Diabetes Volume 64, October 2015 3573

Genaro A. Ramirez-Correa,1 Junfeng Ma,2 Chad Slawson,3 Quira Zeidan,2 Nahyr S. Lugo-Fagundo,1 Mingguo Xu,1 Xiaoxu Shen,4 Wei Dong Gao,4 Viviane Caceres,5 Khalid Chakir,5 Lauren DeVine,2 Robert N. Cole,2 Luigi Marchionni,6 Nazareno Paolocci,5 Gerald W. Hart,2 and Anne M. Murphy1

Removal of Abnormal Myofilament O-GlcNAcylation Restores Ca2+ Sensitivity in Diabetic

Diabetes 2015;64:3573–3587 | DOI: 10.2337/db14-1107

Contractile dysfunction and increased deposition of In diabetic cardiomyopathy, the contractile and electro- O-linked b-N-acetyl-D-glucosamine (O-GlcNAc) in car- physiological properties of the cardiac muscle are altered diac are a hallmark of the diabetic heart. How- (1). Prior studies have mainly focused on alterations in ever, whether and how this posttranslational alteration Ca2+ handling (2–4). However, these perturbations alone contributes to lower cardiac function remains unclear. unlikely account for lower force production and altered COMPLICATIONS fi b Using a re ned -elimination/Michael addition with tan- relaxation typically found in the heart of patients with dem mass tags (TMT)–labeling proteomic technique, we diabetes (5,6). Indeed, the intrinsic properties of cardiac show that CpOGA, a bacterial analog of O-GlcNAcase myofilaments appear to be altered too (7,8). More specif- (OGA) that cleaves O-GlcNAc in vivo, removes site- 2+ 2+ fi specific O-GlcNAcylation from myofilaments, restoring ically, Ca sensitivity (ECa 50), a measure of myo la- 2+ Ca2+ sensitivity in streptozotocin (STZ) diabetic cardiac ment force production at near physiological Ca levels, – muscles. We report that in control rat hearts, O-GlcNAc is reduced in the heart of patients with diabetes (9 11). 2+ and O-GlcNAc transferase (OGT) are mainly localized at Yet the mechanisms responsible for Ca desensitization the Z-line, whereas OGA is at the A-band. Conversely, in in diabetic hearts remain incompletely understood. diabetic hearts O-GlcNAc levels are increased and OGT O-GlcNAcylation is a posttranslational modification (PTM) and OGA delocalized. Consistent changes were found in linked to glucose metabolism and centrally involved in reg- human diabetic hearts. STZ diabetic hearts display in- ulating cellular homeostasis (12). This PTM consists of creased physical interactions of OGA with a-, the addition of single O-linked b-N-acetyl-D-glucosamine , and light chain 1, along with re- (O-GlcNAc) sugar to serine (Ser) and threonine (Thr) res- duced OGT and increased OGA activities. Our study is idues of nuclear and cytoplasmic proteins. The reaction the first to reveal that specific removal of O-GlcNAcylation is catalyzed by O-GlcNAc transferase (OGT), whereas restores myofilament response to Ca2+ in diabetic hearts O-GlcNAc removal is under the control of O-GlcNAcase and that altered O-GlcNAcylation is due to the subcellular (OGA) (12). Excessive O-GlcNAcylation results from redistribution of OGT and OGA rather than to changes in glucose- or other nutrient-induced overload of the hexos- their overall activities. Thus, preventing sarcomeric OGT amine biosynthesis pathway (HBP). Alterations in HBP and OGA displacement represents a new possible strat- are increasingly recognized as a major contributing factor egy for treating diabetic cardiomyopathy. for insulin resistance (12) and “glucose toxicity” during

1Division of Cardiology, Department of Pediatrics, Johns Hopkins University 6Department of Oncology, Johns Hopkins University School of Medicine, Balti- School of Medicine, Baltimore, MD more, MD 2 Department of Biological Chemistry, Johns Hopkins University School of Medi- Corresponding author: Genaro A. Ramirez-Correa, [email protected]. cine, Baltimore, MD Received 1 August 2014 and accepted 14 May 2015. 3Department of Biochemistry and Molecular Biology, University of Kansas Medical Center, Kansas City, KS G.A.R.-C. and J.M. contributed equally to this work. 4Department of Anesthesiology and Critical Care Medicine, Johns Hopkins Uni- © 2015 by the American Diabetes Association. Readers may use this article as versity School of Medicine, Baltimore, MD long as the work is properly cited, the use is educational and not for profit, and 5Division of Cardiology, Department of Medicine, Johns Hopkins University School the work is not altered. of Medicine, Baltimore, MD See accompanying article, p. 3339. 2+ 3574 OGA Restores Cardiac Muscle ECa 50 in Diabetes Diabetes Volume 64, October 2015 diabetes. Similar to phosphorylation, O-GlcNAcylation is Immunofluorescence Confocal Microscopy a widely distributed and highly dynamic PTM (12). How- Cold acetone–fixed cryosections (7–8 mm) from rat or ever, unlike phosphorylation, which is regulated by a myriad human (BioChain Inc., Office of Human Research Protec- of kinases and phosphatases, the extent of O-GlcNAcylation tion registered IRB00008283) myocardium were blocked relies on two only, specifically OGT and OGA. OGT and incubated overnight with a primary antibody against substrate specificity is regulated by transient -to- O-GlcNAc (CTD110.6). In addition, isolated skinned myo- protein interactions that take place primarily at its tetratri- cytes were obtained from flash-frozen myocardium by copeptide repeat (TPR) domain (12) (Fig. 7A). Often homogenization in 0.03% Triton X-100 at low speed as OGT and OGA interact with each other and/or are found previously described (28), seeded on eight chamber slides forming a holoenzyme complex with protein phospha- coated with 40 mg/mL laminin (Invitrogen), and fixed in tases and kinases (12). Modifications of Ser and Thr by 4% formaldehyde-methanol–free ultrapure (Polysciences O-GlcNAc occur in myofilaments, and the addition of ex- Inc.). Cryosections or isolated skinned myocytes were ogenous N-acetyl-D-glucosamine (GlcNAc) alters myofila- blocked and incubated overnight with OGA O-GlcNAcase ment response to Ca2+ (13,14). In addition, manipulation (345), OGT (AL-25) (29), and anti–a- (Sigma-Aldrich) of cardiac O-GlcNAc levels influences Ca2+ cycling kinetics at 1 mg/mL. Secondary antibodies were Alexa 647 goat anti- (4) and mitochondrial rates of respiration in diabetes mouse IgM (m-chain) for O-GlcNAc, Alexa 647 goat anti- (2,15) and functional recovery after ischemia-reperfusion rabbit IgG for OGT, Alexa 594 goat anti-chicken IgY for injury (16,17) or chronic pressure-overload (18–21). OGA, and Alexa 488 goat anti-mouse IgG for a-actinin. Despite all of this evidence, whether and how altered Prolong antifade with DAPI (Invitrogen) was used for O-GlcNAcylation contributes to myofilament dysfunc- mounting. Images were acquired on a Zeiss 710 LSM up- tion in diabetic cardiomyopathy (7,8,22–24) is currently right microscope using a 325 or a 363 water-immersion unclear. objective (Nikon) and analyzed with Zen 9 Leica Zeiss soft- Here we proved that specific removal of O-GlcNAc ex- ware tools. cess from diabetic myofilaments ameliorates contractile Double Immunoelectron Microscopy dysfunction by linking improvement in force-Ca2+ rela- For gold immunolabeling, goat anti-rabbit or goat anti- tionships to site-specific O-GlcNAc changes. We also mouse were labeled with 12-nm-diameter particles to determined a potential mechanism leading to altered detect anti-OGT (AL-34) or anti–O-GlcNAc (CTD 110.6), O-GlcNAcylation by comparing the status and sarcomeric and goat anti-chicken was labeled with 6-nm-diameter par- distribution patterns of OGT and OGA in the heart of rats ticles to detect anti-OGA (345). Labeled ultrathin sections with streptozotocin (STZ)-induced type 1 diabetes with (60- to 90-nm thick) were examined under the transmis- that found in controls. sion electron microscope (Hitachi 7600 TEM, Tokyo, Japan), and 8–11 random field pictures were used for RESEARCH DESIGN AND METHODS quantification of OGT, OGA, and O-GlcNAc immunolabeled All animal protocols in this study were performed in gold particles with ImageJ software (National Institutes of accordance with institutional guidelines and approval of Health [NIH]). the Johns Hopkins University School of Medicine Insti- tutional Animal Care and Use Committee. Coimmunoprecipitation for OGT and OGA For immunoprecipitation studies, anti-OGT (AL-28) or Type 1 Diabetes Rat Model anti-OGA (345) (1 mg total) antibodies were added to Type 1 diabetes was induced in male Sprague-Dawley rats 0.5 mg/mL protein samples. Immunoprecipitates were (Charles River) by an intraperitoneal injection of STZ then separated by SDS-PAGE, transferred to polyvinyli- (65 mg/kg). Control animals were injected with vehicle only. dene fluoride membranes (Millipore, Bedford, MA) for Animals were killed 6–8 weeks after induction of diabetes. antibody probing against a-cardiac actin (Sigma-Aldrich), At the moment of tissue harvest, STZ diabetic animals had a-tropomyosin (Tm; Sigma-Aldrich), and anti-myosin blood glucose of 693.5 6 61.5 mg/dL and controls had light chain (MLC) 1 (Clone 1LC-14, Spectral Diagnos- 123.9 6 9.5 mg/dL. tics). Between different antibodies blots were stripped Isolated Skinned Fiber Studies for 1 h at 25°C in 200 mmol/L glycine (pH 2.5; Sigma- For skinned cardiac muscles studies, muscles were isolated Aldrich) (29). and mounted as previously described (25,26). Varied Ca2+ 2+ OGT and OGA Activity Assays concentrations [Ca ]o were achieved by mixing the relaxing solutionandactivatingsolutioninvariousratios.After OGT Assays 2+ reaching the highest [Ca ]o concentration, trabeculae were Heart homogenates were separated in cytosolic or myo- washed in relaxing solution and incubated at room temper- filament fractions, desalted, and subjected to OGT activity ature 1 h in 1 mg/mL CpOGA (27) diluted in relaxing solu- assays as described (30). Activity counts in disintegration tion.Afterward,anewCa2+ activation protocol was repeated. per minute (dpm) were normalized to total protein con- 2+ Steady-state force-[Ca ]o relationships were determined ex- tent (dpm/mg), and then background activity without CKII perimentally and fittoamodified Hill equation (25,26). peptide was subtracted (Fig. 7C). diabetes.diabetesjournals.org Ramirez-Correa and Associates 3575

