O-linked GlcNAc Modification of Cardiac Myofilament Proteins: A Novel Regulator of Myocardial Contractile Function

Abstract

In addition to O-Phosphorylation, O-linked modifications of serine and threonine by β-N-acetyl-D-glucosamine (GlcNAc) may regulate muscle contractile function. This study assessed the potential role of O-GlcNAcylation in cardiac muscle contractile activation. To identify specific sites of O-GlcNAcylation in cardiac myofilament proteins, a recently developed methodology based on GalNAz-Biotin labeling followed by DTT replacement and LC-MS/MS site mapping was adopted. Thirty-two O-GlcNAcylated peptides from cardiac myofilaments were identified on cardiac myosin heavy chain, actin, myosin light chains, and troponin I. To assess the potential physiological role of the GlcNAc, force-[Ca²⁺] relationships were studied in skinned rat trabeculae. Exposure to GlcNAc significantly decreased calcium sensitivity (pCa50), whereas maximal force (F_max) and Hill coefficient (n) were not modified. Using a pan-specific O-GlcNAc antibody it was determined that acute exposure of myofilaments to GlcNAc induced a significant increase in actin O-GlcNAcylation. This study provides the first identification of O-GlcNAcylation sites in cardiac myofilament proteins and demonstrates their potential role in regulating myocardial contractile function.

INTRODUCTION

Diabetes mellitus is a risk factor for the development of heart failure ¹, and abnormal glucose metabolism may contribute directly to depressed cardiac function. Studies in humans and animal models of diabetes mellitus have demonstrated abnormal myofilament function ² and impaired E-C coupling ^{3, 4}, which may depress myocardial function. Post-translational modifications (PTM) of myofilament proteins regulate cardiac function and phosphorylation of myofilament proteins may result in functional abnormalities in heart failure ^5-7. In addition to O-linked phosphorylation of serine (Ser) and threonine (Thr) residues of proteins, dynamic O-linked β-N-acetyl-D-glucosamine (O-GlcNAc) modifications can also regulate protein structure and function ⁸ , and interplay between O-GlcNAcylation and O-phosphorylation may have an important role in cellular function ^{9, 10}. Although recent studies have suggested that O-linked modifications of Ser/Thr by GlcNAc could be involved in the regulation of myofilament Ca²⁺ activation properties in skeletal muscle ^11-13, cardiac myofilament proteins have not been examined for this modification. Conversely, recent studies on isolated myocytes have directly associated diabetic cardiac dysfunction with increased levels of O-GlcNAcylation of cardiac proteins (>50kD), albeit the specific proteins and residues modified remained unknown ⁹. Those studies prompted us to verify the hypothesis that modification of cardiac myofilament proteins by O-GlcNAc could also regulate cardiac contractile function.

METHODS

Mass spectrometry identification of O-GlcNAc modified proteins

To label the specific sites (further details in online supplement), GlcNAc modified peptides were labeled with GalNAz-Biotin, enriched by avidin chromatography and then DTT was used to replace the GlcNAc-GalNAz-Biotin by β-elimination and Michael addition (BEMAD) as previously described ¹⁴.

Isolated skinned fiber studies

For skinned fiber studies, cardiac trabeculae were isolated and mounted as previously described⁵.

Immunoblots

Myofilament proteins were isolated as previously described¹⁵, with minor modifications. To determine the global GlcNAc modifications of myofilament proteins, a pan-GlcNAc antibody (CTD 110.6, Covance) was used as previously described ¹⁶. To assess cardiac troponin I (cTnI) phosphorylation, a phospho-TnI (Ser23/Ser24) antibody (Cell Signaling, Danvers, MA) was used as previously described ⁵.

RESULTS & DISCUSSION

Myofilament proteins are modified by O-GlcNAc

With the enrichment and BEMAD experiments described, MS data demonstrated that at baseline cardiac myofilament proteins are O-GlcNAcylated at the specific amino acid residues noted in online table I. Thirty-two O-GlcNAcylated peptides from cardiac myofilaments were identified: 21 from cardiac myosin heavy chain, 6 from alpha-sarcomeric actin, 2 from myosin light chain 1, 1 from myosin light chain 2, and 1 from troponin I. As a control for specificity, in a parallel preparation treated extensively with β-N-acetyl-hexosaminidase to remove O-GlcNAc prior to enrichment, no modified sites were detected. Typical MS/MS spectra of O-GlcNAc modified peptides from TnI and actin are shown in Figure 1A and 1B. The DTT modification on Ser 54 of cardiac actin was confirmed by the observation of multiple matched ion pairs that contain the DTT mass (Figure 1A). We also identified O-GlcNAcylation at Ser 150 of Troponin I (Figure 1B). Interestingly, Ser 150 is also phosphorylated by p21-activated kinase 3 (PAK3), a modification that increases calcium sensitivity ¹⁷.

