecipitation (RNA-IP, RIP) of Mhrt in P1 hearts using antibodies against Ezh2 and Suz12. Right panels show immunostaining of Ezh2 and Suz12 in P1 hearts. PRC2: polycomb repressor complex 2. Red: Ezh2 or Suz12. Green: WGA. Blue: DAPI. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

c, Quantitation of Opn mRNA in control and Brg1-null mice after Sham or TAC operation.

d, ChIP of Brg1 on Opn proximal promoter in control and transgenic (Tg779) mice after Sham or TAC operation.

e, Quantitation of Opn in control and transgenic (Tg779) mice after Sham or TAC operation. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

8

Extended Data Figure 7. Induction of Mhrt779 is insufficient to change Brg1 mRNA or protein level

a, qPCR analysis of Brg1 expression in hearts without TAC operation. Ctrl: control mice. Tg779: Tnnt2-rtTA;Tre-Mhrt779 mice. P-value: Student’s t-test. Error bar: SEM.

b-e, Immunostaining of Brg1 (red) in adult heart ventricles 2 weeks after sham or TAC operation. Green: WGA. Blue: DAPI. Ctrl: control. Tg779: Tnnt2-rtTA;Tre-Mhrt779. Scale = 50 μm.

f, Western Blot analysis of Brg1 and Coomassie staining of total proteins in control or Tg779 hearts after 2 weeks of sham or TAC operation. Ctrl: control. Tg779: Tnnt2-rtTA;Tre-Mhrt779.

g, Quantitation of Myh6 and Myh7 in control and Tg779 hearts after 2 weeks of sham or TAC operation. Ctrl: control hearts. Tg779: Tnnt2-rtTA;Tre-Mhrt779 hearts. P-value: Student’s t-test. Error bar: SEM.

9

Extended Data Figure 8. Brg1 sequence alignment and motif analysis

Schematics of the architecture of mouse Brg1 and the sequence alignment of Brg1, Vasa (Fruit fly), Rad54 (Zebrafish, Sulfolobus solfataricus) and Chd1 (Yeast). The motifs were outlined by blue boxes (D1 domain) and purple boxes (D2 domain).

Extended Data Figure 9. Purification of Brg1 helicase core domains, EMSA of naked Myh6 promoter, ChIP and reporter studies in SW13 cells

a, Coomassie blue staining of purified MBP-tagged Brg1 helicase domains. BSA (bovine serum albumin) was loaded as a control. MBP: maltose binding protein. MBP-D1: Brg1 amino acids 774-913 fused to MBP. MBP-D2: Brg1 amino acids 1086-1310 fused to MBP. MBP-D1D2: Brg1 amino acids 774-1310 fused to MBP.

b, EMSA assay of naked Myh6 promoter (-426 to +170) with helicase domains of Brg1. Probe: biotin-labeled Myh6 promoter. 50 μM of MBP, MBP-D1, MBP-D2, and MBP-D1D2 proteins were used for EMSA.

c, d, ChIP (c) and luciferase reporter (d) analysis of Brg1 on chromatinized (episomal) and naked Myh6 promoter in SW13 cells. GFP: green fluorescence protein control. P-value: Student’s t-test. Error bar: SEM.

e, The luciferase reporter of helicase-deficient Brg1 on chromatinized (episomal) Myh6 promoter in SW13 cells. ΔD1: Brg1 lacking amino acid 774-913. ΔD2: Brg1 lacking amino acid 1086-1246. GFP: green fluorescence protein control. ChIP: H-10 antibody recognizing N-terminus, non-disrupted region of Brg1. P-value: Student’s t-test. Error bar: SEM.

10

Extended Data Figure 9. Purification of Brg1 helicase core domains, EMSA of naked Myh6 promoter, ChIP and reporter studies in SW13 cells

a, Coomassie blue staining of purified MBP-tagged Brg1 helicase domains. BSA (bovine serum albumin) was loaded as a control. MBP: maltose binding protein. MBP-D1: Brg1 amino acids 774-913 fused to MBP. MBP-D2: Brg1 amino acids 1086-1310 fused to MBP. MBP-D1D2: Brg1 amino acids 774-1310 fused to MBP.

b, EMSA assay of naked Myh6 promoter (-426 to +170) with helicase domains of Brg1. Probe: biotin-labeled Myh6 promoter. 50 μM of MBP, MBP-D1, MBP-D2, and MBP-D1D2 proteins were used for EMSA.