OGAse Assays spectrometer through 1-mm emitter tip (New Objective) OGAse activity was determined as previously described at 2.2 kV. Survey scans (full MS) were acquired within (31). Briefly, activity was expressed as the amount of 350–1800 m/z with up to eight peptide masses (precursor catalyzing the release of 1 mmol/mg/min pNP ions) individually isolated at IW1.9Da, and fragmented from pNP-GlcNAc, and then background activity in the (MS/MS) using higher-energy collisional dissociation, at 35 presence of the most specific and potent OGA inhibitor activation collision energy. Precursor and the fragment (1mmol/L thiamet-G) was subtracted (Fig. 7C). ions were analyzed at resolution 30,000 and 15,000, re- Myofilament Isolation and Tandem Mass Tags spectively. Dynamic exclusion of 30 s, repeat count 1, “ ” Labeling monoisotopic ion precursor selection (MIPS) on , m/z “ ” “ ” Myofilament proteins were isolated as previously de- option off , lock mass on (silocsane 371 Da) were scribed (32). Protein concentration was determined used. MS/MS spectra were processed by PEAKS Studio by the Bradford assay and equal amounts of protein (Bioinformatics Solutions Inc.) using Rattus norvegicus (;200 mg) were reduced with 5 mmol/L dithiothreitol FASTA as the proteome database, with concatenated de- (DTT), alkylated with 15 mmol/L iodoacetamide, and coy database, specifying all peptide species, trypsin as digested by trypsin (trypsin-to-protein ratio, 1:50). Tryptic enzyme, missed cleavage 2, precursor mass tolerance peptides were labeled for quantitation with tandem mass 10 ppm, fragment mass tolerance 0.03 Da, and oxidation tags (TMT) 10plex labels (127, 128, 129, 130) following the (M), deamidation (NQ), carbamidomethyl (C), and TMT fi manufacturer’s guidelines (Thermo Fisher Scientific). The labels 127, 128, 129, and 130 as variable modi cations. fl labeled peptides were combined and fractioned of ine using TMT and O-GlcNAc Quantification XBridge HPLC column (Waters). The resulting 96 fractions Quantitation function of PEAKS Studio was used to – were combined into 24 fractions for liquid chromatography export the raw intensity values of TMT peptides with or tandem mass spectrometry (LC-MS/MS) runs, while only without O-GlcNAc enrichment. Only peptides with a pos- reserving 10% for the preenriched analysis, and combin- itive identification value or a cutoff value of 20 for pep- ing the remaining 90% for O-GlcNAc peptide enrichment tide score threshold (210 log P), quantification mass (postenrichment). tolerance (0.2 Da), and a 0.1% false discovery rate O-GlcNAc Peptide Enrichment (FDR) were considered. All Ser or Thr residues from The remaining pooled peptides were used to enrich O-GlcNAcylated peptides were identified by a +136 mass O-GlcNAcylated peptides by adapting a method described shift given by DTT during BEMAD and their MS/MS spec- previously (33,34), with some modifications. In brief, pep- tra manually inspected. The experimental design con- tides were treated with alkaline phosphatase (50 units; sisted of 12 biological samples divided into three New England Biolabs) and PNGase F (1,000 units; New experiments with four isobaric mass-tags, each experi- England Biolabs) for 6 h, followed by desalting with ment comprising a STZ diabetic and a control sample a C18 spin column (Nest Group). The dried peptides with and without CpOGA treatment. The TMT preenrich- were resuspended in a buffer containing 20 mmol/L ment samples were analyzed to factor in potential DTT, 20% (v/v) ethyl alcohol, and 1.5% triethanolamine changes in protein expression, the relative total protein (pH 12.5) and incubated at 50°C for 4 h with gentle shak- load, and to refine the comparisons among postenrich- ing. Reaction was quenched by the addition of trifluoro- ment samples by accounting for technical and experimen- acetic acid (final pH ;7.0). Peptides were desalted and tal variation. The median signal value for each modified then incubated with thiol-Sepharose beads (Sigma-Aldrich) peptide of each sample was first determined and con- in PBS containing 1 mmol/L EDTA (PBS/EDTA, pH 7.4) for verted to log2 notation (with 0.0 values excluded as nulls) 4 h. After five washes in a PBS/EDTA buffer supplemented for further processing. Data were then quantile normal- with acetonitrile 40% (v/v), beads were incubated in PBS/ ized to achieve the same median. The difference between EDTA containing 20 mmol/L DTT for 30 min. Released prenormalized and normalized data provided a number peptides were collected and desalted with a C18 spin col- that was further used as a correction factor for the post- umn. Dried peptides were then analyzed by LC-MS/MS for enrichment peptide analysis. For the O-GlcNAc–enriched O-GlcNAc site mapping and quantification. peptides, signals were treated as above, except when con- verted to log2 notation, the 0.0 values were log-converted LC-MS/MS Analysis to 0.001 to be able to express ratios. The unenriched and O-GlcNAc–enriched fractions were analyzed with an LTQ-Orbitrap Velos (Thermo Fisher Sci- Detection of Differential O-GlcNAcylation entific) attached to a Nano Acquity (Waters) chromatog- We used a statistical generalized linear model approach raphy system. Peptides were loaded on a 75-mm 3 2.5-cm for differential gene expression detection. Briefly, a mixed- C18(YMC*GELODS-A12-nmS–10-mm) trap at 600 nL/min effects linear model was fit for each individual modified 0.1% formic acid (solvent A) and fractionated at 300 nL/min peptide to estimate O-GlcNAcylation differences between on a 75-mm 3 150-mm reverse-phase column using the groups of samples being compared (i.e., control vs. STZ a2–90% acetonitrile in 0.1% formic acid gradient over diabetic, control vs. CpOGA control, STZ vs. CpOGA STZ). 90 min. Eluting peptides were sprayed into the mass When distinct peptides for the same O-GlcNAcylated site 2+ 3576 OGA Restores Cardiac Muscle ECa 50 in Diabetes Diabetes Volume 64, October 2015 were available, correlation coefficients were computed and in diabetic skinned fibers (Fig. 1B). Remarkably, the 2+ fi the associated consensus correlation was added to the ECa 50 of CpOGA-treated diabetic skinned bers became model (35). An empirical Bayes approach was applied to similar to that found in control muscles, regardless of moderate standard errors of log2 O-GlcNAcylation fold- CpOGA presence (Fig. 1B). Upon CpOGA administration, 2+ change, as previously described (36). Finally, for each ana- no difference in maximal Ca activated force (Fmax) and lyzed feature, moderated t statistics, log-odds ratios of Hill coefficient (n) was evident between control and diabetic differential expression (B statistics), and raw and adjusted muscles (Fig. 1B). Next, control and diabetic (STZ) rat heart P values (FDR control by the Benjamini and Hochberg myofilaments were compared before and after being method) were obtained. All analyses were performed using incubated with recombinant CpOGA and analyzed for software packages available from the R/Bioconductor for O-GlcNAcylation by MS/MS tagging and LC-MS/MS. In to- statistical computing “limma” (36). tal we found 63 O-GlcNAcylated sites, 39 in myosin heavy a Statistics chain(MHC),9in -sarcomeric actin, 2 in MLC 1, 5 in a-Tm 1, 7 in cardiac I (cTnI), and 1 in myosin The Student t test, one-way ANOVA, and two-way ANOVA binding protein C (Table 1). with repeated measures, followed by post hoc pairwise Because CpOGA removes O-GlcNAc in both control and comparison, when appropriate, was used for statistical diabetic (STZ) rat muscles but only restores ECa2+ in analysis of the data. A value of P , 0.05 was considered 50 to indicate significant differences between groups. Unless diabetic muscles, we focused our proteomic analysis on identifying the site-specific O-GlcNAc changes that are sta- otherwise indicated, data are expressed as mean 6 SEM. tistically significant on CpOGA-treated STZ diabetic versus RESULTS CpOGA-untreated STZ diabetic myofilaments but not Removal of Site-Specific O-GlcNAcylation Excess significant on CpOGA-treated control versus CpOGA- Restores Myofilament Ca2+ Sensitivity in Diabetic untreated control myofilaments (Fig. 1C). Surprisingly, Cardiac Muscle most of the sites that change significantly are located on Using an MS approach, we have previously identified MHC (S740, S844, S1414, S1465, S1471, S1472, S1598, specific O-GlcNAcylation sites on five major cardiac myo- T1601, S1602, S1778, and S1917), a-sarcomeric actin (actin filament proteins on normal hearts (14). We also showed S54, T326), and a-TmS87. Noteworthy, the sites that that incubation of skinned cardiac muscles with GlcNAc are significantly more O-GlcNAcylated in diabetic hearts reduces myofilament Ca2+ sensitivity, thus reproducing (i.e., MHC S844, S1471, S1472, T1601, and S1917 and a hallmark of diabetic cardiac muscle (7,14). Yet, site- actin T326) (Fig. 1C, red rectangles) are also part of the group specific O-GlcNAcylationchangesondiabeticheartsremain of sites that change significantly upon CpOGA treatment of unknown. Nor is it clear whether removing endogenous STZ diabetic myofilaments. Thus, abnormal O-GlcNAcylation O-GlcNAc from diabetic skinned cardiac muscles is suffi- ofcardiacmuscleproteinsissufficient to reduce myofilament cient to restore myofilament function. To address these Ca2+ sensitivity, and its removal is necessary to prevent questions, we first confirmed that O-GlcNAcylation is en- this Ca2+ desensitization. hanced in protein extracts from STZ diabetic hearts (Fig. 1A). Next, we used CpOGA to remove O-GlcNAc from diabetic Diabetic Cardiac Muscle Displays Sarcomeric O-GlcNAc skinned cardiac muscles (Fig. 1B). Finally, we used a re- Signal Increased and Delocalized From Z-Lines fined global quantitative proteomic technique that com- The a-cardiac actin O-GlcNAc signal intensity (by immu- bines TMT labeling with b-elimination/Michael addition noblot) is augmented in animal models of diabetes (14). to quantify the O-GlcNAcylation changes in diabetic Moreover, in normal human myocardium, O-GlcNAc pre- hearts and to identify the key amino acid residues where dominantly modifies a Z-line protein called ZASP (Z-band O-GlcNAc is reduced after CpOGA in STZ diabetic hearts alternatively spliced PDZ motif protein) (38). ZASP but not in controls (Fig. 1C). O-GlcNAcylation increases further in heart failure and OGA is the mammalian enzyme that removes O-GlcNAc hypertrophic cardiomyopathy (38). However, whether in vivo; however, its glycosidase activity is greatly reduced O-GlcNAc–specific subcellular localization changes in di- when expressed as a recombinant protein (31). To circum- abetes is not clear. To fill this gap, cryosections from vent this issue, we used a bacterial glycosidase, CpOGA (27) normal and STZ-induced diabetic rat hearts were fixed (Clostridium perfringens N-acetyl-glycosidase J, provided by in cold acetone and stained for O-GlcNAc (CTD 110.6) Dr. Daan van Aalen, University of Dundee). CpOGA is to determine the relative abundance and localization of highly homologous to human OGA and displays potent O-GlcNAc (Fig. 2A). Consistent with previous findings, and specific activity toward mammalian protein homoge- O-GlcNAc signal intensity was markedly increased in nates (37). Hence, skinned cardiac muscles from diabetic STZ rat diabetic myocardium compared with control (STZ-induced) and control rats were incubated with hearts (Fig. 2B). Also, immunoelectron microscopy CpOGA and their contractile properties determined before revealed that in control myocardium, O-GlcNAc was pre- and after this treatment. Steady-state force-[Ca2+]relation- dominantly localized in clusters near the Z-lines, whereas shipsmeasurementsrevealedthat CpOGA removal of abnor- in STZ-treated rats O-GlcNAc clusters were spread toward mal O-GlcNAcylation restores myofilament Ca2+ sensitivity the A-band (Fig. 2C). Next, we performed a morphometric diabetes.diabetesjournals.org Ramirez-Correa and Associates 3577

