RESEARCH ARTICLE SUMMARY◥

CELL BIOLOGY

E-C coupling structural proteinjunctophilin-2 encodes a stress-adaptivetranscription regulatorAng Guo, Yihui Wang, Biyi Chen, Yunhao Wang, Jinxiang Yuan, Liyang Zhang,Duane Hall, Jennifer Wu, Yun Shi, Qi Zhu, Cheng Chen, William H. Thiel,Xin Zhan, Robert M. Weiss, Fenghuang Zhan, Catherine A. Musselman,Miles Pufall, Weizhong Zhu, Kin Fai Au, Jiang Hong, Mark E. Anderson,Chad E. Grueter, Long-Sheng Song*

INTRODUCTION:Cardiacexcitation-contraction(E-C) coupling refers to acascadeofCa2+-mediatedevents whereby membrane depolarization leadsto cell contraction. At the subcellular level, E-Ccoupling occurs within a microdomain of thecardiomyocyte, termed the cardiac dyad. Invarious forms of heart disease, such as patholog-ical hypertrophy and heart failure, the E-C cou-pling process is abnormal, in part because ofultrastructural remodeling. Abnormal Ca2+

homeostasis (as a result of failed E-C coupling)triggers maladaptive remodeling at the tran-scriptional level, contributing to pathologicalmyocardial remodeling, hypertrophy, and heart

failure. However, it remains unclear whethercardiomyocytes possess a self-protective orhomeostatic mechanism that mitigates ad-verse myocardial remodeling.

RATIONALE: Junctophilin-2 (JP2) is a struc-tural protein that organizes the E-C couplingultrastructural machinery in cardiomyocytes.We previously showed that calpain-mediatedproteolytic cleavage of JP2 is key to its down-regulation in the diseased heart after car-diac stress. This cleavage contributes to lossof ultrastructural integrity at cardiac dyads,E-C uncoupling, and dysfunction of Ca2+ hand-

ling that results in heart failure. Computa-tional analyses predicted that JP2 contains anuclear localization signal (NLS), as well asan alanine-rich region (ARR) with character-istics of a helix-turn-helix structure, a DNAbinding motif. We tested the hypothesis thatJP2 encodes a stress-adaptive transcriptionalregulator, which transduces mechanical infor-mation (E-C uncoupling) into transcriptionalreprogramming in the myocardium in the

setting of cardiac stress.

RESULTS: Biochemical,mutagenesis, and confocalimaging analyses revealedthat stress-induced pro-teolysis of JP2 liberated

an N-terminal fragment (JP2NT) that wasimported into the nucleus through its NLS.Further biochemical and microarray assaysshowed that in the nucleus, JP2NT asso-ciated with chromatin and regulated tran-scription of a wide spectrum of genes via anevolutionarily conserved ARR located in thea-helix region of JP2. Chromatin immuno-precipitation sequencing (ChIP-seq) of JP2NT-overexpressing hearts revealed that it boundpreferentially to the transcription start sites(TSSs) of genes, and gel shift studies definedtheDNAbindingmotifs of JP2NT as the TATAbox and aMEF2-response element (MRE). Ele-vation of JP2NT levels by JP2NT overexpres-sion altered the in vivo genomic binding profileof TATA-boxbindingprotein (TBP) andMEF2C.In addition, JP2NT suppressed the transcrip-tional activity ofMEF2C by competing forMRE.Overexpression of JP2NT in mice led to repro-gramming of the transcriptome in the settingof stress and attenuated hypertrophic remod-eling and the progression of heart failure. Lossof JP2NT function by deletion of the JP2 NLSin mice accelerated the development of hyper-trophy and heart failure after cardiac stress.

CONCLUSION: Our data reveal that calpain-mediated cleavage of JP2 transforms this E-Ccoupling structural protein into a transcrip-tional regulator that is shuttled into the nu-cleus and binds to promoters of target genes,inducing cardioprotective transcriptional re-programming. These data reveal that cardio-myocytes possess a self-protective mechanismthat counters pathological transcriptional re-modeling after cardiac stress. Our findingsalso identify an intrinsic direct connectionbetween ultrastructural remodeling and tran-scriptional reprogramming in the stressedheart.▪

RESEARCH

Guo et al., Science 362, 1375 (2018) 21 December 2018 1 of 1

The list of author affiliations is available in the full article online.*Corresponding author. Email: long-sheng-song@uiowa.eduCite this article as A. Guo et al., Science 362, eaan3303(2018). DOI: 10.1126/science.aan3303

T-tubule

calpain

Ca2+

Overload

ARR NLS

DHPR

JP2

e

jSRjSR

Stresses

JP2NT

Nucleus

CytosolCytosol

Reprogramtranscription

RyR2RyR2

Sarcomere

Ca2+jSR 2

3

4

5

1

TATA MRE

Normal Condition Stressed Condition

Schematic depiction of the mechanism by which JP2NTconverts a mechanical stresssignal to transcriptional reprogramming in the stressed heart. Left: E-C coupling underphysiological conditions. Right: E-C coupling under pathologic conditions. Cardiac stressresults in Ca2+ overload (1), promoting calpain activation (2). The resulting cleavage ofJP2 liberates JP2NT from the T-tubule/sarcoplasmic reticulum junction, disrupting theultrastructure of the E-C coupling machinery (3). JP2NT is shuttled into the nucleus via aconserved NLS (4). JP2NT binds to TATA-box elements via the ARR and associates with aMEF2 response element (MRE) to repress the transcription of genes that control deleteriouscardiac remodeling (5).

ON OUR WEBSITE◥

Read the full articleat http://dx.doi.org/10.1126/science.aan3303..................................................

on March 27, 2020

http://science.sciencem

ag.org/D

ownloaded from

RESEARCH ARTICLE◥

CELL BIOLOGY

E-C coupling structural proteinjunctophilin-2 encodes a stress-adaptivetranscription regulatorAng Guo1, Yihui Wang1,2, Biyi Chen1, Yunhao Wang1, Jinxiang Yuan1, Liyang Zhang3,Duane Hall1, Jennifer Wu1, Yun Shi1, Qi Zhu1,4, Cheng Chen1,2, William H. Thiel1,Xin Zhan1, Robert M. Weiss1, Fenghuang Zhan1, Catherine A. Musselman3,Miles Pufall3, Weizhong Zhu4, Kin Fai Au1*, Jiang Hong2, Mark E. Anderson5,Chad E. Grueter1,6, Long-Sheng Song1,6,7†

Junctophilin-2 (JP2) is a structural protein required for normal excitation-contraction(E-C) coupling. After cardiac stress, JP2 is cleaved by the calcium ion–dependent proteasecalpain, which disrupts the E-C coupling ultrastructural machinery and drives heart failureprogression. We found that stress-induced proteolysis of JP2 liberates an N-terminalfragment (JP2NT) that translocates to the nucleus, binds to genomic DNA, and controlsexpression of a spectrum of genes in cardiomyocytes. Transgenic overexpression ofJP2NT in mice modifies the transcriptional profile, resulting in attenuated pathologicalremodeling in response to cardiac stress. Conversely, loss of nuclear JP2NT functionaccelerates stress-induced development of hypertrophy and heart failure in mutant mice.These data reveal a self-protective mechanism in failing cardiomyocytes that transducemechanical information (E-C uncoupling) into salutary transcriptional reprogrammingin the stressed heart.

