Reduced O-GlcNAcylation diminishes cardiomyocyte Ca²⁺ dependent facilitation and frequency dependent acceleration of relaxation

Highlights

• Ca²⁺ dependent facilitation (CDF) and frequency dependent acceleration of relaxation (FDAR) contribute to cardiomyocyte Ca²⁺ handling.

• Genetic and pharmacologic inhibition of intracellular O-linked glycosylation (O-GlcNAcylation) reduce cardiomyocyte CDF and FDAR.

• The enzyme responsible for O-GlcNAcylation (OGT) interacts with proteins responsible for CDF and FDAR at the SR and/or dyad space.

• Chronic inhibition of O-GlcNAcylation nearly abolishes CaMKII autophosphorylation despite increased CaMKIIδ expression.

• OGT contributes to the basal regulation of CDF and FDAR likely by regulating CaMKII activity.

Abstract

Ca²⁺ dependent facilitation (CDF) and frequency dependent acceleration of relaxation (FDAR) are regulatory mechanisms that potentiate cardiomyocyte Ca²⁺ channel function and increase the rate of Ca²⁺ sequestration following a Ca²⁺-release event, respectively, when depolarization frequency increases. CDF and FDAR likely evolved to maintain EC coupling at increased heart rates. Ca²⁺/calmodulin-dependent kinase II (CaMKII) was shown to be indispensable to both; however, the mechanisms remain to be completely elucidated. CaMKII activity can be modulated by post-translational modifications but if and how these modifications impact CDF and FDAR is unknown. Intracellular O-linked glycosylation (O-GlcNAcylation) is a post-translational modification that acts as a signaling molecule and metabolic sensor. In hyperglycemic conditions, CaMKII was shown to be O-GlcNAcylated resulting in pathologic activity. Here we sought to investigate whether O-GlcNAcylation impacts CDF and FDAR through modulation of CaMKII activity in a pseudo-physiologic setting. Using voltage-clamp and Ca²⁺ photometry we show that cardiomyocyte CDF and FDAR are significantly diminished in conditions of reduced O-GlcNAcylation. Immunoblot showed that CaMKIIδ and calmodulin expression are increased but the autophosphorylation of CaMKIIδ and the muscle cell-specific CaMKIIβ isoform are reduced by 75% or more when O-GlcNAcylation is inhibited. We also show that the enzyme responsible for O-GlcNAcylation (OGT) can likely be localized in the dyad space and/or at the cardiac sarcoplasmic reticulum and is precipitated by calmodulin in a Ca²⁺ dependent manner. These findings will have important implications for our understanding of how CaMKII and OGT interact to impact cardiomyocyte EC coupling in normal physiologic settings as well as in disease states where CaMKII and OGT may be aberrantly regulated.

Introduction

Cardiomyocyte L-type voltage gated Ca²⁺ channels (Ca_vs) control the influx of extracellular Ca²⁺ (I_Ca) and are therefore the main trigger for excitation-contraction (EC) coupling in the heart [1]. In addition to their basal properties, Ca_vs can be dynamically regulated so that the heart can respond to changes in physiologic demand [2,3]. Ca²⁺ dependent facilitation (CDF) is a positive feedback regulatory mechanism that results in increased Ca²⁺ entry through Ca_vs and/or slower Ca_v inactivation when depolarization frequency increases [3]. CDF likely evolved as a response to the negative feedback process of Ca²⁺-dependent inactivation (CDI) that accelerates Ca_v inactivation when [Ca]_i rises to protect the cell from Ca²⁺ overload [4]. Thus, the physiologic role of CDF is likely to maintain EC coupling when heart rate increases. Working in concert with CDF is frequency dependent acceleration of relaxation (FDAR) whereby the sequestration of [Ca]_i following a Ca²⁺ release event is accelerated when stimulation rate increases [5,6].