Figure 1—Removing abnormal O-GlcNAcylation restores myofilament Ca2+ sensitivity in diabetic cardiac muscle. A: O-GlcNAc Western 2+ blots (WB) demonstrate increased O-GlcNAcylation on STZ diabetic heart homogenates. B: Steady-state force-[Ca ]o relationship in skinned diabetic and control cardiac STZ diabetic muscles pre-CpOGA (n = 5) and STZ post-CpOGA (n = 4), control pre-CpOGA 2+ (n = 6) and control post-CpOGA (n = 5). Comparison of ECa 50 for all groups (lower panel). Maximal force (Fmax) and Hill coefficient (n). C: Proteomic identification and quantification of site-specific O-GlcNAc changes that are statistically significant (generalized linear model approach) on CpOGA treatment of STZ diabetic myofilaments but not on CpOGA treatment of control myofilaments. Highlighted in red rectangles are the sites that are significantly more O-GlcNAcylated in diabetic hearts. The horizontal line in the middle of each box indicates the median; the top and bottom borders of the box mark the 75th and 25th percentiles, respectively, the whiskers mark the 90th and 10th percentiles, and the c indicate outliers. Statistical analysis used moderated t-statistics, log-odds ratios of differential expression (B statistics), and raw and adjusted P values (FDR control by the Benjamini and Hochberg method). Con, control.

analysis of nine random fields of control or STZ rat di- intensity was markedly increased in the human diabetic abetic myocardium using immune-gold particles (12 nm specimen (Fig. 2E). diameter) to assess O-GlcNAc distance from the nearest Z-line (Fig. 2D). This approach allowed us to conclude that OGT and OGA Sarcomeric Distribution Is Inverted in overall O-GlcNAc signal is increased in the of Diabetic Myocardium diabetic rats and moves away from Z-lines. To test whether Next, we reasoned that the presence of enhanced similar alterations pertain to human pathology, O-GlcNAc O-GlcNAcylation in specific compartments of immune-fluorescence was investigated in one sample from diabetic hearts could be due to enhanced activity of OGT, a normal (donor) and one from a patient with type 2 di- reduced OGA activity, or abnormal localization of enzymes. abetes. Consistent with findings in the rat, O-GlcNAc signal To address this issue, we used anti-OGT (AL-28) or 2+ 3578 OGA Restores Cardiac Muscle ECa 50 in Diabetes Diabetes Volume 64, October 2015