Calcium ion (Ca2+) signaling affects almostevery aspect of cells (1). In heart muscle,excitation-contraction (E-C) coupling is acascade of Ca2+-mediated processes linkingmembrane depolarization to activation

of cell contraction (2). At the cellular level,E-C coupling in working ventricular myocytesdepends on precise communication betweenvoltage-gated L-type Ca2+ channels locatedmainly on the transverse tubule (T-tubule)membrane and Ca2+-sensitive ryanodine re-ceptors (RyRs) on the terminal cisternae of thesarcoplasmic reticulum (SR) (3–6). Upon mem-brane depolarization, Ca2+ influx through theopening of voltage-gated L-type Ca2+ channelsincreases local Ca2+ concentration. This highconcentration of Ca2+ sensitizes adjacent RyRsto release a much larger amount of Ca2+ from

the SR. The SR-released Ca2+ together with Ca2+

influx activates myofilaments, resulting in myo-cyte contraction. This intermolecular Ca2+ cross-talk between L-type Ca2+ channels and RyRstakes place in a confined spatial microdomain,where T-tubules and terminal cisternae of SRform tight junctional couplings with a gap of12 to 15 nm, termed “cardiac dyads” (7). Cardiacdyads provide the structural basis for E-C cou-pling and are established and maintained byjunctophilin-2 (JP2) (8). JP2 contains eightN-terminal MORN (membrane occupation andrecognition nexus) domains that mediate inter-actions with the T-tubule membrane, a space-spanning a helix that is thought to control thedyad distance, and a C-terminal transmembrane(TM) domain that anchors JP2 in the SR mem-brane (8, 9). Genetic manipulation of JP2 bysilencing, knockout, or overexpression authen-ticated its role as a structural protein respon-sible for the formation of cardiac dyads andmaintenance of normal E-C coupling in the heart(8, 10, 11).Defective E-C coupling is a hallmark of heart

failure (12–16). Recent studies have providedcompelling evidence that the expression levelof JP2 is decreased in failing hearts of multipleetiologies including human heart failure, con-tributing to the loss of ultrastructural integrityof cardiac dyads and E-C coupling dysfunction(16–22). In particular, we discovered that JP2proteolytic cleavage by calpain in response tocardiac stress represents a key mechanism of

JP2 down-regulation, causing E-C uncoupling,Ca2+ mishandling, and heart failure (23–25).Abnormal Ca2+ homeostasis triggers maladap-tive transcriptional remodeling, contributing topathological myocardial remodeling and devel-opment of heart failure (26–32). However, it wasnot clear whether cardiomyocytes undergoingE-C uncoupling possess a self-protective or ho-meostatic mechanism that mitigates adversemyocardial remodeling. It was also unknownwhether there is an intrinsic connection betweencardiac ultrastructural remodeling at E-C cou-pling junctions and transcriptional reprogram-ming in stressed hearts.Here, we show a mechanism in which a JP2

fragment, generated during cardiac stress and amarker of E-C uncoupling, serves as a negativefeedback mechanism to antagonize maladaptivecardiac remodeling. This fragment translocatesto the nucleus and represses transcriptional re-programming, in part through regulating a keymuscle transcription factor, MEF2 (myocyte en-hancer factor 2). Specifically, we found that thea-helix domain of JP2 contains an evolutionarilyconserved DNA binding domain. Under stressconditions, proteolytic processing of JP2 by calpainconverts it from a structural protein to a tran-scriptional regulator, indicating an intrinsic con-nection between cardiomyocyte ultrastructuralremodeling and transcriptional reprogrammingin the heart.

Nuclear localization of JP2NT

We previously reported that JP2 is a substrateof calpain (calpain 1) and identified the primarycalpain proteolysis site in the C-terminal region ofJP2 between residues Arg565 and Thr566 (24). Cal-pain cleavage creates anN-terminal truncate (resi-dues 1 to 565, termed JP2NT) that contains theplasma membrane–binding MORN motifs, and aC-terminal fragment containing the SR mem-brane–anchoring TM domain (Fig. 1A). We per-formed Western blotting with an antibody to aninternal epitope of JP2, which is not destroyedby calpain cleavage of JP2 (fig. S1A). Analysis ofsubcellular fractions of mouse myocardium es-tablished that endogenous JP2NT (75 kDa) ispresent and predominantly enriched in nuclearfractions (Fig. 1B and fig. S1, B and C). Immuno-stainings of human and mouse myocardiumsections using the same antibody also detecteda JP2 product in nuclei (fig. S2, arrows). Incontrast, an antibody to the C terminus of JP2did not detect JP2 signals in the nucleus of myo-cardium sections (fig. S2). JP2NT was markedlyincreased inmyocardium frommice with cardiac-specific overexpression of calpain1 and enrichedin the nuclear fraction (Fig. 1C) [for calpain1-OEmice, see (33)]. In addition, treatment with micro-coccal nuclease (MNASE), which cleaves DNA andreleases chromatin-associated proteins, releasedJP2NT from the chromatin pellet (Fig. 1C); thisresult substantiates the nuclear localization ofendogenous JP2NT in vivo and suggests thatJP2NT is a chromatin-associated protein.We hypothesized that pathological stresses

that activate calpain (34) promote the generation

RESEARCH

Guo et al., Science 362, eaan3303 (2018) 21 December 2018 1 of 9

1Department of Internal Medicine, Abboud CardiovascularResearch Center, Carver College of Medicine, University ofIowa, Iowa City, IA 52242, USA. 2Department of EmergencyMedicine, Shanghai General Hospital, Shanghai Jiao TongUniversity School of Medicine, Shanghai 200080, China.3Department of Biochemistry, Carver College of Medicine,University of Iowa, Iowa City, IA 52242, USA. 4Department ofPharmacology, School of Pharmacy, Nantong University,Nantong, Jiangsu 226001, China. 5Department of Medicine,Johns Hopkins School of Medicine, Baltimore, MD 21205,USA. 6Fraternal Order of Eagles Diabetes Research Center,Carver College of Medicine, University of Iowa, Iowa City,IA 52242, USA. 7Iowa City Veterans Affairs Medical Center,Iowa City, IA 52242, USA.*Present address: Department of Biomedical Informatics, OhioState University, Columbus, OH 43210.†Corresponding author. Email: long-sheng-song@uiowa.edu

on March 27, 2020

http://science.sciencem

ag.org/D

ownloaded from

and nuclear accumulation of JP2NT. Consistentwith this notion, we found that both isoprotere-nol infusion (Fig. 1D and figs. S1B and S2A) andmyocardial infarction (Fig. 1E, fig. S1C, and fig.S2, B and C) increased the amount of JP2NT innuclei of stressed hearts relative to (sham) con-trols. Under pressure overload stress, nuclear

accumulation of JP2NT reached its peak at 2to 3 weeks after transaortic banding (TAB)surgery (Fig. 1F). Conversely, administration ofthe calpain inhibitor MDL-28170 significantlyattenuated stress-induced elevation in nuclearJP2NT (Fig. 1, D to F), further supporting theidea that calpain-mediated proteolysis of full-

length JP2 under cardiac stress results in ac-cumulation of nuclear JP2NT.These data led us to postulate that the post-

translational removal of JP2 C terminus is suf-ficient to promote JP2NT translocation into thenucleus. To recapitulate the process by whichcalpain-mediated proteolysis is associated with

Guo et al., Science 362, eaan3303 (2018) 21 December 2018 2 of 9

eGFP-JP2TRS eGFP-JP2TRS + sTEVp

DMSO 1h

Rapamycin 100 nM 1h

ii

eGFP-JP2TRS eGFP-JP2TRS + sTEVp iv

G i

iii

TEVp-NT

TEVp-CTTEVp-NT

TEVp-CT

Split TEVp (sTEVp)

reconstitutionof proteaseactivity

eGFPeGFP-JP2TRS

Inserted TRS (ENLYFQ\S)

Rapamycin

FR

B

FK

BP

12

FR

B

FK

BP

12

+ + =

T566R565

Control ISO ISO+MDL-28170

Tot

al

Nu-

S

Chr

omat

in

Tot

al

Nu-

S

Chr

omat

in

JP2NT

RyR2

Cav1.2

GAPDH

H3

WT Calpain1-OE

JP2NT(75KD)

RyR2

Cav1.2

H3

Cyt

osol

Mem

bran

e

Nuc

lear

Pos

itive

con

trol

(Clo

ned

JP2N

T)

Pos

itive

con

trol

(Clo

ned

JP2N

T)

TM MORN I~VI MORN VII~VIII AVR565 T566GP

JP2NT JP2CT

Primary calpain cleavage site A

B C

Sham MI MI+MDL-28170

Pos

itive

con

trol

(C

lone

d JP

2NT

) H3

JP2NT(nuclear)

0.0

2.0

4.0

6.0

8.0

10.0***

Control ISO ISO + MDL-28170

JP2N

T/H

3

0.0

2.0

4.0

6.0

8.0

10.0****

***

*

Sham MI MI + MDL-28170

JP2N

T/H

3

D E

15 μm

F

Sha

m

TA

B 1

wk

TA

B 2

wk

TA

B 3

wk

TA

B 5

wk

TA

B 2

wk

+M

DL-

2817

0

JP2N

T-O

E

JP2N

T-O

E

Sham

TAB 1

wkTA

B 2wk

TAB 3

wkTA

B 5wk

TAB 2

wk

+ M

DL-28

170

JP2NT(nuclear)