c, d, ChIP (c) and luciferase reporter (d) analysis of Brg1 on chromatinized (episomal) and naked Myh6 promoter in SW13 cells. GFP: green fluorescence protein control. P-value: Student’s t-test. Error bar: SEM.

e, The luciferase reporter of helicase-deficient Brg1 on chromatinized (episomal) Myh6 promoter in SW13 cells. ΔD1: Brg1 lacking amino acid 774-913. ΔD2: Brg1 lacking amino acid 1086-1246. GFP: green fluorescence protein control. ChIP: H-10 antibody recognizing N-terminus, non-disrupted region of Brg1. P-value: Student’s t-test. Error bar: SEM.

11

Extended Data Figure 10. Brg1 outruns Mhrt to bind to its target Mhrt promoter

a, Assembly of nucleosomes on the Mhrt promoter (a3/4).

b, Amylose pull-down assay: amylose was used to pull down the chromatinized Mhrt promoter that was incubated with various doses of MBP and MBP-Brg1 D1D2. DNA precipitated by amylose was further quantitated by qPCR. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

12

Extended Data Figure 11. Sequence alignment and secondary structure prediction of human and mouse MyHEART, and demography of heart transplantation donors

a, Sequence alignment of human MHRT and mouse Mhrt779.

b, Predicted secondary structure of mouse Mhrt779 and human MHRT, using minimal free energy (MFE) calculation of RNAfold WebServer.

c, Demography of heart donors whose tissues were used for RT-qPCR analysis (Fig. 4l). LVH: left ventricular hypertrophy. ICM: ischemic cardiomyopathy. IDCM: idiopathic cardiomyopathy.

2

Extended Data Figure 1. Mhrts have no or minimal coding potential

a, RNA in situ analysis of Mhrt (blue) in E12.0 heart. The RNA probe targets all Mhrt species. Red: nuclear fast red. Black arrowheads: nuclei of endothelial, endocardial or epicardial cells. Inset: magnified region from the boxed area. RA and RV: right atrium and ventricle; LV: left ventricle; IVS: interventricular septum; Endocardium (endo); epicardium (epi). Scale=100 μm.

b, Codon substitution frequency (CSF) scores of TfIIb and Hprt1 mRNA, as well as full-length Mhrt species. P-value: Student’s t-test. Error bar: SEM.

c, In vitro translation of control Mhrt species (709, 779, 826, 828, 857, 1147) and Luciferase (Luc). Arrow points to the protein product of Luc.

d, Biotin-labeling of Mhrt species (709, 779, 826, 828, 857, 1147) and Luciferase (Luc) in the in vitro translation reactions. Arrow points to the RNA product of Luc.

e, Ribosome profiling relative to whole transcriptome RNA sequencing. X-axis: genomic position at the human GAPDH and the murine Myh7 loci. Y-axis: mapped reads.

f, Scatter plot of RNA in fragments per kilobase per million reads (FPKM). Non-coding RNAs (purple) cluster towards the X-axis; coding RNAs (orange) towards the Y-axis. Mhrt779 locates below both the identity line (dashed, slope=1, intercept=0) and the smooth-fit regression line (in blue). RNA examples are endogenous except that HOTAIR was co-transfected with Mhrt779.

3

Extended Data Figure 2. Quantitation of Myh6/Myh7, Northern blot, and Mhrt779 characterization

a, Quantitation of cardiac Myh6/Myh7 ratio 2-42 days after sham or TAC operation. P-value: Student’s t-test. Error bar: SEM.

b, Northern blot analysis of Mhrt, Myh6 and Myh7. Negative: control RNA from 293T cells. Size control: 826 is recombinant Mhrt826; 643 (not a distinct Mhrt species) contains the 5′ common region of Mhrt. Heart: adult heart ventricles.

c, Un-cropped Northern blots of Mhrt, Myh6 and Myh7.

d, RNA in situ hybridization of Mhrt779 of adult heart ventricles. White arrowheads: nuclei of myocardial cells. Black arrowheads: nuclei of endothelial, endocardial or epicardial cells. Blue: Mhrt779; Red: nuclear fast red. Epi: epicardium. The dashed line separates the epicardium from myocardium. Scale = 50 μm.

e, Quantitation of TfIIb, Hprt1, 28S rRNA, Neat1 and Mhrt779 in nuclear and cytoplasmic fraction of adult heart ventricle extracts. The nuclear/cytoplasmic ratio of TfIIb is set as 1. P-value: Student’s t-test. Error bar: SEM.

