A Novel Cardiac Function of Sumo2/3 and Senp5 Dependent Pathway and Its Physiological Impact on Congestive Cardiomyopathy

The Texas Medical Center Library DigitalCommons@TMC

The University of Texas MD Anderson Cancer Center UTHealth Graduate School of The University of Texas MD Anderson Cancer Biomedical Sciences Dissertations and Theses Center UTHealth Graduate School of (Open Access) Biomedical Sciences

8-2013

A NOVEL CARDIAC FUNCTION OF SUMO2/3 AND SENP5 DEPENDENT PATHWAY AND ITS PHYSIOLOGICAL IMPACT ON CONGESTIVE CARDIOMYOPATHY

Eun Young Kim

Follow this and additional works at: https://digitalcommons.library.tmc.edu/utgsbs_dissertations

Part of the Biochemistry Commons, and the Medicine and Health Sciences Commons

Recommended Citation Kim, Eun Young, "A NOVEL CARDIAC FUNCTION OF SUMO2/3 AND SENP5 DEPENDENT PATHWAY AND ITS PHYSIOLOGICAL IMPACT ON CONGESTIVE CARDIOMYOPATHY" (2013). The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences Dissertations and Theses (Open Access). 390. https://digitalcommons.library.tmc.edu/utgsbs_dissertations/390

This Dissertation (PhD) is brought to you for free and open access by the The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences at DigitalCommons@TMC. It has been accepted for inclusion in The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences Dissertations and Theses (Open Access) by an authorized administrator of DigitalCommons@TMC. For more information, please contact [email protected]. A NOVEL CARDIAC FUNCTION OF SUMO2/3 AND SENP5

DEPENDENT PATHWAY AND ITS PHYSIOLOGICAL

IMPACT ON CONGESTIVE CARDIOMYOPATHY

by Eun Young Kim, MS

APPROVED BY THE DISSERTATION COMMITTEE

______Robert J. Schwartz, Ph.D.

______Pierre McCrea, Ph.D.

______Edward T.H. Yeh, M.D.

______James F. Martin, M.D./Ph.D.

______Hamed Jafar-Nejad, M.D.

APPROVED BY THE DEAN

______Michelle Barton, Ph.D.

______Michaele Blackburn, Ph.D.

The University of Texas Graduate School of Biomedical Sciences at Houston A NOVEL CARDIAC FUNCTION OF SUMO2/3 AND SENP5

DEPENDENT PATHWAY AND ITS PHYSIOLOGICAL

IMPACT ON CONGESTIVE CARDIOMYOPATHY

A THESIS Presented to the Faculty of

The University of Texas Health Science Center at Houston and The University of Texas MD Anderson Cancer Center

Graduate School of Biomedical Sciences

in Partial Fulfillment of the

Requirements for the Degree of

DOCTOR OF PHILOSOPHY

Eun Young Kim, MS

Houston, Texas

August, 2013 Copyright

Table I.1. Motifs of SUMO conjugation.

License : 3203140162941

Table I.2. SENPs localization and substrate specificity

License : 3196670151606, 3200861147118

Figure I.1. Reversible SUMOylation pathway

License : 3196680378058

Figure I.2. Diverse consequences of SUMOylation

License : 3196680378058

Figure I.3. Structural comparison of SENPs

License : 3203140609653, 3203140976954

Figure I.4. Diverse functions of SUMOylation pathway

License : 3196671495501

Figure I.5. Morphology of cardiomyocytes in cardiac hypertrophy

Reproduced with permission from Copyright Massachusetts Medical

Society

Figure I.6. Cardiac hypertrophy

License : 3196690309189

iii Acknowledgments

It is a long journey to stand here. It’s my great pleasure to thank the people who let me make this thesis possible. First of all, to my Ph.D. thesis advisor,

Robert Schwartz, for accepting me as a Ph.D. student, teaching me as a trainee, supporting me, and letting me mature scientifically. With his generosity, patience, inspiration, creativity, knowledge, and experience, he has never stopped providing me with excellent guidance, lessons from the past, encouragement, and opportunities to attend outstanding world-class conferences. He is always a positive mentor. Even when I was afraid of walking forward at certain points, he encouraged me as saying, “I am standing next to you”. This simple comment gave me the power to go forward. His endless concern and careful touch made me to continue and finally to finish this long journey.

I am indebted to my thesis committee members: Pierre McCrea, Ed Yeh, Jim

Martin and Hamed Jafar-Nejad for their availability, unexhausted encouragement, constructive criticism, and helpful suggestions of my thesis project, which allowed me to have a better progress.

I am grateful to the previous and present members of the Schwartz laboratory:

Ling Qian, Li Chen, Wei Yu, Ming Chen, Jose Islas, Kuo-Chan Weng, Allan

Prejusa, Vladimir Potaman and Ann Kong. Special gratitude is extended to the

iv former laboratory member: Jun Wang for his advice to troubleshoot projects as well as to allow me to continue further study. I congratulate on his achievement for own laboratory at Texas Heart Institute. My appreciation also goes to Cindy

Beery, an assistant for Bob. She made me musch easy to contact Bob and helped me for every issue. My appreciation also goes to Elisabeth Lindheim, a

G&D program manager, for kind guidance of administrative things. I deeply thank the GSBS staffs and Victoria P. Knutson, associate dean of academic affair, for her generous advice that allowed me to reach at this momonet.

I was inspired by my beautiful family: Chloe, my lovely 7-year-old daughter,

Allison, little angel 18-month-old daughter, and Jae Man, my husband. Their deliveries were not on “my things to do lists” but it was clear that they made me stronger to finish this journey. I always admire Jae Man’s passion and creativity for the science. Most of them, he held my hands to continue this journey when I wanted to give up. He has always encouraged me with positive mind. That made this moment. I am happy to have a mother-in-law, who understands my scientific career. In addition, I appreciate Jin Goo, my father and Mi Sook, a stepmother for their endless understanding and support. Lastly, I am particularly grateful to my mother. Although she has passed away about 15 years ago, I still feel her warm heart embracing me. I believe that she should be proud of me in heaven.

v A Novel cardiac function of SUMO2/3 and SENP5 dependent

pathway and its physiologic impact on congestive

cardiomyopathy

Publication No.______

Eun Young Kim, M.S.

Supervisory professor: Robert J. Schwartz, Ph.D.

SUMOylation regulates diverse cellular processes including transcription, cell cycle, protein stability, and apoptosis. Although SUMO1 has been extensively studied so far, relevance of SUMO2/3 is unclear, especially in heart. Here we show that failing heart induces SUMO2/3 conjugation. Increased SUMO2/3- dependent modification leads to congestive heart disease such as cardiac hypertrophy by promoting cardiac cell death. Calpain2 and Calpastatin as a novel

SUMO2 targets have been known to be involved in mitochondrial-independent cell death pathway in heart. These SUMOylations of Calpain2 and Calpastatin facilitate activation of Calpain2 by reducing inhibitory role of Calpastatin. These findings identify a SUMO2/3-dependent modification as a novel posttranslational modification that controls Calpain2-Calpastatin system resulting in cardiac protein degradation. On the other hand, transgenic mice overexpressing SENP5,

SUMO2/3-specific deconjugation enzyme in the heart, developed dilated cardiomyopathy. This unexpected phenotype mimics human heart failure, which has elevated SENP5 protein levels. Dilated hearts of SENP5 transgenic mice

vi showed markedly reduced functions of mitochondrial complexes due to almost complete loss of mitochondrial cristae, and compromised cardiac energy metabolism. These hearts also showed increased deSUMOylation of mitochondrial proteins including dynamin related protein (Drp1), which leads to nuclear translocation of mitochondrial apoptosis inducing factor (AIF). This induces massive loss of cardiomyocytes via caspase-independent apoptotic pathway. These findings identify a novel SENP5-mediated deSUMOylation pathway that controls mitochondrial structure and pathological remodeling of the heart. Thus, inhibiting SENP5 might be beneficial to dilated cardiomyopathy and heart failure.

vii Table of Contents

Approvals ...... i

Title……..………………………………………………………………………………ii

Acknowledgments ...... iv

Abstract………………………………………………………………………...……..vi

Table of Contents...... viii

List of Figures ...... x

List of Tables...... xiv

Chapter I: Introduction and Background ...... 1

1.1. Overview of SUMOylation ...... 2

1.2. SUMO2/3...... 15

1.3. SENP5...... 16

1.4. Congestive cardiomyopathy ...... 17

Chapter II: Stress-induced SUMO2/3 conjugation of Calpain-Calpastatin

system Reveals a critical function in myocardial cell death...... 23

viii 2.1. Introduction...... 24

2.2. Materials and Methods...... 26

2.3. Results...... 34

2.4. Discussion ...... 64

Chapter III: SENP5 dependent desumoylation pathway in dilated

cardiomyopathy ...... 71

3.1. Introduction...... 72

3.2. Materials and Methods...... 74

3.3. Results...... 82

3.4. Discussion...... 106

Chapter IV: Summary, Significance and Future Directions ...... 113

Bibliography...... 124

Vita ...... 157

ix List of Figures

Figure I.1 Reversible SUMOylation pathway…………………………..….....6

Figure I.2 Diverse consequences of SUMOylation………..………….….....7

Figure I.3 Structural comparison of SENPs……………………………..….11

Figure I.4 Diverse functions of SUMOylation pathway……………………14

Figure I.5 Morphology of cardiomyocytes in cardiac hypertrophy……..21

Figure I.6 Cardiac hypertrohpy……………...………………………………...22

Figure II.1 Endogeneous SUMO2/3 protein levels in diseased heart.…..35

Figure II.2 Generation of cardiac specific transgenic mice with increased

SUMO2 Expression and conjugation……………………………36

Figure II.3 Cardiac defects caused by cardiac overexpression of

SUMO…………………………………………………………………..38

Figure II.4 SUMO2 overexpression in the heart develops reactivation of

fetal cardiac genes………………………………………………….39

Figure II.5 Abnormal cardiac remodeling in SUMO2 Tg……..…………... 40

Figure II.6 Abnormal cardiac growth in SUMO2 Tg hearts……………….43

x Figure II.7 Abnormal cardiac cell death in SUMO2 Tg hearts……..….....45

Figure II.8 Early onset of cell death in SUMO2 Tg hearts…………...……47

Figure II.9 Cardiac specific cell death on cultured neonatal rat

cardiomyocytes (NRCMs) ……………………….………….……49

Figure II.10 Calpain2 is modified by SUMO2/3 and is involved

in pathologic hearts………………………………………………..50

Figure II.11 Calpain2 induced cell death is enhanced by SUMO2

conjugation……………………………………..…………………...53

Figure II.12 The proteolytic activity of Calpain2 is enhanced by SUMO2

……..………………………………………………….…..………….54

Figure II.13 Calpastatin is modified by SUMO2/3 and is involved in

pathologic hearts…………………………………...... …………….56

Figure II.14 Calpastatin is modified by SUMO……………………..…………59

Figure II.15 SUMOylation promotes Calpastatin degradation……………..61

Figure II.16 SUMOylation changes intracellular Calpastatin

localization…………………………………………………...………63

Figure II.17 Stress incuding SUMO2/3 conjugation in the heart…………..65

Figure II.18 Pathologic human hearts induced active SUMO conjugation

…………………………………………………………………………66

xi Figure II.19 Stress induced abnormal hearts elevated SUMO1-dependent

modification…………………………………………...………..…….68

Figure III.1 Increased SENP5 expression level in human heart failure…..84

Figure III.2 Cardiac defects resulting from cardiac overexpression of

SENP5………………………………………………………………….87

Figure III.3 Cardiac defects resulting from cardiac overexpression of

SENP5………………………………………………………………….88

Figure III.4 Abnormal hearts developed in cardiac SENP5 Tg……..……...90

Figure III.5 Aberrant cardiomyocyte proliferation in SENP5 Tg ……….....91

Figure III.6 Abnormal increase of cell death in SENP5 Tg hearts……...... 92

Figure III.7 Cardiomyocyte death is induced directly by SENP5

overexpression…………………………………………..…………..95

Figure III.8 SENP5 facilitates cell death in NRCMs……………….……..…..96

Figure III.9 Increased desumoylation of mitochondrial protein in cardiac

specific SENP5 Tg heart…………………………………..……….98

Figure III.10 Overexpression of SENP5 causes enlarged mitochondria....99

Figure III.11 Mitochondrial dysfunction in cardiac specific SENP5 Tg

heart …………………………………...... 100

Figure III.12 Mitochondrial defects caused by SENP5 overexpression....102

xii Figure III.13 Myocardial AIF translocation in cardiac overexpression of

SENP5 ……………………………………………………………...... 104

Figure III.14 Impaired cardiac energy metabolism in SENP5 Tg.……..…...109

xiii List of Tables

Table I.1 Motifs of SUMO conjugation …………………………….…………5

Table I.2 SENPs localization and substrate specificity……………...…...10

xiv Chapter I

Introduction and Background

1 1.1. Overview of SUMOylation

SUMO pathway. Precise and specific control of proteins in multicellular organisms is important for proper biologic function during entire life span. Protein modifications called PTM are most widely used to maintain physiologic homeostasis. These include the covalent attachment of sugars, lipids or chemical groups such as phosphate, acetyl or methyl groups into proteins. PTM is also critical for spatial and temporal modulation of protein function. Ubiquitylation has been well known to study as protein-based modification, involving in protein degradation largely by tagging as a mark to be recognized by proteasome. Nearly two decades after discovering Ubiquitin, SUMO, another Ubiquitin-related modification, has been identified. SUMO stands for Small Ubiquitin-like Modifier and the attention on SUMO has not been so long. Its unstable attachment to proteins and loss of association during experimental treatments make it difficult to find its existence. In addition, most of SUMOylation-dependent modification has been found in an association with nuclear pore component Ran GTPase activating protein (RanGAP1) (1, 2), indicating its difficulty to search for various targets. SUMO is evolutionally conserved from yeast to mammals and involved in the reversible and covalent attachment to targets. SUMOylation pathway is very similar to its biochemical analog, Ubiquitylation (3, 4). Human cells have four distinct SUMO isoforms: SUMO1-SUMO4 (5-7). SUMO1 to 3 are ubiquitously expressed. SUMO4 has been detected in limited tissues including kidney, spleen

2 and lymph node. However, substrate and function for SUMO4 has remained unclear (8-11). SUMO isoforms are expressed immaturely, in which their C- terminals are stretched. Mature isoforms expose diGlycine (Gly-Gly) motif, serving as an active acceptor to attach with free amino group (NH2) of lysine within the substrates through isopeptide bond. SUMOylation predominatly occurs at a consensus motif (ΨKXD/E, where Ψ indicates a large hydrophobic residue such as I, V or L, K is the target lysine but X can be any amino acid) in target proteins,

(7, 12). This consensus motif contributes to specificity of SUMO targets and to better regulatory mechanism. More extended SUMOylation motifs are summarized in Table 1. Mature SUMO isoforms show very similar 3D structures of ubiquitin in spite of less amino acid homology (3, 4, 13). This similarity causes competition and/or interchange for substrate affinity. Sequential SUMOylation cascade requires three enzymes (Figure I.1.): an E1 activating enzyme, an E2 conjugating enzyme, and SUMO E3 ligase. The first step is initiated from exposure of C-terminal diGlycine morif of SUMO. SAE1/2, SUMO activating enzyme, then, is able to associate with mature SUMO by consuming ATP, resulting in transfering SUMO to Ubc9, E2 conjugating enzyme. Finally SUMO is facilitated to conjugate with substrates by E3 ligases including PIAS (Protein

Inhibotor of Activated STAT) family, RanBP2, polycomb 2 and MAPL

(Mitochondrial-Anchored Protein Ligase) (14-17). Compared to ubiquitin pathway,

SUMO is conjugated by fewer enzymes including Ubc9, only identified veterbrate

E2 enzyme and about 12 SUMO E3 enzymes (18). This limitation promotes

3 increasing usage of each SUMO components by distinct cellular occupancy, specificity and efficient combinatorial coordination. Emerging evidences suggest that non-covalent interaction of SUMO isoforms with substrates also occurs (19-

21). In that concept, SIM (SUMO-interacting motif) could also modulate targets through conformational changes (19, 22). Short hydrophobic peptides of SIM primarily consisting of acidic and/or phosphorylated serine residiues determine specificity of SUMO binding and control conformation of SUMO-SIM interactions

(7, 20, 23). Diverse ways of SUMOylation allows SUMO-dependent PTM to regulate target proteins more efficiently (Figure I.2).

Table I.1. Motifs of SUMO conjugation. Ψ indicates any large hydrophobic residue, X, any residue, RanGAP, Ran GTPase activating protein; MEF2a, myocyte enhancer factor 2A; AR, androgen receptor; HlC1, hypermethylated in cancer 1; NDSM, negative-charge dependent SUMOylation motif; PDSM, phosphorylation-dependent SUMOylation motif. Diagram was adopted from Henley et al. (2007), permission #3203140162941.

Figure I.1. Reversible Sumoylation pathway. All SUMO paralogues are synthesized as precursors and matured by specific SUMO proteases, SENPs. SENPs also efficiently cleave the isopeptide bond between SUMO and its target. Both the released SUMO and the target become available for additional rounds of SUMOylation. Figure was modified from Melchior et al. (2008), permission #3196680378058.

Figure I.2. Diverse consequences of SUMOylation. SUMOylation modulates target molecules by three different ways. a. SUMO-conjugation to target inhibits the binding of partner. In this case, the target interacts with binding partner only in the absence of SUMOylation. b. SUMO-conjugation triggers binding of partner with target, in which case SUMOylation serves as platform for better binding. SIM is a good example. c. Conformational change of target is induced by SUMOylation. Diagram was adopted from Melchior et al. (2007), permission #3196681235089.

7 deSUMOylation pathway. SUMO-dependent PTM is very reversible by SENPs,

Sentrin/SUMO specific proteases. SENPs have been identified from cleavage assay within bacterial transformants expressing yeast proteins and sequence homology search using Ubiquitin-like protein (Ulp) domain (24). SENP family consists of 6 isoforms (SENP1-3, SENP5-7 in human), representing high homology with yeast Ulp1 and 2 (Figure I.3). Yeast Ulp1 deletion mutant is lethal whereas Ulp2 deletion in yeast can survive but shows abnormal growth and hypersensitivity to DNA damage. Each human isoform has been catagorized by its enzymatic activity in terms of SUMO maturation and isopeptide cleavage, and their activity towards SUMO isoforms (Table I.2). SENP1 and 2 predominatly existing in the nucleus have shown the substrate preferency to all SUMO isoforms whereas SENP3 and 5 show a preference to only SUMO2 and 3. SENP6 and 7 have preferentially modified SUMO2/3 conjugates. Unlike other SENPs, they seem to be involved in SUMOylated chain editing rather than deconjugation pathway. C-terminal of SENPs is relatively conserved whereas N-terminal regions are variable. Thus, less conserved N-terminal region of SENPs plays a key role in subcellular localization and substrate specificity. SENP1 has been typically observed in nucleus although It has both nuclear localization signal (NLS) and nuclear export signal (NES) (25-27). Along with these NLS and NES, SENP1 has also been shown to shuttle between cytoplasm and nucleus in CV-1 cells (28).

SENP2 also contains both NLS and NES, and has been shown in nuclear envelope (29, 30). Later, SENP2 has also been discovered as shuttling between

8 cytoplasm and nucleus. Mutations in NES has shown to block its shuttling.

Interestingly, cytoplasmic SENP2 has been known to be polyubiquitinated, easily leading to 26S proteasome degradation (31). Interestingly, SENP2 splicing variants have been localized in different subcellular areas. Axam, a SENP2 splicing variant, is found in nucleoplasmic face of nuclear pore complx (NPC), and

Axam2 and SuPr-1, other SENP2 splicing variants are able to be localized in the cytoplasm and promyelocytic leukaemia (PML) body of the nucleus (31-33).

These diverse subcellular occupancies upon each alternative splices underscore the notion that each variant contributes to modulating specific substrates and/or to finetuning regulation strategy. SENP3 and SENP5 have been predominantly located in nucleolus but increasing evidence providing no limitation in subcellular localization (34, 35). Shuttling of SENP5 between nucleus and cytoplasm, or between nucleus and mitochondria has been known to be in a cell cycle dependent manner, or for mitochondrial fragmentation (36, 37). SENP6 has been primarily detected in cytoplasm but recent report shows its existence in the nucleus (38-40). Complexity of subcellular localization and distinctive enzymatic activity of SENPs suggest that spatial and temporal process of deSUMOylation may be critical to regulate the function of SUMOylated proteins.

Table I.2. SENPs localization and substrate specificity. Diagram was modified from Baek et al. (2009) and Hochstrasser et al. (2012), permission #3196670151606, #3200861147118.

Figure I.3. Structural comparison of SENPs. Conserved catalytic domain is represented as CD. White empty boxes indicate non-conserved N-terminal region. Specifically, SENP6 and 7 has divided conserved catalytic domain. The function of N-terminal region is unknown but it is involved in distinct subcellular localization. Diagram was adopted from Baek et al. (2009) and Salvesen et al. (2008), permission #3203140609653, #3203140976954.

11 Interestingly, SENP2 splicing variants have been localized in different subcellular areas. Axam, a SENP2 splicing variant, is found in nucleoplasmic face of nuclear pore complx (NPC), and Axam2 and SuPr-1, other SENP2 splicing variants are able to be localized in the cytoplasm and promyelocytic leukaemia (PML) body of the nucleus (31-33). These diverse subcellular occupancies upon each alternative splices underscore the notion that each variant contributes to modulating specific substrates and/or to finetuning regulation strategy. SENP3 and SENP5 have been predominantly located in nucleolus but increasing evidence has provided no limitation in subcellular localization (34, 35). Shuttling of SENP5 between nucleus and cytoplasm, or between nucleus and mitochondria has been known to be in a cell cycle dependent manner, or for mitochondrial fragmentation (36, 37). SENP6 has been primarily detected in cytoplasm but recent report shows its existence in the nucleus (38-40). Complexity of subcellular localization and distinctive enzymatic activity of SENPs suggest that spatial and temporal process of deSUMOylation may be critical to regulate the function of SUMOylated proteins.

Biologic significance of SUMO pathway. Most components of SUMOylation associated with conjugation and deconjugation are localized in nucleus. Therefore their functions were relatively well understood in cell cycle, DNA repair, transcriptional regulation, and chromatin remodeling (41). Recent studies have been proposed additional role of SUMOylation process outside of the nucleus. For instance, GLUT1 and GLUT4, glucose transporters, in plama membrane have been known to be SUMOylated, resulting in opposite outcomes in terms of

12 transporter activity (42). Extranuclear functions of SUMOylation are summarized in Figure I.4. A variety of targets of SUMOylation suggest that SUMO-dependent modification is tighly associated with broad spectrum of biological events including apoptosis, genomic stability, gene expression, metabolism, signaling cascade, cell cycle, ion channel, mitochondrial dynamics (10, 43).

Figure I.4. Diverse functions of SUMOylation pathway. SUMO conjugation is dynamic and reversible. The role can be elevated or reduced by many different factors to modulate the activity of the targets. Thus SUMO-dependent PTM contributes to a variety of cellular phenomena. Diagram was modified from Henley et al. (2007), permission #3196671495501.

14 1.2. SUMO2/3

Each SUMO isoform represents substrate specificities although there is an overlapped preference to substrates. Since identified in yest genetics, genes associated with SUMO and its related components have been proven to be critical for cellular homeostasis in mammals (43, 44). Haploinsufficiency of murine

SUMO1 has suggested its involvement in human cleft lip and palate through modification of Eya1 (45). In contrast, another study has reported that genetic ablation of SUMO1 in mice has not shown any lethality, although there is a variability of lethality depending on genetic background. This may suggest redundant role of SUMO2/3 for SUMO1 target (46). Although it is necessary to examine the role of SUMO2/3 using mice lacking genes encoding SUMO2 and/or

3, these paradoxical results about SUMO1 deficient mice could be explained at least in part by following characteristics of SUMO2/3. In normal condition, compared to SUMO1, SUMO2/3 is detected as unconjugated form, a readily conjugated form upon various stresses including heat shock, oxidative stress, and ethanol exposure (47). This suggests that SUMO2/3 may have more active response upon environmental stimulation than those of SUMO1. Moreover, the

SUMO consensus motif in the N-terminal region (Lys 11) makes SUMO2/3 distinct from SUMO1 (48-51). PolySUMOylation by SUMO2/3 has facilitated ubiquitin-mediated degradation including PML by activating RNF4, ubiquitin E3 ligase (52-54) and BMAL, a core component of circadian rthythm (55). These

15 indicate that SUMO2/3-dependent conjugation mediates distinctive biological functions. A recent finding has provided a new connection of SUMO2/3 with chromatin remodeling (56). SUMOylation of CoREST1, a transcriptional cofactor by SUMO2/3 suggests that SUMO2/3 may participate in gene regulation (56).

