Stereoselective drug-channel interactions may help to elucidate the molecular basis of voltage-gated potassium channel block by local anesthetic drugs. We studied the effects of the enantiomers of bupivacaine on a cloned human cardiac potassium channel (hKv1.5). This channel was stably expressed in a mouse Ltk- cell line and studied using the whole-cell configuration of the patch-clamp technique. Both enantiomers modified the time course of this delayed rectifier current. Exposure to 20 microM of either S(-)-bupivacaine or R(+)-bupivacaine did not modify the activation time constant of the current, but reduced the peak outward current and induced a subsequent exponential decline of current with time constants of 18.7 +/- 1.1 and 10.0 +/- 0.9 ms, respectively. Steady-state levels of block (assessed with 250-ms depolarizing pulses to +60 mV) averaged 30.8 +/- 2.5% (n = 6) and 79.5 +/- 3.2% (n = 6) (p < 0.001), for S(-)- and R(+)-bupivacaine, respectively. The concentration dependence of hKv1.5 inhibition revealed apparent KD values of 27.3 +/- 2.8 and 4.1 +/- 0.7 microM for S(-)-bupivacaine and R(+)-bupivacaine, respectively, with Hill coefficients close to unity, suggesting that binding of one enantiomer molecule per channel was sufficient to block potassium permeation. Analysis of the rate constants of association (k) and dissociation (l) yielded similar values for l (24.9 s-1 vs. 23.6 s-1 for S(-)- and R(+)-bupivacaine, respectively) but different association rate constants (1.0 x 10(6) vs. 4.7 x 10(6) M-1 s-1 for S(-)- and R(+)-bupivacaine, respectively). Block induced by either enantiomer displayed a shallow voltage dependence in the voltage range positive to 0 mV, i.e., where the channel is fully open, consistent with an equivalent electrical distance delta of 0.16 +/- 0.01. This suggested that at the binding site, both enantiomers of bupivacaine experienced 16% of the applied transmembrane electrical field, referenced to the inner surface. Both bupivacaine enantiomers reduced the tail current amplitude recorded on return to -40 mV and slowed their time course relative to control, resulting in a "crossover" phenomenon. These data indicate 1) the charged form of both bupivacaine enantiomers block the hKv1.5 channel after it opens, 2) binding occurs within the transmembrane electrical field, 3) unbinding is required before the channel can close, 4) block of hKv1.5 channels by bupivacaine is markedly stereoselective, with the R(+)-enantiomer being the more potent one, 5) this stereoselective block was associated with a 1.11 -kcal/mol difference in binding energy between both enantiomers, and 6) the stereoselectivity derives mainly from a difference in the association rate constants, suggesting that the S(-)-enantiomer is less likely to access the binding site in an optimal configuration.