H3

0

10

20

30

40

50

JP2N

T/H

3

v

Fig. 1. JP2 N-terminal truncate (JP2NT) accumulates in the nucleus of stressed hearts.(A) Schematic of JP2 and JP2 truncates. (B) JP2NT is primarily present in nuclear fractions ofmurine heart lysates. H3, histone H3 (nuclear marker); CaV1.2, L-type Ca2+ channel (membranemarker); RyR2, ryanodine receptor type 2. (C) Increased endogenous JP2NT in soluble (Nu-S)and chromatin-containing (chromatin) nuclear fractions from 3 weeks postnatal calpain1–overexpressing (OE) hearts. The chromatin fraction was derived by treating a chromatin pelletwith micrococcal nuclease (MNASE), which cleaves DNA and releases chromatin-associatedproteins (see also Fig. 2C). (D to F) Increased endogenous JP2NT levels in chronic cardiac stressmodels: (D) isoproterenol (ISO, 1 week) minipump infusion; (E) myocardial infarction (MI, 1 week);(F) TAB-induced ventricular pressure overload. Calpain inhibitor MDL-28170 attenuated orabolished the elevation of nuclear JP2NT in all these models. JP2NT-OE, transgenic mouse withcardiac specific expression of JP2NT. The overexpression level of JP2NT is in the same order ofmagnitude of the peak level of endogenous JP2NT induced by TAB. N = 3 or 4 for each group; *P <0.05, **P < 0.01 (t test or Kruskal-Wallis rank test when appropriate). (G) Analysis of JP2NTnuclear translocation using the rapamycin-inducible split tobacco etch virus protease (sTEVp)system. (i) Schematic of the sTEVp system. The N- and C-terminal fragments of TEV proteasewere fused to FRB and FKBP12, respectively. Rapamycin induces reconstitution of TEV proteasethrough the FKBP12-rapamycin-FRB complex. A TEVp substrate recognition sequence (TRS) wasinserted into the primary calpain cleavage site (Arg565/Thr566) of JP2 (eGFP-JP2TRS), allowingfor inducible and site-specific rapid cleavage of substrates at the TRS. (ii to v) eGFP-JP2TRS wastransfected into HEK293T cells alone (ii and iv) or with sTEVp system (iii and v), followed bytreatment with DMSO control (ii and iii) or rapamycin (100 nM) for 1 hour (iv and v).

RESEARCH | RESEARCH ARTICLEon M

arch 27, 2020

http://science.sciencemag.org/

Dow

nloaded from

JP2NT translocation, we adapted an induciblesplit tobacco etch virus protease (sTEVp) system(Fig. 1G, i) (35). A TEVp substrate recognitionsequence (TRS) was inserted into an enhancedgreen fluorescent protein–JP2 fusion in theprimary calpain cleavage site (Arg565/Thr566)(eGFP-JP2TRS; Fig. 1G, i). At baseline in hu-man embryonic kidney (HEK) 293T cells, eGFP-JP2TRS was localized at the cell membrane and

an intracellular network-like structure that islikely the endoplasmic reticulum (Fig. 1G, ii).In the absence of rapamycin, cotransfection ofsTEVp did not affect eGFP-JP2TRS localization(Fig. 1G, iii). In cells expressing sTEVp, rapamy-cin treatment rapidly induced nuclear importa-tion of eGFP-JP2TRS N terminus (Fig. 1G, v).These data show that JP2 C terminus anchorsthe intact JP2 protein at the dyad, and that re-

moval of JP2 C terminus is sufficient to inducetrafficking of the N-terminal fragment intonuclei.

JP2NT has a NLS and a chromatin/DNAbinding region

To investigate themolecular mechanism of JP2NTnuclear importation, we performed in silico anal-ysis (36). In JP2NT, we found a monopartite

Guo et al., Science 362, eaan3303 (2018) 21 December 2018 3 of 9

S3 S4 S4P4 P4

MNASE- MNASE+ ii

Cells

S1 P1

S3 P3

S4 P4

S3 = Soluble nuclear proteins

P3 =Chromatin-enriched + cytoskeletal proteins

S4 =Chromatin-associated proteins solubilized by DNA digestion

P4 =Cytoskeletal proteins + remaining chromatin proteins

Lysis buffer with 0.1% Triton X-100 and low speed centrifugation

Hypotonic nuclearlysis and low speedcentrifugation

Bi

MNASE digestionand centrifugation

JP2NTΔMORNs

JP2NTΔMORNs/ΔbNLS

H3

JP2NT

JP2NTΔbNLS/ΔARR

a

b

c

d

e

f

JP2NTΔMORNs/ΔbNLS/ΔARR

JP2NTΔMORNs/ΔbNLS/ΔARR

MORNI~VI

NLS (K488RPRP492)

MORNVII~VIII

K345RRVLPLKSSKVRQK359Alanine-rich region (ARR)A367QRAAAIARQKAEIAASRTSHAKAKAEAAEQAALAA402

JP2NT

Bipartite NLS (bNLS)

A eGFP To-Pro-3 OverlapeG

FP

eGF

P-J

P2

eGF

P-J

P2N

T

eGF

P-J

P2N

TΔN

LSeG

FP

-JP

2NT

ΔbN

LSeG

FP

-JP

2NT

ΔAR

R

i

ii

iii

iv

v

vi

OverlapTo-Pro-3eGFP

15 μm

or JP2(331-565)ΔbNLS/ΔARR

or JP2(360-565)

or JP2(400-565)

or JP2(331-565)

Fig. 2. JP2NT contains a NLS and a chromatin/DNA binding domain.(A) A conserved NLS is essential for nuclear accumulation of JP2NT. (i to iii)Subcellular localization of eGFP and eGFP-fused JP2 (eGFP-JP2) and JP2NT(eGFP-JP2NT) in HEK293Tcells. Full-length JP2 localizes on both plasmamembrane and ER network (ii). JP2NT is highly enriched in nuclei (iii). (iv) Deletionof NLS from JP2NT (JP2NTDNLS) abolished its nuclear localization and restrictedits localization on plasma membrane. (v and vi) A domain containing a bipartiteNLS (bNLS) and an alanine-rich region (ARR) is essential for colocalization ofJP2NTwith DNA (stained with To-Pro-3). eGFP-fused JP2NTmutants withoutthe bNLS [JP2NTDbNLS (v)] or without the alanine-rich domain [JP2NTDARR (vi)]

lost colocalization with DNA.(B) JP2NTassociates with chroma-tin. (i) Schematic of the subcellularfractionation approach [adaptedfrom (38)]. (ii) Subnucleardistribution of JP2NT and frag-ments/mutants. JP2NT is presentin both soluble nuclear (S3) andMNASE-releasable chromatinfractions (S4+MNASE). Deletion ofthe MORN domains (JP2NTDMORNs)had no effect on subnuclear distrib-ution of JP2NT. However, theamount of chromatin-associatedJP2NTwas decreased by deletion ofthe bNLS alone (JP2NTDMORNs/DbNLS)or by deletion of both the bNLSand ARR (JP2NTDMORNs/DbNLS/DARR

and JP2NTDbNLS/DARR). HistoneH3 serves as a chromatin marker.

RESEARCH | RESEARCH ARTICLEon M

arch 27, 2020

http://science.sciencemag.org/

Dow

nloaded from

nuclear localization signal (NLS) at positions488 to 492 (Lys-Arg-Pro-Arg-Pro) and a bipartiteNLS-like peptide (bNLS) at positions 345 to 359(Lys-Arg-Arg-Val-Leu-Pro-Leu-Lys-Ser-Ser-Lys-Val-Arg-Gln-Lys), adjacent to an alanine-richregion (ARR; Ala367 to Ala402) (fig. S3A) thatshows characteristics of a helix-turn-helix struc-ture [GYM 2.0 (37)]. These domains are evolu-tionarily conserved among species (fig. S3, Band C). Fusion of a short peptide containing thismonopartite NLS to mCherry resulted in nu-clear enrichment of the fusion proteins (fig.S3D). Deletion of this sequence (JP2NTDNLS)abolished nuclear localization of eGFP-JP2NTin HEK293T cells (Fig. 2A, iii and iv) and cardio-

myocytes (fig. S3E), indicating that this NLSis indispensable for nuclear localization ofJP2NT. A fusion protein containing mCherryand the bNLS peptide together with the ARRwas imported into nuclei (fig. S3D). However,deletion of the bNLS sequence from JP2NT(eGFP-JP2NTDbNLS) did not prevent its nuclearimportation in HEK293T cells (Fig. 2A, v) andcardiomyocytes (fig. S3E), indicating that thisregion is not necessary for nuclear importationof JP2NT. However, the subnuclear localiza-tion of eGFP-JP2NTDbNLS wasmutually exclusivefrom To-Pro-3 staining, which labels genomicDNA (Fig. 2A, v), suggesting physical dissocia-tion of eGFP-JP2NTDbNLS from genomic DNA.