Table 1—O-GlcNAcylated peptides and their corresponding amino acid sites in normal and diabetic cardiac myofilaments MHC or myosin heavy chain 6, P02563 Actin, a-cardiac, P68035 Peptide ID AA Peptide ID AA R.*TNCFVPDDKEEYVK.A T35 R.AVFP*SIVGR.P S34 K.V*TAETENGK.T T60 K.D*SYVGDEAQSK.R S54 R.ENQ*SILI*TGESGAGK.T S173/T177 K.DSYVGDEAQ*SK.R S62 Q.SILI*TGE*SGAGK.T T177/S180 R.GY*SFVTTAER.E S201 F.A*SIAAIGDR.S S197 A.TAA*SSSSLEK.S S234 K.TVRNDN*S*SR.F S241/S242 K.*SYELPDGQVITIGNER.F S241 Y.A*S ADTGDSGK.G S627 K.EITALAP*S*T MK.I S325/T326 N.PAAIPEGQFID*S R.K S740 K.QEYDEAGP*SIVHR.K S370 K.*SAETEK.E S844 Myosin regulatory light chain 1, P16409 R.IEDEQALG*SQLQK.K S1102 Peptide ID AA R.*S DLTR.E S1139 K.EAEFDA*SK.I S 45 R.ELEEI*SER.L S1149 R.ALGQNP*TQAEVL.R T 93 R.SVNDL*T*SQR.A T1274/S1275 Tropomyosin a 1 K.LQTENGEL*S R.Q S1288 Peptide ID AA K.EALI*SQL*TR.G S1301/T1304 K.ATDAEADVA*SLNR.R S87 K.AN*SEVAQWR.T S1368 K.AADE*SER.G S123 K.C*S*SLEKTK.H S1414/S1415 K.VIE*SR.A S132 R.*SNAAAAALDK.K S1437 R.AEL*SEGK.C S186 K.YEE*SQ*S ELESSQK.E S1465/S1467 K.*SLEAQAEK.Y S206 E.SQSELE*SSQK.E S1469 Cardiac , P23693 E.SQSELE*S*SQK.E S1471/S1472 Peptide ID AA R.VVD*SLQ*T*S LDAETR.S S1598/T1601/S1602 E.*SSDSAGEPQPAPAPVR.R S5 Q.*TSLDAETR.S T1601 S.*SD*SAGEPQPAPAPVR.R S6/S8 R.IA*SEAQK.H S1638 K.ISA*SR.K S43/S45 R.AVVEQ*TER.S T1697 R.VL*STR.C S78 K.LAEQELIE*T*SER.V T1711/S1712 K.E*SLDLR.A S167 K.EQD*T*SAHLER.M T1777/S1778 Cardiac myosin binding protein C, P23693 R.NAE*SVK.G S1838 Peptide ID AA R.ADIAE*SQVNK.L S1917 R.DG*SDIAANDK.Y S47

anti-OGA (345) antibodies and an immunofluorescence frozen slides were triple-antigen stained by immunofluo- and confocal microscopy approach to analyze the signal in- rescence for OGT (AL-28), OGA (345), and a-actinin. As tensity and colocalization of both enzymes. OGT exhibited expected, human control and diabetic myocardium a predominantly sarcomeric localization, whereas OGA dis- exhibited a differential sarcomeric pattern for OGT and played both a sarcomeric and reticular pattern (Fig. 3A and OGA signals. A selected area from control (Fig. 4A, top B). Similar to O-GlcNAc, OGT tended to colocalize mostly panel) or diabetic (Fig. 4A, bottom panel) is shown as with a-actinin near the Z-line (Fig. 3C and D). Although a white square in the merged left picture, followed by detectable at the Z-line too, OGA was instead predomi- enlarged areas showing the regions in single or combined nantly distributed throughout the A-band along the entire channels for linear surface profile plots of approximately sarcomeric unit. To further consolidate this evidence, we six to eight sarcomere units (dashed line). Combined sig- used the colocalization tools of Zen 9 Image Analysis Soft- nals of OGT (magenta) or OGA (red) and a-actinin (green) ware (Leica Zeiss), enabling us to quantify changes in OGT for control and diabetic myocardium indicated that OGT and OGA distribution. Representative images of normal and labeling peaks in control myocardium correspond mostly STZ diabetic hearts were used to generate XY pixel dot plots with the a-actinin signal. Conversely, OGA labeling of laser scanning signals. A series of colocalization quanti- showed variable signals, peaking both at A-band and fication parameters are displayed as bar graphs for OGT and Z-line regions (Fig. 4A, top panel). In diabetic myocardium, a-actinin (Fig. 3E) and for OGT and OGA (Fig. 3F). Analysis however, OGT signal peaks tended to shift away from the of at least three random fields from four replicates on three a-actinin signal, whereas OGA labeling centered more on different hearts showed clear differences between normal the a-actinin signal (Fig. 4A, bottom panel). Because cry- and STZ diabetic myocardium. OGT colocalization with osections fixed with acetone might reflect tissue architec- a-actinin was decreased in STZ myocardium (Fig. 3E1–3), ture alterations, we decided to corroborate the OGT and thus confirming a redistribution apart from the Z-line (P = OGA staining pattern on isolated skinned myocytes 0.0053). Interestingly, OGT/OGA colocalization was also obtained from flash-frozen myocardium derived from reduced in STZ diabetic myocardium (P # 0.05, Fig. 3F1–3). rat samples as described (28). Similar to rat or human a-Actinin, OGT, and OGA immunofluorescence was also eval- myocardium cryosections, isolated skinned myocytes uated in human heart samples (Fig. 4A). Acetone prefixed fixedwith4%formaldehydedisplayedtheexpected diabetes.diabetesjournals.org Ramirez-Correa and Associates 3579

Figure 2—Diabetic myocardium displays increased O-GlcNAcylation. A: Rat control and STZ diabetic myocardium cryosections were immunostained for O-GlcNAc and nuclei (DAPI). B:Quantification of O-GlcNAc immunofluorescence, channel intensity normalized to area (mm2), shows a significant increase in STZ diabetic myocardium. *P < 0.05. C: Immunoelectron microscopy for O-GlcNAc shows signal redistribution from mainly the Z-line (arrowheads) toward the A-band. D: Relation of immunogold particles and distance from the Z-line. E: Human donor and diabetic myocardium cryosections display O-GlcNAc sarcomeric distribution and its increase in the diabetic heart.

OGT and OGA staining pattern in control (Fig. 4B,top cycling rates, thus likely perturbing O-GlcNAc stoichiometry panel) or STZ diabetic (Fig. 4B, bottom panel) myo- in specific sarcomere compartments. cytes. When sarcomeres in control hearts are compared Next, we analyzed OGT and OGA by double immuno- with those found in diabetic hearts, these changes in the electron microscopy in hearts from STZ diabetic and control OGT staining pattern mirror faithfully those found in rat rats. Immunoelectron microscopy data fully corroborated STZ diabetic myocardium (Fig. 3), with an inverted pattern the evidence obtained with the immunofluorescence ap- of OGT and OGA sarcomeric distribution. This inverted pat- proach, confirming the redistribution of both enzymes tern of OGT/OGA localization may influence O-GlcNAcylation within the sarcomere. Indeed, similar to O-GlcNAc, OGT 2+ 3580 OGA Restores Cardiac Muscle ECa 50 in Diabetes Diabetes Volume 64, October 2015

Figure 3—OGT and OGA mislocalization in STZ diabetic myocardium. A and B: Rat control and STZ diabetic myocardium cryosections were immunostained for a-actinin (green), OGT (magenta), and OGA (red). C and D: Regions of interest (white square and dashed line) were used to plot the signals of OGT and a-actinin; the red arrows show Z-line. E: OGT and a-actinin colocalization parameters show significant differences in weighted coefficient (1), overlap coefficient (2), and Pearson correlation R (3). *P = 0.0053. F: OGT and OGA colocalization parameters show significant differences in overlap coefficient (1), Pearson correlation R (2), and correlation R2 (3). *P < 0.05. Images were acquired with a Zeiss 710 LSM Meta confocal microscope. NIH ImageJ and Origin 8.0 software were used to plot OGT or OGA with a-actinin as relative intensity profiles.

immunogold-labeled particles were mainly located at the panel). Finally, we determined the frequency of OGT and Z-disk in normal hearts (Fig. 5A,toppanel),whereasthey OGA immunogold-labeled particles, examining 9–10 ran- were more diffused along the A-band in diabetic hearts (Fig. dom fields of normal or diabetic myocardium and quanti- 5B, top panel). Instead of being localized mainly at the fying OGT (12 nm, purple circle) and OGA (6 nm, red circle) A-band (normal hearts), OGA formed sizable clusters in immunogold-labeled particles confined to the myofilament the vicinity of the Z-disk (diabetic hearts; Fig. 5B, bottom apparatus. Our approach revealed that both O-GlcNAc diabetes.diabetesjournals.org Ramirez-Correa and Associates 3581