1.3. SENP5

SENP5 has both C-terminal hydrolase activity for maturation of SUMO isoforms and isopeptidase activity for SUMO deconjugation. In spite of its dual enzymatic activity, it has preference of SUMO3 hydration and SUMO2/3 deconjugation in vitro (35, 57). Strong endopeptidase activity of SENP5 (58) indicates that SENP5 might play a role for SUMOylation repression. Although it has dual enzymatic activities, SENP5 reveals less isopeptidase activity compared to SENP2. In addition, among SUMO isoforms, SENP5 has SUMO2/3 preference shown in biochemical study (35) but, selected preference does not seem to be restricted in vivo. Athough Drp-1, mitochondrial fission protein, has been shown to be modified by SUMO1, it becomes deconjugated by SENP5 (37, 59). Most components of SUMO pathway have been found in nucleus. Thus, their functional studies have been best characterized for nuclear proteins (39, 57). SENP5 has been also considered as a nucleolar protein based on fluorescent substrates such as SUMO2/3 in nucleus and N-terminal nucleolar localization sequence (49, 50,

57). N-terminal deletion mutant of SENP5 has led to disappear its predominant nucleolar localization and triggered deSUMOylation activity (57). Diverse

16 subcellular localizations of SENP5 have been further supported. Majority of

SENP5 has been detected in cytoplasm of tumor cells, although normal epithelial cells show its equal distribution in both nucleus and cytoplasm (60). Moreover, mitochondrial fraction from COS-7 cells has reasonable level of SENP5 (30)(37).

These suggest that SENP5 shuttle between nucleus and mitochondria to be involved in subcellular specific modulation of targets.

1.4. Congestive cardiomyopathy

The heart is a muscle organ to pump blood via circulatory system for oxygen and nutrients delivery and exchange carbon dioxide back to the lung.

Cardiomyocytes generate most of contractile powers in hearts by well-organized contractile units called sarcomeres. Although there are other cell types consisting of the heart including fibroblasts, endothelial cells and immune cells, cardiomyocytes as contractile cells are terminally differentiated. Therefore, the heart loses cell proliferation ability soon after birth, although a recent report has discovered a very low level of cardiomyocyte turnover throughout life (61-63).

Thus, it could respond only by increasing cardiomyocyte size to improve workloads (Figure I.5). Initially cardiac hypertrophy is beneficial because it is an adaptive response to reduce overloaded stress, but sustained hypertrophic growth finally leads to heart failure and sudden death. Heart develops hypertrophic growth in response to pathologic stimuli, resulting in fibrosis, dilation, ventricular remodeling and impaired cardiac output ((64-66). In constrast, there is

17 a physiologic hypertrophy that occurs as adaptive response to continuous physiologic stresses such as exercise and pregnancy (Figure I.6). Physiologic hypertrophy is fundamentally different from pathologic hypertrophy, which does not lead to disease state. Rather, physiologic hypertrophy undergoes more beneficial response to maintain stable cardiac function through specific molecular pathways. Pathologic hypertrophy could be characterized by several alterations.

First, sarcomeric proteins including α-actinin are rearranged to strengthen contractile power, of which failing hearts have reduced level. Second, expression of fetal cardiac genes is reactivated. Fetal gene reprogramming includes atrial natriuretic factor (ANF), brain natriuretic peptide (BNP), β-myosin heavy chain (β-

MHC) and skeletal α-actin (Sk-actin) (67). Thus, fetal cardiac gene expression is used as indicator of pathologic hypertrophy in spite of altered expression levels of genes encoding calcium handling proteins and collagen I. Third, energy source is shifted from fatty acid to glucose. Cardiomyocytes use glycolysis to obtain main energy for proliferation during embryogenesis but after birth, fatty acid oxidation becomes metabolic pathway due to high mitochondrial capacity. This switch is broken in pathologic cardiac hypertrophy, in which glucose is rebounced as primary energy source (68, 69). Protein synthesis rate is also increased to satisfy the requirement of hypertrophied cardiomyocytes. All alterations are the process to adapt to increased workloads as compensatory mechanism. However, prolonged adaptive responses lead to decompensatory dilated cardiomyopathy, and heart failure, which is associated with thin ventricular walls through massive

18 cardiac muscle cell loss. This view is controversial but, a recent finding supports that cardiomyocyte loss is sufficient to induce heart failure. It has been shown that human heart failure has resulted in increased apoptotic cardiomyocyte death (70,

71). Genetically enhanced caspase-8 in mouse heart has been shown dilated cardiomyopathy with exaggerated apoptosis (72). A well-known apoptotic pathway is caspases-dependent and is further stimulated by cytoplasmic release of cytochrome C. However, mitochondrial independent cell death pathway is also involves in heart failure, which utitlizes Calpain-Calpastatin system. Calpains are

Ca2+-activated cysteine proteases (73, 74), leading to protein degradation.

Calpains affect many biologic processes including cell cycle, migration, differentiation and apoptosis (73, 75-79). Stress-induced Ca2+ influx in cardiomyocytes activates Calpains leading to cleavage of pro-caspase 12. Finally activated caspase12 by Calpains facilitate further caspase cascade including caspase 9 and 3 (80, 81).

Caspase-indenpendet cell death pathway is also emerging and apoptosis inducing factor (AIF) is one of key proteins in this pathway. Originally, it has been thought that AIF in mitochondrial inner membrane is critical for mitochondrial respiratory complex I integrity, representing NADH oxidoreductase and peroxide scavenging activities (82-84). N-terminal domain of AIF contains NADH and FAD, critical for mitochondrial respiratory function. Interestingly, AIF is also associated with caspase-independent cell death. AIF-meidated death function depends on increased intracellular Ca2+ concentration that triggers mitochondrial memebrane

19 depoloarization. This leads to Calpain1 activation, resulting in proteolytic cleavage. C-terminal domain of AIF is important for integrity and nuclear translocation (85). Nuclear AIF has not been studied well but known that it has involved in chromosome breakage and chromatin condensation (86-89).

Figure I.5. Morphology of cardiomyocytes in cardiac hypertrophy. Distinct morphologic alteration of cardiomyocytes is characterized by increased cell size in response to various stimuli. Exercise induced physiologic hypertrophy is usually not bad, in which resversely cardiomyocytes come back to normal status. However, pathologic cardiac hypertrophy easily enters into unrecovered state with sarcomere rearrangement, fetal gene reactivation. Concentric hypertrophy is caused by relative increase in the width of cardiac mucle cells. On the other hands, great increase in the lenghth than in the width causes eccentric hyperophy. This type of cell growth leads to dilated cardiomyopathy. Diagram was reproduced from Chien et al. (1999) with permission from copyright Massachusetts Medica Society.

Figure I.6. Cardiac hypertrophy. The heart enlarges cardiac muscle cell size in response to stimulus to decrease ventricular wall stress since cardiac muscle cells lose their ability of proliferation soon after birth. Cardiac hypertrophy can be classified by stimuli. Physiologic hypertrophy is induced by exercise or pregnancy, in which ventricular volume is increased by both wall and septum thickness as a result of concentric cardiomyocyte growth. Mostly, this is not harmful because it is reversed easily upon reduced stress. Thus any pathologic hypertrophic markers such as disorganized contractile proteins and upregulated fetal cardiac genes are not detected in physiologic hypertrophy. However, concentric cardiomyocyte growth by pathologic stimuli makes thick wall and septum, in which left ventricular chamber volume is reduced. Sustained and decompensated concentric hypertrophy is not helpful any more to heart although initial hypertrophic growth is the result of increased workload demand. It finally leads to unreversible dilated cardiomyopathy with loss of cardiac mucle cells. Ventricular chamber size and volume is increased along with thinner wall and septum thickness. Diagram was adopted from Molkentin et al. (2013), permission #3196690309189.

22 Chapter II

Stress-induced SUMO2/3 conjugation of

Calpain-Calpastatin system reveals

a critical function

in myocardial cell death

23 2.1. Introduction

Since many transcription factors involving cardiovascular development such as Tbx2 (90), GATA4(91), Nkx2.5(92), Mef2 (93) and SRF(94) have been modified by SUMOs, most of known functional studies focuse on SUMO1.

SUMO1-dependent conjugated those transcription factors activate cardiac genes.

A recent interesting study has first suggested the correlation of SUMO1 with heart. The calcium-transporting ATPase, SERCA2a plays a critical role in Ca2+ reuptake to maintain cardiac muscle contraction. Impaired Ca2+ re-uptake has been caused by altered expression and activity of SERCA, resulting in heart failure

(95). In this study, novel regulation has been proposed that SERCA2a is modified by SUMO1, improving cardiac function through preserving ATPase activity and stability (96). Since they have observed that human and animal failured hearts revealed decreased SUMO1 protein level, it was concluded that SUMO1 is required for normal cardiac contraction. On the same concept, SUMO1 has been also suggested as an anti-fibrotic factor. SUMO1 expression has been promoted in the patients with systemic sclerosis (SSc) to trigger TGF-β signaling, involving fibrosis (97). SUMO1 has been given as a novel therapeutic target for fibrotic disease. However, SUMO1 has been also considered as a protective factor.

SUMO1 has been increased in response to focal cerebral brain ischemia as well as overexpressed SUMO1 has been shown to reduce infarction size in brain (98).

Although active investigation for SUMO1 and disease, especially heart is currently

24 focused, SUMO2/3 is little vailed. One interesting report has been shown that

SUMO2/3 specific conjugation modulates the chromain modification complex,

CoREST-LSD1 (56). It is very interesting since cardiac fetal gene reactivation has depended on REST (84)(99). After birth, the expression of fetal cardiac genes is shut down by REST, recruiting CoREST-LSD1 complex, finally HDAC1 or 2 (85,

86)(100, 101). HDAC1 and 2 have been involved in regulation of genes encoding cardiac specific myofibrillar protein and calcium channels such as T-type calcium channel Cav3.2 and troponin I (TnI) (99, 100). Thus abolished HDAC1/2 has caused high sensitivity to stress induced cardiac hypertrophy by altered cardiac specific gene expression and GSKβ signaling. Those support that there is another modulation layer, different from SUMO1 to release repressive REST-CoREST-

LSD1 complex for stress-induced reactivation of fetal cardiac genes in adult.

Moreover, Calpain2, a calcium dependent protease system component has been

SUMO2/3-ylated, involving cell migration (102). Increased Calpain1 and

Calpastatin has been involved in heart disease caused by uncontrolled protein degradation (88)(103), suggesting that SUMO2/3 dependent modification of

Calpain2 might influence the cardiac function. Further, the first study of SUMO- dependent PTM with human cardiovascular disease has shown that sumoylation defect lamin A mutation is involved in familiar dilated cardiomyopathy (104-106).

Especially, decreased lamin A function has been related to SUMO2/3 dependent modification and resulted in cell death facilitation and abnormal nuclear envelop integrity (105). The stress induces high increase in SUMO2/3 conjugation but not

25 in SUMO1 although the function of conjugated SUMO2/3 is little known. SUMO1 and SUMO2/3 have different specificities to targets although somehow they overlappe target proteins. Collectively, although SUMO2/3 has been studied little compared with SUMO1, it could play a role to maintain healthy heart state.

Studies about SUMO2/3 in the heart were undertaken to examine a unique role for SUMO2/3 during congenital heart disease.

2.2. Materials and Methods

Human heart samples. The explanted human failing hearts were collected during transplantation at the St. Luke Episcopal Hospital, Houston, TX. All hearts were frozen in liquid nitrogen and stored at -80°C until use. Failing hearts were from adults diagnosed with dilated cardiomyopathy (DCM). Normal human hearts used as controls were obtained from donors who died from motor-vehicle accidents. All studies were performed using left ventricles.

Angiotensin-II (Ang-II) administration. Animal model was made by the lab.

Member, Ling Qian as following. 8- to 12-week-old mice were used for either sham-operation or Ang-II (AnaSpec Inc.) administration as described (107).

Anesthetized mice with avertin (240mg/kg, injected intraperitoneally (IP injection) were intubated and ventilated for the procedure. Ang-II was delivered by implantation of a miniosmotic pump (Alzet, model 1007D) dorsally, containing either Ang-II (3mg/kg/d) in 0.9% NaCl. 0.9% NaCl solution alone will be delivered

26 to utilize as control. Hearts were harvested 7 days after Ang-II infusion (108, 109).

Heart abnormality was evaluated by comparing HW/BW of sacrificed mice.

Generation of transgenic mice. The α-MHC-SUMO2 transgene was constructed as described (93)(110) under the control of mouse α-MHC promoter

(provided by Dr. J. Robbins, University of Cincinnati). DNA sequecing was used to confirm the orientation of inserted transgene. The pronucleus microinjection of transgene construct into the fertilized eggs from FVB mice to generate founder

(F0) SUMO2 transgenic (SUMO2Tg) mice. Foudners were crossed back with

C57Bl/6 mice for pure background. Genomic DNA was isolated from tail biopsies performed on weaned animals (approximately 3-week old pups) and screened by

PCR. The expression of transgene in SUMO2Tg mouse hearts was verified by

Western blot and/or qPCR. Animal work was followed by accordance with IACUC approval.

Echocardiography. Mice of interest were anesthetized by inhalation of 1% isofluorane and rested on a warm pad during transthoracic measurements of cardiac function using two dimensional M-mode of a Vevo 770 in vivo micro- imaging system (Visual Sonics, Toronto, Canada). The probe contacted with hair- removed chest to record cardiac function indices. The anlaysis of cardiac function and heart dimensions was performed and evaluated under genotype-blinded condition. Cardiac functional analysis using echocardiography was performed by lab. member, Wei Yu.

27 Histopathology. Mice hearts were dissected and fixed overnight in 4% paraformaldehyde (PFA). Hematoxylin and eosin (H&E) or Masson’s trichrome staining was performed on heart sections (10 µm) according to standard protocols.

Transmission Electron Microscopy. Left ventricle tissue was fixed in 2% paraformaldehyde (PFA) and 3% glutaraldehyde in 0.1M cacodylate buffer, pH7.3 and prepared according to standard protocol. Electron microscopy was examined in a JEM 1010 transmission electron microscope (JEOL, USA, Inc.) at an accelerating voltage of 80kV. Digital images were obtained using AMT Imaging

System (Advanced Microscopy Techniques Corp). This work was done by MD

Anderson HREM facility.

RNA Isolation and mRNA quantification. RNA was extracted from age-matched

WT or SUMO2Tg mouse hearts using Trizol according to the manufacturer’s protocol. Reverse transcription reaction was performed using 1 µg total RNA per reaction and cloned reverse transcriptase (Invitrogen), followed by qPCR

(7900HT, Applied Biosystems) using the following gene-specific primers: SUMO-

1: forward, 5’ TCTGACCAGGAGGCAAAACC 3’; reverse, 5’

CTAAACCGTCGAGTGACCCC 3’. SUMO-2: forward, 5’

GACGAGAAACCCAAGGAAGG 3’; reverse, 5’ CTCCAGTCTGCTGCTGGAAC 3’.

SUMO-3: forward, 5’ CCAAGGAGGGTGTGAAGACA 3’; reverse, 5’

TCAATAGCACAGGTCAGGACA 3’. ANF: forward, 5’

GTGTCCAACACAGATCTGATGGAT 3’; reverse, 5’

28 GCCTAGTCCACTCTGGGCTCCAAT 3’. BNP: forward, 5'

GGCCAACACCAACCTGTCCAAGTT 3’; reverse, 5'

TGCAAAGGCTCCAGGTCTGAGGGC 3’. α-MHC: forward, 5’

GGAAGAGTGAGCGGCGCATCAAGGA 3’; reverse, 5’

TCTGCTGGAGAGGTTATTCCTCGT 3’. β-MHC: forward, 5’

GGCCAACACCAACCTGTCCAAGTT 3’; reverse, 5’

TGCAAAGGCTCCAGGTCTGAGGGC 3’. skeletal α-actin: forward, 5’

AGACACCATGTGCGACGAAGA 3’; reverse, 5’

CCGTCCCCAGAATCCAACACGA 3’. GAPDH: forward, 5’

ATGTTCCAGTATGACTCCACTCAC 3’; reverse, 5’

GAAGACACCAGTAGACTCCACGA 3’.

Neonatal rat cardiomyocytes culture. Enzymatically dissociated cardiomyocytes from 1 to 2-day-old Sprague-Dawley rats were enriched using

Percoll (GE Healthcare) gradient centrifugation. Isolated cardiomyocytes were then plated on collagen-coated dishes (50µg/ml) at a density of 200cells/mm2 and maintained in cardiomyocytes culture medium (DMEM supplemented with 10%

FBS and 2mM of L-glutamine).

Generation of the recombinant adenoviral vector. Recombinant adenoviral vectors were generated by using the AdEasy XL Adenoviral Vector System

(Stratagene). A DNA fragment containing the full length human SUMO2 and mutant human SUMO2 which deletes two guanidine amino acids at C-terminus as

29 well as mouse Calpain2, SUMO2 fused mouse Calpain2 (SUMO-GA-Calpain2) and sumoylation site mutated Calpain2 (Calpain2-K390R) were subcloned into the pShuttle-IRES-hrGFP1 vector. The entire expression cassettes from the resulting vectors were recombined in BJ5183 bacterial strain with serotype 5 first- generation adenoviral backbone, AdEasy-1 (Stratagene). These recombinant adenoviral backbones were transfected into AD293 cells (Stratagene) to generate infectious viral particles. Viral titer was determined by the tissue culture infectious dose (TCID50) method (111). Cardiomyocytes were infected with recombinant adenoviruses for 2hrs at a multiplicity of infection (M.O.I) of 100 particles/cell and incubated for additional 24-48 hrs to ensure transgene expression.

Annexin-V staining. Annexin V staining was performed using Annexin V Alexa

Fluor 594 (Invitrogen). Neonatal rat cardiomyocytes on cover slides were added by 5µl of the Annexin V Alexa Fluor 594 and incubated for 15 minutes at room temperature. Then cells were washed and mounted with Vectashield (Vector

Laboratory).

Cardiomyocyte size measurement. WGA-TRITC staining was used on heart sections of P5 or P90 WT and SUMO2Tg mice to distinguish sarcolemmal membrane. 10-20 randomly selected fields from each individual heart sample was used to measure surface areas of cardiomyocytes using Software ImageJ

(http://rsbweb.nih.gov). The average surface area calculated was normalized by

WT cardiomyocytes. All analysis was performed under blindness.

30 In vivo Calpain2 activity assay. In vivo Calpain2 activity was analyzed by using synthetic substrate t-BOC-Leu-Met-chloromethylaminocoumarin (t-BOC-LM-

CMAC, Invitrogen). HeLa cells were plated at 70-80% confluence in 22mm square coverglass (Coring) and grown in complete media for 24hrs. Cells were transfected with plasmids encoding the Calpain2, SUMO2 fused Calpain2

(SUMO-GA-Calpain2) and sumoylation site mutated Calpain2 (Calpain2-K390R) in pShuttle-IRES-hrGFP-1. pShuttle-IRES-hrGFP-1 was used as empty vector.

After two more days incubation, cells were incubated for 30 min in the presence of

10µM t-BOC-LM-CMAC. Then cells were washed with PBS before covering with a glass cover slide. Both GFP and t-BOC-LM-CMAC positive cells were scored for analysis. Fluorescence was visualized using a fluorescence microscope (Olympus fluorescence microscope). Intensity of fluorescence caused by cleavage of synthetic substrate was measured using Image J. Because backgrounds between each coverslips slightly varied, background measurements were also taken and subtracted. Calpain2 activity was expressed as CTCF (corrected total cell fluorescence), which is calculated as follows: integrated density − (area of selected cell × mean fluorescence of background readings).

Immunostaining. HeLa cells were cultured in DMEM containing 10% FBS and

1% penicillin/streptomycin antibiotics. 70% confluent cells were replated into two well Lab-Tek II chamber slide (Nunc) with 4x104 cells density were transfected with mouse Calpastatin (mCAST, 0.5ug/well) and SUMO2 fused mCAST

(SUMO2-mCAST, 0.5ug/well). Transient transfections were performed using

31 Lipofectamine 2000 (Invitrogene) according to the manufacturer’s protocol.

Briefly, Plasmid applied with 0.5µg were mCAST expression vector and SUMO2 fused mCAST (SUMO2-mCAST) expression vector. Media were changed at next day. Cells were maintained additional day. Transfections were done in triplicate.

HeLa cells were fixed in 3.5% PFA and permeabilized with 0.5% Triton X-100.

Each transfected construct staining was conducted with anti-V5 antibody (1:200,

Bethyl Laboratories, Inc.). Cells were sequentially incubated with secondary antibodies, Alexa Fluoro® 488 anti-goat antibody (Invitrogen) and then mounted with Vectashield with DAPI (Vector Laboratory).

To detect DNA synthesis in heart tissues, BrdU was injected intraperitoneally into age-matched mice at a dose of 100µg/g body weight. Sixteen hours after injection, mouse hearts were isolated, fixed in 4% PFA, dehydrated in increasing concentrations of ethanol (up to final of 100%), and sectioned at 5 µm thickeness.

Paraffin embedded heart sections went further into antigen-unmasking step by boiling in 10mM sodium citrate buffer (pH6.8) for 15min. BrdU staining was conducted with BrdU in situ detection kit according to the protocol provided by the manufacturer (BD Pharmingen). Ki67 or phosphoylated Histone H3 staining was performed by using anti-Ki67 antibody (1:200, Santa Cruz) or anti-phopho Histone

3-Ser10 (1:200, Millipore). Anti-caveolin 3 antibody (1:8000, Santa Cruz) was used as a cardiomyocyte marker. Sections were then incubated with the following secondary antibodies: anti-rabbit TRITC antibody or anti-mouse FITC antibody

(Jackson ImmunoResearch). Sections were mounted with Vectashield containing

32 DAPI (Vector Laboratory). BrdU and Ki67 labeling indices were calculated based on scores obtained from at least 6 different hearts (3 for each group), with each at least 5 fields (adult hearts) randomly selected for evaluation. To analyze the apoptosis, ApopTag® Red In situ Apoptosis Detection Kit (Milipore) was applied.

TUNEL staining followed to the standard protocol provided by the manufacturer.

Immunoblotting. Mouse hearts were removed and stored at -80°C until use. For protein extraction, 100 µg of protein lysates from mouse hearts or 50 µg of neonatal rat cardiomyocytes lysates containing overexpressed proteins were purified in a cold lysis buffer (50mM Tris-HCl, pH7.4; 0.5% sodium deoxycholate;

1% NP-40; 150mM MaCl; 1mM EDTA; 1mM DTT; 0.5mM PMSF) containing protease inhibitor cocktail (Roche) and presence of 20 mM isopeptidase inhibitor

N-ethylmaleimide (NEM), which prevents desumoylation of SUMO-conjugated substrates. The tissue lysates were solubilized for 20 min on the ice. Then the supernatants were separated by centrifugation at 13,000 rpm for 15 min at 4°C.

The supernatants were used for immunoblot analyses and immunoprecipitations.

For immunoprecipitation, protein lysates from frozen mouse hearts were prepared. 400 µg of total protein was diluted to a final concentration of 0.4 mg/ml in lysis buffer. 4 µg of bead-conjugated anti-SUMO-2/3 antibody (Santa Cruz) was applied to the sample and incubated for 2 hours at 4°C on a rotary platform. The beads were subsequently pelleted by centrifugation, washed five times with lysis buffer. These protein lysates were subsequently subjected to 4-12% NuPAGE, transferred to PVDF membrane, detected with the desired

33 antibody, which was then visualized with chemiluminescence.