Deletion of the adjacent ARR from JP2NT(eGFP-JP2NTDARR) induced greater separationof eGFP-JP2NT from DNA and was accompa-nied by accumulation of DNA at the nuclearperiphery (Fig. 2A, vi). These data indicate thatbNLS and ARR are involved in DNA or chro-matin binding.To further confirm the association of JP2NT

with chromatin, we applied a biochemical frac-tionation procedure (38) (Fig. 2B, i). JP2NT wasdetected in both soluble (S3) and chromatin-containing insoluble (P4) nuclear fractions (Fig. 2B,ii, a).MNASE-mediated DNA digestion releasedJP2NT from the insoluble chromatin fraction(MNASE+/S4) (Fig. 2B, ii, a). Deletion of the eight

Guo et al., Science 362, eaan3303 (2018) 21 December 2018 4 of 9

-2.5

Kb

2.5K

b

TT

S

TS

S

-2.5

Kb

2.5K

b

TT

S

TS

S

-2.5

Kb

2.5K

b

TT

S

TS

S

Genomic binding profile of JP2NT

Promoter

Exon

Intron

Intergenic

TSS

Other

A

JP2NTΔMORNs/ΔARR

GGGATCCTGAGTCGCAGTATAAAAGAAGCTTTTCGGGCGTT

GGGATCCTGAGTCGCAGTATATTAGAAGCTTTTCGGGCGTT

GGGATCCTGAGTCGCAGTATATAAGAAGCTTTTCGGGCGTT

GGGATCCTGAGTCGCAGTATTAAAGAAGCTTTTCGGGCGTTGGGATCCTGAGTCGCAGTATAAAAGAAGCTTTTCGGGCGTT

Freeprobes

Fre

epr

obes

Shiftedprobes

Shi

fted

prob

es

GST-JP2NT

Bla

nk

GST

200

ng

50 n

g

10

0 ng

200

ng

GST-JP2331-405

Bla

nk

GS

T

50 n

g

200

ng

100

ng

200

ng

GST-JP2331-405

Bla

nk

GST

50 n

g

200

ng

100

ng

200

ng

Consensus TATA BOX(promoter

cMyc)Mutant TATA BOX

(promoter cMyc)

WT TATA BOX(promoter

cMyc)Mutant TATA BOX

(promoter cMyc)

WT TATA BOX (promoter cMyc)

D i ii

iv v

Bla

nkG

ST

200

ng

50 n

g

100

ng

200

ng

Bla

nk

GST

200

ng

50 n

g

100

ng

200

ng

GST- GST-JP2NTΔMORNs

GST-JP2NT

Bla

nk

GST

200

ng

50 n

g

10

0 ng

200

ng

iiiTATA box core

sequence variationTATAAATATAATTATATATATATTTATTAATTTAAA

Bindingby JP2NT

yesyesyesnonono

CJP2NT

(p-value<10-10)

(p-value<10-10)

471 8943 22354

H3K9ac

JP2NT

2109 7305 19889

Pol II

0123456789

10

ChI

P-s

eq s

igna

lC

hIP

-seq

sig

nal d

ensi

ty

B Input_mm10 IP replicate1_mm10 IP replicate2_mm1018

14

10

6

2

-2.5

Kb

2.5K

b

TT

S

TS

S

-2.5

Kb

2.5K

b

TT

S

TS

S

-2.5

Kb

2.5K

b

TT

S

TS

S

Fig. 3. JP2NT is a TATA-box bindingprotein enriched at transcription startsites and interacts with basic tran-scription machinery. (A) Genomic DNAbinding profile of JP2NT in cardiomyocytesas revealed by ChIP-seq. (B) JP2NT ispreferentially localized around transcrip-tion start sites (TSSs), as revealed by tworeplicates of ChIP-seq. (C) Overlap chart ofDNA binding peaks of indicated proteins.(D) JP2NT binds to TATA box DNA sequencesin vitro. (i and ii) Gel shift assays of purifiedGST-JP2NT binding to wild-type(WT) (i) or mutant TATA box–containingsequences (ii) derived from the cMycpromoter. (iii) Summary of the results ofgel shift assays with various TATA boxvariants or mutants. Mutation of the coreTATA sequences abolished the interactionwith GST-JP2NT. (iv) Deletion of the ARR,but not the N-terminal MORN domains,eliminated JP2NT binding to the TATA boxsequence. (v) The peptide JP2331-405

containing the ARR specifically binds tothe consensus TATA box sequence.

RESEARCH | RESEARCH ARTICLEon M

arch 27, 2020

http://science.sciencemag.org/

Dow

nloaded from

MORNdomains from JP2NT (JP2NTDMORNs) didnot influence its distribution in the nucleus or itsassociation with chromatin (Fig. 2B, ii, b). Incontrast, deletion of the bNLS-like sequencefrom this construct (JP2NTDMORNs/DbNLS) signifi-cantly reduced the association of JP2NT withchromatin (Fig. 2B, ii, c). Deletion of the alanine-rich domain in combination with the bNLS(JP2NTDMORNs/DbNLS/DARR or JP2NTDbNLS/DARR)completely prevented presence of JP2NT inthe MNASE-releasable chromatin fraction (Fig.2B, ii, d to f). On the basis of these data, weconclude that JP2NT associates with chromatinvia a domain located at residues ~345 to 402,which is highly evolutionarily conserved in mam-

malian species as well as in vertebrates such asfish and birds (fig. S3C).

JP2NT is enriched at transcriptionstart sites

To systematically study the genomic targets ofJP2NT, we generated transgenic mice with cardiac-specific overexpression of hemagglutinin (HA)–tagged JP2NT. In these mice, JP2NT is predomi-nantly localized in the nuclei of cardiomyocytes(fig. S4A). The level of JP2NT transgene in nu-clear fraction is of the same order of magnitude,relative to the peak level of endogenous JP2NTinduced by TAB (Fig. 1F). Hearts of the JP2NT-overexpressing mice (JP2NT-OE) were subjected

to chromatin immunoprecipitation sequenc-ing (ChIP-seq) analysis using antibody to HA.Two replicates of ChIP-seq analyses were per-formed using different batches of JP2NT heartsamples. Only the DNA peaks (P < 10−10) thatwere detected in both replicates of ChIP-seqanalyses were considered as JP2NT-binding DNAregions. We identified 9414 JP2NT-binding ge-nomic DNA regions encompassing 7398 genes.The DNA binding profile revealed that JP2NTis concentrated in gene-enriched regions, espe-cially the promoter and 5′ untranslated regions(Fig. 3A). Moreover, JP2NT is preferentially en-riched at transcription start sites (Fig. 3B), acharacteristic of transcription regulators. We

Guo et al., Science 362, eaan3303 (2018) 21 December 2018 5 of 9

Fig. 4. JP2NT represses MEF2-mediated transcription by competingfor the MEF2 response element (MRE). (A) Enrichment of MEF2 bindingmotifs in ChIP-seq dataset. (B) Gel shift assay of JP2NT binding to adesmin enhancer–derived DNA sequence containing a WTor mutant MREin vitro. (C) Coimmunoprecipitation of HA-tagged JP2NT binding to MEF2Cor Histone H3 in HEK293T cells transfected with JP2NT and MEF2C as wellas in JP2NT-OE hearts (postnatal 2 weeks). (D and E) MEF2 activity assays in

HEK293T cells cotransfected with PMEF2-firefly, PSV40-renilla, MEF2C,and WT JP2NT (D) or a mutant lacking the ARR (E, JP2NTDARR) (n ≥ 3independent batches of experiments, three transfection replicates in-cluded in each batch; **P < 0.01 versus MEF2C+empty pcDNA). (F) Overlapchart of MEF2C (left) and TBP (right) DNA binding peaks comparedbetween control and JP2NT-OE hearts, as well as compared with JP2NTbinding peaks from JP2NT-OE hearts.

RESEARCH | RESEARCH ARTICLEon M

arch 27, 2020

http://science.sciencemag.org/

Dow

nloaded from

compared JP2NT occupancy of genomic lociwith that of cardiac polymerase II (Pol II) andacetylated histone H3 Lys9 (H3K9ac, a markerfor active promoters and enhancers) bindingpeaks from published data (39). We found that95% of JP2NT peaks overlap with H3K9ac peaksand 78% overlap with Pol II peaks (Fig. 3C;example tracks shown in fig. S4B, a). These datastrongly suggest that JP2NT regulates activecardiac transcription.