Figure 4—OGT and OGA distribution profile in human and rat control and diabetic hearts. A: Human donor and diabetic myocardium cryosections were immunostained for a-actinin (green), OGT (magenta), and OGA (red). Regions of interest (white square and line) were chosen to plot OGT, OGA, and a-actinin signals. Signals from donor or diabetic myocardium are displayed as double y-axisgraphsto illustrate the inversed distribution profile of OGT/a-actinin and OGA/a-actinin, respectively. B: Rat control and diabetic skinned myocytes were isolated from flash-frozen myocardium, fixed in 4% ultrapure formaldehyde, and immunostained as above. Regions of interest and OGT, OGA, and a-actinin signals were plotted as above and confirmed the abnormal distribution profile of OGT and OGA in diabetic hearts. cycling enzymes were detected at higher frequency in diabetic OGT and OGA are physically associated with myofilament myocardium (Fig. 5C and D). Taken together, these data proteins and determining possible changes in their myofil- suggest that, in analogy to cardiac kinases and phosphatases ament abundance imparted by diabetes. Fresh whole-heart (39), mislocalized OGT and OGA activities can affect the homogenates from control or diabetic (STZ-treated) rats function of contractile or regulatory proteins, or both. were immunoprecipitated with anti-OGT (AL-28) or anti- OGA (345) antibodies (29), resolved by SDS-PAGE, and Differential Interactions of OGT and OGA With analyzed by Western blots against several myofilament Myofilament Proteins in the Diabetic Heart proteins. OGT and OGA were both associated with a-cardiac Assessing OGT by Western blot in myofilament prepara- actin, a-Tm,andMLC1innormalhearts(Fig.5E and F). tions (14) involves the use of high (1%) concentrations of A representative coimmunoprecipitation of OGT and OGA Triton X-100 (32). This procedure may preclude the possi- with interacting myofilament proteins is provided in Fig. 5E bility of detecting weak interactions between OGT, OGA, and F. In the diabetic hearts, the OGA-immunoprecipitated and their potential transient binding partner proteins, thus interactions with a-cardiac actin, a-Tm, and MLC 1 were underestimating the extent of OGT and OGA abundance in increased several fold compared with the control interac- the myofilaments. To properly evaluate these interactions, tions (Fig. 5H)(n =5vs.n =4,P , 0.05), whereas inter- we used a coimmunoprecipitation approach, testing whether actions with OGT were normal. On the other hand, 2+ 3582 OGA Restores Cardiac Muscle ECa 50 in Diabetes Diabetes Volume 64, October 2015

Figure 5—Differential OGT and OGA subcellular localization and myofilament interactions in control and diabetic myocardium. Representative transmission electron microscopy images of control (A) and STZ diabetic rat myocardium (B). Ultrathin sections were examined with immunoelectron microscopy with primary antibodies (anti-OGT AL-28 and anti-OGA), and secondary antibodies gold-labeled (anti-rabbit, 12 nm; anti-chicken, 6 nm). Z-line, green arrowhead; OGT, purple circles; OGA, red circles; OGT and OGA in close vicinity, pink circles; OGT, purple arrowhead; and OGA, red arrowhead. Quantification of OGT (C)andOGA(D) number of particles/field in nine fields shows an increase for OGT (2 6 0.6 vs. 8.6 6 2.6, *P # 0.028) and OGA (9.4 6 2.9 vs 37.6 6 12.6, *P # 0.05) immunoelectron microscopy in STZ diabetic myocardium. Images were analyzed in ImageJ software (NIH). Representative coimmunoprecipitations (co-IP) of OGT (E) and OGA (F), followed by Western blots (WB) for a-sarcomeric actin, Tm, and MLC 1. A fraction of the inputs from controls (C1, C2), STZ (S1, S2), and agarose beads with no antibody (M) or isotype-specific normal antibody + agarose beads (1°) were included. G: Analysis of integrated signal density of myofilament immunoreactivity normalized to total IP OGT displayed a trend toward increased interactions with Tm and MLC 1. H: OGA co-IP shows that diabetic STZ rats display several fold increased associations with a-actin, a-Tm, and MLC 1. *P # 0.05.

although the OGT-immunoprecipitated interactions with OGT and OGA Activity Are Altered in Diabetic Hearts myofilaments tended to have increased interactions in di- In diabetic hearts, overall O-GlcNAclevelshavebeen abetic samples, they did not reach statistical significance. reported to change with or without reciprocal changes in The heterogeneity of the signals for OGT and OGA coim- OGT expression and uridine diphosphate (UDP)–GlcNAc munoprecipitations on the control heart homogenates levels (40). Protein specific O-GlcNAcylation, however, can might be related to the feeding status (ad libitum) of the change in either direction (40). With this paradox in mind, animals or to a potential cyclic nature of these protein-to- we hypothesized that the observed changes in OGT/OGA protein interactions. Hence, OGA interactions with myofil- sarcomeric distribution and myofilament protein interac- ament proteins are more abundant in diabetic hearts, and tions might influence their enzymatic activity. To test they are likely contributing to reduce contractility. this, fresh whole hearts from control or STZ diabetic rats diabetes.diabetesjournals.org Ramirez-Correa and Associates 3583 were homogenized in extraction buffer containing 1% Tri- vs. 1.14 6 0.29 for controls, n = 3 each) (Fig. 6C). OGA ton X-100. Total tissue homogenates were separated in cy- activity was assessed by the release of p-nitrophenol tosolic and myofilament fractions by centrifugation. For (pNP) from pNP-GlcNAc, a synthetic substrate. In addition, OGT assays, both cytosolic and myofilament fractions our approach accounted for OGA activity background more were desalted into OGT desalting buffer (for activity assays) strictly because we subtracted OGA activity detected in the or stored at 280°C for SDS-PAGE and Western blots. After presence of thiamet-G, the most potent and specificOGA OGT expression was normalized to cTnI content, neither inhibitor available (41) (Fig. 7C). OGA-specificactivitywas subcellular fraction showed significant differences between significantly increased in homogenates from STZ diabetic control (n =3)andSTZdiabetic(n = 3) hearts (myofilament hearts compared with controls (Fig. 6F). Thus, changes in and cytosolic fractions are shown in Fig. 6A and B,respec- OGT/OGA protein-to-protein interactions and subcellular tively). OGT activity was assayed by the incorporation of localization may influence OGT/OGA enzymatic activity. UDP–[H3]-labeled GlcNAc into a synthetic peptide (CKII peptide); disintegration per minute counts were normalized DISCUSSION to micrograms of protein, and then OGT activity without This study establishes that excessive O-GlcNAcylation, in a spe- peptide was subtracted (31) (Fig. 7C). This approach cificsubsetofmyofilament sites of MHC, a-sarcomeric actin, revealed a significant reduction in myofilament-associated and Tm, is sufficient to produce myofilament dysfunction in OGT activity in STZ diabetic hearts (Fig. 6D,obtainedin diabetic cardiomyopathy, and it does so by affecting Ca2+ sen- myofilament fractions). Conversely, OGT activity in the cy- sitivity. These perturbations in myofilament O-GlcNAcylation tosolic fractions of STZ diabetic hearts was not significantly result from OGT and OGA displacement within the sarcomere different (Fig. 6E). rather than from variations in enzymatic activity per se, result- For OGA assays, myofilament fractions retained traces ing in an overall perturbed O-GlcNAc cycling. Important, from of detergent; therefore, they were not suitable for a translational point of view, is the finding that similar alter- measurement of enzymatic activity. However, total tissue ations occur in experimental and human diabetic hearts. homogenates became properly usable after protein pre- Depressed myofilament Ca2+ sensitivity is a hallmark cipitation with 30–50% ammonium sulfate and desalting of myofilament dysfunction in diabetic cardiomyopathy into OGA assay buffer. After the latter procedure, OGA and heart failure (7,11). Alterations in myofilament phos- expression normalized to cTnI content was not signifi- phorylation are now recognized as important negative cantly different between groups (0.54 6 0.02 for STZ modulators of cardiac function, along with perturbations