Statistics. Numbers of mice for each group used in each experiment are indicated in the Figure legends. 2-tailed student’s t-test was used to determine statistical significance between groups and error bars represent means± s.e.m.

P value < 0.05 was considered statistically significant and P<0.005 as highly significant.

2.3 Results

Enhanced expression of SUMO2 causes cardiomyopathy. Human failing hearts made elevated SUMO2/3 conjugation compared to normal hearts (Figure

II.1.a). This result was also observed in the artificial cardiac stress induced animal model in which cardiac stress was introduced through angiotensin-II (Ang-II) administration (Figure II.1.b). To determine whether SUMO2 is sufficient to induce pathologic hearts, heart specific SUMO2 transgenic mice under the control of α-

MHC promoter directing cardiomyocyte-specific expression of transgene were generated. Several independent Tg lines were established based on immunoblot analysis, and we chose lines with high (9606 or 9610 lines), medium (9592 line) and low (9608 line) expression for characterization. Figure II.2.a represented

SUMO2 protein level of 9592 line. The change of SUMO2/3 conjugated protein level (HMW) from human diseased and SUMO2 Tg (9592 line) hearts was comparable (Figure II.2.b). Thus heart specific SUMO2 Tg can be used to study the function of SUMO2/3 in the pathologic condition.

Figure II.1. Endogeneous SUMO2/3 protein levels in diseased heart. a, SUMO2/3 conjugation in human heart tissues (n=3 per non-failing heart (cont), n=6 per heart failure (HF, n=3 for non-ischemic heart failure (NIS-HF), n=3 for ischemic heart failure (IS-HF)) are represented. b, Representative immunoblot for SUMO2/3 in experimental animal model which is Ang-II (3mg/kg/d, 1wk) stimulated heart disease model. GAPDH, glyceraldehyde 3-phosphate dehydrogenase, HMW, high molecular weight group.

Figure II.2. Generation of cardiac specific transgenic mice with increased SUMO2 expression and conjugation. a, Heart specific SUMO2 transgenic mice were generated using α-MHC promoter. Transgene, SUMO2 was constructed and modified to result in high conjugation activity contained SUMO2 by expose of diglycine motif at C-termini. Immunoblot analysis of SUMO2/3 protein in the mouse heart confirmed that both conjugated and free SUMO2/3 was highly increased in the Tg. GAPDH was used as a loading control. WT, wild type littermate, SUMO2 Tg, heart specific SUMO2 Tg, HMW, high molecular weight SUMO2 conjugates. b, Quantification of HMW from immunoblot analysis between human and Tg mouse was done using an image process program (ImageJ). Error bars represent mean ± SEM. (**P<0.005 vs cont for human, *P<0.05 vs WT for mouse, n=5 for human, n=3 for mouse/group)

36 SUMO2 overexpression caused cardiac remodeling. Elevated SUMO2 expression in murine hearts revealed enlarged hearts at P80 (Figure II.3.a).

Cardiac function was assessed by echocardiography in SUMO2 Tg at P80. Heart specific SUMO2 Tg showed significant reduction in contractility as indicated by a reduced EF, FS and increased HW/BW ratio (Figure II.3.b). Fetal cardiac genes, such as atrial natriuretic factor (ANF), b-type natriuretic protein (BNP), β-myosin heavy chain (β-MHC) and skeletal actin (SK-Act) re-expressed during cardiac stress were assessed by qPCR. Those stress marker genes were upregulated, but adult form of myosin heavy chain (α-MHC) was downregulated (Figure II.4).

These results indicate that elevated SUMO2 expression in murine hearts causes cardiac stress. To address cardiomyopathy induced by SUMO2 overexpression, we analyzed size of cardiomyocyte by weat germ agglutinin (WGA) staining at the same age of Figure II.4 (Figure II.5.a). Image J quantification of cardiomyocyte cross sectional size of SUMO2 Tg was normalized by one of WT, indicating significantly large myocyte from SUMO2 Tg (Figure II.5.b). α-MHC promoter

SUMO2 Tg exhibited marked cardiac hypertrophy with enlarged heart, upregulated fetal cardiac genes, increased cardiac cell size and apparent fibrosis

(blue) (Figure II.5.c). Some of Tg died suddenly at one week of age, in which the level of transgene expression was high such as 9606 or 9610 lines. Those Tgs represented smaller body size but enlarged heart, indicating high HW/BW ration

(Figure III.6.a, b, e). Notably

Figure II.3. Cardiac defects caused by cardiac overexpression of SUMO2. a, Gross morphological change of hearts from WT and Tg at P80. b, Cardiac function was measured through echocardiography. Two-month-old α-MHC- SUMO2 Tg and their control littermates were analyzed (n=3 for each group). Graphs represented a decrease EF, FS and an increase LV mass, which indicating hypertrophied hearts. Left ventricle weight was normalized by body weight for WT and Tg (two months of age). EF, ejectional fraction. FS, fractional shortening. Error bars indicate SEM. (*P<0.05, **P<0.001 vs WT)

Figure II.4. SUMO2 overexpression in the heart develops reactivation of fetal cardiac genes. Reactivation of fetal cardiac genes was determined in 1year-old α-MHC-SUMO2 Tg and WT mice (n=3 for each group). Total heart RNA was isolated and prepared for cDNA. Fetal cardiac genes involved in cardiac stress were analyzed. Expression levels were normalized by GAPDH. Error bar indicates mean±SEM. (*P<0.05, **P<0.0005 vs WT)

Figure II.5. Abnormal cardiac remodeling in SUMO2 Tg. 3-month-old Tg and WT littermates were stained. n=3 for each group. a, b, Wheat germ agglutinin (WGA) staining was applied to measure the cardiomyocyte size. Quantification of cardiomyocyte size from WGA positive heart section by Image J indicated increased cardiomyocyte size in α-MHC-SUMO2 Tg (b). Bar, 100µm. Error bars indicate the SEM. (**P<0.001 vs WT) c, Masson’s Trichrome staining demonstrated fibrosis (blue) was apparently enhanced in cardiac SUMO2 Tg compared to WT. Bar, 200µm.

41 ventricular dilation was observed in Tgs. It suggests sudden death is caused by dilated cardiomyopathy. Although the suddenly dead Tg revealed dilated ventricular wall, cardiomyocyte cross sectional size was significantly increased

(Figure II.6.c, d), supporting hypertrophic response was common in enhanced

SUMO2 expressed hearts. To investigate the basic mechanism for dilation,

TUNEL staining was applied. An increase in apoptotic cell death was prominent around one week of age Tgs (Figure II.7.a,b). Moreover, transmission electron microscopy from the same aged Tgs showed sarcomere rearrangement and disorganized mitochondria (Figure II.7.c).

Aberrant cardiomyocyte apoptosis in SUMO2 Tg hearts. To test further if increased cell death is the primary cause of cardiomyopathy in SUMO2 Tg hearts,

TUNEL staining was performed in SUMO2 mutant hearts when HW/BW ration was not different compare to WT (Figure II.8.c). Even if cardiac stress indicator such as HW/BW ratio was not evident, an aberrantly prominent increase of apoptosis was also observed in SUMO2 Tg hearts at P4 (Figure II.8.a,b).

To evaluate whether increased cell death was caused by SUMO2, AdSUMO2-

GG (constitutive active SUMO2 expressing recombinant adenovirus) was generated. Adenoviral infection was controlled as MOI=100 condition, resulting in infection of more than 90% of cultured primary neonatal rat cardiomyocytes

(NRCMs). At 36hrs after infection, NRCMs were stained with Annexin-V to see early event of cell death. In normal condition, phosphatidylserine (PS) is located on the cytoplasmic surface of the cell membrane; however, in apoptotic cells, PS

Figure II.6. Abnormal cardiac growth in SUMO2 Tg hearts. Heart specific SUMO2 Tg revealed smaller body size (a) whereas, enlarged heart size (b) around one week old. The enlarged hearts from SUMO2 Tg developed dilated cardiomyopathy with sudden death. Dramatically severe harm phenotype representing SUMO2 Tg expressed higher transgene. c, WGA staining showed increased the cardiomyocyte size in α-MHC-SUMO2 Tg around one week old (n=3 for each group). Bar, 50µm. d, Cardiomyocyte size was measured using Image J, indicating hypertrophic growth was also occurred. Error bars indicate the SEM (**P<0.0001 vs WT). e, HW/BW ratio (±SEM, mg/g) of WT and SUMO2 Tg mice were determined around one week old. Heart index was significantly changed in SUMO2 Tg hearts.

Figure II.7. Abnormal cardiac cell death in SUMO2 Tg hearts. Immunohistochemistry of WT and Tg mouse hearts around one week old. (n=3 for each group) a, TUNEL staining showed highly increased apoptosis in α-MHC- SUMO2 Tg. TUNEL positive cells were represented as red dots. Cardiomyocyte was chosen through Cav3 (green) staining. Nuclear counter staining was done using DAPI (blue). Bar, 200µm. b, Only TUNEL-positive and Cav3 positive cells were scored at least five randomly selected fields. Error bars indicate the SEM (**P<0.005 vs WT). c, TEM image showed disarrayed myofibrils and irregular mitochondria in SUMO2 Tg. Bar indicates 500nm.

Figure II.8. Early onset of cell death in SUMO2 Tg hearts. Immunohistochemistry of WT and Tg mouse hearts at P4. (n=3 for each group) a, TUNEL staining showed increased apoptosis in α-MHC-SUMO2 Tg at P4. TUNEL positive cells were stained as red dots. Cav3 (green) was used as the cardiomyocyte specific marker. Nuclear counter staining was done using DAPI (blue). Bar, 200µm. b, Only TUNEL-positive and Cav3 positive cells were counted in at least five randomly selected fields. Error bars indicate the SEM (**P<0.005 vs WT). c, HW/BW ratio (±SEM, mg/g) of WT and SUMO2 Tg mice were determined at P4. Although heart index at P4 was not changed, apoptotic cell death was significantly increased in SUMO2 Tg hearts.

47 is translocated from the inner to the outer plasma membrane. Annexin-V has the high affinity to PS. Thus it can be used to verify the cell death event. Constitutive active SUMO2 containing adenovirus infection significantly induced cell death. In contrast, this result was totally diminished under SUMO2 conjugation activity dead mutant by deleting two glycines at c-termini (AdSUMO2-ΔGG) (Figure II.9). Taken together, we conclude that SUMO2 is sufficient for cardiac cell death.

Calpain2 is modified by SUMO2 in the heart. As the calcium-dependent protease, Calpain2 is modified by SUMO and is involved in cell death (102, 112).

In addition, sumoylation of Calpain2 results in the activation of its enzymatic reaction (102). Heart specific overexpression of Calpain2 causes reduced cardiac function including LVESD (left ventricular end-systolic dimension), FS (fractional shortening) just one week after induction (103). There are mainly three different isotypes of Calpain in the heart : Calpain1, 2 and 3 (113). Calpain1 and 2 are expressed ubiquitously but Calpain3 is found only in skeletal muscle. We examined sumo-dependent modification in vitro for all three isoforms. Only

Calpain2 showed SUMO dependent modification. Thus SUMO-conjugation of

Calpain2 in murine hearts was tested under the enhanced expression of SUMO2.

Using immunoblot and immunoprecipitation assay, SUMO2 specific modification of Calpain2 was observed in the SUMO2 Tg (Figure II.10.a, b). Because human heart failure promoted SUMO2/3 conjugation (Figure II.1.a), we also examined

Calpain2 modification in the human hearts. Multiple additional bands of Calpain2

Figure II.9. Cardiac specific cell death on cultured neonatal rat cardiomyocytes (NRCMs). a, Immunostaining for Annexin V (red) on NRCMs after 36hrs after active SUMO2 (AdSUMO2-GG), diglycine deleted SUMO2 (AdSUMO2-ΔGG) or empty (AdGFP) adenovirus infection at MOI=100. Overexpression of active SUMO2 in NRCMs resulted in increased Annexin V positive cells. It indicated that cell death was facilitated. Bar, 50µm. b, Quantification of apoptosis in NRCMs. Annexin V positive cells were scored at least 100 cells from tree independent experiments and averaged. (**P<0.001 vs AdGFP)

Figure II.10. Calpain2 is modified by SUMO2/3 and is involved in pathologic hearts. a, b, WT and α-MHC-SUMO2 Tg heart tissue homogenates were assessed for Calpain2 in immunoblot (IB) (a) or immunoprecipitation (IP) analysis (b). Enhanced expression of SUMO2 caused Calpain2 modification as shown a. Those additional bands were confirmed as sumoylation, especially by SUOMO2 (b). c, d, Human failing heart also exhibited additional upper band of Calpain2 in IB (c). Those modified bands were also proved as SUMO2/3-dependent conjugation of Calpain2 in IP (d). SUMO2/3ylated Calpain2 was clearly increased in dilated cardiomyopathic human hearts. IP was performed using agarose- conjugated anti-SUMO2/3 antibody for pulldown, followed by immunoblot using anti-Calpain2 antibody.

51 in failing human hearts were determined as SUMO2-specific (Figure II.10.c,d).

These data suggest that increase in SUMOylation of Calpain2 is the conserved observation in diseased heart both in murine and human.

Calpain-2 is involved in caspase-independent cell death pathway along with

ER stress (114). To examine whether the SUMO2-conjugated Calpain2 associates directly with apoptosis, we used adenoviral mediated gene delivery system in NRCMs. Through Annexin-V staining assay, a 2.5-fold increase was observed in the Calpain2 (AdCalpain2) overexpressed NRCMs respectively.

Interestingly, this increase was accelerated by SUMO2 fused Calpain-2

(AdSUMO2-GA-Calpain2) introduction. The number of Annexin-V positive NRCM were abrogated by forced expression of SUMO2 conjugation site mutated

Calpain2 (AdCalpain2-K390R) (Figure II.11), suggesting that SUMO2 dependent modification of Calpain2 induces cardiac muscle death directly.

We examined that SUMO2-dependent modification influences activity of

Calpain2 alternatevely using synthetic substrate, t-BOC. Calpain2 expressing

HeLa cells showed 1.8-fold increased activity in HeLa cells (Figure II.12).

Interestingly, even higher enzymatic activity was observed in SUMO2 fused

Calpain2 (AdSUMO2-GA-Calpain2) expressing cells whereas, mutant Calpain2 deleting the site for SUMO2 conjugation (Calpain2-K390R) revealed decreased enzymatic activity. This result indicates that SUMO2-conjugation facilitates catalytic activity of Calpain2. Of note, SUMO2 conjugation mutant Calpain2

Figure II.11. Calpain2 induced cell death is enhanced by SUMO2 conjugation. a, Adenoviral delivery of Calpain2 induced cell death on NRCMs based on Annexin V staining. SUMO2 conjugation promoted cardiomyocyte death (AdSUMO2-GA-Calpain2). This effect was diminished on sumoylation site mutated Calpain2 overexpressed NRCMs (AdCalpain2K390R). Bar, 50µm. b, Quantitative analysis of Annexin V labeling. The number of double positive (Annexin V/GFP) NRCMs was chosen from at least 100 cells on randomly selected field. Data were collected 3 independent experiments. All adenoviral constructs were delivered at MOI=100. (**P<0.001 vs AdGFP, # P<0.005 vs AdCalpain2)

Figure II.12. The proteolytic activity of Calpain2 is enhanced by SUMO2 conjugation. HeLa cells were transfected by cDNA encoding each construct, Calpain2, SUMO2 fused Calpain2 (SUMO2-GA-Calpain2) and mutant Calpain2 (Calpain2-K390R). Trasfected cells were selected by GFP expression. Synthetic substrate, t-BOC-LM-CMAC was visualized cleavage-dependently by Calpain. Cells were scored by double positive signals of GFP and t-BOC-LM-CMAC. SUMO2 enhanced proteolytic activity of Calpain2. However this increase was attenuated by mutant Calpain2, to which SUMO2 was not conjugated. Data were collected independent experiments. (**P<0.0005 vs GFP, # P<0.005 vs Calpain2)

54 (Calpain2-K390R) still showed significantly high proteolytic activity compared with control (GFP only). It might be explained by several reasons. Since the synthetic substrate, t-BOC-LM-CMAC responds to both Calpain1 and Calpain2, it would be hindered. Also, the endogenous Calpain1 would be affected by activated

Calpain2 since this analysis was performed without any Calpain1 inhibitor.

Calpastatin is modified by SUMO2 in pathologic hearts. Homeostatic protein turn-over including contractile machinery is critical to maintain the beating ability in the heart. Calpain-dependent proteolytic pathway is proposed as one of system especially, in the skeletal muscle (75, 115, 116). The Calpastatin, an endogenous inhibitor prevents Calpains proteolytic activation by binding (80). Unbalanced expression of Calpastatin resulted in Calpainopathy including dilated chamber, reactivation of fetal genes and decreased cardiac contractility with distorted myofibrillar architecture (103). We hypothesized that increased SUMO2 protein level hits not only Calpain2 but also its binding partner, Calpastatin under stress condition. To identify Calpastatin protein level in the diseased heart model, an immunoblot analysis was performed in SUMO2 Tg and failured human hearts.

Both diseased hearts represented additional bands above Calpastatin band (90 kDa) respectively (Figure II.13. a, c). Immunoprecipitation analysis was used to confirm that Calpastatin was at least partially conjugated by SUMO2 in the pathologic hearts (Figure II.13. b, d).

55 56

Figure II.13. Calpastatin is modified by SUMO2/3 and is involved in pathologic hearts. a, b, WT and heart specific SUMO2 Tg heart tissue homogenates were used for calpastastin in immunoblot (IB) (a) or immunoprecipitation (IP) analysis (b). Increase in SUMO2 level resulted in Calpastatin modification as shown a. Those additional slow migration bands were turned as SUMO-conjugated one, especially by SUOMO2 (b). c, d, Human failing heart also showed additional upper band of Calpastatin in IB (c). Those modified bands were also proved as SUMO2/3-dependent modified Calpastatin in IP (d). SUMO2/3ylated Calpastatin was visibly increased in dilated cardiomyopathic human hearts. The agarose-conjugated anti-SUMO2/3 antibody was utilized for pulldown, followed by immunoblot using anti-Calpastatin antibody. NS, none specific.

57 Calpastatin is a novel SUMO2 target. Protein kinase A (PKA) and protein kinase C (PKC) have been reported to modify and to modulate Calpastatin activity

(117, 118). Here we showed that pathologic hearts stimulated SUMO-dependent modification of Calpastatin (Figure II.13). To define the molecular basis, in vitro system was applied along with transfection method. Although SUMO1 was even aggressively conjugated, both SUMO1 and SUMO2 were responsible for

Calpastatin modification (Figure II.14.a). To verify that those slowly migrating bands were SUMO-dependent, conjugation defective SUMO (SUMO-∆GG-flag), which served as the binding platform was deleted. Indeed, cotransfection of mutant SUMO (SUMO-∆GG-flag) and Calpastatin (CAST-V5) into HeLa cells abolished upper defined but slower migrating band of WT SUMO (SUMO-flag)

(Figure II.14.b). Taken together, these results suggest that SUMOylation is a novel post-translational modification of Calpastatin. Next we asked how SUMO- dependent modification affects Calpastatin. SUMOylation has been reported essential to stabilize HIF1α during hypoxia (119) and critical subcellular localization of target such as Drp-1 (59), leading to the hypothesis that SUMO- dependent conjugation of Calpastatin is critical for its stability, cellular localization and interaction ability to Calpain2.

Figure II.14. Calpastatin is modified by SUMO. a. cDNAs coding Calpastatin (CAST) with V5 tag and SUMO1 or SUMO2 in which flag was conjugated were cotransfected into HeLa cells. Ni-NTA pull down showed that Calpastatin is sensitive to both SUMO1 and SUMO2 although SUMO1 dependent conjugation was active than SUMO2 dependent one. b. Modified upper bands detected in a were confirmed as SUMO-dependent conjugated Calpastatin. Mutant SUMO cDNA coding SUMO without diglycine (ΔGG) acting as immatured SUMO, fused Calpastatin cotransfected HeLa cells completely lost SUMO-conjugated upper band. HMW, high molecular weight globally modified SUMO targets.

59 To evaluate the ability of SUMO2 conjugation on Calpastatin stability, a protein half-life analysis was performed using cycloheximide (CHX). Because CHX blocks protein synthesis, it is useful to study already synthesized protein degradation rate. Clear and slow migrating Calpastatin band indicating SUMO conjugated

Calpastatin was attenuated upon cotransfection of Calpastatin (CAST-V5) along with SUMO (SUMO-Flag) in the CHX treatment, as indicated in Figure II.15.a.

Notably, free Calpastatin protein level was not affected even if protein synthesis was inhibited for 12 hrs by CHX, in accordance in the strong stability of

Calpastatin against the heat. This result was reproducible in the presence of

SUMO fused Calpastatin (SUMO-CAST-V5) whereas, not in the unSUMOylated

Calpastatin (CAST-V5) expressed HeLa cells (Figure II.15.b). These results support our hypothesis that SUMO-conjugation makes the calpastastin unstable.

Calpastatin itself is also controlled by several proteases such as caspases,

Calpains, which is critical for proteolytic activity of Calpains (74, 80, 114, 120).

Together, next prompted question was which protease is involved in SUMO- dependent degradatin of Calpastatin. SUMO fused Calpastatin expressing vector was transfected into HeLa cells with each protease inhibitors including Calpain1,

2, cathepsins and caspases under blocking protein synthesis. SUMO fused

Calpastatin was quite resistant to pan caspase inhibitor treatment (Figure II.15.c), demonstrating that SUMO conjutaed Calpastatin is sensitive to caspase mediated degradation.

Figure II.15. Sumoylation promotes Calpastatin degradation. a, b. SUMO conjugated Calpastatin (a) or SUMO fused Calpastatin (b) showed unstability. In block of newly synthesized protein through cycloheximide treatment, SUMOylated Calpastatin was degraded time-dependently compared to free calpastsatin. c. Among known Calpastatin degrading enzymes, SUMO-conjugation made Calpastatin sensitive to caspase, leading to release from Calpain2.

61 It is quite consistent with the concept that caspase mediated degradation of

Calpastatin contributes to apoptosis although activated Calpains also tear down

Calpastatin (114).

To establish whether SUMOylation promotes subcellular translocation of

Calpastatin, free Calpastatin (CAST-V5) or SUMO2 fused Calpastatin (SUMO2-

CAST-V5) was overexpressed in HeLa Cells. Capastatin with or without SUMO2 was visualized through immunofluorescence staining using anti-V5 Ab and counted as aggregated or cytoplasmic distribution (Figure II.16), indicating that

Calpastatin changes its subcellular localization dependent on SUMO2 conjugation.

Recent studies demonstrated that phosphorylation regulates subcellular localization of Calpastatin and that cAMP mediated dephosphorylated Calpastatin enters aggregated inactive distribution (117). It is consistent with our finding

(Figure II.16), demonstrating that it is possible that increased aggregate pattern of

SUMO2 fused Calpastatin is inactive. All data support that the inhibitory role of

Calpastatin against Calpain2 is modulated by SUMO2-dependent modification.

Figure II.16. SUMOylation changes intracellular Calpastatin localization. a, CAST or SUMO2 fused CAST in HeLa cells were visualized by immunostaining through V5 tag (green fluorescence). DAPI staining (blue fluorescence) was used as nuclear counter staining. Diffused pattern (left panel) in cytoplasm of CAST is totally changed under SUMO2 conjugation (right panel). Bar indicates 50µm. b, Bar graph indicates the percentage of cells with each localization. Data were collected from 4 independent experiments. (**P<0.001).

63 2.4 Discussion

Dynamic SUMOylation in congestive heart disease. SUMO conjugation is one of growing reversible posttranslational modification (PTM), involving many different biologic processes such as cell death, transcription, cell cycle, protein stability etc. Although its role in cardiac development is well established especially with SUMO1, very little is known about the role of SUMO2/3 particularly in congestive heart disease. Here we have identified that SUMO2/3 is upregulated in human and experimental animal failing hearts. It is correlated that SUMO2/3 is more sensitive to environmental stress (47). This elevated SUMO2/3 level is not protective but detrimental for cardiac function based on heart specific SUMO2 Tg.