The overwhelming evidence indicates that CDF and FDAR are attributed to the activity of the Ca²⁺ and calmodulin dependent kinase, CaMKII. This was largely gleaned from studies showing abolishment and/or significant reductions in CDF and FDAR following pharmacologic inhibition of CaMKII [[6], [7], [8], [9], [10], [11]]. Our current understanding of how CaMKII mediates CDF is by direct phosphorylation of Ca_vs likely at Ser1512(1517) and/or Ser1570 of the α subunit [12,13]. This alters the gating properties of individual Ca_vs resulting in longer and more frequent openings. Despite these findings, much remains unknown. For example, in a mouse model where both putative CDF-related targets of CaMKII (Ser1512 and Ser1570 of mouse Ca_vα) were mutated to Ala, CDF was reduced but not abolished [12]. Additionally, while there are several CaMKII isoforms found in the heart, CaMKIIδ is the predominant isoform [14,15]; however, genetic ablation of CaMKIIδ also reduced but did not abolish CDF [16]. These results raise the questions that other mechanisms and multiple CaMKII isoforms may contribute to CDF and FDAR. Despite these discrepancies, in a mouse model where CaMKII inhibition was targeted to the sarcoplasmic reticulum (SR) of cardiomyocytes, CDF was abolished and FDAR was markedly reduced [5] indicating that localization to the SR and/or the dyad space is critical to regulating CDF and FDAR. In addition to these unresolved issues regarding the mechanisms by which CaMKII contributes to CDF and FDAR and what isoforms may be responsible, it is also largely unknown whether other signaling pathways that can impact CaMKII activity [17] may also contribute to CDF and FDAR.

CaMKII can be post-translationally modified by oxidation, S-nitrosylation, autophosphorylation and, discovered only recently, intracellular O-linked glycosylation (O-GlcNAcylation) [18,19]. These modifications typically cause CaMKII to be trapped in an autonomous state that result in increased activity [18,19]. If and how these modifications impact CDF and FDAR in vivo is not known. O-GlcNAcylation is characterized by the dynamic shuttling of N-acetylglucosamines (GlcNAc) to intracellular serine and threonine residues, akin to phosphorylation [20]. O-GlcNAc transferase (OGT) and O-GlcNAcase (OGA) are the only two enzymes known to add (OGT) or remove (OGA) O-GlcNAc residues [21]. In addition to functioning as a signaling molecule, OGT is considered a nutrient-sensor because production of UDP-GlcNAc, the substrate of OGT, requires processing of glucose or other metabolites through the hexosamine biosynthesis pathway [22]. This metabolite-sensing property of OGT is likely why aberrant O-GlcNAcylation is implicated in diseases like diabetes mellitus and heart failure, where metabolic substrate utilization is altered [[23], [24], [25]]. While not as well understood as phosphorylation, the complexity and scope of O-GlcNAc signaling in the heart is emerging at an accelerating rate [26]. By creating a cardiomyocyte-specific OGT-null mouse strain (OGTKO), we recently showed for the first time that OGT is a critical and direct regulator of multiple aspects of Ca_v function [27]. Cardiomyocytes from OGTKO mice showed marked reductions in I_Ca density and post-transcriptional expression, depolarizing shifts in voltage-dependent gating and increased efficacy of adrenergic stimulation [27]. Heart and cardiomyocyte EC coupling were also consistently reduced in the OGTKO [27]. While these findings were a critical first step in uncovering a new role for O-GlcNAc signaling in the heart, they are likely only a primer given the extensive cross-talk O-GlcNAcylation displays with other signaling molecules that regulate Ca_vs and cardiomyocyte EC coupling such as CaMKII.

Work by the Bers group indicated that CaMKII O-GlcNAcylation occurs in conditions of hyperglycemia and that the resulting O-GlcNAcylated CaMKII demonstrates increased and, in most cases, pathologic activity [[28], [29], [30], [31]]. This activity is like the activity incurred by increased CaMKII autophosphorylation. Their data also suggest that these two post-translational modifications occur independently of each other [28,29]. This spurred us to question whether OGT and CaMKII demonstrate a broader interaction that could be observed in a pseudo-physiologic setting. Thus, in the present study, we tested whether OGT inhibition could impact CDF and FDAR in the absence of any other pathophysiologic stimuli such as hyperglycemia. Our data indicate that in conditions of chronic and acute OGT inhibition, CDF and FDAR are both significantly depressed. We also observed that chronic reductions in O-GlcNAcylation result in increased CaMKIIδ and calmodulin expression but nearly abolished CaMKII autophosphorylation. The data indicate that the relationship between OGT and CaMKII is more complex than previously thought and can occur in non-hyperglycemic conditions. These findings will have important ramifications for our understanding of how CaMKII and OGT interact to impact cardiomyocyte EC coupling in normal physiologic settings as well as in other disease states where CaMKII and OGT may be aberrantly regulated.