Figure 6—Subcellular expression and activity of OGT and OGA in diabetic hearts. Analysis of OGT expression in myofilament fractions (A) and in cytosolic fractions (B) showed no significant differences between STZ diabetic and control hearts. C: OGA expression tended to be lower in myocardium total homogenates from STZ diabetic hearts but did not reach statistical significance. D: OGT activity showed a significant decrease in myofilament fractions from STZ diabetic hearts. *P < 0.05. E: OGT activity assays in cytosolic fractions showed no significant difference. F: OGA assays for specific activity in total heart homogenates showed a significant increase in STZ diabetic hearts. For all quantifications, bar graphs represent the mean of n =36 SEM. *P < 0.05. 2+ 3584 OGA Restores Cardiac Muscle ECa 50 in Diabetes Diabetes Volume 64, October 2015

Figure 7—Model of O-GlcNAc role in diabetic cardiomyopathy pathogenesis. A: General characteristics of OGT and OGA. B: OGT/OGA localization in normal and diabetic sarcomeres; functional effect of O-GlcNAc-specific removal by OGA-like CpOGA. C: Graphical repre- sentation of OGT activity toward CKII peptide or endogenous targets and OGA total or specific activity when normalized to residual activity in the presence of thiamet-G (TMG).

in Ca2+ handling (39,42–44). That alterations in cardiac sensitivity (14). Moreover, adenoviral-based or inducible protein O-GlcNAcylation are present in models of type 1 transgenic overexpression of human OGA, for which and type 2 diabetes (2,4,24,40,45) and cardiac disease CpOGA is a very close homolog, is known to ameliorate (21,38) is well established; however, no studies have contractile and energetic deficits associated with diabetic addressed yet the functional effect of cardiac myofila- cardiomyopathy (4,45). Here, we add important novel ments O-GlcNAcylation during the course of diabetic car- evidence that in diabetic myofilaments there are selective diac dysfunction (2–4). We have previously identified proteins, such as MHC, a-sarcomeric cardiac actin (actin), numerous sites within the cardiac myofilament are and cardiac a-Tm, and some of their specific sites such as O-GlcNAcylated under normal conditions (14). We also MHC S844, S1471, S1472, T1601 and S1917, actin T326, showed that incubation of normal skinned cardiac and TmS87, that indeed have excessive O-GlcNAcylation. muscles with GlcNAc reduced myofilament Ca2+ More importantly, we demonstrated that specific removal diabetes.diabetesjournals.org Ramirez-Correa and Associates 3585 of excessive O-GlcNAcylation, from a subset of myofilaments and phosphorylation signaling cascades have extensive cross- sites via CpOGA (Fig. 1C) rapidly restores myofilament Ca2+ talk (12,56), both at the level of site occupancy and at the sensitivity, thus correcting myofilament dysfunction in STZ level of modifying enzymes, further dissecting this phenom- diabetic skinned muscles. The advantage of our approach enon in the heart would be important. (i.e., the use of recombinant CpOGA on skinned cardiac Previously, OGT expression (40,57,58) and activity muscles) resides in the ability of assessing the functional (59) in diabetic hearts have been found to be normal, consequences of O-GlcNAcylation removal from myofila- whereas OGA activity is reduced, with or without recip- ments independently from other major adverse effects even- rocal changes in expression (57–59). Here, we analyzed tually imposed by O-GlcNAcylation on Ca2+ handling and myofilament and cytosolic fractions separately. We found mitochondrial key proteins (2,4,15,24,45–48). Thus, here that OGT expression in diabetic hearts is not different we cement the view that excessive O-GlcNAcylation is between diabetic and control hearts. Owing to OGA in- a novel, negative modulator of myofilament function in hibition by traces of Triton X-100, OGA expression and the diabetic myocardium. At the same time, our data sug- activity were analyzed only in total homogenates in which gest, for the first time, that removing excess of O-GlcNAc OGA expression was not different between groups. In from diabetic myofilaments specific sites mimics what de- addition, our study accounts for OGT and OGA activity phosphorylation does on Ca2+ sensitivity in experimental background in a more rigorous way. Indeed, OGT activity (49) and human heart failure (50–53). without CKII peptide and OGA activity with thiamet-G Cardiac myofilament phosphorylation and function is (a specific OGAse inhibitor) were subtracted. Our myofila- regulated by multiple kinases and phosphatases (54,55). ment versus cytosolic compared data showed that OGT They strategically dock at the Z-line and modify their enzymatic activity is reduced in the myofilament in di- activity and/or localization in response to mechanical abetic hearts but not in the cytosolic fraction. In con- forces and neurohumoral stimuli (54,55). The displace- trast, OGA activity in total homogenates is significantly ment of these enzymatic activities may have important increased in diabetic hearts (Fig. 6F). This apparent par- functional repercussions on the contractile apparatus adoxical decrease of OGT activity in diabetic hearts could (39). For instance, at baseline, protein phosphatase 2A beexplainedbythewayweaccountedfortheback- forms a complex with protein phosphatase 2B and p38– ground activity, subcellular fractionation, or by changes mitogen-activated protein kinase, but it moves away from in OGT substrate specificity in response to metabolic the Z-line upon b-adrenergic stimulation (55). Whether cues (30). Increased OGA activity may reflect compen- similar effects can be attained by changes in the sarco- sation for excessive O-GlcNAcylation. Supporting this meric localization of OGT and OGA is currently unknown. possibility is previous evidence showing that increased Here we report that, in normal hearts, OGT and OGA erythrocyte protein O-GlcNAcylation and OGA activity distribute mainly at the Z-line and A-band, respectively. are present in samples from patients with prediabetes However, in diabetic hearts, OGT drifts away from the and with type 2 diabetes (60). Z-line, whereas OGA appears even more clustered at this Cardiac contractility is regulated on a beat-to-beat basis site. Importantly, these changes are similarly evident in by the Ca2+-dependent modulation of myosin cross-bridge rat and human diabetic myocardium. It is already well binding on actin by the Tm-troponin complex. Cross-bridge known that OGT localization and catalytic activity cycling occurs at the A-band (61), whereas most of the changes in response to metabolic cues (12,56). OGT and signaling arises from the Z-line (54,55,62). Thus, retarget- OGA often occur in transient protein-to-protein complexes ing of OGT and OGA may increase a-cardiac actin containing kinases and phosphatases (12,56). These O-GlcNAcylation at the A-band, potentially interfering include kinases and phosphatases such as AMPK with actin-Tm modulation of cross-bridge cycling (Fig. 7B). (56), Ca2+/-dependent protein kinase IV, p38– Altogether, these data suggest that in diabetic hearts, OGT mitogen-activated protein kinase, protein phosphatase 1, and OGA displacement rather than their catalytic activity is and myosin phosphatase targeting 1, which are key regu- the key in modulating excessive O-GlcNAcylation. lators of cardiac metabolism and function (12). Although Therearelimitationsinthecurrentstudy.First,more these protein interactions have not been confirmed in work is needed to define the functional effect of myofila- the heart, a recent report elegantly shows that CaMKII is ment O-GlcNAcylation at the site-specific level, along O-GlcNAcylated during acute hyperglycemia and diabetes, with the interplay with phosphorylation. Second, future leading to chronic activation, which contributes to diabetes studies should address whether excessive a-cardiac actin cardiac mechanical dysfunction and arrhythmias (24). Pres- O-GlcNAcylation indeed perturbs cross-bridge cycling ki- ent data show that, in the heart, OGT and OGA form netics, actin-Tm interactions, or the actin rate of poly- complexes with a-cardiac actin, a-Tm, and MLC 1, and merization.Finally,itisnotknownwhetherO-GlcNAc that these protein-to-protein complexes are enhanced for modifications or phosphorylation of specificsubstratesor OGA in diabetes. These alterations are potentially relevant in OGT/OGA contribute to their altered pattern of localiza- to the pathogenesis of diabetic systolic and diastolic dysfunc- tion. Answering this question would define more in detail tion, given that transient protein-to-protein complexes regu- the effect of this PTM on the function of diabetic hearts; late OGT substrate specificity (12). Because O-GlcNAcylation however, it requires fully dedicated, separate studies. 2+ 3586 OGA Restores Cardiac Muscle ECa 50 in Diabetes Diabetes Volume 64, October 2015

