Background: Worldwide, diabetes mellitus and heart failure represent frequent comorbidities with high socioeconomic impact and steadily growing incidence, calling for a better understanding of how diabetic metabolism promotes cardiac dysfunction. Paradoxically, some glucose-lowering drugs have been shown to worsen heart failure, raising the question of how glucose mediates protective versus detrimental cardiac signaling. Here, we identified a histone deacetylase 4 (HDAC4) subdomain as a molecular checkpoint of adaptive and maladaptive signaling in the diabetic heart. Methods: A conditional HDAC4 allele was used to delete HDAC4 specifically in cardiomyocytes (HDAC4-knockout). Mice were subjected to diabetes mellitus either by streptozotocin injections (type 1 diabetes mellitus model) or by crossing into mice carrying a leptin receptor mutation (db/db; type 2 diabetes mellitus model) and monitored for remodeling and cardiac function. Effects of glucose and the posttranslational modification by β-linked N -acetylglucosamine (O-GlcNAc) on HDAC4 were investigated in vivo and in vitro by biochemical and cellular assays. Results: We show that the cardio-protective N-terminal proteolytic fragment of HDAC4 is enhanced in vivo in patients with diabetes mellitus and mouse models, as well as in vitro under high-glucose and high–O-GlcNAc conditions. HDAC4-knockout mice develop heart failure in models of type 1 and type 2 diabetes mellitus, whereas wild-type mice do not develop clear signs of heart failure, indicating that HDAC4 protects the diabetic heart. Reexpression of the N-terminal fragment of HDAC4 prevents HDAC4-dependent diabetic cardiomyopathy. Mechanistically, the posttranslational modification of HDAC4 at serine (Ser)-642 by O-GlcNAcylation is an essential step for production of the N-terminal fragment of HDAC4, which was attenuated by Ca 2+ /calmodulin–dependent protein kinase II–mediated phosphorylation at Ser-632. Preventing O-GlcNAcylation at Ser-642 not only entirely precluded production of the N-terminal fragment of HDAC4 but also promoted Ca 2+ /calmodulin–dependent protein kinase II–mediated phosphorylation at Ser-632, pointing to a mutual posttranslational modification cross talk of (cardio-detrimental) phosphorylation at Ser-632 and (cardio-protective) O-GlcNAcylation at Ser-642. Conclusions: In this study, we found that O-GlcNAcylation of HDAC4 at Ser-642 is cardio-protective in diabetes mellitus and counteracts pathological Ca 2+ /calmodulin–dependent protein kinase II signaling. We introduce a molecular model explaining how diabetic metabolism possesses important cardio-protective features besides its known detrimental effects. A deeper understanding of the here-described posttranslational modification cross talk may lay the groundwork for the development of specific therapeutic concepts to treat heart failure in the context of diabetes mellitus.