Enhanced dispersion of repolarization (DR) was proposed as a unifying mechanism, central to arrhythmia genesis in the long QT (LQT) syndrome. In mammalian hearts, K(+) channels are heterogeneously expressed across the ventricles resulting in 'intrinsic' DR that may worsen in long QT. DR was shown to be central to the arrhythmia phenotype of transgenic mice with LQT caused by loss of function of the dominant mouse K(+) currents. Here, we investigated the arrhythmia phenotype of mice with targeted deletions of KCNE1 and KCNH2 genes which encode for minK/IsK and Merg1 (mouse homolog of human ERG) proteins resulting in loss of function of I(Ks) and I(Kr), respectively. Both currents are important human K(+) currents associated with LQT5 and LQT2. Loss of minK, a protein subunit that interacts with KvLQT1, results in a marked reduction of I(Ks) giving rise to the Jervell and Lange-Nielsen syndrome and the reduced KCNH2 gene reduces MERG and I(Kr). Hearts were perfused, stained with di-4-ANEPPS and optically mapped to compare action potential durations (APDs) and arrhythmia phenotype in homozygous minK (minK(-/-)) and heterozygous Merg1 (Merg(+/-)) deletions and littermate control mice. MinK(-/-) mice has similar APDs and no arrhythmias (n = 4). Merg(+/-) mice had prolonged APDs (from 20 +/- 6 to 32 +/- 9 ms at the base, p < 0.01; from 18 +/- 5 to 25 +/- 9 ms at the apex, p < 0.01; n = 8), longer refractory periods (RP) (36 +/- 14 to 63 +/- 27 at the base, p < 0.01 and 34 +/- 5 to 53 +/- 21 ms at the apex, p < 0.03; n = 8), higher DR 10.4 +/- 4.1 vs. 14 +/- 2.3 ms, p < 0.02) and similar conduction velocities (n = 8). Programmed stimulation exposed a higher propensity to VT in Merg(+/-) mice (60% vs. 10%). A comparison of mouse models of LQT based on K(+) channel mutations important to human and mouse repolarization emphasizes DR as a major determinant of arrhythmia vulnerability.