In conclusion, the present work demonstrates that 2. Clark RJ, McDonough PM, Swanson E, et al. Diabetes and the accompa- abnormal O-GlcNAcylation is sufficient to cause myofila- nying hyperglycemia impairs cardiomyocyte cycling through increased ment functional deficit in diabetic cardiomyopathy. In- nuclear O-GlcNAcylation. J Biol Chem 2003;278:44230–44237 creased OGA interactions with sarcomeric proteins 3. Davidoff AJ, Ren J. Low insulin and high glucose induce abnormal re- laxation in cultured adult rat ventricular myocytes. Am J Physiol 1997;272:H159– (a-cardiac actin, a-Tm, and MLC 1) are likely central to H167 these alterations, showing that normal sarcomeric OGT fi 4. Hu Y, Belke D, Suarez J, et al. Adenovirus-mediated overexpression of and OGA subcellular localization is lost in myo laments O-GlcNAcase improves contractile function in the diabetic heart. Circ Res 2005; from diabetic hearts. This abnormal redistribution of 96:1006–1013 O-GlcNAc cycling enzymes is similarly present in experi- 5. Zhang L, Cannell MB, Phillips AR, Cooper GJ, Ward ML. Altered calcium mental and human diabetic cardiomyopathy. On these homeostasis does not explain the contractile deficit of diabetic cardiomyopathy. grounds, here we propose that, in addition to direct Diabetes 2008;57:2158–2166 targeting of abnormal site-specific phosphorylation of 6. Flagg TP, Cazorla O, Remedi MS, et al. Ca2+-independent alterations in myofilament regulatory subunits, removing abnormal diastolic sarcomere length and relaxation kinetics in a mouse model of lipotoxic site-specific O-GlcNAcylation in myofilaments and/or pre- diabetic cardiomyopathy. Circ Res 2009;104:95–103 venting changes in O-GlcNAc cycling should be considered 7. Malhotra A, Penpargkul S, Fein FS, Sonnenblick EH, Scheuer J. The effect of as another promising new therapeutic avenue for treating streptozotocin-induced diabetes in rats on cardiac contractile proteins. Circ Res 1981;49:1243–1250 diabetic cardiomyopathy. 8. Malhotra A, Sanghi V. Regulation of contractile proteins in diabetic heart. Cardiovasc Res 1997;34:34–40 Acknowledgments. For expert technical assistance, the authors thank 9. Akella AB, Ding XL, Cheng R, Gulati J. Diminished Ca2+ sensitivity of John Robinson from Pediatric Cardiology and the JHMI Microscope Facility, and skinned cardiac muscle contractility coincident with -band shifts in the – Marina Allary (Johns Hopkins School of Medicine) for helpful suggestions. diabetic rat. Circ Res 1995;76:600 606 Funding. This work has been supported by the National Institutes of Health 10. Hofmann PA, Menon V, Gannaway KF. Effects of diabetes on isometric (NIH)/National Heart, Lung, and Blood Institute (NHLBI) Proteomics Center Con- tension as a function of [Ca2+] and pH in rat skinned cardiac myocytes. Am J – tract HHSN268201000032C (A.M.M., G.W.H.), R01-DK-61671 (G.W.H.); Program Physiol 1995;269:H1656 H1663 fi of Excellence in Glycobiology NHLBI P01-HL-107153 (G.W.H., N.P.), R01-HL- 11. Jweied EE, McKinney RD, Walker LA, et al. Depressed cardiac myo lament 091923 (N.P.); American Heart Association (AHA) and the Lawrence and Florence function in human diabetes mellitus. Am J Physiol Heart Circ Physiol 2005;289: – A. DeGeorge Charitable Trust Scientist Developing Grant (AHA-12SDG9140008) H2478 H2483 for G.A.R.-C., AHA (AHA-0855439E) for W.D.G.; and Institutional Development 12. Hart GW, Slawson C, Ramirez-Correa G, Lagerlof O. Cross talk between Award (IDeA) from National Institute of General Medical Sciences/NIH (P20-GM- O-GlcNAcylation and phosphorylation: roles in signaling, transcription, and – 12345 and R01-DK100595) for C.S. In addition, this study was supported by the chronic disease. Annu Rev Biochem 2011;80:825 858 Johns Hopkins Institute for Clinical and Translational Research, which is funded 13. Hedou J, Cieniewski-Bernard C, Leroy Y, Michalski JC, Mounier Y, Bastide in part by Grant Number UL1-TR-001079 from National Center for Advancing B. O-linked N-acetylglucosaminylation is involved in the Ca2+ activation prop- – Translational Sciences/NIH for L.M. erties of rat . J Biol Chem 2007;282:10360 10369 fi Duality of Interest. Under a licensing agreement between Johns 14. Ramirez-Correa GA, Jin W, Wang Z, et al. O-linked GlcNAc modi cation of fi Hopkins University and several companies, including Covance Research cardiac myo lament proteins: a novel regulator of myocardial contractile function. – Products,Sigma-Aldrich,andSantaCruzBiotechnology,G.W.H.receives Circ Res 2008;103:1354 1358 royalties from the sale of the CTD110.6 O-GlcNAc antibody. The terms are 15. Makino A, Suarez J, Gawlowski T, et al. Regulation of mitochondrial mor- managed by Johns Hopkins University in accordance with conflict of interest phology and function by O-GlcNAcylation in neonatal cardiac myocytes. Am J – policies. No other potential conflicts of interest relevant to this article were Physiol Regul Integr Comp Physiol 2011;300:R1296 R1302 reported. 16. Fülöp N, Zhang Z, Marchase RB, Chatham JC. Glucosamine cardioprotection Author Contributions. G.A.R.-C. designed and performed research, an- in perfused rat hearts associated with increased O-linked N-acetylglucosamine fi alyzed data, and wrote the manuscript. J.M. and C.S. designed and performed protein modi cation and altered p38 activation. Am J Physiol Heart Circ Physiol – research, analyzed data, and edited and reviewed the manuscript. Q.Z. performed 2007;292:H2227 H2236 research, analyzed data, and edited and reviewed the manuscript. N.S.L.-F., 17. Jones SP, Zachara NE, Ngoh GA, et al. Cardioprotection by N- – M.X., V.C., K.C., L.D., and R.N.C. performed research and analyzed data. X.S. acetylglucosamine linkage to cellular proteins. Circulation 2008;117:1172 1182 b performed research. L.M., N.P., and G.W.H. assisted with experimental discus- 18. Watson LJ, Facundo HT, Ngoh GA, et al. O-linked -N-acetylglucosamine sion and with critical evaluation and editing of the manuscript. W.D.G. analyzed transferase is indispensable in the failing heart. Proc Natl Acad Sci U S A 2010; – data and edited and reviewed the manuscript. A.M.M. designed research, inter- 107:17797 17802 preted data, and assisted with experimental discussion and with critical evalu- 19. Laczy B, Hill BG, Wang K, et al. Protein O-GlcNAcylation: a new signaling ation and editing of the manuscript. G.A.R.-C. is the guarantor of this work and, paradigm for the cardiovascular system. Am J Physiol Heart Circ Physiol 2009; – as such, had full access to all data in the study and takes responsibility for the 296:H13 H28 fi integrity of the data and the accuracy of the data analysis. 20. Ngoh GA, Facundo HT, Za r A, Jones SP. O-GlcNAc signaling in the car- – Prior Presentation. Parts of this study were presented at the Basic diovascular system. Circ Res 2010;107:171 185 Cardiovascular Sciences 2011 Scientific Sessions of the American Heart Asso- 21. Lunde IG, Aronsen JM, Kvaløy H, et al. Cardiac O-GlcNAc signaling is in- – ciation, New Orleans, LA, 18–21 July 2011, and at the American Heart Associ- creased in hypertrophy and heart failure. Physiol Genomics 2012;44:162 172 ation 2012 Scientific Sessions, Los Angeles, CA, 3–7 November 2012. 22. Fein FS, Strobeck JE, Malhotra A, Scheuer J, Sonnenblick EH. Reversibility of diabetic cardiomyopathy with insulin in rats. Circ Res 1981;49:1251–1261 References 23. Malhotra A, Lopez MC, Nakouzi A. Troponin subunits contribute to altered 1. Boudina S, Abel ED. Diabetic cardiomyopathy revisited. Circulation 2007; myosin ATPase activity in diabetic cardiomyopathy. Mol Biochem 1995;151: 115:3213–3223 165–172 diabetes.diabetesjournals.org Ramirez-Correa and Associates 3587