Figure 1. Representative MS/MS fragmentation spectra.

A, MS/MS spectrum of Ser 54 modified peptide from Actin. B, MS/MS spectrum of Ser 150 modified peptide from cardiac TnI.

This is the first report to define specific sites of O-linked GlcNAc of cardiac myofilament proteins. Although studies of skeletal muscle suggested GlcNAcylation of myosin heavy chain, actin and myosin light chains, they did not define specific sites ¹¹. Most of the newly identified O–GlcNAcylated sites in cardiac myofilaments were not previously described as phosphorylated, with the exception of cTnI Ser 150 and MLC2 at Ser 15. Interestingly, the O-GlcNAc targets in MLC1 at Thr 93/Thr164 are different from phosphorylation sites at Thr69 and Ser 200, previously found in pharmacological preconditioning ¹⁸.

GlcNAc desensitized myofilaments to calcium without altering phosphorylation of TnI Ser23/24

In order to characterize the effects of O-linked GlcNAc in cardiac muscle steady-state force-[Ca²⁺] relationships were studied in freshly skinned rat trabeculae. Baseline force-[Ca²⁺] relationships were established in individual trabeculae, thereafter the trabeculae were washed in relaxing solution and exposed to either 200 mM Glycerol (n=5), 200 mM GlcNAc (n=7) or 5 mM GlcNAc (n=3) in a relaxing solution. After 1h incubation at room temperature, force-[Ca²⁺] relationships were determined. As shown in Figure 2A, 2B and 2C, a significant desensitization of the force-calcium relationship was observed in GlcNAc but not glycerol exposed trabeculae (pCa 50 1.81 ± 0.13 μM for Control vs. 3.83 ± 0.44 μM for 200mM GlcNAc, n=7, p =0.001, pCa50 1.86 ± 0.51 μM for Control vs. 3.24 ± 0.44 μM for 5mM GlcNAc, n=3, p < 0.05), whereas maximal force (Fmax) and Hill coefficient (n) were not significantly modified. Notably, despite washing out GlcNAc and bathing for 1h with fresh relaxing solution, pCa50 did not return to basal levels (data not shown). These results indicate that the changes observed are not dependent on simple chemical artifacts or osmotic alterations reflected on myofilament lattice, the later usually sensitize rather than desensitize myofilaments¹⁹. Brief (∼5 to 10 minutes) skinning in relaxing solution containing 1% Triton X-100 allows permeation of membranous system to calcium but preserves endogenous O-GlcNAc transferase (OGT) activity. Skinned trabeculae showed immuno-reactivity to anti-OGT antibody and possessed metabolically active enzymes to incorporate [³H]-GlcNAc into proteins of a wide range of molecular weight extracted from freshly isolated trabeculae using UDP-[³H]-GlcNAc as substrate, as shown in online Figure 2A and 2B. In support of this notion, other endogenous enzymatic activities, i.e. pyruvate kinase, have been demonstrated in rat skeletal fibers skinned with Triton X-100 in comparable experimental conditions ²⁰.

Figure 2. GlcNAc desensitizes the myofilaments to calcium.

Steady-state Force-[Ca²⁺] relationship in skinned trabeculae, in control conditions and after incubation with 200 mM of, Glycerol A (n= 5), GlcNAc B (n=7) or 5mM of GlcNAc C (n= 3). Note the rightward shift after exposure to 200 or 5 mM GlcNAc. D, To quantify phospho-TnI, three trabecuale per group were pooled and immunobloted for phospho-TnI (Ser23/Ser24), then stripped and probed for sarcomeric alpha-actin. E, The intensity of phsophoTnI/ sarcomeric alpha-actin bands was calculated by NIH ImageJ. There was no significant difference in the phosphorylation status of TnI between GlcNac and Glycerol samples. F, Shows immuno-reactive bands to Anti-OGT antibody on whole trabeculae crude protein extracts at the expected size of ∼105 to 110 kDa. Lane 1, ∼20 μg of Trabecula 18, lane 2, ∼60 μg of Trabecula 19, lane 3, ∼20 μg of Trabecula 23, lane 4 empty, lane 5, ∼70 ng of recombinant O-GlcNAc transferase (EC-OGT).