Increased global SUMO2/3 conjugation level (high molecular weight, HMW) is even critical for modulation of their targets to affect on cardiac function. Since one of SUMO2 Tg lines showed no HMW level change in spite of increased free

SUMO2/3, which did not reveal any cardiac defects under normal condition whereas Ang-II stimulation caused significantly increased global SUMO2/3 conjugation and sudden death (Figure II.17). Consistently, SUMO E3 ligase, protein inhibitor of activated STAT 2 (PIAS2) was up-regulated in human failing hearts, indicating that SUMO conjugation is facilitated actively in stressed hearts

(Figure II.18). It suggests that SUMO2/3 conjugation is more important for cardiac remodeling responding to stress. Alternatively, the contribution of SUMO1 conjugation to heart is very controversial.

Figure II.17. Stress inducing SUMO2/3 conjugation in the heart. a. Homogenates from WT and SUMO2 Tg hearts with or without Ang-II administration were applied to immunoblot using anti SUMO2/3 antibody. Ang-II administration induced SUMO2/3 conjugation in heart specific SUMO2 Tg, in which global SUMO2/3 conjugation was not observed. Interestingly, elevated SUMO2 conjugation level correlated with mortality rate (b). Exaggeratedly increased SUMO2 conjugation would result in fatal phenotypes.

Figure II. 18. Pathologic human hearts induced active SUMO conjugation. Human hearts from normal or failing were prepared to obtain protein lysates. Immunoblot analysis was applied for PIAS2 protein level. Human Failing hearts and experimental animal model revealed active conjugation of SUMO2/3 (Figure II.1). SUMO E3 ligase, PIAS2 was also significantly enhanced in human failing hearts. Since SUMO E3 ligase accelerates SUMO conjugation in the SUMO pathway, it supports that increased global SUMO2/3 conjugation from human failing heart is caused by not only SUMO2/3 but also active conjugation pathway. Normal, healthy heart; NIS-HF, none ischemic heart failure (dilated cardiomyopathy) ; IS-HF, ischemic heart failure.

66 Ca2+ reuptake-involving ATPase, SERCA2a has been modified by SUMO1. Both

SUMO1 and SUMO1 dependent PTM of SERCA2a have been decreased during heart failure (96). However the others reported that increase in SUMO1 has been induced in pathologic hearts in purpose of protection (98, 121) or pathological progress (97). Our approach has identified that global SUMO1 conjugation level is increased or unchanged in human failing hearts and Ang-II stimulated murine hearts (Figure II.19). Thus increased SUMO1 conjugation is more related to protective function.

SUMO2 inducing cell death. The prominent outcome upon SUMO2 overexpression in the neonatal rat cardiomyocytes (NRCMs) was apoptotic cell death, indicating that primary effect of SUMO2 enhancement in the hearts is the myocyte death. This detrimental effect was abolished by conjugation activity dead

SUMO2 mutant (AdSUMO2-ΔGG) that ablates diglycine residues on SUMO2. In accordance, clinically increased cardiomyocyte death was reported as a clear phenomenon in human heart failure (70, 71). Several animal studies also supported that heart failure was induced sufficiently by heart specific apoptosis

(72, 122, 123). Apparently, SUMO2/3 can be sufficient to control cell death in cardiac muscle tissues.

Cardiac cell death caused by upregulated SUMO2 is possibly associated with

Ca2+ -dependent cysteine protease, Calpain 2 based on our observation indicating its increased SUMO2 specific sumoylation in SUMO2 Tg and human

Figure II.19. Stress induced abnormal hearts elevated SUMO1-dependent modification. Immunoblot analysis was performed by utilizing whole heart homogenates to check SUMO1 alteration. Stress induced overload by Ang-II treatment showed similar pattern of SUMO1 compared to SUMO2/3 conjugation (Figure II.1). Stress induced pathologic condition triggers SUMO1 conjugation also. However, SUMO1 conjugation seems the protective effect than deteriorative one.

68 failing hearts (Figure II.10). Calpain has been known to be involved in mitochondria independent cell death pathway (80). Moreover, Calpain 2 has been identified to be regulated by SUMO2-dependent modification (102). Here we confirmed that cardiac muscle cells are also affected by enhanced SUMO2 conjugated Calpain2, especially apoptotic cell death. It is intriguing that the more

Ca2+-sensitive Calpain, Calpain1, has not shown a SUMO dependent modification in our trial. Only Calpain2 showed SUMO2 dependent modification, resulting in enhanced proteolytic activity. The little effect of Calpain2 overexpression for cardiac remodeling in a previous study (103) might result from no consideration of

SUMO2 in their study. Interestingly we have found that its endogenous inhibitor,

Calpastatin is also modified by SUMO2 specifically in human failing heart and animal model. SUMO2 dependent modification modulates the stability and subcellular localization of Calpastatin. It supports our idea that enhanced SUMO2 protein level hits both Calpain2 and Calpastatin, resulting in the enhancement of enzymatic activity of Calpain2 and weakening Calpastatin’s inhibitory function by altered stability and localization within cell. Although Calpastatin has been reported that its subcellular localization is changed by phosphorylation (117, 118,

124), little is known about how phosphorylated Calpastatin affects on association with Calpain. Moreover, under normal Ca2+ condition, there is no evidence to explain how Calpain2 and Calpastatin regulate each other. Here our study strongly supports that enhanced global SUMO2/3 conjugation at least, in partial involves in Calpain2-Calpastatin system to induce mitochondria-independent

69 apoptosis pathway by activating Calpain2 but by releasing its inhibitory partner,

Calpastatin. However still remaining question is the relationship between phosphorylation and SUMOylation for control of the same system.

Cardiomyocyte survival and SUMO2. Along with increased cell death, cardiomyocyte proliferation has been also decreased remarkably in SUMO2 Tg.

Very recently, AMP-activated protein kinase (AMPK) has been discovered as a novel SUMO2 target (125). Moreover, AMPK has been activated specifically by

SUMO2 conjugation to AMPKβ2. Activated AMPK by reduced mitochondrial ATP production negatively regulates cell cycle checkpoint in Drosophila (126, 127), indicating mitochondrial complex mediated regulation of cell proliferation. Since

SUMO2-dependently modified AMPKβ2 has resulted in increased activity, it provides obvious possibility that decreased cardiomyocyte proliferation in heart specific SUMO2 Tg is due to activated AMPK via enhanced SUMO2 conjugation.

Furthermore, human familiar mutation in AMPK mimic animal model, PRKAG2-

N488I Tg under α-MHC promoter has developed hypertrophic heart with elevated

AMPK activity (128). Collectively, these emphasize that the important role of

SUMO2 dependent modification in cell proliferation contributes to cardiomyocyte survival during cardiac reprogramming.

70 Chapter III.

SENP5-dependent deSUMOylation pathway

in dilated cardiomyopathy

71 3.1. Introduction

SENP1 representing preferences to all SUMO isoforms, has shown that its knock-out (SENP1 KO) mouse model reveals the sensitivity to hypoxia, suggesting SENP1 involves in HIF1α stability (119). In another instance of even preferences for SUMO1 and 2/3, SENP2 knock-out (SENP2 KO) mice have developed congenital heart disease with growth retardation (129). The primary cause of embryonic or neonatal death in SENP2 KO mice is atrial septum defects

(ASD) and/or ventricular septum defects (VSD) induced by significantly decreased cell proliferation. Those studies suggest that each SENP might involve in a distinct biologic pathway despite their similar target preferences for SUMO isoforms.

SENP3 has been involved in protein turnover via stable interaction with CHIP

(co-chaperone/ubiquitin ligase carboxyl terminus of Hsc 70-interacting protein) and Hsp90 (heat shock protein 90) (130). SENP7 has regulated epithelial to mesenchymal transition by SUMO2/3-dependnet modification of heterochromatin protein 1α (HP1α) (131, 132). In that study, larger metastatic lesions in lung were formed upon SENP7 silencing (131). Both SENP3 and SENP7 prefer SUMO2/3 rather than SUMO1 as their substrates but they have different enzymatic activity:

SENP3 tends to mediate deconjugation more whereas SENP7 mediates SUMO chain editing (133). Collectively, those suggest that each SENP has a specific role

72 to control biologic phenomena as variable manners including substrate specificity, different subcellular occupancy and distinct enzymatic activity and so on.

SENP5 favoring SUMO2/3-deconjugation has revealed the distinct subellular pattern to control cell proliferation and mitochondrial dynamics by shuttling between nucleolus and mitochondria (36, 37, 57). During mitotic cell division,

SENP5 has escaped from nucleus to move to mitochondria (36). Mitochondria- staying SENP5 could involve in mitochondrial fragmentation to distribute mitochondria evenly into daughter cells. Additionaly, SENP5 has been shown to be required for cell proliferation (57). Mitochondrial fission protein containing

GTPase activity, Drp-1 has been reported to be deconjugated as SENP5- dependent manner although its SUMOylation seems SUMO1-dependent manner

(59). Ectopic expression of SENP5 induced mitochondrial fission with active cell death. More interestingly, Drp-1 mutation by ENU has result in dilated cardiomyopathy with low mitochondrial function in the mouse (134). It is the first report to link SENP5 with disease, particularly heart disease. Additionally, heart is very energy demanding organ to maintain contraction continuously, thus it habors mitochondria abundantly. Given its potential biological activity and unexplored physiological significance, this study could provide valuable evidence to explain the possible role of SENP5 for cardiac disease development.

73 3.2. Material and Methods

Human heart samples. Failing human left ventricular samples obtained from St

Luke Episcopal Hospital were removed during cardiac transplantation. Failing hearts were from adults diagnosed with dilated cardiomyopathy (DCM). Normal, non-failing adult human heart samples were collected from motor-vehicle accidents dead donors. The left ventricular samples were used for studies.

Animal studies. The α-MHC-SENP5 transgene was constructed as described

(93)(110) under the control of mouse α-MHC promoter (provided by Dr. J.

Robbins, University of Cincinnati). DNA sequecing was used to confirm the orientation of inserted transgene. The pronucleus microinjection of transgene construct into the fertilized eggs from FVB mice to generate founder (F0) SENP5 transgenic (SENP5Tg) mice. Foudners were crossed back with C57Bl/6 mice for pure background. Genomic DNA was isolated from tail biopsies performed on weaned animals (approximately 3-week old pups) and screened by PCR. The expression of transgene in SUMO2Tg mouse hearts was verified by Western blot and/or qPCR. Animal work was followed by accordance with IACUC approval.

74 and heart dimensions was performed and evaluated under genotype-blinded condition. Lab. member, Wei Yu helped animal cardiac functionality analysis through echocardiography.

System (Advanced Microscopy Techniques Corp). The process for EM was done by MD Anderson HREM facility and Texas Heart Institute pathology core.

Mitochondrial enzyme activity assay. Mice hearts mitochondrial enzyme activities were determined spectrophotometrically as described previously with slight modification (135, 136). Mitochondria were isolated from mice cardiac ventricular tissues by homogenization in lysis buffer (100mM potassium phosphate buffer, 1mM EDTA, 1mM β-mercaptoethanol (pH7.4), protease inhibitor and 10mM NEM). Tissue extracts were incubated for 60min at 4°C to ensure complete enzyme extraction. 0.1% Triton X-100 was added to permeabilize cell membranes. The mitochondrial fraction was separated by further differential centrifugation. Before assay, homogenates or mitochondrial fraction were thawed and frozen three times. All mitochondrial enzyme activities were measured using quartz cuvette. Complex I activity was determined by monitoring the reduction of ferricyanide at 420nm. Homogenates were added to

75 reaction buffer (120mM Triethanolamine (pH7.8), 0.5mM ferricyanide, 5mM rotenone) to start the reaction at 30°C for 2min. The NADH-dependent ferricyanide reduction was measured. Complex II activity was measured as the decrease rate of 2,6-dichlorophenolidophenol (DCPIP) when coupled to the complex II catalyzed reduction of decylubiquinone (DB) at 600 or 750nm.

Homogenates (80µg) or mitochondria (30µg) were added to reaction reagents

(0.1M potassium phosphate buffer (pH7.4), 20mM succinate. After this mixture was incubated at 30°C for 5min, antimycin A (2µg), rotenone (2µg), KCN (2mM),

DCPIP (50µM) were added. The reaction was started by DB (50µM) addition. The change of absorbance was monitored during 5min reaction. Complex IV activity represents the activity of cytochrome c oxidase by monitoring the oxidation of reduced cytochrome c at 550nm. Protein extracts (80µg) or mitochondria (30µg) were added to reaction reagents (20mM potassium phosphate (pH7.4) and 20mM succinate). 10µM cytochrome c was added to initiate the reaction at 30°C for 1-

2min. Citrate synthase activity was examined by monitoring the absorbance at

412nm using especially unsonicated mitochondrial proteins. The frozen cardiac tissues were homogenized in 20% (wt/vol) lysis buffer (20mM HEPES, 10mM

EDTA, 1mM EGTA, 1mM DTT, 5mM MgCl2, 0.1% Triton X-100). Extracts were incubated at 4°C for 60min to ensure complete enzyme extraction. Heart homogenates reacted with 0.05mM oxaloacetate in reaction buffer containing

20mM HEPES, 1mM EGTA, 220mM sucrose, 40mM KCl, 0.1mM DTNB and

0.1mM Acetyl-CoA (pH7.4) at RT. The reaction mixture excluding oxaloacetate

76 was used as reference. All measured activities were normalized by WT record and expressed as percentile.

Histology. Hearts were removed and fixed overnight in 4% paraformaldehyde

(PFA), embedded in paraffin. Hematoxylin and eosin (H&E) or Masson’s trichrome staining was performed on heart sections (10 µm) according to standard protocols. Ca2+ deposits from paraffin sections (5 µm) were assessed using von Kossa staining according to the manufacturer’s protocol (Diagnostic

BioSystems Inc.) with nuclear fast red staining as counter staining.

Immunohistochemistry. To detect DNA synthesis in mice hearts, BrdU was injected intraperitoneally into pregnant female mice bearing embryos or postnatal pups at a dose of 100µg/g body weight. Sixteen hours after injection, hearts were collected, fixed in 4% PFA, dehydrated in increasing concentrations of ethanol (up to final of 100%), and sectioned with 5 µm thickeness. BrdU staining was conducted with BrdU in situ Detection Kit according to the manufacturer’s protocol

(BD Pharmingen). For staining, commonly antigen retrieval was performed by boiling slides in 10 mM sodium citrate (pH 6.0, 15 min) followed by blocking in

10% normal goat serum (NGS) for 1 h at room temperature (RT). Ki67 or phosphoylated Histone H3 staining was performed with anti-Ki67 antibody (1:200,

Santa Cruz) or anti-phopho Histone 3-Ser10 (1:200, Millipore). Anti-caveolin 3 antibody (1:8000, Santa Cruz) was used as a cardiomyocyte marker. Sections were then incubated with the following secondary antibodies: anti-rabbit TRITC antibody or anti-mouse FITC antibody (Jackson ImmunoResearch). Sections

77 were mounted with Vectashield including DAPI (Vector Laboratory). BrdU and

Ki67 labeling indices were calculated based on scores obtained from 6 different embryonic or adult hearts (at least 3 for each group), with each at least 100 cells

(embryonic hearts) or 5 fields (adult hearts) randomly selected for evaluation.

Assessment of TUNEL was conducted using ApopTag® Red In situ Apoptosis

Detection Kit (Milipore) based on the standard protocol provided by the manufacturer.

Generation of the recombinant adenoviruses. Recombinant adenoviral vectors were generated by using the AdEasy XL Adenoviral Vector System (Stratagene).

A DNA fragment containing the full length human SENP5 which was the gift of

Dr.Ed Yeh, The University of Texas MD Anderson Cancer Center and mutant human SENP5 which does not have any enzymatic activity coding sequences were subcloned into the pShuttle-IRES-hrGFP1 vector. The entire expression cassettes from the resulting vectors were recombined in BJ5183 bacterial strain with serotype 5 first-generation adenoviral backbone, AdEasy-1 (Stratagene).

These recombinant adenoviral backbones were transfected into AD293 cells

(Stratagene) to generate infectious viral particles. Viral titer was determined by the tissue culture infectious dose (TCID50) method (111). Cardiomyocytes were infected with recombinant adenoviruses for 2hrs at a multiplicity of infection

(M.O.I) of 100 particles/cell and incubated for additional 24-48 hrs to ensure transgene expression.

78 Neonatal rat ventricular cardiomyocytes culture. Neonatal rat cardiomyocytes were prepared as described (137). Breifly, primary cultures of cardiac myocytes from 1- to 3-day- old Sprague-Dawley rats were prepared. Ventricles were enzymatically dissociated, and the cells were collected for only cardiomyocytes using Percoll (GE Healthcare) gradients. Cells were plated onto the collagen- coated cover slips at a density of 200cells/mm2 and maintained in cardiomyocytes culture medium (DMEM supplemented with 10% FBS and 2mM of L-glutamine).

Immunofluorescence staining. Cardiomyocytes grown on collagen-coated cover slips or two well Lab-Tek II chamber slide (Nunc) were fixed in 3.5% PFA and permeabilized with 0.5% Triton X-100. Then slips were incubated with 10% normal goat serum for blocking for 1hr at room temperature. Sequentially the cells were treated with anti-caspase 3 antibody (1:200, Santa Cruz) and TRITC- conjugated goat anti-rabbit IgG (Jackson ImmunoResearch) and then mounted with Vectashield containing DAPI (Vector Laboratory). Annexin V staining was performed using Annexin V Alexa Fluor 594 (Invitrogen). Neonatal rat cardiomyocytes on Lab-Tek II chamber slides were added by 5ul of the Annexin V

Alexa Fluor 594 and incubated for 15 minutes at room temperature. Then cells were washed and mounted with Vectashield containing DAPI (Vector Laboratory).

AIF (hAIF) point mutations, K85R (conversion of lysine 85 to arginine) and K442R

(conversion of lysine 442 to arginine) were done using Stratagene mutagenesis kit based on manufacturere’s protocol. Mutations were verified by sequencing.

Transient transfections were performed on HeLa cells through Lipofectamine

79 2000 (Invitrogen). Briefly, plasmids applied with 0.5µg were point mutated hAIFs (

AIF-K85R, AIF-K442R) containing p3xFLAG-CMV-10 expression vectors (Sigma-

Aldrich). Media were changed to complete media at next day. Cells were maintained additional day. Transfections were done in triplicate. HeLa cells were fixed in 3.5% PFA and permeabilized with 0.5% Triton X-100. After blocking with

10% NGS, HeLa cells were treated with anti -Flag antibody (1:1000, Invitrogen) and subsequently FITC-conjugated goat anti-mouse IgG (Jackson

ImmunoResearch) and then mounted with Vectashield containing DAPI (Vector

Laboratory).

Immunoblotting & Immunoprecipitation. Briefly, 100 µg of protein lysates from mouse hearts or 50 µg of NRCM containing overexpressed proteins were purified in the presence of 25 mM isopeptidase inhibitor N-ethylmaleimide (NEM), which prevents substrate desumoylation. These protein lysates were subsequently runned to 4-12% NuPAGE, transferred to PVDF membrane, detected with the desired antibody, which was then visualized with chemiluminescence.

For immunoprecipitation, protein lysates from frozen mouse hearts were prepared. 400 µg of total protein was diluted to a final concentration of 0.4 mg/ml in lysis buffer. 5 µg of bead-conjugated anti-SUMO-2/3 antibody (Santa Cruz) was applied to the sample and incubated for 2 hours at 4°C on a rotary platform. The beads were subsequently pelleted by centrifugation, washed five times with lysis buffer. Specifically bound proteins were eluted by 4x SDS sample buffer, subjected to SDS- PAGE, and visualized by autoradiography.

80 Subcellular Fractionation. Appropriate amount of mice hearts was minced in 5 volumes of mitochondrial isolation buffer (220mM mannitol, 68mM sucrose, 80mM

KCl, 0.5mM EGTA, 2mM MgAc2, 10mM HEPES (pH7.4), protease inhibitor cocktail and 20mM NEM) as describe (59). Heart tissues were ruptured and homogenized under speed of middle stroke to fast stroke at least fourtimes. The lysates were centrifuged at 500×g for 5min at 4°C. The resuspended pellet in the mitochondrial isolation buffer was homogenized again and separated twice more.

Collected supernatants were pooled and centrifuged at 1000×g for 5min at 4°C.

Result pellets were collected as the nuclear fraction. Cleared supernatants were centrifuged at 12,000×g for 15min at 4°C. Mitochondrial pellets were washed and resuspended in mitochondrial isolation buffer. Remained supernatants were saved as cytosolic fraction. Samples were further resuspended in SDS sample buffer and probed with anti-SENP5 antibody (Abcam), anti -Drp-1 antibody (BD

Transduction Lab.) and anti-TOM 20 antibody (Santa Cruz) for immunoblotting analaysis. For immunoprecipitation, only mitochondrial fractions were used. 70µg of mitochondrial fraction was diluted in mitochondrial isolation buffer and an appropriate amount (5ug) of bead-conjugated anti-SUMO-2/3 antibody was added to the sample and incubated for 2 hours at 4°C. The beads were subsequently pelleted by centrifugation, washed five times with mitochondrial isolation buffer.

Bound proteins were eluted by 4x SDS sample buffer, subjected to SDS- PAGE, and visualized by autoradiography.

RNA isolation and mRNA quantification. Total RNA was isolated from age-

81 matched snap-frozen hearts using Trizol reagent (Invitrogen). cDNA was prepared by cloned reverse transcriptase (Invitrogen). Cardiac stress indicating gene expression was determined by qPCR (7900HT, Applied Biosystems) using

FastStart SYBR Green master (ROX)(Roche). mRNA levels were normalized by the housekeeping gene, GAPDH.

P value < 0.05 was considered statistically significant and P<0.01 as highly significant.

3.3 Results

SENP5 is upregulated during human heart failure. Sumoylation process is very dynamic and is known to involve in cardiovascular development. Although enhanced expression of SENP2 results in congenital heart disease (129), very little is known about other SUMO pathway components during heart disease. To identify the SENP5 expression level, immunoblot analysis was performed using human hearts. Human heart failure revealed highly elevated SENP5 protein level

(Figure III.1.a). However the function of elevated SENP5 protein level in human heart failure is not clear. Thus human SENP5 transgene was applied to express under the α-MHC promoter (SENP5 Tg). Three lines of Tg, #2723, #2737 and

#2739 showed forced expression of SENP5 in hearts (Figure III.1.b). We also

82 confirmed transgene expression through Southern blot, RT-PCR and qPCR (data not shown). Those supported SENP5 is overexpressed under the α-MHC promoter. SENP5 transgene expression level was compared with diseased hearts from human based on immunoblot analysis (Figure III.1.a, b). Cardiac specific

SENP5 Tg represented comparably and similarly high protein level to human

Cardiac specific SENP5 Tg represented comparably and similarly high protein level to human SENP5 in diseased hearts (Figure III.1.c). It suggested that cardiac specific SENP5 Tg could be used as a good animal model to study the function of increased SENP5 protein level in diseased human hearts. Further,

SENP5 Tg revealed SUMO1 conjugation was not changed (Figure III.1.d) but

SUMO2/3 conjugation was decreased (Figure III.1.e), indicating SUMO2/3 specificity of SENP5 Tg mice.

83 84

Figure III.1. Increased SENP5 expression level in human heart failure. a, SENP5 in human heart tissues (n=3 per non-failing heart (cont), n=3 for non- ischemic heart failure (NIS-HF), n=3 for ischemic heart failure (IS-HF)) are represented. b, Cardiac specific (α-MHC promoter) SENP5 Tg were generated. Proteins were prepared from ventricles of Tg at 2-month-old age for Immonoblot analysis using α-SENP5 antibody. GAPDH was used as a loading control. c, Comparison of SENP5 increased level from Immunoblot analysis (a, b) by quantification using ImageJ. Error bars represent mean ± SEM. (**P<0.05 vs cont for human, vs WT for mouse, n=5-7 for human, n=3 for mouse/group) d, e, Sumo- conjugation was admitted using Immunoblot analysis on hearts from b. Increased unconjugated SUMO2/3 (arrow) resulted from SENP5 transgene (d). SUMO-1 conjugation was not affected by SENP5 overexpression (e).