24. Erickson JR, Pereira L, Wang L, et al. Diabetic hyperglycaemia activates 44. Takimoto E, Soergel DG, Janssen PM, Stull LB, Kass DA, Murphy AM. CaMKII and arrhythmias by O-linked glycosylation. Nature 2013;502:372–376 Frequency- and afterload-dependent cardiac modulation in vivo by troponin I with 25. Gao WD, Backx PH, Azan-Backx M, Marban E. Myofilament Ca2+ sensitivity constitutively active protein kinase A phosphorylation sites. Circ Res 2004;94: in intact versus skinned rat ventricular muscle. Circ Res 1994;74:408–415 496–504 26. Gao WD, Murray CI, Tian Y, et al. Nitroxyl-mediated disulfide bond formation 45. Fricovsky ES, Suarez J, Ihm SH, et al. Excess protein O-GlcNAcylation and between cardiac myofilament cysteines enhances contractile function. Circ Res the progression of diabetic cardiomyopathy. Am J Physiol Regul Integr Comp 2012;111:1002–1011 Physiol 2012;303:R689–R699 27. Rao FV, Dorfmueller HC, Villa F, Allwood M, Eggleston IM, van Aalten DM. 46. Ren J, Davidoff AJ. Diabetes rapidly induces contractile dysfunctions in Structural insights into the mechanism and inhibition of eukaryotic O-GlcNAc isolated ventricular myocytes. Am J Physiol 1997;272:H148–H158 hydrolysis. EMBO J 2006;25:1569–1578 47. Ren J, Gintant GA, Miller RE, Davidoff AJ. High extracellular glucose impairs 28. Kirk JA, Holewinski RJ, Kooij V, et al. Cardiac resynchronization sensitizes the cardiac E-C coupling in a glycosylation-dependent manner. Am J Physiol 1997; sarcomere to calcium by reactivating GSK-3b. J Clin Invest 2014;124:129–138 273:H2876–H2883 29. Slawson C, Lakshmanan T, Knapp S, Hart GW. A mitotic GlcNAcylation/ 48. Zhu-Mauldin X, Marsh SA, Zou L, Marchase RB, Chatham JC. Modification phosphorylation signaling complex alters the posttranslational state of the cy- of STIM1 by O-linked N-acetylglucosamine (O-GlcNAc) attenuates store- toskeletal protein . Mol Biol Cell 2008;19:4130–4140 operated calcium entry in neonatal cardiomyocytes. J Biol Chem 2012;287: 30. Cheung WD, Hart GW. AMP-activated protein kinase and p38 MAPK activate 39094–39106 O-GlcNAcylation of neuronal proteins during glucose deprivation. J Biol Chem 49. Belin RJ, Sumandea MP, Allen EJ, et al. Augmented protein kinase C-alpha- 2008;283:13009–13020 induced myofilament protein phosphorylation contributes to myofilament dys- 31. Zachara NE, Cole RN, Hart GW, Gao Y. Detection and analysis of proteins function in experimental congestive heart failure. Circ Res 2007;101:195–204 modified by O-linked N-acetylglucosamine. Curr Protoc Protein Sci 2001;Chapter 50. Kooij V, Saes M, Jaquet K, et al. Effect of troponin I Ser23/24 phosphor- 12:Unit 12.18 ylation on Ca2+-sensitivity in human myocardium depends on the phosphory- 32. Murphy AM, Solaro RJ. Developmental difference in the stimulation of cardiac lation background. J Mol Cell Cardiol 2010;48:954–963 myofibrillar Mg2(+)-ATPase activity by calmidazolium. Pediatr Res 1990;28:46–49 51. Kooij V, Zhang P, Piersma SR, et al. PKCa-specific phosphorylation of the 33. Wang Z, Park K, Comer F, Hsieh-Wilson LC, Saudek CD, Hart GW. Site- troponin complex in human myocardium: a functional and proteomics analysis. specific GlcNAcylation of human erythrocyte proteins: potential biomarker(s) for PLoS One 2013;8:e74847 diabetes. Diabetes 2009;58:309–317 52. Wijnker PJ, Boknik P, Gergs U, et al. Protein phosphatase 2A affects 34. Wells L, Vosseller K, Cole RN, Cronshaw JM, Matunis MJ, Hart GW. Mapping myofilament contractility in non-failing but not in failing human myocardium. J sites of O-GlcNAc modification using affinity tags for serine and threonine post- Muscle Res Cell Motil 2011;32:221–233 translational modifications. Mol Cell Proteomics 2002;1:791–804 53. Wijnker PJ, Foster DB, Tsao AL, et al. Impact of site-specific phosphory- 35. Smyth GK, Michaud J, Scott HS. Use of within-array replicate spots for lation of protein kinase A sites Ser23 and Ser24 of cardiac troponin I in human assessing differential expression in microarray experiments. Bioinformatics 2005; cardiomyocytes. Am J Physiol Heart Circ Physiol 2013;304:H260–H268 21:2067–2075 54. Solaro RJ. Multiplex kinase signaling modifies cardiac function at the level 36. Smyth GK. Linear models and empirical bayes methods for assessing dif- of sarcomeric proteins. J Biol Chem 2008;283:26829–26833 ferential expression in microarray experiments. Stat Appl Genet Mol Biol 2004;3: 55. Solaro RJ. Sarcomere control mechanisms and the dynamics of the cardiac Article3 cycle. J Biomed Biotechnol 2010;2010:105648 37. Schimpl M, Borodkin VS, Gray LJ, van Aalten DM. Synergy of peptide and 56. Bullen JW, Balsbaugh JL, Chanda D, et al. Cross-talk between two essential sugar in O-GlcNAcase substrate recognition. Chem Biol 2012;19:173–178 nutrient-sensitive enzymes: O-GlcNAc transferase (OGT) and AMP-activated 38. Leung MC, Hitchen PG, Ward DG, Messer AE, Marston SB. Z-band protein kinase (AMPK). J Biol Chem 2014;289:10592–10606 alternatively spliced PDZ motif protein (ZASP) is the major O-linked b-N- 57. Bennett CE, Johnsen VL, Shearer J, Belke DD. Exercise training mitigates acetylglucosamine-substituted protein in human heart myofibrils. J Biol Chem aberrant cardiac protein O-GlcNAcylation in streptozotocin-induced diabetic mice. 2013;288:4891–4898 Life Sci 2013;92:657–663 39. Zhang P, Kirk JA, Ji W, et al. Multiple reaction monitoring to identify site- 58. Cox EJ, Marsh SA. Exercise and diabetes have opposite effects on the specific troponin I phosphorylated residues in the failing human heart. Circulation assembly and O-GlcNAc modification of the mSin3A/HDAC1/2 complex in the 2012;126:1828–1837 heart. Cardiovasc Diabetol 2013;12:101 40. Fülöp N, Mason MM, Dutta K, et al. Impact of Type 2 diabetes and aging on 59. Jensen RV, Zachara NE, Nielsen PH, Kimose HH, Kristiansen SB, Bøtker HE. cardiomyocyte function and O-linked N-acetylglucosamine levels in the heart. Am Impact of O-GlcNAc on cardioprotection by remote ischaemic preconditioning in J Physiol Cell Physiol 2007;292:C1370–C1378 non-diabetic and diabetic patients. Cardiovasc Res 2013;97:369–378 41. Yuzwa SA, Macauley MS, Heinonen JE, et al. A potent mechanism-inspired 60. Park K, Saudek CD, Hart GW. Increased expression of beta-N- O-GlcNAcase inhibitor that blocks phosphorylation of tau in vivo. Nat Chem Biol acetylglucosaminidase in erythrocytes from individuals with pre-diabetes and 2008;4:483–490 diabetes. Diabetes 2010;59:1845–1850 42. Bilchick KC, Duncan JG, Ravi R, et al. Heart failure-associated alterations in 61. Ibanez-Garcia D, Requejo-Isidro J, Webb MR, West TG, French P, troponin I phosphorylation impair ventricular relaxation-afterload and force- Ferenczi MA. Fluorescence lifetime imaging reveals that the environment of frequency responses and systolic function. Am J Physiol Heart Circ Physiol 2007; the ATP binding site of myosin in muscle senses force. Biophys J 2010;99: 292:H318–H325 2163–2169 43. Ramirez-Correa GA, Cortassa S, Stanley B, Gao WD, Murphy AM. Calcium 62. Hoshijima M. Mechanical stress-strain sensors embedded in cardiac cy- sensitivity, force frequency relationship and cardiac troponin I: critical role of PKA toskeleton: Z disk, , and associated structures. Am J Physiol Heart Circ and PKC phosphorylation sites. J Mol Cell Cardiol 2010;48:943–953 Physiol 2006;290:H1313–H1325