JP2NT is a TATA-box binding proteinand interacts with transcription machinery

On the basis of these results, we hypothesizedthat JP2NT may directly associate with core cis-regulatory elements that regulate transcriptioninitiation. Cross-linking reversal coimmuno-precipitation experiments in JP2NT transgeniccardiomyocytes demonstrated that JP2NT as-sociates with RNA Pol II (RPB1) and TATA-boxbinding protein (TBP), both of which are com-ponents of the basic transcriptional machinery(fig. S5A). TBP specifically binds to TATA boxes,eukaryotic core cis-regulatory elements localizedat transcription start sites. Subsequent in vitroanalysis with purified recombinant glutathioneS-transferase (GST)–JP2NT revealed that JP2NTdirectly binds to theTATAboxor variants (TATAAA,TATAAT, and TATATA) from the cMyc (Fig. 3D,i, and fig. S5B) or the CMV promoter (fig. S5C).This interaction was abrogated by mutation ofthe TATA-box elements (Fig. 3D, ii, and fig. S5B).We conclude that JP2NT is a DNA binding pro-tein, binding to the consensus TATA box repre-sented as TATAA(A/T) or TATATA (Fig. 3D, iii).Deletion of N-terminal MORN domains alone

from JP2NT (GST-JP2NTDMORNs) did not alterthe interaction of JP2NT with the TATA-boxoligonucleotide (Fig. 3D, iv, and fig. S5C). How-ever, deletion of the ARR from this construct(GST-JP2NTDMORNs/DARR) completely abrogatedthe association of JP2NTwith TATA-box elements(Fig. 3D, iv, and fig. S5C). Conversely, a purifiedpeptide containing the ARR (GST-JP2331-405)specifically bound to the consensus but not themutant TATA box (Fig. 3D, v, and fig. S5D).Together, these data indicate that the ARR isresponsible for TATA-box binding.

JP2NTmodulates MEF2-mediatedtranscription via competingfor MEF2 binding sites

The MEF2 family, master regulators of hyper-trophic genes in cardiomyocytes, binds to the A/Tenriched consensus sequence (C/T TA(A/T)4TAG/A), which shares the same core sequence withthe TATA box. Thus, we hypothesized that JP2NTdirectly interacts with the MEF2 response ele-ment (MRE). Consistent with this hypothesis,MEF2 bindingmotifs were significantly enrichedin the ChIP-seq dataset (Fig. 4A and table S1). Gelshift assay demonstrated that purified JP2NT orpurified DNA binding domain of JP2NT (GST-JP2331-405) interacts with aMRE from the desminenhancer (Fig. 4B and fig. S6A). The specificity ofthis interaction was further shown by mutantMRE oligos (Fig. 4B) and cold competitor assay

Guo et al., Science 362, eaan3303 (2018) 21 December 2018 6 of 9

Fig. 5. JP2NTdrives broad-spectrum transcriptional reprogramming in cultured cardiomyo-cytes. (A) Heat map of significantly altered genes in cultured adult murine cardiomyocytesexpressing JP2 or JP2NT by adenovirus (Ad). (B) IPA pathway enrichment analysis ofsignificantly altered transcripts induced by JP2NT. (C) RT-qPCR validation of genes that weresignificantly down-regulated by JP2NT as compared to Ad-Empty control. Note that deletionof the ARR (JP2NTDARR) prevented JP2NT-mediated transcriptional repression. Data werecalculated as the log2 fold change relative to cells transfected with Ad-Empty. Each transcriptwas assayed in n ≥ 4 batches of cells. (D) Transcriptional activity assays in which luciferaseis under the control of the indicated promoters [n ≥ 3 independent batches of HEK293T cells,three transfection replications included in each batch; *P < 0.05, **P < 0.01 versus control(empty virus or plasmid)]. Under each bar graph are the JP2NT binding tracks at these gene lociand location of cloned promoters for luciferase construct, respectively.

RESEARCH | RESEARCH ARTICLEon M

arch 27, 2020

http://science.sciencemag.org/

Dow

nloaded from

(fig. S6B). Biotinylated oligo DNA pull-downassay confirmed that the JP2NT transgene aswell as endogenous full-length JP2 from heartlysis can interact with MRE (fig. S6C). Coimmu-noprecipitation of HEK293T cells transfectedwith Myc-tagged MEF2C and HA-tagged JP2NTdemonstrated an interaction of the two proteins(Fig. 4C). The interaction of endogenous MEF2Cwith JP2NT transgene was confirmed in JP2NTtransgenic hearts (Fig. 4C).To examine whether JP2NT regulates MEF2-

mediated transcription, we used a luciferase re-porter systemwith firefly cDNA driven by desminenhancer–derived MRE (PMEF2-firefly) (40). Co-transfection of a plasmid expressing MEF2Cand PMEF2-firefly in HEK293T cells significantlyincreased the firefly luciferase signal relative toconstitutive PSV40-renilla (Fig. 4, D and E). Co-transfection of JP2NT attenuated the MEF2-responsive signal in a dose-dependent manner(Fig. 4D). By contrast, MEF2C-mediated tran-scriptional activity was not significantly alteredin cells expressing a JP2NT construct lackingthe ARR (JP2NTDARR, Fig. 4E), which we foundto be required for its association with chromatinand TATA box sequences. These data suggest thatJP2NT competes with MEF2 for direct interac-tion with its consensus sequence at promotersto block MEF2-mediated transcription.

JP2NT alters DNA binding profilesof TBP and MEF2C in cardiomyocytes

To understand whether JP2NT influences DNAoccupancy of MEF2 in cardiomyocytes, weanalyzed the DNA binding profile of MEF2Cin control and JP2NT-OEmyocardium. ChIP-seqanalyses revealed that JP2NT-OE suppressedMEF2C interactions at 42% of the endogenousMEF2C binding sites. JP2NT-OE created 2386new MEF2C binding peaks that are absent incontrol hearts, and 65% of them overlap withJP2NT binding sites. In addition, 33% of theMEF2C binding peaks inhibited by JP2NT-OEoverlap with JP2NT binding sites (Fig. 4F). ChIP-seq analyses of TBP revealed that ~25% of theendogenous TBP binding sites were suppressedby JP2NT-OE, and 22% of them overlap withJP2NT binding sites. JP2NT-OE created 2192new TBP binding peaks, and 28% of them over-lap with JP2NT binding sites (Fig. 4F; exampletracks shown in fig. S4B, b to e). These data in-dicate that JP2NT can affect MEF2C and TBPDNA binding, either by competing for the en-dogenous binding sites or by recruitingMEF2Cand TBP to new binding sites.

Overexpression of JP2NT in culturedcardiomyocytes induces profoundchanges in transcriptional profile

The association of JP2NT with DNA and tran-scription machinery led us to investigate whetherJP2NT directly modulates the transcriptionalprofile in cardiomyocytes. Compared with car-diomyocytes infected with empty adenovirus(Ad-Empty control), Affymetrix GeneChip analy-sis revealed that the expression of 574 and 1996known genes were significantly induced or

repressed (P < 0.01), respectively, in JP2NT-expressing cardiomyocytes (Fig. 5A and fig.S7A; see table S2 for complete list of differentiallyexpressed genes). Conversely, only 96 signifi-cantly induced and 264 significantly repressedgenes were detected in cardiomyocytes withoverexpression of full-length JP2 (Fig. 5A andfig. S7B). Of the differentially expressed genesinduced by JP2NT, ~60% mapped to genomicloci where JP2NTwas found to bind by ChIP-seq;this proportion is significantly larger than the~46% (P < 10−15) of JP2NT binding genes amongthose genes whose expression was detectable incardiomyocytes but was not significantly influ-enced by JP2NT. Notably, the regulatory effectof JP2NT on gene expression appeared to be de-pendent onbinding of JP2NT to gene loci: Amongthe differentially expressed genes for which JP2NTbound to the genomic loci, the vast majority (84%)were down-regulated in the presence of JP2NToverexpression; this percentage is decreased to67% (P < 10−15) in differentially expressed genesnot bound by JP2NT. We interpret these findingsto indicate that JP2NT represses transcriptionby binding to genomic regions, either directlythroughbinding toTATAbox andMREor throughinteractions with transcription factors such asMEF2C and TBP.Many of the JP2NT–down-regulated genes en-

code nuclear proteins (fig. S7C) and proteins thatare functionally enriched in nuclear events suchas transcriptional regulation and chromatinmodification (fig. S7D). Ingenuity Pathway Anal-ysis (IPA; Qiagen) of the differentially expressedgenes induced by JP2NT identified pathwaysand regulators implicated in cardiac hypertro-phy, fibrosis, cell growth, and differentiation aswell as inflammation. Specifically, ERK/MAPK,NF-kB, TGF-b, and integrin signaling pathwayswere predicted to be inhibited in response toJP2NT overexpression (Fig. 5B and table S3).Confirming these GeneChip findings, reversetranscription quantitative polymerase chain re-action (RT-qPCR) revealed that mRNA levels ofgenes including those that encode KLF4, KLF6,Myc, TGFbR1, NFKBIA, FOXO1, and PI3KR1 weresignificantly decreased in cardiomyocytes express-ing JP2NT relative to Ad-Empty infected cells(Fig. 5C). Deletion of the DNA binding regionfrom JP2NT (JP2NTDbNLS/DARR) largely preventedthe repressive effect of JP2NT (Fig. 5C). No-tably, the cardiac hypertrophy markers ANP andBNP were not altered by JP2NT expression butwere significantly increased by JP2NTDbNLS/DARR

