Purpose: Diabetes can impact on cardiovascular health by causing a distinct condition termed "diabetic cardiomyopathy". Its characteristic left ventricular (LV) diastolic dysfunction has been associated with interstitial fibrosis, reduced NO availability, and abnormal calcium handling. However, the early triggers and the underlying cellular mechanisms remain unknown. Here, we investigate changes in vascular and myocardial reactive oxygen species (ROS) and NO availability in a murine model of type 1 diabetes, and evaluate potential beneficial effects of inducing a myocardial-specific increase in the NOS cofactor tetrahydrobiopterin (BH4) on the development of LV dysfunction. Methods: Diabetes was induced in male mice by daily intraperitoneal streptozotocin (STZ) injection (43mg/kg, 5 consecutive days). To augment myocardial BH4 and increase NOS activity, transgenic mice were generated with myocardial-specific overexpression of the rate-limiting enzyme for BH4 synthesis, GTP cyclohydrolase 1 (mGCH1 Tg). Vascular function in isolated aortas was evaluated by isometric tension studies (myograph), NOS activity and biopterins by HPLC, and superoxide production by lucigenin-enhanced chemiluminescence. High-resolution echocardiography was used to assess LV function. Results: After 12 weeks of diabetes, WT and mGCH1 Tg mice showed impaired aortic endothelium-dependent vasodilatation, in association with increased superoxide production and reduced BH4 bioavailability (n=6-10 per group). By contrast, diabetic LV homogenates showed no increase in superoxide generation or reduced BH4:BH2 ratio and no reduction in NOS activity (n=9-12 per group). Nevertheless, in vivo echocardiography revealed significant LV diastolic dysfunction in WT diabetic mice, which was prevented in mGCH1 Tg mice (E'/A' diabetic vs control: 1.52±0.08 vs 1.53±0.08 in mGCH1 Tg; 0.89±0.07 vs 1.35±0.06 in WT, n=9 per group, P<0.01 for the interaction between genotype and diabetes). In line with these results, isolated LV myocytes from WT diabetic mice displayed prolonged relaxation, which was prevented in diabetic mGCH1 Tg (t50 relaxation, diabetic vs control: 105.3±2.8 vs 95.6±2.4 in WT and 77.3±3 vs 73.3±2 in mGCH1 Tg; n=25 cells from 2-5 hearts per group, P<0.05 for the interaction between genotype and diabetes). Conclusions: Impaired LV diastolic function in diabetic mice can be prevented by myocardial GCH1 overexpression in the absence of NOS dysfunction or increased oxidative stress, suggesting that GCH1/BH4 protect the diabetic myocardium by mechanisms other than redressing the local nitroso-redox balance.