1. Voltage-gated Na+ channels undergo two types of inactivation in response to depolarization. One type, fast inactivation, occurs with a time scale of milliseconds. The other, slow inactivation, occurs over seconds to minutes. In addition, these two processes appear to be distinct at the molecular level. However, the molecular mechanism of Na+ channel slow inactivation is unknown. 2. We used patch clamp techniques to study slow inactivation, activation and fast inactivation in alpha-subunit cDNA clones for wild-type human heart Na+ channels (hH1) and rat skeletal muscle Na+ channels (mu1) transiently expressed in human embryonic kidney (HEK) cells. Our experiments showed that the Na+ channel slow inactivation phenotype (development, steady state and recovery) differed dramatically between hH1 and mu1. Slow inactivation in mu1 had a faster onset, a steeper voltage dependence, and was more complete compared with hH1. In addition, recovery from slow inactivation was much slower for mu1 than for hH1. Activation and fast inactivation kinetics were also different in hH1 and mu1. In hH1, fast inactivation was slower and V values of activation and steady-state fast inactivation (hthorn ) were more negative than in mu1. 3. To better understand the molecular basis of Na+ channel slow inactivation, Na+ channel chimaeras were constructed with domains from hH1 and mu1. The slow inactivation phenotype in the chimaeras (domains denoted by subscripts) mu1(1)hH1(2,3,4), mu1(1,2)hH1(3,4) and mu1(1,2,3)hH1(4) was intermediate compared with that of wild-type. However, the chimaera mu1(1)hH1(2,3,4) was more like wild-type hH1, while the chimaeras mu1(1,2)hH1(3,4) and mu1(1,2,3)hH1(4) were more similar to wild-type mu1. In the chimaeras, activation resembled that of mu1, fast inactivation resembled that of hH1, and steady-state fast inactivation fell between that of hH1 and mu1. 4. The data demonstrate that all four domains can modulate the Na+ channel slow inactivation phenotype. However, domains D1 and D2 may play a more prominent role in determining Na+ channel slow inactivation phenotype than D3 and D4. The results also support previous conclusions that D3 and D4 (and the D3-D4 linker) play an important role in Na+ channel fast inactivation, and that activation may require non-equivalent contributions from all four domains.