85 SENP5 Tg mice developed cardiac pathology. To address whether elevated

SENP5 is detrimental to the heart, 2723 line of heart specific SENP5 Tg was selected. This Tgs grew without any apparent abnormalities including behavior, fertility for several months. However α-MHC promoter bearing SENP5 Tg showed sudden death starting 8 days after birth, which was most severe impairment. Two- month-old hearts from SENP5 Tg were enlarged (approximately 1.5 times increase compared to WT) and had significantly decreased cardiac function

(Figure III.2) with extensive fibrosis (Figure III.3.b). α-MHC promoter carrying

SENP5 Tg revealed the remarkable reduction in cardiac contractility at diastole and systole, a significant increase in left ventricular internal diameter at diastole

(LVIDd) and at systole (LVIDs), as well as reduced ejectional fraction (EF) along with impaired fractional shortening (FS) (Figure II.2.b). Another consequence of the cardiac stress response is upregulation of cardiac fetal genes. As shown in

Figure III.3.a, reactivation of markers, atrial natriuretic peptide (ANF), brain natriuretic peptide (BNP), β-myosin heave chain (β-MHC) were notably observed in SENP5 overexpressed murine hearts (Figure III.3.a). Expression of connective tissue growth factor (CTGF) and collagen type I (Col1) were also induced (data not shown), correlating with Masson’s trichrome staining (Figure III.3.b).

Interestingly, no difference in LV posterior wall thickness in diastole (LVPWd) indicated that SENP5 Tg hearts were not hypertrophic. This was confirmed by histological sections stained by H&E of Tgs which were dead suddenly around

P10 (Figure III.4.a).

Figure III.2. Cardiac defects resulting from cardiac overexpression of SENP5. a, Whole-mount pictures of the hearts at two-month-old. b, Analysis of cardiac function of Tg and WT littermates (n=5-6) by echocardiography. LVIDd: left ventricular internal diameter at end-diastole; LVIDs: left ventricular internal diameter at end-systole; LVPWd: left ventricle posterior wall thickness at end- diastole; EF: ejectional fraction; FS: fractional shortening; LV mass/BW: left ventricle mass-to body wight ratio (* P<0.05, **P<0.005 vs WT).

Figure III.3. Cardiac defects resulting from cardiac overexpression of SENP5. a, Total heart RNA was isolated from 2-month-old mice and prepared for the cDNA. Indicated markers of cardiac stress were determined using qPCR. mRNA levels are expressed as relative value to GAPDH. Error bars indicate SEM. (*P<0.05, **P<0.005 vs WT). b, sections stained with Masson’s Trichrome to address fibrosis. Bar, 10µm. Enhanced expression of SENP5 in heart resulted in intensive fibrosis (blue).

88 Sudden dead SENP5 Tg displayed dilated cardiomyopathy with thinner ventricular wall and dilated chamber volume accompanying large thrombi in ventricles (Figure III.4). All together, these results suggest that elevated SENP5 protein level observed in human failing hearts is involved in progress of heart disease, especially, dilated congestive heart failure.

Abberant cardiomyocyte proliferation and death upon forced expression of

SENP5. To investigate the basis of obviously reduced wall thickness, which is one of characters of dilated cardiomyopathy, cell proliferation was analyzed by Ki67 staining. SENP5 overexpressing animals selected at P8 showed a 3.5-fold decrease in cell proliferation compared with wild type (WT) (data not shown).

Consistently, cell death was also clearly increased at the same stage (data not shown). It was expected because the aged Tg already developed sudden death with dilated cardiomyopathy. Thus we selected 5-day-old time window at which heart specific SENP5 Tg did not represent any cardiac defects such as HW/BW ratio (Figure III.6.c). Unexpectedly, this time point (P5) already displayed decreased cardiomyocyte proliferation based on Ki67 staining (Figure III.5). Since

Ki67 staining did not distinguish DNA synthesis, S-phase and mitotic cell division, to gain further insight, DNA synthesis was first evaluated through BrdU staining.

The number of BrdU positive cardiac cells, integrating into DNA during S-phase was significantly decreased. Next, staining for phospho Histone H3, a marker of mitotic cell division was applied. The number of phospho Histon H3 (pH3) positive

Figure III.4. Abnormal hearts developed in cardiac SENP5 Tg. a, Cardiac specific SENP5 Tg died suddenly around P10. Whole-mount pictures of the hearts from WT and cardiac specific SENP5 Tg mice at P8 are shown in the top panel. Bottom panels show histological sections stained with H&E. SENP5 Tg develops big heart with thin chamber wall in ventricle. Bar, 200µm (bottom panel). b, Heart weight (mg) to body weight (g)(HW/BW) ratio in mice of a. SENP5 Tg reveals increased heart index. Error bar indicates SEM. (**P<0.001 vs WT)

Figure III.5. Aberrant cardiomyocyte proliferation in SENP5 Tg. Immunohistochemistry for cell proliferation markers at P5. a, Phospho- histone H3 (red) and DAPI (blue) staining show increased cell proliferation, especially mitotic cardiomyocytes. Statistical analysis in b represents the scores of at least 5 randomly selected fields per section. c, Ki67 staining represents decreased cell proliferation in SENP5 Tg hearts. d, statistical results represent average scores of 200 cells per section. Quantification of each positive staining is done at least three different samples. Caveolin-3 (Cav3, green) was used as a cardiomyocyte specific marker. Bar, 50µm. (**P<0.0005 vs WT)

Figure III.6. Abnormally increase in cell death in SENP5 Tg hearts. a, Double- immunofluorescence analysis of TUNEL(red) and Cav3(green) showed increased apoptosis in Tg at P5. Bar represents 100µm. b, The number of double positive (TUNEL+:Cav3+) cells were counted from at least three randomly selected field. Error bar represents mean ± SEM. (**P<0.005 vs WT, n=4) One of cardiac disease indicator, HW/BW ratio was not changed in SENP5 Tg at the same age of increased cell death (c). DATA were collected from 72 animals from 6 litters.

92 cardiomyocytes were remarkably reduced at SENP5 Tgs. These results suggest that Increased SENP5 protein level affects cardiomyoctye proliferation.

SUMO-dependent modification is known to involve in apoptotic cell death (44), indicating possibly desumoylation enzyme, SENP5 protein level might influence cell death. Also, dilated ventricle from SENP5 Tg might be possibly caused by cell death. Therefore TUNEL assay was applied to murine hearts at P5 (Figure III.6).

TUNEL positive dead cardiomyocytes were dramatically facilitated by heart specific SENP5 Tg compared to WT, in spite of no evidence in heart stress index.

Collectively, thinning ventricular wall of SENP5 Tg was caused by increased cell death and decreased cardiomyocyte division.

Increased cardiomyocyte death is SENP5-dependent. This in vivo observation provided a valuable clue that overexpression of transgene, SENP5 in murine hearts impacts on cell survival. Thus we wanted to examine whether significantly increased apoptosis is the direct result of SENP5 overexpression especially in the hearts. To address this, adenovirus carrying wild type SENP5 (AdSENP5-wt) and desumoyation activity dead SENP5 (AdSENP5-mut) were generated and transduced in neonatal rat ventricular cardiomyocytes (NRCMs). Annexin-V is a calcium-dependent phospholipid binding protein and is used to detect cell death.

During cell death, inner plasma membrane located phosphatidylserine (PS) is exposed to the external environment. Due to high affinity to PS, main phospholipid in the plasma membrane, Annexin-V is useful for cell death

93 analysis. Transiently overexpressed SENP5 in primary cardiomyocyte culture substantially increased cell death (Figure III.7). However it was totally abrogated upon introduction of deconjugation enzymatic activity mutated SENP5.

Quantification of Annexin-V positive cardiomyocytes was consistent with immunostating analysis (Figure III.7.b).

Additional cell death properties including cytoplasmic release of cytochome C and nuclear accumulation of active caspase-3 were determined to confirm.

Cytoplasmic fraction from both NRCMs and murine hearts revealed cytochrome C release upon SENP5 overexpression (Figure III.8.a). Nuclear translocation of caspase-3 was prominent in AdSENP5-wt transduced NRCMs but not in

AdSENP5-mut transduced ones (Figure III.8.b), suggesting that enhanced SENP5 expression in the heart is specific and sufficient to induce cardiomyocyte death.

The mitochondrial fission protein, Drp-1 is sensitive to SENP5. Recent reports indicate that mitochondria also harbor SUMO pathway components, suggesting that potential SUMO targets could be located in mitochondria (17, 36,

37). Among them, we tested mitochondrial fission protein, Drp-1, which was initially reported as a SENP5 target in COS cells (37). In addition, its ENU mutagenesis study showed the same cardiac structural defect, dilated cardiomyopathy of heart specific SENP5 Tg (134). Cardiac Drp-1 was examined in carefully isolated subcellular fractions including nuclear, cytosolic and mitochondrial fraction from murine hearts. Most of Drp-1 was detected as a free form with remarkably reduced level of SUMO conjugated form in SENP5

Figure III.7. Cardiomyocyte death is induced directly by SENP5 overexpression. a, Immunostaining for Annexin V (red) on NRCMs after 24 or 36hrs after wild type SENP5 (AdSENP5-wt), enzymatic activity dead mutant SENP5 (AdSENP5-mut) or empty (AdGFP) adenovirus infection at MOI=100. Enhanced expression of SENP5 in NRCMs resulted in increased Annexin V positive cells. Bar, 50µm. b, Annexin V positive NRCMs were scored and normalized by GFP positive NRCMs at least 100 cells from tree independent experiments and averaged. (**P<0.001 vs AdGFP)

Figure III.8. SENP5 facilitates cell death in NRCMs. a, Cytoplasmic release of cytochrome c. Cytosolic fraction from NRCM or mouse heart was applied to analyze by immunoblot. Cytosolic accumulation of cytochrome c was observed in both in vitro (left panel) and in vivo (right panel) SENP5 overexpression condition. b, Nuclear translocation of caspase-3 in SENP5 overexpressed NRCMs. Cytoplasmic remained caspase-3 in normal condition (AdGFP) was translocated into the nucleus under SENP5 overexpression (AdSENP5-wt). Accumulated caspase-3 induced by SENP5 overexpression was diminished by desumoylation activity dead SENP5 (AdSENP5-mut). Bar indicates 50µm.

96 overexpressed hearts (Figure III.9.a). Interestingly, this pattern was prominent at the mitochondrial fraction. SENP5 was also observed in the mitochondrial fraction, which correlated with increased desumoylated drp-1 in the mitochondrial fraction. Mitochondrial Drp-1 was sensitively desumoylated by SENP5 based on immunoprecipitation analysis. Particularly, SENP5 overexpression caused

SUMO2/3-conjugated Drp-1 to be desumoylated (Figure III.9.b).

Since SUMOylated Drp-1 facilitates mitochondrial fission (30)(37), it was examined that mitochondrial dynamics is influenced by desumoylation of Drp-1 upon SENP5 especially in the heart. Mitochondrial size was measured from TEM images. Mitochondria were remarkably enlarged in heart specific SENP5 Tg mice

(Figure III.10). Interestingly, SENP5 Tg already developed enlarged mitochondria although heart index including HW/BW ratio was not changed. These results suggested that SENP5 overexpression might involve in mitochondria. Thus our next question was whether mitochondrial enzymatic activities were affected by

SENP5 overexpression at the same time point of Figure III.10. As complex II consists of only nucleus-encoded subunits, therefore measurement of complex II activity can be useful for comparison with activities of the other complexes. In addition, citrate synthase is a sensitive indicator for mitochondrial activity, because decreased level of citrate synthase indicates that mitochondrial function was declined. In accordance with results from mitochondria size and apoptosis analysis of SENP5 Tg, reduced mitochondrial enzymatic activity already started at

P5 (Figure III.11). Moreover, heart specific SENP5 enhancement influcenced not

Figure III.9. Increased desumoylation of mitochondrial protein in cardiac specific SENP5 Tg heart. a, SENP5 target candidates were examined on desumoylation through Western blot analysis. Mitochondrial fission protein, Dynamin- related protein (Drp-1) was desumoylated in Tg heart mitochondrial extracts. Mitochondrial fraction showed decreased sumoylated Drp-1 but increased free Drp-1. Arrows indicate unconjugated Drp-1 (bottom) and sumolyated Drp-1(upper). TOM20 was used to verify mitochondrial fractionation. . (N, Nuclear fraction; C, cytosolic fraction; M, mitochondrial fraction) b, Sumoylated Drp-1 was confirmed by immunoprecipitation (IP) analysis. Mitochondrial fraction from WT and SENP5 Tg hearts were used for IP with anti Drp-1 antibody and for IB with anti SUMO2/3 antibody.

Figure III.10. Overexpression of SENP5 causes enlarged mitochondria. a, Electron microscopy images were collected from WT and SENP5 Tg at P5, which showed increased cell death but no change in HW/BW ratio (Figure III.5). Bar indicates 500nm. b, Mitochondrial size were measured by ImageJ and statistically analyzed. Size was measured from at least four randomly selected regions (20,000x magnification) and normalized by WT. Error bar represents mean ±SEM. (**P<0.0001 vs WT, n=4 for each group). Data revealed mitochondria were enlarged by SENP5 overexpression.

Figure III.11. Mitochondrial dysfunction in cardiac specific SENP5 Tg heart. The activity of mitochondrial respiratory complex II, IV and citrate synthase were determined from WT and α-MHC-SENP5 Tg at P5 spectrophotometrically. SENP5 overexpression in heart resulted in decreased mitochondrial respiratory complex activity. Error bar indicates SEM. (*P<0.05 vs WT, n=20 for each group).

100 only the nuclear DNA encoding mitochondrial complex but also the mtDNA encoding complexes, suggesting that decreased mitochondrial respiratory chain complex activities were primarily due to SENP5 overexpression.

Necrotic mitochondrial defects. Unexpectedly, we found severely degenerative mitochondrial structures in TEM of SENP5 Tg hearts at P10 (Figure III.12.a-b’).

SENP5 Tg hearts from P10 showed enlarged mitochondria and prominently loss of cristae (Figure III.12.b’), indicating necrotic mitochondrial structures. At this time point, α-MHC promoter carrying SENP5 Tg died suddenly accompanied with dilated ventricles. Also cardiac defects such as high HW/BW ratio, remarkable increase of cell death and declined respiratory chain activities (data not shown) at the same time point led to necrotic cell death in SENP5 Tgs. Thus we analyzed

Ca2+ deposition via von Kossa staining to address whether detrimental mitochondrial structure is caused by necrosis. Remarkable Ca2+ deposition was observed starting from P10 but more prominent at P15, corresponding with cardiomyocyte necrosis (Figure III.12.c-d). It supports that necrosis causes sudden death with severely dilated ventricle starting from one week after birth upon elevated SENP5 level.

AIF subcellular localization depends on SENP5. Necrosis is known to occur passively. However recent works provide evidence that some necrosis was programmed (138-141), especially mitochondrial-dependent necrosis (142, 143).

Mitochondrial structure, activity and dynamics defects were not enough to explain

101

Figure III.12. Mitochondrial defects caused by SENP5 overexpression. a, b, Transmission electron micrographs of WT and Tg mice at P10 showed that the majority of mitochondria were damaged, swollen and their cristae in disarray. a’, b’, higher magnification of a, b represents clear abnormal patterns of cristae in Tg. M, mitochondrion c, d, Histological assessment of Ca2+ deposition representing as brown dots in the ventricle by von Kossa staining. Bar, 500nm (a- b’), 50µm (c,d).

102 that the SENP5 Tg induced necrosis. Here we focused on apoptosis inducing factor (AIF) as a new target of SENP5 to mediate both apoptosis and necrosis.

Initially, it has been known that nuclear synthesized AIF moves into mitochondria by its mitochondria leading sequence (MLS), at which it involves in apoptotic cell death. However recent rising evidences supported that mitochondrial localized

AIF can be translocated reversely into the nucleus where it is involved in necrosis.

Several candidate sites (ψKXE) for SUMO conjugation were found within human

AIF (hAIF) through bioinformatic analysis. To examine whether hAIF is modified

SUMO-dependently, HeLa cells were co-transfected by vectors expressing human AIF (hAIF) with/without SUMO contained vector or with/without SENP5 carrying vector. It was determined that hAIF was showing clear and slow migrated bands corresponding to the SUMO-conjugated form. However, this modulation was diminished by desumoylation enzyme, SENP5 (data not shown). Based on this observation, we hypothesized that SENP dependent desumoylated AIF is released from mitochondria and moves to nucleus to induce necrosis. In accordance with the results of in vitro SUMO assay, nuclear localized AIF was increased in heart specific SENP5 Tg compared with WT (Figure III.13.a). To verify this different subcellular localization of AIF carefully, immunostaining using anti-AIF Ab was performed under acute SENP5 overexpression through adenovirus in NRCMs. Endogenous AIF was detected in the cytoplasm whereas, enhanced SENP5 expression induced endogenous AIF translocation into the nucleus in NRCMs (Figure III.13.b). However, nuclear translocated AIF was

103

104

Figure III.13. Myocardial AIF translocation in cardiac overexpression of SENP5. a, Nuclear AIF level was detected by immunoblot analysis from WT and α-MHC-SENP5 Tg hearts. Histone 3 specific antibody was applied to verify nuclear fraction. N, nuclear extract. b, Nuclear accumulation of AIF is SENP5 dependent. The redistribution of AIF (red) into the nuclei was explained by the overlap of AIF with nuclear staining accessed by DAPI (blue). Nuclear translocation of AIF induced by SENP5 was reduced under desymoylation activity mutated SENP5 (AdSENP5-mut) overexpressed NRCMs. Bar, 20µm. c, SUMO- conjugation site mutated AIF (AIF-K58R) was observed in nucleus in HeLa cells. Bar, 50µm.

105 reduced by enzymatic activity dead mutated SENP5 (AdSENP5-mut) (Figure

III.13.b), indicating nuclear translocation of AIF from mitochondria is influenced by

SENP5 directly. As an alternative way to test this hypothesis, one of SUMO site candidates within hAIF was mutated to block its SUMOylation and then its expression vector was tranasfected into the HeLa cells. Although only one candidate (K58R) was examined, AIF-K58R was prominently observed in nucleus

(Figure III.13.c), indicating SUMO-dependent modulation is critical to determine its subcelular localization. Collectively, these data suggest that SENP5 is sufficient and necessary to induce different subcellular localization of AIF.

3.4. Discussion

The detrimental role of SENP5 in congestive heart disease. SENP5 is known to express during embryogenesis and its expression declines after birth despite ubiquitous expression in the body. It has been shown that SENP5 is upregulated upon diabetes and oral squamous cell carcinoma (60). However, until now, there is little attention focused on SENP5 in the heart, especially congestive heart disease. Enhanced SENP2 expression in the murine hearts developed congenital heart defects including VSD and ASD (129). Moreover its loss of function study also showed embryonic lethality caused by suppression of critical genes required

106 for cardiac development (144). Combined with those studies, the differences of

SENP5 including substrate specificity and subcellular localization let us make the hypothesis that SENP5 may have unique functions in hearts, compared to SENP2.

Interestingly, SENP5 was induced in human idiopathic heart failure, typically downregulated postnatally, which leads to the idea that SENP5 re-induction could be involved in pathologic heart development. In current study, we provide strong in vivo evidence for its linkage with dilated cardiomyopathy, demonstrating that elevated protein level of SENP5 is not protective but detrimental and pushes to pathologic cardiac responses.

Mitochondrial dysfuction connecting with SUMO. Mitochondria are well known as an energy generator, indicating its metabolic importance in the energetic organ, heart. In many cases of congestive heart disease such as dilated and hypertrophic cardiomyopathy, mitochondrial structural and functional abnormalities are observed. Moreover, several point mutations in mitochondrial

DNA coding genes have been reported as associated with cardomyopathy (145).

Along with those genetic mutations, mitochondrial respiratory chain deficiency is a main functional defect (145). Cardiac SENP5 Tg showing decreased mitochondrial enzymatic activities was followed by mitochondrial pathology in dilated cardiomyopathy. Interestingly, complex IV activity was decreased in

SENP5 Tg, which consisted of proteins encoded by mtDNA, but complex II activity was also reduced, which was encoded by nuclear DNA. It suggested that

107 SENP5 induced mitochondrial respiratory chain deficiency is not caused by genetic mutations but possibly by SENP5 mediated deSUMOylation of targets.

In energy metabolic view, pathologic hearts change their main energy source from fatty acid to glucose, reprogramming back to the early post-natal energy metabolism (146, 147). Consistently, SENP5 overexpressed murine hearts disrupted their metabolic properties. Some metabolic genes were down-regulated in SENP5 Tg hearts compared to WT (Figure III.14). Destroyed fatty acid oxidation along with increased glucose metabolism was consistent with dysfunctional mitochondria from the same age of SENP5 Tg. It is also supported by pathologically unbalanced heart, in other words, decompensated heart showing totally decreased metabolic gene profiles (47). Together, increased

SENP5 protein level induces the pathologic heart.

Mitochondrial dynamic protein sensitive to SENP5. Mitochondrial dynamics such as fusion and fission are correlated with heart function. In human hearts, one of mitochondrial fusion protein, optic atrophy 1 (Opa1) is down-regulated or unchanged upon ischemia or failing (148). Decreased respiratory and contractile function from cardiac specific double knock-out of another mitochondrial fusion protein, mitofusins (Mfn) 1,2 indicates that interrupting mitochondrial fusion results in lethal cardiac function (149). Although in human failured heart, Mfn is upregulated (148) it seems to be a compensatory response because Opa1 is remarkably reduced in the same context. Recently, the first direct relevance of

108

Figure III.14. Impaired cardiac energy metabolism in SENP5 Tg. In increasing need of glucose utilization, glucose metabolism related genes such as glucose transporter 4(GLUT4) and pyruvate dehydrogenase kinase 4 (PDK4) were increased but fatty acid oxidation is down-regulated through carnitine acyl transferase 1 (CPT1), peroxisome proliferator-activated receptor α (PPARα) and Estrogen-related receptors (ERRs). Those gene expression patterns are correlated with metabolic property of heart failure.

109 mitochondrial fission with cardiomyopthy is reported. A mutant dynamin-related protei-1 (Drp-1), mitochondrial fission protein, which loses the ability of self- oligomerization causes dilated cardiomyopathy (134). Also, the strong supportive studies have reported that Drp-1 is SUMO-dependently modified and its

SUMOylation is sensitive to SENP5 (36, 37, 59). Unconjugated mitochondrial Drp-

1 was significantly increased but SUMO2/3 conjugated Drp-1 was decreased in mitochondrial fraction from SENP5 Tg. Usually, Drp-1 stays in the cytoplasm as a phosphorylated form. Once a phosphatase is activated such as calcineurin then dephosphorylated Drp-1 finally can move to mitochondria (150). SUMOylation is thought to make the stable attachment of Drp-1 to mitochondria. Thus, forced

SENP5 expression desumoylated Drp-1, resulting in its move away from mitochondria. Finally, the cytoplasmic Drp-1 cannot mediate mitochondrial fission and causes a dramatic increase in mitochondrial size. Interestingly, dilated cardiomyopathy suffering patients represent remarkable number of gigantic mitochondria (151). It is supportive to our observation in human and murine study

(Figure III.1) that increased SENP5 is detrimental to heart and results in dilated cardiomyopathy.

Uncontrolled fission or fusion is involved in apoptosis which is a critical mechanism leading to cardiomyocyte loss in heart failure. The involvement of mitochondrial fusion in apoptosis is controversial. Down-regulated Mfn2 induced apoptosis (152, 153) but overexpresed Mfn2 also causes apoptosis by attenuating Akt signaling (154, 155). Conversely, the mitochondrial fission

110 triggers apoptosis, the proapoptotic factor, Bax facilitates sumoylation of Drp-1 as well. However, increased sumoylated Drp-1 by Bax is related to mitochondrial permeability transition pore (MPTP) opening rather than mitochondrial fission (156,

157). Loss of Drp-1 in Drp-1-/- murine brain significantly increases apoptosis (158).

