Tes of the VSD. Also, the extent of the contact surface between S0 1379592 and TM2 below the first helical turns is not known. The paths of the three helices, TM2, S0, and S4 through the membrane would be revealed by the propensities of substituted Cys at their intracellular ends to form disulfide bonds with each other and with the other TM helices in a and in b1. Other regions of b1, such as the N-terminal and Cterminal tails [48?0] and the extracellular loop [51,52], also contribute to the modulation of the channel. The functional roles of the interfaces between TM1 and each of S1 and S2 [25] remain unexplored.Voltage-dependent movements of the extracellular ends of S0 and SSeveral groups using substituted-Cys-accessibility methods [29], crosslinking, or fluorometry on other voltage-dependent K+ channels and Na+ channels have inferred movements of S4 relative to the bilayer or relative to other TM helices [30?9]. In BK channels, as well, voltage-dependent movements of S4 [40-42] and, in particular, get SPDP Crosslinker separation of S0 and S4 [26] have been inferred from perturbations of fluorescent reporters. In addition, Horrigan and co-workers observed voltage-dependent Cu2+ binding to residues in S2, S3 and S4, consistent with some rearrangement of the BK voltage-sensor TM helices [43]. We found, however, that nearly complete disulfide crosslinking of R20C in S0 to W203C in S4, two positions that Pantazis et al. [26] inferred 
separated on activation, had no effect on the V50 for opening. This implies that separation of these two residues and, hence, of the extracellular ends of S0 and S4 might occur but is not required for activation. Furthermore, our finding that there was no difference in the rate constants in the activated and deactivated states for the induced reformation of the disulfide between W22C and W203C at the cell surface argues against much relative movement (Fig. 3). We did not evaluate the degree or significance of the interaction between the native W22 and W203 and their change in relative positions, or lack of it, during activation. Despite the similar relative dispositions of W22C and W203C in the activated and deactivated states, paradoxically the crosslink shifted the V50 for opening negatively and hence stabilized the open state at Ca2+ of 1 mM or greater. It is possible that this interface is 
normally more constrained in the activated state than in the inactivated state and that by pre-constraining the interface, the disulfide removes an entropic cost of activation and thereby stabilizes the open state. It is not readily apparent why the V50’s for W22C and W22C/W203C are not shifted to more positive and negative potentials respectively at nominally 0 Ca2+. The size of the shift may increase with increasing Ca2+ even with a change only in the voltage sensor equilibrium constant in the context of the dual allosteric model [44]. Some but not all crosslinks across the S0-S4 interface favor the open state. For example, the 1317923 crosslinks of both R20C and M21C in S0 to W203C in S4 favored the closed state. Also, in the CP21 flanking regions, some crosslinks between S0 and the four-residue S3 4 loop stabilized the open state, some stabilized the closed state, and only a few had no effect at all [22]. Crosslinks can of course distort the structures of the crosslinked segments as well as constrain their relative movement. Nevertheless, that even a few crosslinks have little effect on V50 indicate that either the top of S4 does not move much or that S0 an.Tes of the VSD. Also, the extent of the contact surface between S0 1379592 and TM2 below the first helical turns is not known. The paths of the three helices, TM2, S0, and S4 through the membrane would be revealed by the propensities of substituted Cys at their intracellular ends to form disulfide bonds with each other and with the other TM helices in a and in b1. Other regions of b1, such as the N-terminal and Cterminal tails [48?0] and the extracellular loop [51,52], also contribute to the modulation of the channel. The functional roles of the interfaces between TM1 and each of S1 and S2 [25] remain unexplored.Voltage-dependent movements of the extracellular ends of S0 and SSeveral groups using substituted-Cys-accessibility methods [29], crosslinking, or fluorometry on other voltage-dependent K+ channels and Na+ channels have inferred movements of S4 relative to the bilayer or relative to other TM helices [30?9]. In BK channels, as well, voltage-dependent movements of S4 [40-42] and, in particular, separation of S0 and S4 [26] have been inferred from perturbations of fluorescent reporters. In addition, Horrigan and co-workers observed voltage-dependent Cu2+ binding to residues in S2, S3 and S4, consistent with some rearrangement of the BK voltage-sensor TM helices [43]. We found, however, that nearly complete disulfide crosslinking of R20C in S0 to W203C in S4, two positions that Pantazis et al. [26] inferred separated on activation, had no effect on the V50 for opening. This implies that separation of these two residues and, hence, of the extracellular ends of S0 and S4 might occur but is not required for activation. Furthermore, our finding that there was no difference in the rate constants in the activated and deactivated states for the induced reformation of the disulfide between W22C and W203C at the cell surface argues against much relative movement (Fig. 3). We did not evaluate the degree or significance of the interaction between the native W22 and W203 and their change in relative positions, or lack of it, during activation. Despite the similar relative dispositions of W22C and W203C in the activated and deactivated states, paradoxically the crosslink shifted the V50 for opening negatively and hence stabilized the open state at Ca2+ of 1 mM or greater. It is possible that this interface is normally more constrained in the activated state than in the inactivated state and that by pre-constraining the interface, the disulfide removes an entropic cost of activation and thereby stabilizes the open state. It is not readily apparent why the V50’s for W22C and W22C/W203C are not shifted to more positive and negative potentials respectively at nominally 0 Ca2+. The size of the shift may increase with increasing Ca2+ even with a change only in the voltage sensor equilibrium constant in the context of the dual allosteric model [44]. Some but not all crosslinks across the S0-S4 interface favor the open state. For example, the 1317923 crosslinks of both R20C and M21C in S0 to W203C in S4 favored the closed state. Also, in the flanking regions, some crosslinks between S0 and the four-residue S3 4 loop stabilized the open state, some stabilized the closed state, and only a few had no effect at all [22]. Crosslinks can of course distort the structures of the crosslinked segments as well as constrain their relative movement. Nevertheless, that even a few crosslinks have little effect on V50 indicate that either the top of S4 does not move much or that S0 an.