4

Extended Data Figure 3. Wheat germ agglutinin staining, time course, molecular marker, studies of the stressed Tg779 mice

a, Wheat germ agglutinin (WGA) immunostaining 6 weeks after the sham or TAC operation. Green: WGA stain, outlining cell borders of cardiomyocytes. Blue: DAPI. Ctrl: control mice. Tg779: Tnnt2-rtTA;Tre-Mhrt779 mice. Scale = 50 μm.

b, Time course of fractional shortening (FS) in control and Tg779 mice. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

c, Quantification of Anf, Bnp, Serca2 and Tgfb1 in control and Tg779 mice 2 weeks after sham or TAC operation. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

d, Experimental design for treatment study and time course of left ventricular fractional shortening changes.

e, Fractional shortening of the left ventricle 8 week after the operation. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

f, Ventricular weight-body weight ratio of hearts harvested 8 weeks after sham or TAC operation.

5

Extended Data Figure 4. Regulation of the Mhrt promoter

a, Sequence alignment of Mhrt promoter loci from mouse, human and rat. Peak heights indicate degree of sequence homology. Black boxes (a1-a4) are sequences of high homology, which were used for further ChIP analysis. Green box region between Myh6 and Mhrt is the putative Mhrt promoter. Red: promoter regions. Salmon: introns. Yellow: untranslated regions.

b-d, ChIP-qPCR analysis of Mhrt promoter using antibodies against Pol II (b), H3K4me3 (c), and H3K36me3 (d) in tissues of adult mice. P-value: Student’s t-test. Error bar: SEM.

e, RT-qPCR quantitation of Mhrt in control and Brg1-null hearts after 7 days of TAC. Ctrl: control. Brg1-null: Tnnt2-rtTA;Tre-Cre;Brg1^f/f. P-value: Student’s t-test. Error bar: SEM.

f, Luciferase reporter assay of Mhrt promoter in SW13 cells. Ctrl: DMSO. TSA: trichostatin (HDAC inhibitor). PJ-34: PARP inhibitor. P-value: Student’s t-test. Error bar: SEM.

g, ChIP analysis of BRG1, HDAC2, HDAC9 and PARP1 in SW13 cells. The cells were transfected with episomal Mhrt promoter cloned in pRep4. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

h, Deletional analyses of the Mhrt promoter in luciferase reporter assays in SW13 cells. Luciferase activity of full-length Mhrt promoter were set up as 1. P-value: Student’s t-test. Error bar: SEM.

6

Extended Data Figure 5. Mhrt doesn’t affect Myh expression by direct RNA sequence interference

a, qPCR analysis of Mhrt779, Myh6 and Myh7 in mice without TAC operation. Expression levels were normalized to TfIIb, and the control is set as 1. Ctrl: control mice. Tg779: Tnnt2-rtTA;Tre-Mhrt779 mice. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

b, c, RNA quantification of Mhrt (b) and HOTAIR (c) in SW13 cells transfected with Vector (pAdd2), HOTAIR (pAdd2-HOTAIR) or Mhrt (pAdd2-Mhrt779). Expression in vector-transfected cells is set as 1. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

d, e, RNA quantification of Myh6 (d) and Myh7 (e) in SW13 cells relative to GAPDH. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

f, g, Western blot analysis of Myh6 (f) and Myh7 (g) in SW13 cells. Constructs containing Myh6 and Myh7 coding sequences were tagged with FLAG and co-transfected with Vector, HOTAIR or Mhrt779. GAPDH was used as the loading control. FLAG-D1 was used as a positive control for the FLAG antibody.

h, i, Protein quantification of Myh6 (h) and Myh7 (i) in control and transfected SW13 cells relative to GAPDH. Signals of Myh6 and Myh7 from major bands or the entire lanes were quantitated. P-value: Student’s t-test. Error bar: standard error of the mean (SEM).

j, Luciferase reporter assay of Mhy6 and Myh7 promoters in SW13 cells transfected with Vector (pAdd2) or Mhrt (pAdd2-Mhrt779). P-value: Student’s t-test. Error bar: SEM.

Figure 5.