Although sumoyation of Drp-1 has not been explored in this study, the number of highly connected mitochondria is dominant in Drp-1-/-. Thus increased apoptosis by SENP5 overexpression is correlated with mitochondrial dynamics, especially enhanced desumoylation of Drp-1. Apparently, SENP5 is an emerging new member, which can control mitochondrial physiology in cardiac muscle tissues.

A novel target of SENP5 is involved in necrosis. Ultrastructural images from

SENP5 Tg exhibited abnormal mitochondrial features including catastrophic mitochondria destruction and extensive swelling with loss of cristae, which is called as cristolysis. Those characteristic phenotypes match to necrosis. Necrosis is initiated in the mitochondria by breakage of Ca2+ homeostasis and oxidative damage (138, 159, 160). Mitochondrial Ca2+ overflow follows mitochondrial membrane permeability reduction and finally induces necrosis. In early study, necrosis is known as spontaneous cell death compared to apoptosis, programmed cell death. However, increasing evidence suggests that it also occurrs in a programmed way. Cyclophilin D has been reported as one of regulators in the programmed Ca2+-dependent necrosis. Its genetic deletion causes Ca2+-dependent necrosis because it regulates Ca2+ influx (138, 140, 143).

Along with that concept, apoptosis inducing factor (AIF) is also proposed as a

111 necrotic factor. Although its initial role is known to regulate apoptosis, its nuclear translocation results in large scale DNA breakdown and chromatin condensation and leads to necrosis (82, 161, 162). Nevertheless, little is known about the mechanism to explain its nuclear translocation. Here we found that AIF was modified by SUMO and SENP5. Elevated SENP5 protein level both in human diseased heart and experimental animal heart such as SENP5 Tg reduced SUMO conjugation level of AIF. Consistently, nucleus-located AIF was increased in immunoblot and immunostaining analysis (Figure III.14). These results suggest several possibilities. First, SENP5 desumoylates AIF, leading to conformational change. It results in release of protector binding to AIF and finally results in easy access of proteases such as Calpain-1 into AIF. It is evident that Increased Ca2+ level stimulates ROS production, leading to carbonylation of AIF, which is associated with Calpain-1 mediated cleavage (163). Another one is that desumoylation by SENP5 makes high accessibility of nucleus leading sequence

(NLS) and/or masking of mitochondria leading sequence (MLS) of AIF. The reason is that AIF-K59R overexpression caused high nuclear translocated AIF, to which MLS is adjacent (Figure III.14.c). Sequentially, AIF is able to be released from mitochondria, it is involved in necrosis in the nucleus as well. However it is still unknown largely where AIF is sumoylated and whether desumoylated AIF can induce necrosis and so on. Nevertheless, it should provide valuable information for therapeutic or preventative application of AIF in the heart failure.

112 Chapter IV

Summary, Significance and Future Directions

113

4.1. Specificity for targets: SUMO conjugation vs deconjugation.

Among causative genes for dilated cardiomyopathy, lamin A/C has been identified as a SUMO target associated with familiar dilated cardiomyopathy (105,

106). We tested whether lamin A/C is sensitive to SENP5 using immunoprecipitation analysis since its PTM is SUMO2/3-dependent and its mutation correlates to dilated cardiomyopathy (164). Unexpectedly, cardiac lamin

A/C sumoylation was not changed in heart specific SENP5 Tg, compared to WT.

However, mitochondrial fission protein, Drp-1 was responsive to SENP5, known as SUMO-dependently modification. In calpastastin study, there were additional sumoylated bands not responsive to SUMO2/3. We have poor understanding about that but it could be due to SUMO1 conjugation to Calpastatin since stress condition induces SUMO1 upregulation along with SUMO2/3 in the hearts based on our observation. Still it is unclear whether increased SUMO1 functions as protective or detrimental especially to the heart. SUMOylation is tightly regulated by cellular availability, time and place of SUMO components. Also there are many different SUMO components for SUMO conjugation and deconjugation, suggesting maximal possible combination for efficient regulation. Recent study supports this idea, in which SENP3 has modulated Drp-1, promoting brain cell death after ischemia (165). In that study, ischemia has up-regulated SENP3 but not SENP5 in the brain. This is a totally different response, compared to our

114 observation in the hearts. Thus it indicates that variable SUMO pathway components makes different preferences to the same target in a tissue- or time- dependent manner for fine-tune regulation.

4.2. Cardiomyocyte death by SUMO2/3.

Previously, among death factors, caspase-2, 7 and 8 have found to be modified by SUMO1 for their nuclear localization and self-activation (166-168).

Also, it has been suggested that reactive oxygen species (ROS) influence desumoylation enzymes such as SENP3 (130). Oxidative cell death is very sensitive to cellular sumoylation level and their targets such as JNK and homeodomain-interacting protein kinase 2 (HIPK2) (169, 170). Furthermore,

SUMO1 has conjugated to nuclear factor κB (NFκB) regulator inhibitory κBα

(IκBα) for protective function but in contrast, associated with NFκB essential modulator (NEMO), for activation (171-173). Those all indicate that SUMO dependent modification correlates with cell death.

Nevertheless, it is largely unknown whether SUMO2/3 is involved in cardiac muscle cell death or not. We observed that human and experimental animal heart failures have enhanced SUMO2/3 level (Figure II.1) and E3 ligation enzyme,

PIAS2 (Figure II.17), suggesting that hypersumoylation response is possibly explained by altered fortification of SUMOylation components. Apparently SUMO2 overexpressed murine hearts have undergone to exaggerated apoptosis. It has long been suggested that cardiomyocyte death is sufficient to facilitate heart

115 failure by cardiac muscle cell loss. Therefore, current studies demonstrate that

SUMO2/3 is critical for modulating cell death pathway. Moreover, they propose the new concept of SUMO2/3 for Ca2+ dependent protease system, Calpain2-

Calpastatin. In this context, elevated SUMO2/3 conjugation activity hits both

Calpain2 and Calpastatin then, induces cell death through synergistic effects by releasing active Calpain2 from its inhibitor and by suppression of inhibitory

Calpastatin via changed stability, subcellular localization and interaction ability.

Additionally, SUMO2/3 induced Calpain2 activation has been further confirmed by accumulation of cleaved caspase-12, leading to activation of caspase-9 and -3

(80, 81). Calpain2 and Calpastatin -mediated cell death could cause further degradation of cardiac structural proteins including the actin and the tropon T, leading to contractile dysfunction. But there are still remained questions that exact

SUMO2 conjugation sites on Calpastatin and definition of additional slow running bands not matching with SUMO2/3 –dependent modification for both Calpain2 and Calpastatin remain to be identified. Careful identification of those unsolved questions would help to define the mechanism and to design drugs for heart failure.

4.3. Mitochondrial cell death with SUMO protease, SENP5.

Prominent alterations observed in α-MHC promoter driven SENP5 overexpression have been swollen mitochondria with disorganized cristae, accompanied by reduced mitochondrial enzymatic activities. Those impaired

116 mitochondrial respiratory chain with mitochondrial morphologic defects has been reported as apparent properties of pathologic hearts including failing rat hearts after aortic constriction and hypertrophic cardiomyopathy of pig (174, 175). Also specific PTM such as nitrosylation and O-linked glycosylation in mitochondria have been suggested that reversible modification of mitochondrial proteins results in pathologic hearts (176, 177). Here, we proposed SUMO-dependent PTM of mitochondrial proteins to be critical for mitochondrial function. Especially, recent findings have supported that it is possible that mitochondrial proteins would be modified by SUMO pathway through mitochondrial SUMO E3 ligase, MAPL (17),

SUMO1 and SENP5 (36, 37, 59). Mitochondrial fission protein, Drp-1 has been altered by many different modifications including PKA, CaM kinase Iα and calcineurin (150, 178-182). In addition, SUMO1 conjugation to Drp-1 results in mitochondrial fragmentation by stabilizing Drp-1 to localize in mitochondria. We found that heart specific overexpressed SENP5 caused decreased mitochondrial fission and finally led to giant mitochondria. The huge mitochondrial size is one of the prominent phenotypes in early cardiomyopathy. Although end stage of heart failure developed small fragmented mitochondria, early adaptive response such as hypertrophied heart has revealed enlarged mitochondria. Enlarged mitochondria in the hypertrophied heart have been further supported. Drp-1 has been increased in diseased hearts (148) although relative SUMO association level of Drp-1 was not explored. We found that SUMO-deconjugated Drp-1 has been also remarkably increased in SENP5 overexpressed murine hearts.

117 Together, impaired SUMOylataion especially for Drp-1 by forced SENP5 expression directs mitochondrial dynamics into less pro-fission. Consistent with previous Drp-1 deficent animal studies representing giant mitochondria and dilated cardiomyopathy (134), this finding makes us conclude that dysregulated mitochondrial fission fails to maintain mitochondrial function, finally leading to pathologic hearts. At this moment, we still do not understand whether uncontrolled

SUMOylation in Drp-1 is sufficient to cause dilated cardiomyopathy, also increased Drp-1 expression has resulted in controversial cell death outcomes

(156, 183, 184). However, at least our study could add one possibility that mitochondrial SUMO conjugation cycle regulates mitochondrial dynamic protein by pulling Drp-1 into mitochondria and likely associating Drp-1 with fission prone membrane-receptor, Fis1 rather than Mfn (mitofusin).

We also have a poor understanding of how unbalanced mitochondrial SUMO conjugation by SENP5 resulted in cell death. Here, AIF has been proposed as a novel SENP5 target. Especially, AIF is attractive as a target because it has been known to involve in caspase-independent apoptosis in many different tissues (162,

185). Sequential subcellular localization of AIF is important for its biologic functions including respiratory chain stabilization and pro-apoptotic properties (83,

186) but processing for different subcellular localization of AIF has been largely unknown. Calpain1 and cathepsin B, L or S have been well known proteases mediating cleavage of AIF (187-189) but still their regulation has not been applied yet. Cardiac AIF has been modified SUMO-dependently in murine hearts (Figure

118 III.13). In addition, nuclear accumulated AIF has been predominant upon enhanced SENP5 expression in the hearts. At this moment, we still do not understand exact association of SUMO conjugation with AIF but we suggest the following mechanisms. First, decreased SUMO conjugation of AIF makes its conformational change, which might cause easy access for proteases. Second, different level of SUMO conjugation could block mitochondrial leading sequence

(MLS) or expose nuclear leading sequence (NLS) of AIF for induction of nuclear accumulation. Third, SUMOylation intensity in AIF might affect not only its subcellular localization but also its nuclear partners such as Endonuclease G mediating DNA degradation or chromatic condensation.

We have a poor understanding of how and where AIF was modified by

SUMO-dependent PTM, although it seems to occur in mitochondria because AIF usually exists in mitochondria for mitochondrial complex I maintenance in normal state (190). A recent interesting study showed that necrosis can also be programmed (191) but little is discovered as required trigger molecules. Moreover, our preliminary data indicate that AIF seems to be poly and/or multiple- sumoylated and is also sensitive to SENP2 in vitro, indicating that SUMO- dependent PTM of AIF is regulated as complicated ways and their involvement in necrosis is tightly controlled by SUMO pathway. Further investigation into the effectiveness of SUMOylation or deSUMOylation may provide important insights into AIF mediated programmed necrosis.

119 4.4. Cardiomyocyte proliferation and SUMO.

Increased SUMO2 or SENP5 level in the hearts results in decreased cardiac muscle cell proliferation accompanied by increased cell death, correlating with congenital heart disease. It has been shown that SUMO is also critical for cell division. In those studies, SUMO-dependent PTM has repressed cyclin D1 and arrested G1 cell cycle through modification of Stra13 (192) or activated cell cycle by enhancing nuclear protein, RanGAP1-RanBP2 interaction (193). However,

SUMO2/3-dependent PTM mediated cell proliferation is less studied although

SUMO protease, SENP5 has been suggested to be required for cell proliferation

(36, 57). Intriguingly, these results are quite different from our observation, lower cardiomyocytes proliferation upon promoting SENP5 expression (Figure III.5).

The discrepancy might be because our observation has been from in vivo animal model and tissue-specific effect, as well as depended on its threshold level.

Recently, a very interesting study has provided a critical clue that AMPKβ is specifically modified by SUMO2 (125) and that activated AMPK has suppressed cell proliferation (127). It is possible that heart specific SUMO2 enhancement might affect AMPK activity, resulting in cell cycle arrest.

In SENP5 study, decreased cardiac muscle cell division has been also shown

(Figure III.5). In Drosophila study (126, 127), reduced mitochondrial complex I activity has caused ROS, finally leading FOXO activation. Activated FOXO has been suggested as a regulator of cell cycle to repair DNA damage by turn-on of

120 proteins involved in tolerance to oxidative stress (194, 195). Here, we proposed

AIF as a novel target of SUMOylation. AIF has dual function including mitochondrial respiratory chain and caspase-independent cell death. Its loss of function study has revealed unstabilized mitochondrial respiratory complexes and higher ROS. Based on this result, it could be assumed that the liberal AIF detached from mitochondria increase through SENP5 overexpression might promote unstable mitochondrial complex I assembly. Disarrayed complexes should cause mitochondrial function drop, leading to cardiomyocyte division stall.

This idea awaits further investigation.

In summary, we provide the first evidence that covalent conjugation of

SUMO2/3 or deconugation by SENP5 is a new regulatory strategy for cardiomyocyte survival and congestive heart disease. Especially, these observations open up new roads not only for understanding SUMO2/3 conjugation in stressed hearts, but also for identifying how SUMO2/3-dependent conjugation can provide valuable evidence for adult cardiac disorders.

4.5. Implication for human disease.

Cardiac hypertrophy is initially beneficial for heart against overloaded work.

However, sustained responses finally lead to heart failure and sudden death.

During adaptive response, postnatal hearts adapt to those stresses via a variety of reactions including increasing cardiomyocyte size and reactivation of fetal cardiac genes. Thus, those kinds of changes are used as cardiac stress indicators

121 for diagnosis. Here, we demonstrate that cardiac muscle loss is accelerated by covalent SUMO2/3 attachment to targets such as Calpain-Calpastatin system.

Also, overexpression of the deconjugation protease, SENP5 contributes to failed mitochondrial function, leading to dilated cardiomyopathy. Similar observation was proposed in cancer cells and those patterns have been used as diagnostic markers (196). Thus it should be also used as a new diagnostic biomarker for stressed heart.

SUMO-dependent PTM has been applied in a number of cellular processes as well as diseases such as neurodegenerative disease (197), and cancers including breast and colon cancer, and myeloma (197, 198). Moreover, a recent study has indicated that SUMOylation is required for c-Myc driven tumorigenesis (199) and heat shock response (200). In spite of these attractive properties, there is little report related to the inhibitor. To date, ginkgolic acid and kerriamycin B are the most famous inhibitors of SUMO E1 enzyme (201, 202). Nevertheless, recent emerging studies have been interested in inhibitor development with more specificity and variety targets (203, 204), as well as high throughput screening assay to seek inhibitors efficient development of inhibitors (205). However, the difficulty to solve is how each inhibitor involving specific SUMOylation provides specificity for substrates in whole proteome because the same SUMO isoform conjugation does not make the same effect of protective or pathological. For example, SUMO2 is required for nuclear integrity through conjugation of lamin

A/C, but not for cell survival because increased SUMO2 conjugation to Calpain-

122 Calpastatin system causes cell death. Our heart specific SUMO2 or SENP5 Tg will allow a detailed analysis to find its targets in diseased state. Furthermore the information obtained from those cardiac specific Tg should be valuable to design potential therapeutic drug with no side effect to ameliorate heart disease progression.

123 Bibliography

1. Mahajan, R., C. Delphin, T. Guan, L. Gerace, and F. Melchior. 1997. A

small ubiquitin-related polypeptide involved in targeting RanGAP1 to nuclear pore complex protein RanBP2. Cell 88:97-107.

2. Matunis, M. J., E. Coutavas, and G. Blobel. 1996. A novel ubiquitin-like modification modulates the partitioning of the Ran-GTPase-activating protein RanGAP1 between the cytosol and the nuclear pore complex. J Cell Biol 135:1457-1470.

3. Bedford, L., J. Lowe, L. R. Dick, R. J. Mayer, and J. E. Brownell. 2011. Ubiquitin-like protein conjugation and the ubiquitin-proteasome system as drug targets. Nat Rev Drug Discov 10:29-46.

4. Johnson, E. S. 2004. Protein modification by SUMO. Annu Rev Biochem 73:355-382.

5. Guo, D., M. Li, Y. Zhang, P. Yang, S. Eckenrode, D. Hopkins, W. Zheng, S.

Purohit, R. H. Podolsky, A. Muir, J. Wang, Z. Dong, T. Brusko, M. Atkinson,

P. Pozzilli, A. Zeidler, L. J. Raffel, C. O. Jacob, Y. Park, M. Serrano-Rios, M. T. Larrad, Z. Zhang, H. J. Garchon, J. F. Bach, J. I. Rotter, J. X. She, and C. Y. Wang. 2004. A functional variant of SUMO4, a new I kappa B alpha modifier, is associated with type 1 diabetes. Nat Genet 36:837-841.

6. Melchior, F. 2000. SUMO--nonclassical ubiquitin. Annu Rev Cell Dev Biol 16:591-626.

124 7. Wilkinson, K. A., and J. M. Henley. 2010. Mechanisms, regulation and consequences of protein SUMOylation. Biochem J 428:133-145.

8. Owerbach, D., E. M. McKay, E. T. Yeh, K. H. Gabbay, and K. M. Bohren. 2005. A proline-90 residue unique to SUMO-4 prevents maturation and sumoylation. Biochem Biophys Res Commun 337:517-520.

9. Bohren, K. M., V. Nadkarni, J. H. Song, K. H. Gabbay, and D. Owerbach. 2004. A M55V polymorphism in a novel SUMO gene (SUMO-4) differentially activates heat shock transcription factors and is associated with susceptibility to type I diabetes mellitus. J Biol Chem 279:27233- 27238.

10. Martin, S., K. A. Wilkinson, A. Nishimune, and J. M. Henley. 2007. Emerging extranuclear roles of protein SUMOylation in neuronal function and dysfunction. Nat Rev Neurosci 8:948-959.

11. Meulmeester, E., and F. Melchior. 2008. Cell biology: SUMO. Nature 452:709-711.

12. Rodriguez, M. S., C. Dargemont, and R. T. Hay. 2001. SUMO-1 conjugation in vivo requires both a consensus modification motif and nuclear targeting. J Biol Chem 276:12654-12659.

13. Johnson, E. S., and G. Blobel. 1997. Ubc9p is the conjugating enzyme for the ubiquitin-like protein Smt3p. J Biol Chem 272:26799-26802.

125 14. Rytinki, M. M., S. Kaikkonen, P. Pehkonen, T. Jaaskelainen, and J. J. Palvimo. 2009. PIAS proteins: pleiotropic interactors associated with SUMO. Cell Mol Life Sci 66:3029-3041.

15. Pichler, A., A. Gast, J. S. Seeler, A. Dejean, and F. Melchior. 2002. The nucleoporin RanBP2 has SUMO1 E3 ligase activity. Cell 108:109-120.

16. Kagey, M. H., T. A. Melhuish, and D. Wotton. 2003. The polycomb protein Pc2 is a SUMO E3. Cell 113:127-137.

17. Braschi, E., R. Zunino, and H. M. McBride. 2009. MAPL is a new mitochondrial SUMO E3 ligase that regulates mitochondrial fission. EMBO Rep 10:748-754.

18. Cohen, P., and M. Tcherpakov. 2010. Will the ubiquitin system furnish as many drug targets as protein kinases? Cell 143:686-693.

19. Kerscher, O. 2007. SUMO junction-what's your function? New insights through SUMO-interacting motifs. EMBO Rep 8:550-555.

20. Song, J., L. K. Durrin, T. A. Wilkinson, T. G. Krontiris, and Y. Chen. 2004. Identification of a SUMO-binding motif that recognizes SUMO-modified proteins. Proc Natl Acad Sci U S A 101:14373-14378.

21. Hannich, J. T., A. Lewis, M. B. Kroetz, S. J. Li, H. Heide, A. Emili, and M. Hochstrasser. 2005. Defining the SUMO-modified proteome by multiple approaches in Saccharomyces cerevisiae. J Biol Chem 280:4102-4110.

126 22. Minty, A., X. Dumont, M. Kaghad, and D. Caput. 2000. Covalent modification of p73alpha by SUMO-1. Two-hybrid screening with p73 identifies novel SUMO-1-interacting proteins and a SUMO-1 interaction motif. J Biol Chem 275:36316-36323.

23. Hecker, C. M., M. Rabiller, K. Haglund, P. Bayer, and I. Dikic. 2006. Specification of SUMO1- and SUMO2-interacting motifs. J Biol Chem 281:16117-16127.

24. Li, S. J., and M. Hochstrasser. 2000. The yeast ULP2 (SMT4) gene encodes a novel protease specific for the ubiquitin-like Smt3 protein. Mol Cell Biol 20:2367-2377.

25. Gong, L., S. Millas, G. G. Maul, and E. T. Yeh. 2000. Differential regulation of sentrinized proteins by a novel sentrin-specific protease. J Biol Chem 275:3355-3359.

26. Bailey, D., and P. O'Hare. 2004. Characterization of the localization and proteolytic activity of the SUMO-specific protease, SENP1. J Biol Chem 279:692-703.

27. Bailey, D., and P. O'Hare. 2002. Herpes simplex virus 1 ICP0 co-localizes with a SUMO-specific protease. J Gen Virol 83:2951-2964.

28. Kim, Y. H., K. S. Sung, S. J. Lee, Y. O. Kim, C. Y. Choi, and Y. Kim. 2005. Desumoylation of homeodomain-interacting protein kinase 2 (HIPK2) through the cytoplasmic-nuclear shuttling of the SUMO-specific protease SENP1. FEBS Lett 579:6272-6278.

127 29. Hang, J., and M. Dasso. 2002. Association of the human SUMO-1 protease SENP2 with the nuclear pore. J Biol Chem 277:19961-19966.

30. Zhang, H., H. Saitoh, and M. J. Matunis. 2002. Enzymes of the SUMO modification pathway localize to filaments of the nuclear pore complex. Mol Cell Biol 22:6498-6508.

31. Itahana, Y., E. T. Yeh, and Y. Zhang. 2006. Nucleocytoplasmic shuttling modulates activity and ubiquitination-dependent turnover of SUMO-specific protease 2. Mol Cell Biol 26:4675-4689.

32. Nishida, T., F. Kaneko, M. Kitagawa, and H. Yasuda. 2001. Characterization of a novel mammalian SUMO-1/Smt3-specific isopeptidase, a homologue of rat axam, which is an axin-binding protein promoting beta-catenin degradation. J Biol Chem 276:39060-39066.

33. Best, J. L., S. Ganiatsas, S. Agarwal, A. Changou, P. Salomoni, O. Shirihai, P. B. Meluh, P. P. Pandolfi, and L. I. Zon. 2002. SUMO-1 protease-1 regulates gene transcription through PML. Mol Cell 10:843-855.

34. Nishida, T., H. Tanaka, and H. Yasuda. 2000. A novel mammalian Smt3- specific isopeptidase 1 (SMT3IP1) localized in the nucleolus at interphase. Eur J Biochem 267:6423-6427.

35. Gong, L., and E. T. Yeh. 2006. Characterization of a family of nucleolar SUMO-specific proteases with preference for SUMO-2 or SUMO-3. J Biol Chem 281:15869-15877.

128 36. Zunino, R., E. Braschi, L. Xu, and H. M. McBride. 2009. Translocation of SenP5 from the nucleoli to the mitochondria modulates DRP1-dependent fission during mitosis. J Biol Chem 284:17783-17795.