(Fig. 5C).To further test whether JP2NT regulates tran-

scription of these genes, we constructed lucifer-ase reporters controlled by promoters of Myc,KLF6, TGFbR1, and NFKBIA. For all genes, pro-moter activity was significantly attenuated bycoexpression of JP2NT in HEK293T cells (Fig.5D). Consistent with the changes in mRNA lev-els, expression of JP2NTDbNLS/DARR either hadno effect on baseline firefly luciferase signal orincreased promoter activity as compared toempty vector control (Fig. 5D), supporting theidea that the DNA binding domain of JP2NT is

important for its function as a transcriptionalrepressor.

JP2NTattenuates hypertrophic responseand heart failure development in mice

Because JP2NT represses transcription of keyregulators of hypertrophy, fibrosis, and inflam-mation, we predicted that JP2NT would exert aprotective effect on stress-induced pathologi-cal cardiac remodeling. At baseline, JP2NT over-expression had no effect on cardiac morphologyor function (Fig. 6, A to C). The E-C coupling func-tionat the single-cell level (e.g., L-typeCa2+ channeldensities, amplitudeandkinetics ofCa2+ transients,SR Ca2+ content, and the gain function of E-Ccoupling)was not altered by the JP2NT transgene(fig. S8). Under stress conditions induced by TABsurgery to produce pressure overload hypertrophyand heart failure, JP2NT-OE mice had improvedcardiac function (Fig. 6A), lower heartweight/bodyweight ratio (Fig. 6B), and reduced lung edemaindicated by the lung weight/body weight ratio(Fig. 6C) relative to controls. These results sug-gest that JP2NT-OE protects the heart againststress-induced pathological remodeling.RNA-seq demonstrated minor differences in

the cardiac transcription profile of JP2NT-OEmice at baseline relative to control littermates(fig. S9A), with only 220 significantly altered genes(P< 0.01). TAB promoted amarked change in thetranscriptome of control hearts as compared tosham surgery, with 4636 significantly altered tran-scripts derived from 3580 genes (fig. S8B). Over-expression of JP2NT significantly modified thetranscriptional response to cardiac stress: We de-tected a significant difference in 1082 transcriptsderived from 954 known genes based on a linearregression model (P < 0.01, Fig. 6D). Among these,540 transcripts (mapped to 481 knowngenes)werenegatively influenced and 542 transcripts (mappedto476knowngenes) positively influencedbyJP2NToverexpression (see table S4), with a predicted in-hibition of ERK, TGF-b, CREB, and NF-kB signal-ing pathways (Fig. 6E and table S5). These findingsare in line with observations in cultured cardio-myocytes (Fig. 5B) and substantiate a pivotal rolefor JP2NT in the cardiac response to stress byinhibiting transcriptional reprogramming.

Loss of function of nuclear JP2NTexacerbates cardiac dysfunction inresponse to stress

To investigate the function of endogenous JP2NT,we developed a knock-in (KI) mouse line withNLS deleted in the JP2 coding sequence (JP2DNLS-KI) (fig. S10A). NLS deletion abolished the nu-clear accumulation of JP2NT after cardiac stress(fig. S10B). The homozygous JP2DNLS-KImice showno difference in cardiac morphology, contract-ile function, and cellular E-C coupling / Ca2+

handling function compared to wild-type litter-mates under baseline condition (Fig. 6, F to I,and fig. S10, C to E). However, when subjectedto TAB, JP2DNLS-KI mice developed more se-vere cardiac hypertrophy and worsened heartfunction relative to wild-type littermates. Threeweeks after TAB, echocardiography detected

Guo et al., Science 362, eaan3303 (2018) 21 December 2018 7 of 9

RESEARCH | RESEARCH ARTICLEon M

arch 27, 2020

http://science.sciencemag.org/

Dow

nloaded from

larger myocardium mass (Fig. 6F), higher enddiastolic volume (EDV) (Fig. 6H) and end systolicvolume (ESV) (Fig. 6I), and lower ejection frac-tion (Fig. 6G) in JP2DNLS-KImice than in wild-typelittermates. These findings suggest that JP2NTfunctions as an endogenous cardiac protectoragainst pathological challenges.

Discussion

Our study provides compelling evidence suggest-ing that an E-C coupling structural protein can

also act as a transcriptional regulator. We haveshown that regulated cleavage of JP2 converts itfrom a structural protein to a nuclear transcrip-tional regulator via an NLS and an ARR con-tained within JP2NT. JP2NT is enriched in thepromoter region of genes in cardiomyocytes andprimarily acts as a transcriptional repressor ofgenes implicated in cell growth and differentia-tion, hypertrophy, inflammation, and fibrosis,with evidence for a specific interaction withthe transcription factor MEF2. Cardiac-specific

transgenic overexpression of JP2NT attenuatespressure overload induced development of heartfailure, identifying JP2NT generation as a self-protective homeostatic mechanism that safe-guards against the deleterious effects of cardiacstress. These discoveries reveal a signaling path-way that transduces membrane stresses intotranscriptome changes in the setting of E-C un-coupling after cardiac stress.JP2 was initially discovered as a structural

protein with dual membrane-anchoring domains

Guo et al., Science 362, eaan3303 (2018) 21 December 2018 8 of 9

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

HW

/BW

Rat

io (

mg/

g)

Eje

ctio

n F

ract

ion

(EF

)

Baseline

**** **

TAB

Control JP2NT-OE Control JP2NT-OE

Baseline TAB

Control JP2NT-OE Control JP2NT-OE

Baseline TAB

Control JP2NT-OE Control JP2NT-OE0.0

0.2

0.4

0.6

0.8

1.0

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

LW/B

W R

atio

(m

g/g)

A CB

EGF

-4.5 -2.5 -0.5

IGF1

FGF2

VEGF

ERK1/2

TGF-β

CREB1

HIF1AE

Z-score

F

0.0

50.0

100.0

150.0

200.0

Mas

s (m

g)

WT

WT

Baseline TAB JP

2ΔNLS -K

I

JP2ΔNLS -K

I

*

WT

WT

Baseline TAB JP

2ΔNLS -K

I

JP2ΔNLS -K

I0.0

25.0

50.0

75.0

100.0

EF

*G

WT

WT

Baseline TAB JP

2ΔNLS -K

I

JP2ΔNLS -K

I0.0

30.0

60.0

90.0

120.0

ED

V (

um)

*H

JP2ΔNLS -K

I

JP2ΔNLS -K

IW

TW

T

Baseline TAB

0

15

30

45.0

60.0

ES

V(u

m)

*I

-log 10

(p v

alue

)

SignificantFalseTrue

20

15

10

5

0

-8 -4 0 4

Coefficient

D

Fig. 6. JP2NToverexpression protects against pressure overload-induced hypertrophy and heart failure. (A) Cardiac specific over-expression of JP2NT (JP2NT-OE) preserved cardiac ejection fraction (EF)in mice 3 weeks after transverse aortic banding (TAB). (B) JP2NToverexpression attenuated TAB-induced cardiac hypertrophy as evidencedby a decreased heart weight/body weight (HW/BW) ratio. (C) Lungweight/body weight (LW/BW) ratio is significantly reduced in JP2NT-OEmice after TAB; n = 5, 5, 22, 13 for each group respectively. (D) Volcano

plot of the effect of JP2NT overexpression on TAB-induced transcriptionalremodeling. JP2NT-OE significantly changed the transcriptional responseto TAB. Red dots indicate the transcripts whose response to TAB wassignificantly changed by JP2NT-OE. (E) IPA pathway enrichment analysisof significantly altered transcripts in JP2NT-OE mice after TAB ascompared to littermate controls. (F to I) Echocardiogram measurement ofheart mass (F), ejection fraction (G), EDV (H), and ESV (I) in WT andJP2DNLS-KI mice under baseline or TAB condition.

