Arrhythmogenic action potential alternans (APD-ALT) is thought to arise from beat to beat alteration in cellular Ca(2+) cycling. Previously, we found that spatial heterogeneity in APD-ALT between ventricular myocytes is key to the mechanism linking APD-ALT to cardiac arrhythmogenesis. However, the cellular and molecular basis for APD-ALT is poorly understood. To test the hypothesis that spatial heterogeneities in expression and function of calcium cycling proteins underlies heterogeneities in APD-ALT, endocardial and epicardial myocytes were isolated from left ventricular free wall of 20 guinea pig hearts. APD-ALT and Ca(2+) transient alternans (Ca-ALT) were measured simultaneously as stimulus rate was increased progressively. Endocardial myocytes exhibited greater susceptibility to cellular alternans than epicardial myocytes as evidenced by a significantly lower pacing rate threshold for APD-ALT (113 +/ -9 bpm vs. 151 +/- 8 bpm, respectively, P < 0.05) and for Ca-ALT (110 +/- 8 bpm vs. 149 +/- 8 bpm, respectively, P < 0.05). APD-ALT never occurred without Ca-ALT, whereas Ca-ALT was readily induced in the absence of APD-ALT by repetitive constant action potential waveform, suggesting that Ca-ALT was not secondary to APD-ALT. Importantly, there were significant voltage-independent differences in Ca(2+) cycling between endocardial and epicardial myocytes as evidenced by weaker Ca(2+) release (32% lower Ca(2+) amplitude, and 16% longer rise time), and slower Ca(2+) reuptake (24% larger Ca(2+) decay time constant, and 9% longer Ca(2+) transient duration) in endocardial compared to epicardial myocytes. Taken together these data indicate that myocytes that are most susceptible to APD-ALT exhibit impaired Ca(2+) release and reuptake. Moreover, transmural differences in Ca(2+) cycling function was associated with significantly reduced endocardial expression of ryanodine release channel (by 22%) and SERCA2 (by 40%), suggesting a potential molecular basis for spatially heterogeneous APD-ALT. Moreover, transmural differences in expression and function of key SR Ca(2+) cycling proteins may underlie spatial heterogeneity of APD-ALT that has been closely linked to cardiac arrhythmogenesis.