37. Zunino, R., A. Schauss, P. Rippstein, M. Andrade-Navarro, and H. M. McBride. 2007. The SUMO protease SENP5 is required to maintain mitochondrial morphology and function. J Cell Sci 120:1178-1188.

38. Kim, K. I., S. H. Baek, Y. J. Jeon, S. Nishimori, T. Suzuki, S. Uchida, N. Shimbara, H. Saitoh, K. Tanaka, and C. H. Chung. 2000. A new SUMO-1- specific protease, SUSP1, that is highly expressed in reproductive organs. J Biol Chem 275:14102-14106.

39. Choi, S. J., S. S. Chung, E. J. Rho, H. W. Lee, M. H. Lee, H. S. Choi, J. H. Seol, S. H. Baek, O. S. Bang, and C. H. Chung. 2006. Negative modulation of RXRalpha transcriptional activity by small ubiquitin-related modifier (SUMO) modification and its reversal by SUMO-specific protease SUSP1. J Biol Chem 281:30669-30677.

40. Mukhopadhyay, D., F. Ayaydin, N. Kolli, S. H. Tan, T. Anan, A. Kametaka,

Y. Azuma, K. D. Wilkinson, and M. Dasso. 2006. SUSP1 antagonizes formation of highly SUMO2/3-conjugated species. J Cell Biol 174:939-949.

41. Wang, J., and R. J. Schwartz. 2010. Sumoylation and regulation of cardiac gene expression. Circ Res 107:19-29.

42. Giorgino, F., O. de Robertis, L. Laviola, C. Montrone, S. Perrini, K. C. McCowen, and R. J. Smith. 2000. The sentrin-conjugating enzyme mUbc9

129 interacts with GLUT4 and GLUT1 glucose transporters and regulates transporter levels in skeletal muscle cells. Proc Natl Acad Sci U S A 97:1125-1130.

43. Geiss-Friedlander, R., and F. Melchior. 2007. Concepts in sumoylation: a decade on. Nat Rev Mol Cell Biol 8:947-956.

44. Hayashi, T., M. Seki, D. Maeda, W. Wang, Y. Kawabe, T. Seki, H. Saitoh, T. Fukagawa, H. Yagi, and T. Enomoto. 2002. Ubc9 is essential for viability of higher eukaryotic cells. Exp Cell Res 280:212-221.

45. Alkuraya, F. S., I. Saadi, J. J. Lund, A. Turbe-Doan, C. C. Morton, and R. L. Maas. 2006. SUMO1 haploinsufficiency leads to cleft lip and palate. Science 313:1751.

46. Evdokimov, E., P. Sharma, S. J. Lockett, M. Lualdi, and M. R. Kuehn. 2008. Loss of SUMO1 in mice affects RanGAP1 localization and formation of PML nuclear bodies, but is not lethal as it can be compensated by SUMO2 or SUMO3. J Cell Sci 121:4106-4113.

47. Saitoh, H., and J. Hinchey. 2000. Functional heterogeneity of small ubiquitin-related protein modifiers SUMO-1 versus SUMO-2/3. J Biol Chem 275:6252-6258.

48. Tatham, M. H., E. Jaffray, O. A. Vaughan, J. M. Desterro, C. H. Botting, J. H. Naismith, and R. T. Hay. 2001. Polymeric chains of SUMO-2 and SUMO-3 are conjugated to protein substrates by SAE1/SAE2 and Ubc9. J Biol Chem 276:35368-35374.

130 49. Mukhopadhyay, D., and M. Dasso. 2007. Modification in reverse: the SUMO proteases. Trends Biochem Sci 32:286-295.

50. Hay, R. T. 2007. SUMO-specific proteases: a twist in the tail. Trends Cell Biol 17:370-376.

51. Hickey, C. M., N. R. Wilson, and M. Hochstrasser. 2012. Function and regulation of SUMO proteases. Nat Rev Mol Cell Biol 13:755-766.

52. Lallemand-Breitenbach, V., M. Jeanne, S. Benhenda, R. Nasr, M. Lei, L. Peres, J. Zhou, J. Zhu, B. Raught, and H. de The. 2008. Arsenic degrades PML or PML-RARalpha through a SUMO-triggered RNF4/ubiquitin- mediated pathway. Nat Cell Biol 10:547-555.

53. Tatham, M. H., M. C. Geoffroy, L. Shen, A. Plechanovova, N. Hattersley, E. G. Jaffray, J. J. Palvimo, and R. T. Hay. 2008. RNF4 is a poly-SUMO- specific E3 ubiquitin ligase required for arsenic-induced PML degradation. Nat Cell Biol 10:538-546.

54. Weisshaar, S. R., K. Keusekotten, A. Krause, C. Horst, H. M. Springer, K.

Gottsche, R. J. Dohmen, and G. J. Praefcke. 2008. Arsenic trioxide stimulates SUMO-2/3 modification leading to RNF4-dependent proteolytic targeting of PML. FEBS Lett 582:3174-3178.

55. Lee, J., Y. Lee, M. J. Lee, E. Park, S. H. Kang, C. H. Chung, K. H. Lee, and K. Kim. 2008. Dual modification of BMAL1 by SUMO2/3 and ubiquitin promotes circadian activation of the CLOCK/BMAL1 complex. Mol Cell Biol 28:6056-6065.

131 56. Ouyang, J., Y. Shi, A. Valin, Y. Xuan, and G. Gill. 2009. Direct binding of CoREST1 to SUMO-2/3 contributes to gene-specific repression by the LSD1/CoREST1/HDAC complex. Mol Cell 34:145-154.

57. Di Bacco, A., J. Ouyang, H. Y. Lee, A. Catic, H. Ploegh, and G. Gill. 2006. The SUMO-specific protease SENP5 is required for cell division. Mol Cell Biol 26:4489-4498.

58. Mikolajczyk, J., M. Drag, M. Bekes, J. T. Cao, Z. Ronai, and G. S. Salvesen. 2007. Small ubiquitin-related modifier (SUMO)-specific proteases: profiling the specificities and activities of human SENPs. J Biol Chem 282:26217-26224.

59. Harder, Z., R. Zunino, and H. McBride. 2004. Sumo1 conjugates mitochondrial substrates and participates in mitochondrial fission. Curr Biol 14:340-345.

60. Ding, X., J. Sun, L. Wang, G. Li, Y. Shen, X. Zhou, and W. Chen. 2008. Overexpression of SENP5 in oral squamous cell carcinoma and its association with differentiation. Oncol Rep 20:1041-1045.

61. Bernardo, B. C., K. L. Weeks, L. Pretorius, and J. R. McMullen. 2010. Molecular distinction between physiological and pathological cardiac hypertrophy: experimental findings and therapeutic strategies. Pharmacol Ther 128:191-227.

62. Bergmann, O., R. D. Bhardwaj, S. Bernard, S. Zdunek, F. Barnabe-Heider, S. Walsh, J. Zupicich, K. Alkass, B. A. Buchholz, H. Druid, S. Jovinge, and

132 J. Frisen. 2009. Evidence for cardiomyocyte renewal in humans. Science 324:98-102.

63. Kajstura, J., M. Rota, D. Cappetta, B. Ogorek, C. Arranto, Y. Bai, J. Ferreira-Martins, S. Signore, F. Sanada, A. Matsuda, J. Kostyla, M. V. Caballero, C. Fiorini, D. A. D'Alessandro, R. E. Michler, F. del Monte, T. Hosoda, M. A. Perrella, A. Leri, B. A. Buchholz, J. Loscalzo, and P. Anversa. 2012. Cardiomyogenesis in the aging and failing human heart. Circulation 126:1869-1881.

64. Heineke, J., and J. D. Molkentin. 2006. Regulation of cardiac hypertrophy by intracellular signalling pathways. Nat Rev Mol Cell Biol 7:589-600.

65. Maillet, M., J. H. van Berlo, and J. D. Molkentin. 2013. Molecular basis of physiological heart growth: fundamental concepts and new players. Nat Rev Mol Cell Biol 14:38-48.

66. van Berlo, J. H., M. Maillet, and J. D. Molkentin. 2013. Signaling effectors underlying pathologic growth and remodeling of the heart. J Clin Invest 123:37-45.

67. Hunter, J. J., and K. R. Chien. 1999. Signaling pathways for cardiac hypertrophy and failure. N Engl J Med 341:1276-1283.

68. Lopaschuk, G. D., and J. S. Jaswal. 2010. Energy metabolic phenotype of the cardiomyocyte during development, differentiation, and postnatal maturation. J Cardiovasc Pharmacol 56:130-140.

133 69. Abel, E. D., and T. Doenst. 2011. Mitochondrial adaptations to physiological vs. pathological cardiac hypertrophy. Cardiovasc Res 90:234- 242.

70. Narula, J., N. Haider, R. Virmani, T. G. DiSalvo, F. D. Kolodgie, R. J. Hajjar, U. Schmidt, M. J. Semigran, G. W. Dec, and B. A. Khaw. 1996. Apoptosis in myocytes in end-stage heart failure. N Engl J Med 335:1182- 1189.

71. Olivetti, G., R. Abbi, F. Quaini, J. Kajstura, W. Cheng, J. A. Nitahara, E. Quaini, C. Di Loreto, C. A. Beltrami, S. Krajewski, J. C. Reed, and P. Anversa. 1997. Apoptosis in the failing human heart. N Engl J Med 336:1131-1141.

72. Wencker, D., M. Chandra, K. Nguyen, W. Miao, S. Garantziotis, S. M. Factor, J. Shirani, R. C. Armstrong, and R. N. Kitsis. 2003. A mechanistic role for cardiac myocyte apoptosis in heart failure. J Clin Invest 111:1497- 1504.

73. Huang, Y., and K. K. Wang. 2001. The calpain family and human disease.

Trends Mol Med 7:355-362.

74. Goll, D. E., V. F. Thompson, H. Li, W. Wei, and J. Cong. 2003. The calpain system. Physiol Rev 83:731-801.

75. Dargelos, E., S. Poussard, C. Brule, L. Daury, and P. Cottin. 2008. Calcium-dependent proteolytic system and muscle dysfunctions: a possible role of calpains in sarcopenia. Biochimie 90:359-368.

134 76. Suzuki, K., S. Hata, Y. Kawabata, and H. Sorimachi. 2004. Structure, activation, and biology of calpain. Diabetes 53 Suppl 1:S12-18.

77. Sato, K., and S. Kawashima. 2001. Calpain function in the modulation of signal transduction molecules. Biol Chem 382:743-751.

78. Carafoli, E., and M. Molinari. 1998. Calpain: a protease in search of a function? Biochem Biophys Res Commun 247:193-203.

79. Sorimachi, H., S. Ishiura, and K. Suzuki. 1997. Structure and physiological function of calpains. Biochem J 328 ( Pt 3):721-732.

80. Ma, Y., and L. M. Hendershot. 2004. The role of the unfolded protein response in tumour development: friend or foe? Nat Rev Cancer 4:966- 977.

81. Nakagawa, T., and J. Yuan. 2000. Cross-talk between two cysteine protease families. Activation of caspase-12 by calpain in apoptosis. J Cell Biol 150:887-894.

82. Daugas, E., D. Nochy, L. Ravagnan, M. Loeffler, S. A. Susin, N. Zamzami, and G. Kroemer. 2000. Apoptosis-inducing factor (AIF): a ubiquitous mitochondrial oxidoreductase involved in apoptosis. FEBS Lett 476:118- 123.

83. Miramar, M. D., P. Costantini, L. Ravagnan, L. M. Saraiva, D. Haouzi, G. Brothers, J. M. Penninger, M. L. Peleato, G. Kroemer, and S. A. Susin.

135 2001. NADH oxidase activity of mitochondrial apoptosis-inducing factor. J Biol Chem 276:16391-16398.

84. Delettre, C., V. J. Yuste, R. S. Moubarak, M. Bras, N. Robert, and S. A. Susin. 2006. Identification and characterization of AIFsh2, a mitochondrial apoptosis-inducing factor (AIF) isoform with NADH oxidase activity. J Biol Chem 281:18507-18518.

85. Boujrad, H., O. Gubkina, N. Robert, S. Krantic, and S. A. Susin. 2007. AIF- mediated programmed necrosis: a highly regulated way to die. Cell Cycle 6:2612-2619.

86. Susin, S. A., E. Daugas, L. Ravagnan, K. Samejima, N. Zamzami, M. Loeffler, P. Costantini, K. F. Ferri, T. Irinopoulou, M. C. Prevost, G. Brothers, T. W. Mak, J. Penninger, W. C. Earnshaw, and G. Kroemer. 2000. Two distinct pathways leading to nuclear apoptosis. J Exp Med 192:571-580.

87. Susin, S. A., H. K. Lorenzo, N. Zamzami, I. Marzo, B. E. Snow, G. M. Brothers, J. Mangion, E. Jacotot, P. Costantini, M. Loeffler, N. Larochette,

D. R. Goodlett, R. Aebersold, D. P. Siderovski, J. M. Penninger, and G. Kroemer. 1999. Molecular characterization of mitochondrial apoptosis- inducing factor. Nature 397:441-446.

88. Susin, S. A., N. Zamzami, N. Larochette, B. Dallaporta, I. Marzo, C. Brenner, T. Hirsch, P. X. Petit, M. Geuskens, and G. Kroemer. 1997. A

136 cytofluorometric assay of nuclear apoptosis induced in a cell-free system: application to ceramide-induced apoptosis. Exp Cell Res 236:397-403.

89. Susin, S. A., N. Zamzami, M. Castedo, T. Hirsch, P. Marchetti, A. Macho, E. Daugas, M. Geuskens, and G. Kroemer. 1996. Bcl-2 inhibits the mitochondrial release of an apoptogenic protease. J Exp Med 184:1331- 1341.

90. Roy Chowdhuri, S., T. Crum, A. Woollard, S. Aslam, and P. G. Okkema. 2006. The T-box factor TBX-2 and the SUMO conjugating enzyme UBC-9 are required for ABa-derived pharyngeal muscle in C. elegans. Dev Biol 295:664-677.

91. Wang, J., X. H. Feng, and R. J. Schwartz. 2004. SUMO-1 modification activated GATA4-dependent cardiogenic gene activity. J Biol Chem 279:49091-49098.

92. Wang, J., H. Zhang, D. Iyer, X. H. Feng, and R. J. Schwartz. 2008. Regulation of cardiac specific nkx2.5 gene activity by small ubiquitin-like modifier. J Biol Chem 283:23235-23243.

93. Riquelme, C., K. K. Barthel, and X. Liu. 2006. SUMO-1 modification of MEF2A regulates its transcriptional activity. J Cell Mol Med 10:132-144.

94. Matsuzaki, K., T. Minami, M. Tojo, Y. Honda, Y. Uchimura, H. Saitoh, H. Yasuda, S. Nagahiro, H. Saya, and M. Nakao. 2003. Serum response factor is modulated by the SUMO-1 conjugation system. Biochem Biophys Res Commun 306:32-38.

137 95. Meyer, M., W. Schillinger, B. Pieske, C. Holubarsch, C. Heilmann, H. Posival, G. Kuwajima, K. Mikoshiba, H. Just, G. Hasenfuss, and et al. 1995. Alterations of sarcoplasmic reticulum proteins in failing human dilated cardiomyopathy. Circulation 92:778-784.

96. Kho, C., A. Lee, D. Jeong, J. G. Oh, A. H. Chaanine, E. Kizana, W. J. Park, and R. J. Hajjar. 2011. SUMO1-dependent modulation of SERCA2a in heart failure. Nature 477:601-605.

97. Khodzhigorova, A., A. Distler, V. Lang, C. Dees, H. Schneider, C. Beyer, K. Gelse, O. Distler, G. Schett, and J. H. Distler. 2012. Inhibition of sumoylation prevents experimental fibrosis. Ann Rheum Dis 71:1904-1908.

98. Lee, Y. J., Y. Mou, D. Maric, D. Klimanis, S. Auh, and J. M. Hallenbeck. 2011. Elevated global SUMOylation in Ubc9 transgenic mice protects their brains against focal cerebral ischemic damage. PLoS One 6:e25852.

99. Kuwahara, K., Y. Saito, M. Takano, Y. Arai, S. Yasuno, Y. Nakagawa, N. Takahashi, Y. Adachi, G. Takemura, M. Horie, Y. Miyamoto, T. Morisaki, S. Kuratomi, A. Noma, H. Fujiwara, Y. Yoshimasa, H. Kinoshita, R.

Kawakami, I. Kishimoto, M. Nakanishi, S. Usami, M. Harada, and K. Nakao. 2003. NRSF regulates the fetal cardiac gene program and maintains normal cardiac structure and function. EMBO J 22:6310-6321.

100. Montgomery, R. L., C. A. Davis, M. J. Potthoff, M. Haberland, J. Fielitz, X. Qi, J. A. Hill, J. A. Richardson, and E. N. Olson. 2007. Histone

138 deacetylases 1 and 2 redundantly regulate cardiac morphogenesis, growth, and contractility. Genes Dev 21:1790-1802.

101. Trivedi, C. M., Y. Luo, Z. Yin, M. Zhang, W. Zhu, T. Wang, T. Floss, M. Goettlicher, P. R. Noppinger, W. Wurst, V. A. Ferrari, C. S. Abrams, P. J. Gruber, and J. A. Epstein. 2007. Hdac2 regulates the cardiac hypertrophic response by modulating Gsk3 beta activity. Nat Med 13:324-331.

102. Wang, H. C., Y. S. Huang, C. C. Ho, J. C. Jeng, and H. M. Shih. 2009. SUMO modification modulates the activity of calpain-2. Biochem Biophys Res Commun 384:444-449.

103. Galvez, A. S., A. Diwan, A. M. Odley, H. S. Hahn, H. Osinska, J. G. Melendez, J. Robbins, R. A. Lynch, Y. Marreez, and G. W. Dorn, 2nd. 2007. Cardiomyocyte degeneration with calpain deficiency reveals a critical role in protein homeostasis. Circ Res 100:1071-1078.

104. Malhotra, R., and P. K. Mason. 2009. Lamin A/C deficiency as a cause of familial dilated cardiomyopathy. Curr Opin Cardiol 24:203-208.

105. Zhang, Y. Q., and K. D. Sarge. 2008. Sumoylation regulates lamin A function and is lost in lamin A mutants associated with familial cardiomyopathies. J Cell Biol 182:35-39.

106. Fatkin, D., C. MacRae, T. Sasaki, M. R. Wolff, M. Porcu, M. Frenneaux, J. Atherton, H. J. Vidaillet, Jr., S. Spudich, U. De Girolami, J. G. Seidman, C. Seidman, F. Muntoni, G. Muehle, W. Johnson, and B. McDonough. 1999. Missense mutations in the rod domain of the lamin A/C gene as causes of

139 dilated cardiomyopathy and conduction-system disease. N Engl J Med 341:1715-1724.

107. Hill, J. A., M. Karimi, W. Kutschke, R. L. Davisson, K. Zimmerman, Z. Wang, R. E. Kerber, and R. M. Weiss. 2000. Cardiac hypertrophy is not a required compensatory response to short-term pressure overload. Circulation 101:2863-2869.

108. Frey, N., T. Barrientos, J. M. Shelton, D. Frank, H. Rutten, D. Gehring, C. Kuhn, M. Lutz, B. Rothermel, R. Bassel-Duby, J. A. Richardson, H. A. Katus, J. A. Hill, and E. N. Olson. 2004. Mice lacking calsarcin-1 are sensitized to calcineurin signaling and show accelerated cardiomyopathy in response to pathological biomechanical stress. Nat Med 10:1336-1343.

109. Yang, X., Q. Li, X. Lin, Y. Ma, X. Yue, Z. Tao, F. Wang, W. L. McKeehan, L. Wei, R. J. Schwartz, and J. Chang. 2012. Mechanism of fibrotic cardiomyopathy in mice expressing truncated Rho-associated coiled-coil protein kinase 1. FASEB J 26:2105-2116.

110. Kim, E. Y., L. Chen, Y. Ma, W. Yu, J. Chang, I. P. Moskowitz, and J. Wang.

2011. Expression of sumoylation deficient Nkx2.5 mutant in Nkx2.5 haploinsufficient mice leads to congenital heart defects. PLoS One 6:e20803.

111. LaBarre, D. D., and R. J. Lowy. 2001. Improvements in methods for calculating virus titer estimates from TCID50 and plaque assays. J Virol Methods 96:107-126.

140 112. Paquet-Durand, F., S. Azadi, S. M. Hauck, M. Ueffing, T. van Veen, and P. Ekstrom. 2006. Calpain is activated in degenerating photoreceptors in the rd1 mouse. J Neurochem 96:802-814.

113. Patterson, C., A. L. Portbury, J. C. Schisler, and M. S. Willis. 2011. Tear me down: role of calpain in the development of cardiac ventricular hypertrophy. Circ Res 109:453-462.

114. Orrenius, S., B. Zhivotovsky, and P. Nicotera. 2003. Regulation of cell death: the calcium-apoptosis link. Nat Rev Mol Cell Biol 4:552-565.

115. Kramerova, I., E. Kudryashova, G. Venkatraman, and M. J. Spencer. 2005. Calpain 3 participates in sarcomere remodeling by acting upstream of the ubiquitin-proteasome pathway. Hum Mol Genet 14:2125-2134.

116. Zatz, M., and A. Starling. 2005. Calpains and disease. N Engl J Med 352:2413-2423.

117. Averna, M., R. de Tullio, M. Passalacqua, F. Salamino, S. Pontremoli, and E. Melloni. 2001. Changes in intracellular calpastatin localization are

mediated by reversible phosphorylation. Biochem J 354:25-30.

118. Melloni, E., M. Averna, R. Stifanese, R. De Tullio, E. Defranchi, F. Salamino, and S. Pontremoli. 2006. Association of calpastatin with inactive calpain: a novel mechanism to control the activation of the protease? J Biol Chem 281:24945-24954.

141 119. Cheng, J., X. Kang, S. Zhang, and E. T. Yeh. 2007. SUMO-specific protease 1 is essential for stabilization of HIF1alpha during hypoxia. Cell 131:584-595.

120. Porn-Ares, M. I., A. Samali, and S. Orrenius. 1998. Cleavage of the calpain inhibitor, calpastatin, during apoptosis. Cell Death Differ 5:1028-1033.

121. Mikkonen, L., J. Hirvonen, and O. A. Janne. 2013. SUMO-1 regulates body weight and adipogenesis via PPARgamma in male and female mice. Endocrinology 154:698-708.

122. Adams, J. W., Y. Sakata, M. G. Davis, V. P. Sah, Y. Wang, S. B. Liggett, K. R. Chien, J. H. Brown, and G. W. Dorn, 2nd. 1998. Enhanced Galphaq signaling: a common pathway mediates cardiac hypertrophy and apoptotic heart failure. Proc Natl Acad Sci U S A 95:10140-10145.

123. Hirota, H., J. Chen, U. A. Betz, K. Rajewsky, Y. Gu, J. Ross, Jr., W. Muller, and K. R. Chien. 1999. Loss of a gp130 cardiac muscle cell survival pathway is a critical event in the onset of heart failure during biomechanical stress. Cell 97:189-198.

124. Averna, M., R. De Tullio, F. Salamino, E. Melloni, and S. Pontremoli. 1999. Phosphorylation of rat brain calpastatins by protein kinase C. FEBS Lett 450:13-16.

125. Rubio, T., S. Vernia, and P. Sanz. 2013. Sumoylation of AMPKbeta2 subunit enhances AMP-activated protein kinase activity. Mol Biol Cell 24:1801-1811.

142 126. Mandal, S., P. Guptan, E. Owusu-Ansah, and U. Banerjee. 2005. Mitochondrial regulation of cell cycle progression during development as revealed by the tenured mutation in Drosophila. Dev Cell 9:843-854.

127. Owusu-Ansah, E., A. Yavari, S. Mandal, and U. Banerjee. 2008. Distinct mitochondrial retrograde signals control the G1-S cell cycle checkpoint. Nat Genet 40:356-361.