RESEARCH | RESEARCH ARTICLEon M

arch 27, 2020

http://science.sciencemag.org/

Dow

nloaded from

that connect T-tubules and the SRmembrane (8).Here, we discovered that JP2 contains additionalregulatory domains that extend beyond its role asa structural protein. An NLS in the N-terminalregion of JP2 is necessary for nuclear import ofthe calpain-generated JP2NT truncate. Thus,under stress conditions, calpain-mediated cleav-age of JP2 serves two purposes: (i) It impairs thebridging of T-tubules with the SR membrane[contributing to cardiomyocyte ultrastructuralremodeling and E-C uncoupling (23)], and (ii) itliberates JP2NT, allowing JP2NT to translocateto the nucleus and mediate transcriptional re-programming. In addition, we found that thea-helix region of JP2 contains a previously un-appreciated DNA binding domain that medi-ates selective binding to canonical TATA boxmotifs and MRE. This DNA binding domain isevolutionarily conserved, suggestive of a dualfunction for JP2 as a structural protein and tran-scriptional regulator in other species.The development and progression of heart

failure involves diverse cellular and molecularmechanisms (41, 42). Our ChIP-seq and tran-scriptomic profiling data suggest that JP2NTsuppresses gene transcription by targetingmulti-ple signaling pathways such as inflammatory re-sponses, fibrosis, myocyte hypotrophy, and celldeath among others. Taken with the protectiveeffect of JP2NT overexpression in the setting ofcardiac stress, this study indicates that JP2NT isan endogenous self-protective stress transducerthat conveys the E-C uncoupling signal to the nu-cleus, regulates transcriptional reprogramming,and ultimately attenuates the progression of heartfailure. As JP2 is abundant in all muscle cells(cardiac, skeletal, and smooth muscle), JP2NTmay serve as a general protective mechanismantagonizing stress-induced pathological remod-eling related to many diseases.

REFERENCES AND NOTES

1. D. E. Clapham, Calcium signaling. Cell 131, 1047–1058 (2007).doi: 10.1016/j.cell.2007.11.028; pmid: 18083096

2. D. M. Bers, Cardiac excitation-contraction coupling. Nature415, 198–205 (2002). doi: 10.1038/415198a; pmid: 11805843

3. H. Cheng, W. J. Lederer, M. B. Cannell, Calcium sparks:Elementary events underlying excitation-contraction couplingin heart muscle. Science 262, 740–744 (1993). doi: 10.1126/science.8235594; pmid: 8235594

4. M. B. Cannell, H. Cheng, W. J. Lederer, The control of calciumrelease in heart muscle. Science 268, 1045–1049 (1995).doi: 10.1126/science.7754384; pmid: 7754384

5. J. R. López-López, P. S. Shacklock, C. W. Balke, W. G. Wier,Local calcium transients triggered by single L-type calciumchannel currents in cardiac cells. Science 268, 1042–1045(1995). doi: 10.1126/science.7754383; pmid: 7754383

6. S. Q. Wang, L. S. Song, E. G. Lakatta, H. Cheng, Ca2+ signallingbetween single L-type Ca2+ channels and ryanodine receptorsin heart cells. Nature 410, 592–596 (2001). doi: 10.1038/35069083; pmid: 11279498

7. E. Page, M. Surdyk-Droske, Distribution, surface density, andmembrane area of diadic junctional contacts between plasmamembrane and terminal cisterns in mammalian ventricle.Circ. Res. 45, 260–267 (1979). doi: 10.1161/01.RES.45.2.260;pmid: 376173

8. H. Takeshima, S. Komazaki, M. Nishi, M. Iino, K. Kangawa,Junctophilins: A novel family of junctional membrane complexproteins. Mol. Cell 6, 11–22 (2000). pmid: 10949023

9. M. Nishi, A. Mizushima, Ki. Nakagawara, H. Takeshima,Characterization of human junctophilin subtype genes.

Biochem. Biophys. Res. Commun. 273, 920–927 (2000).doi: 10.1006/bbrc.2000.3011; pmid: 10891348

10. R. J. van Oort et al., Disrupted junctional membrane complexesand hyperactive ryanodine receptors after acute junctophilinknockdown in mice. Circulation 123, 979–988 (2011).doi: 10.1161/CIRCULATIONAHA.110.006437; pmid: 21339484

11. A. Guo et al., Overexpression of junctophilin-2 does not enhancebaseline function but attenuates heart failure development aftercardiac stress. Proc. Natl. Acad. Sci. U.S.A. 111, 12240–12245(2014). doi: 10.1073/pnas.1412729111; pmid: 25092313

12. A. M. Gómez et al., Defective excitation-contraction coupling inexperimental cardiac hypertrophy and heart failure. Science276, 800–806 (1997). doi: 10.1126/science.276.5313.800;pmid: 9115206

13. S. E. Litwin, D. Zhang, J. H. Bridge, Dyssynchronous Ca2+ sparksin myocytes from infarcted hearts. Circ. Res. 87, 1040–1047(2000). doi: 10.1161/01.RES.87.11.1040; pmid: 11090550

14. L. S. Song et al., Orphaned ryanodine receptors in the failingheart. Proc. Natl. Acad. Sci. U.S.A. 103, 4305–4310 (2006).doi: 10.1073/pnas.0509324103; pmid: 16537526

15. M. Xu et al., Intermolecular failure of L-type Ca2+ channel andryanodine receptor signaling in hypertrophy. PLOS Biol. 5, e21(2007). doi: 10.1371/journal.pbio.0050021; pmid: 17214508

16. A. Guo, C. Zhang, S. Wei, B. Chen, L. S. Song, Emergingmechanisms of T-tubule remodelling in heart failure.Cardiovasc. Res. 98, 204–215 (2013). doi: 10.1093/cvr/cvt020;pmid: 23393229

17. S. Wei et al., T-tubule remodeling during transition fromhypertrophy to heart failure. Circ. Res. 107, 520–531 (2010).doi: 10.1161/CIRCRESAHA.109.212324; pmid: 20576937

18. M. Xu et al., Mir-24 regulates junctophilin-2 expression incardiomyocytes. Circ. Res. 111, 837–841 (2012). doi: 10.1161/CIRCRESAHA.112.277418; pmid: 22891046

19. H. D. Wu et al., Ultrastructural remodelling of Ca2+ signallingapparatus in failing heart cells. Cardiovasc. Res. 95, 430–438(2012). doi: 10.1093/cvr/cvs195; pmid: 22707157

20. M. Jiang et al., JPH-2 interacts with Cai-handling proteins andion channels in dyads: Contribution to premature ventricularcontraction-induced cardiomyopathy. Heart Rhythm 13, 743–752(2016). doi: 10.1016/j.hrthm.2015.10.037; pmid: 26538326

21. S. Minamisawa et al., Junctophilin type 2 is associated withcaveolin-3 and is down-regulated in the hypertrophic anddilated cardiomyopathies. Biochem. Biophys. Res. Commun.325, 852–856 (2004). doi: 10.1016/j.bbrc.2004.10.107;pmid: 15541368

22. H. B. Zhang et al., Ultrastructural uncoupling betweenT-tubules and sarcoplasmic reticulum in human heart failure.Cardiovasc. Res. 98, 269–276 (2013). doi: 10.1093/cvr/cvt030; pmid: 23405000

23. C. Y. Wu et al., Calpain-dependent cleavage of junctophilin-2and T-tubule remodeling in a mouse model of reversible heartfailure. J. Am. Heart Assoc. 3, e000527 (2014). doi: 10.1161/JAHA.113.000527; pmid: 24958777

24. A. Guo et al., Molecular Determinants of Calpain-dependentCleavage of Junctophilin-2 Protein in Cardiomyocytes.J. Biol. Chem. 290, 17946–17955 (2015). doi: 10.1074/jbc.M115.652396; pmid: 26063807

25. Y. Wang et al., Targeting Calpain for Heart Failure Therapy:Implications From Multiple Murine Models. JACC Basic Transl.Sci. 3, 503–517 (2018). doi: 10.1016/j.jacbts.2018.05.004;pmid: 30175274