It is well known that PKA phosphorylation of TnI Ser 23/24 results in desensitization of the myofilaments to calcium ²¹. Therefore we sought to exclude the possibility that GlcNAc exposure altered calcium sensitivity by modifying phosphorylation levels at cTnI Ser23/Ser24. Using an antibody specific for phosphorylation at Ser23/Ser24 we concluded that GlcNAc exposure had no effects on phosphorylation at Ser23/Ser24 of TnI (Figure 2D). It is plausible that a balance between O-GlcNAcylaton and O-Phosphorylation at cTnISer150 and MLC2 Ser15 could have an important role in regulating cardiac contractility; however that was not directly investigated in this work. Taken together, these results strongly suggest that excess of GlcNAc levels in myofilaments from cardiac trabeculae decreases cardiac myofilament sensitivity to calcium, through a mechanism that does not involve TnI phosphorylation at Ser 23/Ser24.

Exposure to GlcNAc Alters Cardiac Myofilament O-GlcNAcylation pattern

To further assess the effects of acute GlcNAc exposure, fresh cardiac myofilament preparations were examined using the pan-specific O-GlcNAc antibody CTD 110.6. We determined the pattern of myofilament GlcNAcylation at baseline (Figure 3A). Proteins of a wide range of molecular weight, from ∼15 to ∼250 kD were detected (Figure 3A, lane 1). Next, the effects of pre-incubating myofilament preparations in vitro with 0.2 M GlcNAc on GlcNAcylation pattern were studied (Figure 3A, lane 2). Interestingly, GlcNAc exposure was associated with an increase in actin GlcNAcylation levels (from 27.6±4.2% to 35.1±2.36%, n=4, p<0.05, Figure 3B). There were no significant differences in total signal level by densitometry after GlcNAc exposure. Treatment with PNGaseF (an enzyme that removes specifically N-linked glycans) did not significantly alter the staining pattern or content (data not shown), suggesting that most of the observed GlcNAcylation signal derived from O-linked GlcNAc. In order to explore the relevance of this phenomenon in relevant pathophysiologic models, GlcNAcylation pattern was examined in similar fashion on cardiac myofilament preparations from ob/ob mice and streptozotocin treated rats and their respective controls. Interestingly, ob/ob mice and STZ rats myofilaments showed an increase in actin GlcNAcylation levels (36.4±7.8% to 53.2±2.1%, n=3 vs n=3 *p<0.05 for ob/ob mice and 42.7±2% to 50.9±3.2%, n=3 vs n=4 *p<0.05 for STZ rats, online Figure 1). Current evidence shows that acute enhancement of O-GlcNAcylation of specific myocardial protein targets may have cardioprotective effects ²², however it is also plausible that chronic increase of O-GlcNAcylation may impair cardiac function in diabetic cardiomyopathy ^{9, 23}. These data suggest that O-GlcNAcylation of actin can be increased by GlcNAc exposure in vitro and that this phenomenon is also present in vivo in relevant models of diabetes mellitus. O-GlcNAcylation of actin and/or other myofilament proteins may be responsible for the decreased submaximal force development observed in diabetes mellitus, however increased ROS production and other factors may contribute ²⁴. The O-GlcNAc antibody may not be sensitive enough to detect subtle changes in GlyNAcation of other myofilament proteins.

Figure 3. Exposure to GlcNAc alters cardiac myofilament O-GlcNAcylation pattern.

Myofilament preparations were exposed to relaxing solution only (control) or containing 0.2 M GlcNAc for 1 h at room temperature. Densitometry analysis of the blots was performed using ImageJ Software. A, GlcNacylation pattern of Control and GlcNAc exposed myofilaments with respective densitometry plots (Top panels) and actin signals, as loading control (Bottom panels). B, The contribution of actin peak to total GlcNAcylation staining pattern, where actin peak is increased from 27.6±4.2% to 35.1±2.36% (n=4 *p<0.05).

In summary, this study has identified specific sites of O-GlcNAc modification of myofilament proteins and provides evidence that this PTM can regulate myofilament function. This also raises the possibility that there could be a dynamic interaction between O-GlcNAc and O-phosphate modification of MLC2 and TnI as is well described for other proteins ⁸. Further studies of how O-GlcNAcylation of myofilament proteins regulates cardiac contractility may reveal novel and useful therapeutic targets in heart failure, especially in diabetic cardiomyopathy.