128. Arad, M., I. P. Moskowitz, V. V. Patel, F. Ahmad, A. R. Perez-Atayde, D. B. Sawyer, M. Walter, G. H. Li, P. G. Burgon, C. T. Maguire, D. Stapleton, J. P. Schmitt, X. X. Guo, A. Pizard, S. Kupershmidt, D. M. Roden, C. I. Berul, C. E. Seidman, and J. G. Seidman. 2003. Transgenic mice overexpressing mutant PRKAG2 define the cause of Wolff-Parkinson-White syndrome in glycogen storage cardiomyopathy. Circulation 107:2850-2856.

129. Kim, E. Y., L. Chen, Y. Ma, W. Yu, J. Chang, I. P. Moskowitz, and J. Wang. 2012. Enhanced desumoylation in murine hearts by overexpressed SENP2 leads to congenital heart defects and cardiac dysfunction. J Mol Cell Cardiol 52:638-649.

130. Yan, S., X. Sun, B. Xiang, H. Cang, X. Kang, Y. Chen, H. Li, G. Shi, E. T. Yeh, B. Wang, X. Wang, and J. Yi. 2010. Redox regulation of the stability of the SUMO protease SENP3 via interactions with CHIP and Hsp90. EMBO J 29:3773-3786.

131. Bawa-Khalfe, T., L. S. Lu, Y. Zuo, C. Huang, R. Dere, F. M. Lin, and E. T. Yeh. 2012. Differential expression of SUMO-specific protease 7 variants

143 regulates epithelial-mesenchymal transition. Proc Natl Acad Sci U S A 109:17466-17471.

132. Maison, C., K. Romeo, D. Bailly, M. Dubarry, J. P. Quivy, and G. Almouzni. 2012. The SUMO protease SENP7 is a critical component to ensure HP1 enrichment at pericentric heterochromatin. Nat Struct Mol Biol 19:458-460.

133. Kim, J. H., and S. H. Baek. 2009. Emerging roles of desumoylating enzymes. Biochim Biophys Acta 1792:155-162.

134. Ashrafian, H., L. Docherty, V. Leo, C. Towlson, M. Neilan, V. Steeples, C. A. Lygate, T. Hough, S. Townsend, D. Williams, S. Wells, D. Norris, S. Glyn-Jones, J. Land, I. Barbaric, Z. Lalanne, P. Denny, D. Szumska, S. Bhattacharya, J. L. Griffin, I. Hargreaves, N. Fernandez-Fuentes, M. Cheeseman, H. Watkins, and T. N. Dear. 2010. A mutation in the mitochondrial fission gene Dnm1l leads to cardiomyopathy. PLoS Genet 6:e1001000.

135. Zhou, S., S. Kachhap, and K. K. Singh. 2003. Mitochondrial impairment in p53-deficient human cancer cells. Mutagenesis 18:287-292.

136. Boudina, S., S. Sena, B. T. O'Neill, P. Tathireddy, M. E. Young, and E. D. Abel. 2005. Reduced mitochondrial oxidative capacity and increased mitochondrial uncoupling impair myocardial energetics in obesity. Circulation 112:2686-2695.

137. Jeong, D., H. Cha, E. Kim, M. Kang, D. K. Yang, J. M. Kim, P. O. Yoon, J. G. Oh, O. Y. Bernecker, S. Sakata, T. T. Le, L. Cui, Y. H. Lee, H. Kim do,

144 S. H. Woo, R. Liao, R. J. Hajjar, and W. J. Park. 2006. PICOT inhibits cardiac hypertrophy and enhances ventricular function and cardiomyocyte contractility. Circ Res 99:307-314.

138. Baines, C. P., R. A. Kaiser, N. H. Purcell, N. S. Blair, H. Osinska, M. A. Hambleton, E. W. Brunskill, M. R. Sayen, R. A. Gottlieb, G. W. Dorn, J. Robbins, and J. D. Molkentin. 2005. Loss of cyclophilin D reveals a critical role for mitochondrial permeability transition in cell death. Nature 434:658- 662.

139. Christofferson, D. E., and J. Yuan. 2010. Necroptosis as an alternative form of programmed cell death. Curr Opin Cell Biol 22:263-268.

140. Schinzel, A. C., O. Takeuchi, Z. Huang, J. K. Fisher, Z. Zhou, J. Rubens, C. Hetz, N. N. Danial, M. A. Moskowitz, and S. J. Korsmeyer. 2005. Cyclophilin D is a component of mitochondrial permeability transition and mediates neuronal cell death after focal cerebral ischemia. Proc Natl Acad Sci U S A 102:12005-12010.

141. Vandenabeele, P., L. Galluzzi, T. Vanden Berghe, and G. Kroemer. 2010.

Molecular mechanisms of necroptosis: an ordered cellular explosion. Nat Rev Mol Cell Biol 11:700-714.

142. Elrod, J. W., R. Wong, S. Mishra, R. J. Vagnozzi, B. Sakthievel, S. A. Goonasekera, J. Karch, S. Gabel, J. Farber, T. Force, J. H. Brown, E. Murphy, and J. D. Molkentin. 2010. Cyclophilin D controls mitochondrial

145 pore-dependent Ca(2+) exchange, metabolic flexibility, and propensity for heart failure in mice. J Clin Invest 120:3680-3687.

143. Nakayama, H., X. Chen, C. P. Baines, R. Klevitsky, X. Zhang, H. Zhang, N. Jaleel, B. H. Chua, T. E. Hewett, J. Robbins, S. R. Houser, and J. D. Molkentin. 2007. Ca2+- and mitochondrial-dependent cardiomyocyte necrosis as a primary mediator of heart failure. J Clin Invest 117:2431- 2444.

144. Kang, X., Y. Qi, Y. Zuo, Q. Wang, Y. Zou, R. J. Schwartz, J. Cheng, and E. T. Yeh. 2010. SUMO-specific protease 2 is essential for suppression of polycomb group protein-mediated gene silencing during embryonic development. Mol Cell 38:191-201.

145. Marin-Garcia, J., M. J. Goldenthal, and G. W. Moe. 2001. Mitochondrial pathology in cardiac failure. Cardiovasc Res 49:17-26.

146. Aubert, G., R. B. Vega, and D. P. Kelly. 2013. Perturbations in the gene regulatory pathways controlling mitochondrial energy production in the failing heart. Biochim Biophys Acta 1833:840-847.

147. Huss, J. M., and D. P. Kelly. 2005. Mitochondrial energy metabolism in heart failure: a question of balance. J Clin Invest 115:547-555.

148. Chen, L., Q. Gong, J. P. Stice, and A. A. Knowlton. 2009. Mitochondrial OPA1, apoptosis, and heart failure. Cardiovasc Res 84:91-99.

146 149. Chen, Y., Y. Liu, and G. W. Dorn, 2nd. 2011. Mitochondrial fusion is essential for organelle function and cardiac homeostasis. Circ Res 109:1327-1331.

150. Cereghetti, G. M., A. Stangherlin, O. Martins de Brito, C. R. Chang, C. Blackstone, P. Bernardi, and L. Scorrano. 2008. Dephosphorylation by calcineurin regulates translocation of Drp1 to mitochondria. Proc Natl Acad Sci U S A 105:15803-15808.

151. Baandrup, U., R. A. Florio, F. Roters, and E. G. Olsen. 1981. Electron microscopic investigation of endomyocardial biopsy samples in hypertrophy and cardiomyopathy. A semiquantitative study in 48 patients. Circulation 63:1289-1298.

152. Parra, V., V. Eisner, M. Chiong, A. Criollo, F. Moraga, A. Garcia, S. Hartel, E. Jaimovich, A. Zorzano, C. Hidalgo, and S. Lavandero. 2008. Changes in mitochondrial dynamics during ceramide-induced cardiomyocyte early apoptosis. Cardiovasc Res 77:387-397.

153. Neuspiel, M., R. Zunino, S. Gangaraju, P. Rippstein, and H. McBride.

2005. Activated mitofusin 2 signals mitochondrial fusion, interferes with Bax activation, and reduces susceptibility to radical induced depolarization. J Biol Chem 280:25060-25070.

154. Guo, X., K. H. Chen, Y. Guo, H. Liao, J. Tang, and R. P. Xiao. 2007. Mitofusin 2 triggers vascular smooth muscle cell apoptosis via mitochondrial death pathway. Circ Res 101:1113-1122.

147 155. Shen, T., M. Zheng, C. Cao, C. Chen, J. Tang, W. Zhang, H. Cheng, K. H. Chen, and R. P. Xiao. 2007. Mitofusin-2 is a major determinant of oxidative stress-mediated heart muscle cell apoptosis. J Biol Chem 282:23354- 23361.

156. Wasiak, S., R. Zunino, and H. M. McBride. 2007. Bax/Bak promote sumoylation of DRP1 and its stable association with mitochondria during apoptotic cell death. J Cell Biol 177:439-450.

157. Yu, T., S. S. Sheu, J. L. Robotham, and Y. Yoon. 2008. Mitochondrial fission mediates high glucose-induced cell death through elevated production of reactive oxygen species. Cardiovasc Res 79:341-351.

158. Ishihara, N., M. Nomura, A. Jofuku, H. Kato, S. O. Suzuki, K. Masuda, H. Otera, Y. Nakanishi, I. Nonaka, Y. Goto, N. Taguchi, H. Morinaga, M. Maeda, R. Takayanagi, S. Yokota, and K. Mihara. 2009. Mitochondrial fission factor Drp1 is essential for embryonic development and synapse formation in mice. Nat Cell Biol 11:958-966.

159. Danial, N. N., and S. J. Korsmeyer. 2004. Cell death: critical control points.

Cell 116:205-219.

160. Nakagawa, T., S. Shimizu, T. Watanabe, O. Yamaguchi, K. Otsu, H. Yamagata, H. Inohara, T. Kubo, and Y. Tsujimoto. 2005. Cyclophilin D- dependent mitochondrial permeability transition regulates some necrotic but not apoptotic cell death. Nature 434:652-658.

148 161. Daugas, E., S. A. Susin, N. Zamzami, K. F. Ferri, T. Irinopoulou, N. Larochette, M. C. Prevost, B. Leber, D. Andrews, J. Penninger, and G. Kroemer. 2000. Mitochondrio-nuclear translocation of AIF in apoptosis and necrosis. FASEB J 14:729-739.

162. Zhang, X., J. Chen, S. H. Graham, L. Du, P. M. Kochanek, R. Draviam, F. Guo, P. D. Nathaniel, C. Szabo, S. C. Watkins, and R. S. Clark. 2002. Intranuclear localization of apoptosis-inducing factor (AIF) and large scale DNA fragmentation after traumatic brain injury in rats and in neuronal cultures exposed to peroxynitrite. J Neurochem 82:181-191.

163. Cao, G., J. Xing, X. Xiao, A. K. Liou, Y. Gao, X. M. Yin, R. S. Clark, S. H. Graham, and J. Chen. 2007. Critical role of calpain I in mitochondrial release of apoptosis-inducing factor in ischemic neuronal injury. J Neurosci 27:9278-9293.

164. Nikolova, V., C. Leimena, A. C. McMahon, J. C. Tan, S. Chandar, D. Jogia, S. H. Kesteven, J. Michalicek, R. Otway, F. Verheyen, S. Rainer, C. L. Stewart, D. Martin, M. P. Feneley, and D. Fatkin. 2004. Defects in nuclear structure and function promote dilated cardiomyopathy in lamin A/C- deficient mice. J Clin Invest 113:357-369.

165. Guo, C., K. L. Hildick, J. Luo, L. Dearden, K. A. Wilkinson, and J. M. Henley. 2013. SENP3-mediated deSUMOylation of dynamin-related protein 1 promotes cell death following ischaemia. EMBO J 32:1514-1528.

149 166. Besnault-Mascard, L., C. Leprince, M. T. Auffredou, B. Meunier, M. F. Bourgeade, J. Camonis, H. K. Lorenzo, and A. Vazquez. 2005. Caspase-8 sumoylation is associated with nuclear localization. Oncogene 24:3268- 3273.

167. Shirakura, H., N. Hayashi, S. Ogino, K. Tsuruma, T. Uehara, and Y. Nomura. 2005. Caspase recruitment domain of procaspase-2 could be a target for SUMO-1 modification through Ubc9. Biochem Biophys Res Commun 331:1007-1015.

168. Hayashi, N., H. Shirakura, T. Uehara, and Y. Nomura. 2006. Relationship between SUMO-1 modification of caspase-7 and its nuclear localization in human neuronal cells. Neurosci Lett 397:5-9.

169. Leitao, B. B., M. C. Jones, and J. J. Brosens. 2011. The SUMO E3-ligase PIAS1 couples reactive oxygen species-dependent JNK activation to oxidative cell death. FASEB J 25:3416-3425.

170. de la Vega, L., I. Grishina, R. Moreno, M. Kruger, T. Braun, and M. L. Schmitz. 2012. A redox-regulated SUMO/acetylation switch of HIPK2

controls the survival threshold to oxidative stress. Mol Cell 46:472-483.

171. Desterro, J. M., M. S. Rodriguez, and R. T. Hay. 1998. SUMO-1 modification of IkappaBalpha inhibits NF-kappaB activation. Mol Cell 2:233-239.

150 172. Lin, X., M. Liang, Y. Y. Liang, F. C. Brunicardi, and X. H. Feng. 2003. SUMO-1/Ubc9 promotes nuclear accumulation and metabolic stability of tumor suppressor Smad4. J Biol Chem 278:31043-31048.

173. Huang, T. T., S. M. Wuerzberger-Davis, Z. H. Wu, and S. Miyamoto. 2003. Sequential modification of NEMO/IKKgamma by SUMO-1 and ubiquitin mediates NF-kappaB activation by genotoxic stress. Cell 115:565-576.

174. Bugger, H., M. Schwarzer, D. Chen, A. Schrepper, P. A. Amorim, M. Schoepe, T. D. Nguyen, F. W. Mohr, O. Khalimonchuk, B. C. Weimer, and T. Doenst. 2010. Proteomic remodelling of mitochondrial oxidative pathways in pressure overload-induced heart failure. Cardiovasc Res 85:376-384.

175. Lin, C. S., Y. L. Sun, and C. Y. Liu. 2003. Structural and biochemical evidence of mitochondrial depletion in pigs with hypertrophic cardiomyopathy. Res Vet Sci 74:219-226.

176. Turko, I. V., L. Li, K. S. Aulak, D. J. Stuehr, J. Y. Chang, and F. Murad. 2003. Protein tyrosine nitration in the mitochondria from diabetic mouse

heart. Implications to dysfunctional mitochondria in diabetes. J Biol Chem 278:33972-33977.

177. Hu, Y., J. Suarez, E. Fricovsky, H. Wang, B. T. Scott, S. A. Trauger, W. Han, M. O. Oyeleye, and W. H. Dillmann. 2009. Increased enzymatic O- GlcNAcylation of mitochondrial proteins impairs mitochondrial function in cardiac myocytes exposed to high glucose. J Biol Chem 284:547-555.

151 178. Cribbs, J. T., and S. Strack. 2007. Reversible phosphorylation of Drp1 by cyclic AMP-dependent protein kinase and calcineurin regulates mitochondrial fission and cell death. EMBO Rep 8:939-944.

179. Chang, C. R., and C. Blackstone. 2007. Cyclic AMP-dependent protein kinase phosphorylation of Drp1 regulates its GTPase activity and mitochondrial morphology. J Biol Chem 282:21583-21587.

180. Han, X. J., Y. F. Lu, S. A. Li, T. Kaitsuka, Y. Sato, K. Tomizawa, A. C. Nairn, K. Takei, H. Matsui, and M. Matsushita. 2008. CaM kinase I alpha- induced phosphorylation of Drp1 regulates mitochondrial morphology. J Cell Biol 182:573-585.

181. Sandebring, A., K. J. Thomas, A. Beilina, M. van der Brug, M. M. Cleland, R. Ahmad, D. W. Miller, I. Zambrano, R. F. Cowburn, H. Behbahani, A. Cedazo-Minguez, and M. R. Cookson. 2009. Mitochondrial alterations in PINK1 deficient cells are influenced by calcineurin-dependent dephosphorylation of dynamin-related protein 1. PLoS One 4:e5701.

182. Dagda, R. K., A. M. Gusdon, I. Pien, S. Strack, S. Green, C. Li, B. Van

Houten, S. J. Cherra, 3rd, and C. T. Chu. 2011. Mitochondrially localized PKA reverses mitochondrial pathology and dysfunction in a cellular model of Parkinson's disease. Cell Death Differ 18:1914-1923.

183. Frank, S., B. Gaume, E. S. Bergmann-Leitner, W. W. Leitner, E. G. Robert, F. Catez, C. L. Smith, and R. J. Youle. 2001. The role of dynamin-related

152 protein 1, a mediator of mitochondrial fission, in apoptosis. Dev Cell 1:515- 525.

184. Szabadkai, G., A. M. Simoni, M. Chami, M. R. Wieckowski, R. J. Youle, and R. Rizzuto. 2004. Drp-1-dependent division of the mitochondrial network blocks intraorganellar Ca2+ waves and protects against Ca2+- mediated apoptosis. Mol Cell 16:59-68.

185. Bae, S., P. M. Siu, S. Choudhury, Q. Ke, J. H. Choi, Y. Y. Koh, and P. M. Kang. 2010. Delayed activation of caspase-independent apoptosis during heart failure in transgenic mice overexpressing caspase inhibitor CrmA. Am J Physiol Heart Circ Physiol 299:H1374-1381.

186. Loeffler, M., E. Daugas, S. A. Susin, N. Zamzami, D. Metivier, A. L. Nieminen, G. Brothers, J. M. Penninger, and G. Kroemer. 2001. Dominant cell death induction by extramitochondrially targeted apoptosis-inducing factor. FASEB J 15:758-767.

187. Polster, B. M., G. Basanez, A. Etxebarria, J. M. Hardwick, and D. G. Nicholls. 2005. Calpain I induces cleavage and release of apoptosis-

inducing factor from isolated mitochondria. J Biol Chem 280:6447-6454.

188. Yuste, V. J., R. S. Moubarak, C. Delettre, M. Bras, P. Sancho, N. Robert, J. d'Alayer, and S. A. Susin. 2005. Cysteine protease inhibition prevents mitochondrial apoptosis-inducing factor (AIF) release. Cell Death Differ 12:1445-1448.

153 189. Takano, J., M. Tomioka, S. Tsubuki, M. Higuchi, N. Iwata, S. Itohara, M. Maki, and T. C. Saido. 2005. Calpain mediates excitotoxic DNA fragmentation via mitochondrial pathways in adult brains: evidence from calpastatin mutant mice. J Biol Chem 280:16175-16184.

190. Vahsen, N., C. Cande, J. J. Briere, P. Benit, N. Joza, N. Larochette, P. G. Mastroberardino, M. O. Pequignot, N. Casares, V. Lazar, O. Feraud, N. Debili, S. Wissing, S. Engelhardt, F. Madeo, M. Piacentini, J. M. Penninger, H. Schagger, P. Rustin, and G. Kroemer. 2004. AIF deficiency compromises oxidative phosphorylation. EMBO J 23:4679-4689.

191. Golstein, P., and G. Kroemer. 2007. Cell death by necrosis: towards a molecular definition. Trends Biochem Sci 32:37-43.

192. Wang, Y., V. K. Rao, W. K. Kok, D. N. Roy, S. Sethi, B. M. Ling, M. B. Lee, and R. Taneja. 2012. SUMO modification of Stra13 is required for repression of cyclin D1 expression and cellular growth arrest. PLoS One 7:e43137.

193. Werner, A., A. Flotho, and F. Melchior. 2012. The

RanBP2/RanGAP1*SUMO1/Ubc9 complex is a multisubunit SUMO E3 ligase. Mol Cell 46:287-298.

194. Kops, G. J., T. B. Dansen, P. E. Polderman, I. Saarloos, K. W. Wirtz, P. J. Coffer, T. T. Huang, J. L. Bos, R. H. Medema, and B. M. Burgering. 2002. Forkhead transcription factor FOXO3a protects quiescent cells from oxidative stress. Nature 419:316-321.

154 195. Tran, H., A. Brunet, J. M. Grenier, S. R. Datta, A. J. Fornace, Jr., P. S. DiStefano, L. W. Chiang, and M. E. Greenberg. 2002. DNA repair pathway stimulated by the forkhead transcription factor FOXO3a through the Gadd45 protein. Science 296:530-534.

196. Bawa-Khalfe, T., and E. T. Yeh. 2010. SUMO Losing Balance: SUMO Proteases Disrupt SUMO Homeostasis to Facilitate Cancer Development and Progression. Genes Cancer 1:748-752.

197. Janer, A., A. Werner, J. Takahashi-Fujigasaki, A. Daret, H. Fujigasaki, K. Takada, C. Duyckaerts, A. Brice, A. Dejean, and A. Sittler. 2010. SUMOylation attenuates the aggregation propensity and cellular toxicity of the polyglutamine expanded ataxin-7. Hum Mol Genet 19:181-195.

198. Driscoll, J. J., D. Pelluru, K. Lefkimmiatis, M. Fulciniti, R. H. Prabhala, P. R. Greipp, B. Barlogie, Y. T. Tai, K. C. Anderson, J. D. Shaughnessy, Jr., C. M. Annunziata, and N. C. Munshi. 2010. The sumoylation pathway is dysregulated in multiple myeloma and is associated with adverse patient outcome. Blood 115:2827-2834.

199. Kessler, J. D., K. T. Kahle, T. Sun, K. L. Meerbrey, M. R. Schlabach, E. M. Schmitt, S. O. Skinner, Q. Xu, M. Z. Li, Z. C. Hartman, M. Rao, P. Yu, R. Dominguez-Vidana, A. C. Liang, N. L. Solimini, R. J. Bernardi, B. Yu, T. Hsu, I. Golding, J. Luo, C. K. Osborne, C. J. Creighton, S. G. Hilsenbeck, R. Schiff, C. A. Shaw, S. J. Elledge, and T. F. Westbrook. 2012. A SUMOylation-dependent transcriptional subprogram is required for Myc- driven tumorigenesis. Science 335:348-353.

155 200. Golebiowski, F., I. Matic, M. H. Tatham, C. Cole, Y. Yin, A. Nakamura, J. Cox, G. J. Barton, M. Mann, and R. T. Hay. 2009. System-wide changes to SUMO modifications in response to heat shock. Sci Signal 2:ra24.

201. Fukuda, I., A. Ito, G. Hirai, S. Nishimura, H. Kawasaki, H. Saitoh, K. Kimura, M. Sodeoka, and M. Yoshida. 2009. Ginkgolic acid inhibits protein SUMOylation by blocking formation of the E1-SUMO intermediate. Chem Biol 16:133-140.

202. Fukuda, I., A. Ito, M. Uramoto, H. Saitoh, H. Kawasaki, H. Osada, and M. Yoshida. 2009. Kerriamycin B inhibits protein SUMOylation. J Antibiot (Tokyo) 62:221-224.

203. Kim, Y. S., K. Nagy, S. Keyser, and J. S. Schneekloth, Jr. 2013. An electrophoretic mobility shift assay identifies a mechanistically unique inhibitor of protein sumoylation. Chem Biol 20:604-613.

204. Madu, I. G., A. T. Namanja, Y. Su, S. Wong, Y. J. Li, and Y. Chen. 2013. Identification and Characterization of a New Chemotype of Noncovalent SENP Inhibitors. ACS Chem Biol.

205. Yang, W., L. Wang, and W. Paschen. 2013. Development of a high- throughput screening assay for inhibitors of small ubiquitin-like modifier proteases. J Biomol Screen 18:621-628.

156 Vita

Eun Young Kim was born in Busan, South Korea on March 26, 1976, the

Daughter of Jin Goo Kim and In Ja Yoo. After completing her work at Pusan

National University in Busan, South Korea, receving the degree of Bachelor of

Science with a major in microbiology in Feb. 1999. She entered Gwangju Institute of Science and Technology (GIST), Gwangu, South Korea. She earned the degree of Master of Science with a major in life science from GIST in Feb. 2002.

For the next two years, she worked as a research technician in the same institute

(GIST). She matriculated at The University of Texas Health Science Center at

Houston Graduate School of Biomedical Sciences in August of 2005.

Permanent address:

7675 Phoenix Dr. #516

Houston, Texas 77030

157