26. J. D. Molkentin et al., A calcineurin-dependent transcriptionalpathway for cardiac hypertrophy. Cell 93, 215–228 (1998).doi: 10.1016/S0092-8674(00)81573-1; pmid: 9568714

27. N. Frey, T. A. McKinsey, E. N. Olson, Decoding calcium signalsinvolved in cardiac growth and function. Nat. Med. 6,1221–1227 (2000). doi: 10.1038/81321; pmid: 11062532

28. R. Passier et al., CaM kinase signaling induces cardiachypertrophy and activates the MEF2 transcription factorin vivo. J. Clin. Invest. 105, 1395–1406 (2000). doi: 10.1172/JCI8551; pmid: 10811847

29. J. Backs, K. Song, S. Bezprozvannaya, S. Chang, E. N. Olson,CaM kinase II selectively signals to histone deacetylase 4during cardiomyocyte hypertrophy. J. Clin. Invest. 116,1853–1864 (2006). doi: 10.1172/JCI27438; pmid: 16767219

30. X. Wu et al., Local InsP3-dependent perinuclear Ca2+

signaling in cardiac myocyte excitation-transcription coupling.J. Clin. Invest. 116, 675–682 (2006). doi: 10.1172/JCI27374;pmid: 16511602

31. M. Colella et al., Ca2+ oscillation frequency decoding in cardiaccell hypertrophy: Role of calcineurin/NFAT as Ca2+ signal

integrators. Proc. Natl. Acad. Sci. U.S.A. 105, 2859–2864(2008). doi: 10.1073/pnas.0712316105; pmid: 18287024

32. S. R. Houser, J. D. Molkentin, Does contractile Ca2+ controlcalcineurin-NFAT signaling and pathological hypertrophy incardiac myocytes? Sci. Signal. 1, pe31 (2008). doi: 10.1126/scisignal.125pe31; pmid: 18577756

33. A. S. Galvez et al., Cardiomyocyte degeneration with calpaindeficiency reveals a critical role in protein homeostasis.Circ. Res. 100, 1071–1078 (2007). doi: 10.1161/01.RES.0000261938.28365.11; pmid: 17332428

34. C. Patterson, A. L. Portbury, J. C. Schisler, M. S. Willis, Tear medown: Role of calpain in the development of cardiac ventricularhypertrophy. Circ. Res. 109, 453–462 (2011). doi: 10.1161/CIRCRESAHA.110.239749; pmid: 21817165

35. D. J. Williams, H. L. Puhl 3rd, S. R. Ikeda, Rapid modification ofproteins using a rapamycin-inducible tobacco etch virusprotease system. PLOS ONE 4, e7474 (2009). doi: 10.1371/journal.pone.0007474; pmid: 19830250

36. C. J. Sigrist et al., PROSITE: A documented databaseusing patterns and profiles as motif descriptors. Brief. Bioinform.3, 265–274 (2002). doi: 10.1093/bib/3.3.265; pmid: 12230035

37. G. Narasimhan et al., Mining protein sequences for motifs.J. Comput. Biol. 9, 707–720 (2002). doi: 10.1089/106652702761034145; pmid: 12487759

38. J. Wysocka, P. T. Reilly, W. Herr, Loss of HCF-1-chromatinassociation precedes temperature-induced growth arrest oftsBN67 cells. Mol. Cell. Biol. 21, 3820–3829 (2001).doi: 10.1128/MCB.21.11.3820-3829.2001; pmid: 11340173

39. D. Sayed, M. He, Z. Yang, L. Lin, M. Abdellatif, Transcriptionalregulation patterns revealed by high resolution chromatinimmunoprecipitation during cardiac hypertrophy. J. Biol. Chem.288, 2546–2558 (2013). doi: 10.1074/jbc.M112.429449;pmid: 23229551

40. F. J. Naya, C. Wu, J. A. Richardson, P. Overbeek, E. N. Olson,Transcriptional activity of MEF2 during mouse embryogenesismonitored with a MEF2-dependent transgene. Development126, 2045–2052 (1999). pmid: 10207130

41. J. O. Mudd, D. A. Kass, Tackling heart failure in the twenty-firstcentury. Nature 451, 919–928 (2008). doi: 10.1038/nature06798; pmid: 18288181

42. J. H. van Berlo, M. Maillet, J. D. Molkentin, Signaling effectorsunderlying pathologic growth and remodeling of the heart.J. Clin. Invest. 123, 37–45 (2013). doi: 10.1172/JCI62839;pmid: 23281408

ACKNOWLEDGMENTS

We thank M. J. Welsh, K. P. Campbell, E. D. Abel, B. London,L. Yang (University of Iowa), and S. R. W. Chen (University ofCalgary) for reading the manuscript and providing constructivecomments, S. R. Ikeda (NIAAA/NIH) for providing TEVp plasmids,and W. Kutschke for technique aids in animal surgery. Funding:This work was funded by NIH R01 HL090905, HL130346, VA1I01BX002334 (L.S.S.), AHA 16SDG30820003 (A.G.); NIH R01HL125436 (C.G.), OD019941 (B.W.), and China NSF 81570293 (J.H.).Author contributions: L.S.S. supervised the project; A.G. andL.S.S. designed the study; A.G., Y.W., B.C., J.Y., L.Y.Z., D.H., J.W.,Y.S., Q.Z., C.C., R.W., and X.Z. performed the experiments anddata analysis; Y.W. (Yunhao) and A.G. performed bioinformaticsanalyses; C.G., M.E.A, F.Z., K.F.A., C.M., M.P., W.Z., and J.H.participated in supervision of experiments, data analysis,interpretations and revision of the manuscript. A.G. and L.S.S.wrote the manuscript. All authors reviewed the results and editedand approved the final version of the manuscript. Competinginterests: The authors have no conflicting financial interests. Dataand materials availability: Microarray and Sequencing data areavailable at NCBI Gene Expression Omnibus (GEO) with accessionnumbers GSE121545, GSE121546, and GSE121547. All other dataare available in the manuscript or the supplementary materials.

SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/362/6421/eaan3303/suppl/DC1Materials and MethodsFigs. S1 to S10Tables S1 to S5References (43–55)

8 April 2017; resubmitted 10 May 2018Accepted 24 October 2018Published online 8 November 201810.1126/science.aan3303

Guo et al., Science 362, eaan3303 (2018) 21 December 2018 9 of 9

RESEARCH | RESEARCH ARTICLEon M

arch 27, 2020

http://science.sciencemag.org/

Dow

nloaded from

regulatorE-C coupling structural protein junctophilin-2 encodes a stress-adaptive transcription

Zhu, Kin Fai Au, Jiang Hong, Mark E. Anderson, Chad E. Grueter and Long-Sheng SongWeizhongCheng Chen, William H. Thiel, Xin Zhan, Robert M. Weiss, Fenghuang Zhan, Catherine A. Musselman, Miles Pufall,

Ang Guo, Yihui Wang, Biyi Chen, Yunhao Wang, Jinxiang Yuan, Liyang Zhang, Duane Hall, Jennifer Wu, Yun Shi, Qi Zhu,

originally published online November 8, 2018DOI: 10.1126/science.aan3303 (6421), eaan3303.362Science 

, this issue p. eaan3303; see also p. 1359Sciencetranscriptional reprogramming.Furthermore, failing cardiomyocytes in stressed myocardium transduced mechanical information (E-C uncoupling) intoprotein into a transcriptional regulator that shuttled to the nucleus (see the Perspective by Padmanabhan and Haldar).

found that under stress conditions, calpain-mediated cleavage converted full-length junctophilin-2 from a structuralet al.junctophilin-2, disrupting the E-C coupling machinery and calcium ion signaling, which compromises cell contraction. Guorequired for formation of the E-C coupling machinery. During heart disease, stress-activated calpain cleaves

Excitation-contraction (E-C) coupling is fundamental to heart contraction. Junctophilin-2 is a structural proteinProtecting the heart

ARTICLE TOOLS http://science.sciencemag.org/content/362/6421/eaan3303

MATERIALSSUPPLEMENTARY http://science.sciencemag.org/content/suppl/2018/11/07/science.aan3303.DC1

CONTENTRELATED http://science.sciencemag.org/content/sci/362/6421/1359.full

REFERENCES

http://science.sciencemag.org/content/362/6421/eaan3303#BIBLThis article cites 55 articles, 22 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.ScienceScience, 1200 New York Avenue NW, Washington, DC 20005. The title (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

Science. No claim to original U.S. Government WorksCopyright © 2018 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of

on March 27, 2020

http://science.sciencem

ag.org/D

ownloaded from