Background. Oxaliplatin is a platinum-based chemotherapeutic drug. Neurotoxicity is the dose-limiting side effect. Previous investigations have reported that acute neurotoxicity could be mediated via voltage-gated ion channels. A possible mechanism for some of the effects is a modification of surface charges around the ion channel, either because of chelation of extracellular Ca2+, or because of binding of a charged biotransformation product of oxaliplatin to the channel. To elucidate the molecular mechanism, we investigated the effects of oxaliplatin and its chloride complex [Pt(dach)oxCl]- on the voltage-gated Shaker K channel expressed in Xenopus oocytes. The recordings were made with the two-electrode and the cut-open oocyte voltage clamp techniques. Conclusion. To our surprise, we did not see any effects on the current amplitudes, on the current time courses, or on the voltage dependence of the Shaker wild-type channel. Oxaliplatin is expected to bind to cysteines. Therefore, we explored if there could be a specific effect on single (E418C) and double-cysteine (R362C/F416C) mutated Shaker channels previously shown to be sensitive to cysteine-specific reagents. Neither of these channels were affected by oxaliplatin. The clear lack of effect on the Shaker K channel suggests that oxaliplatin or its monochloro complex has no general surface-charge effect on the channels, as has been suggested before, but rather a specific effect to the channels previously shown to be affected.

The size of the movement and the molecular identity of the moving parts of the voltage sensor of a voltage-gated ion channel are debated. In the helical-screw model, the positively charged fourth transmembrane segment S4 slides and rotates along negative counter charges in S2 and S3, while in the paddle model, S4 carries the extracellular part of S3 (S3b) as a cargo. Here, we show that S4 slides 16-26 Å along S3b. We introduced pairs of cysteines in S4 and S3b of the Shaker K channel to make disulfide bonds. Residue 325 in S3b makes close and state-dependent contacts with a long stretch of residues in S4. A disulfide bond between 325 and 360 was formed in the closed state, while a bond between 325 and 366 was formed in the open state. These data are not compatible with the voltage-sensor paddle model, but support the helical-screw model. © 2008 Elsevier Inc. All rights reserved.

Voltage-gated K channels are regulated by extracellular divalent cations such as Mg²⁺ and Sr²⁺, either by screening of fixed negative surface charges, by binding directly or close to the voltage sensor, or by binding to the pore. Different K channels display different sensitivity to divalent cations. For instance, 20 mM MgCl₂ shifts the conductance versus voltage curve, G(V), of the Kv1-type Shaker channel with 14 mV, while the G(V) of Kv2.1 is shifted only with 7 mV. This shift difference is paralleled with different working ranges. Kv1-type channels open at −20 mV and Kv2.1 channel open at +5 mV. The aim of this study was to identify critical residues for this Mg²⁺-induced G(V) shift by introducing Kv2.1 channel residues in the Shaker K channel. The K channels were expressed in Xenopus laevis oocytes and studied with the two-electrode voltage-clamp technique. We found that three neutral-to-positive amino-acid residue exchanges in the extracellular loops connecting transmembrane segments S5 and S6 transferred the Mg²⁺-shifting properties. The contributions of the three residues were additive, and thus independent of each other, with the contributions in the order 425 > 419 > 451. Charging 425 and 419 not only affect the Mg²⁺-induced G(V) shift with 5–6 mV, but also shifts the G(V) with 17 mV. Thus, a few strategically placed surface charges clearly modulate the channel’s working range. Residue 425, located at some distance away from the voltage sensor, was shown to electrostatically affect residue K427, which in turn affects the voltage sensor S4—thus, an electrostatic domino effect.

Voltage-gated ion channels are fundamental for electrical signaling in living cells. They are composed of four subunits, each holding six transmembrane helices, S1-S6. Each subunit contains a voltage-sensor domain, S1-S4, and a pore domain, S5-S6. S4 contains several positively charged amino-acid residues and moves in response to changes in membrane voltage. This movement controls the opening and closing of the channel. The structure of the pore domain is solved and demonstrates principles of channel selectivity. The molecular mechanism of how the voltage sensor regulates the opening of the channel is still under discussion. Several models have been discussed. One of the models is the paddle model where S3b and S4 move together. The second one is the helical-twist where S4 makes a small rotation in order for the channel to open. The third one is the helical-screw model where S4 twists around its axis and moves diagonally towards the extracellular side of the channel.

The aim of this PhD project was to study the molecular movement of the voltage sensor in the depolarization-activated Shaker K channel. Cloned channels were expressed in Xenopus laevis oocytes, and investigated with several electrophysiological techniques.

1. We show that S4 moves in relation to both S3b and S5. The formation of some disulfide bonds between S4 and neighboring positions, in only the open state, shows that the paddle model cannot be correct. Furthermore, electrostatic and steric effects of residues in S3b suggest that S3b is tilted, with the intracellular part close to S4.

2. We show that the relatively Mg-sensitive Shaker K channel is changed into the less Mg-sensitive Kv2.1 K channel with respect to its sensitivity to extracellularly applied Mg2+ by changing the charge of three extracellularly positioned amino acid residues. One of the residues, F425C, mediates its effect through the neighboring residue K427.

3. We show that oxaliplatin, an anti-cancer drug, has no effect on the Shaker K channel. It has been suggested that a negatively charged monochloro complex of oxaliplatin is the active substance, and also causes the neurotoxic side effects. Neither this complex shows any effect on the channel.

Our experiments point towards the helical-screw model. The other models for voltage-sensor movements are incompatible with the results in this study.