Tetraphenylporphyrin derivative specifically blocks members of the voltage-gated potassium channel subfamily Kv1

Abstract

Tetraphenylporphyrin derivatives represent a promising class of high-affinity ligands for voltage-gated potassium (Kv) channels. Herein, we investigated the mode of Kv channel block of one tetraphenylporphyrin derivative, por3, using electrophysiological methods, structure-based mutagenesis, and solid-state NMR spectroscopy. The combined data showed that por3 specifically blocks Kv1.x channels. Unexpectedly, 2 different por3 binding modes lead to Kv1.x channel block exerted through multiple por3 binding sites: first, por3 interacts in a highly cooperative and specific manner with the voltage sensor domain stabilizing closed Kv1 channel state(s). Therefore, stronger depolarization is needed to activate Kv1.x channels in the presence of por3. Second, por3 bind to a single site at the external pore entrance to block the ion conduction pathway of activated Kv1.x channels. This block is voltage-independent. Por3 appears to have equal affinities for voltage-sensor and pore. However, at negative voltage and low por3 concentration, por3 gating modifier properties prevail due to the high cooperativity of binding. By contrast, at positive voltages, when Kv1.x channels are fully activated, por3 pore blocking properties predominate.

Keywords: Voltage-gated potassium-channel; gating modifier; liposomes; porphyrin; solid-state NMR.

Figure 1. Specific inhibition of Kv1.x channels by por3. (A–F) G–V relations obtained in absence (black dots) and presence of 0.5 µM por3 (red dots) by using normalized Kv1.x tail current amplitudes measured at −80 mV. (G) G−V relation of por3 insensitive hERG (Kv11.1) channels. All Kv channels were expressed in the Xenopus laevis oocyte expression system as described. Solid curves are Boltzmann fits to the data (n = 4 – 7). (H) Dependence of Kv1.1 (black squares) and Kv1.3 (black circles) channel block on Por3 concentration. Fractional inhibition data were obtained measuring tail current amplitudes after a preceding test potential to +80 mV (n = 5). A Hill equation was fit to the data with a Hill coefficient of 3.8 and K_d = 0.15 µM in the case of Kv1.1 (solid line) and K_d = 0.24 µM in the case of Kv1.3 (dashed line).

Figure 2. Kinetics and voltage-dependence of por3 block. (A) Block of Kv1.3 current by 0.5 µM por3. Currents were recorded on outside-out patches of Xenopus laevis oocytes expressing the Kv1.3 channel in the absence (black trace) and presence of por3 (red trace). Holding potential was –80 mV and test potential at −20 mV. Bars indicate current amplitude and test pulse duration. (B) Exemplary wash-in and washout kinetics of por3 block shown in (A). Red bar indicates duration of 0.5 µM por3 application by bath perfusion. Current amplitude (black dots) was recorded every 20 s at –20 mV. Kinetics of por3 block was fit with 1 time constant (τ_on,por3 = 16.5 s; dashed curve) and those of por3 unblock with 2 time constants (τ_1,off, = 34.8 s and τ_2off = 295 s; dashed curve). (C) Activation and deactivation kinetics of Kv1.3 current recorded on Kv1.3 expressing Xenopus laevis oocytes in absence (black trace) and presence of 0.5 µM por3 (red trace) at indicated test potential. Holding potential was –80 mV, test potential +60 mV, tail potential –80 mV. For comparison currents were scaled. Horizontal bar indicates pulse duration. (D) Kv1.3 I–V relation in absence (black dots) and presence (red dots) of 0.5 µM por3. Data were obtained from normalized tail current amplitudes measured at –80 mV. Solid curves are Boltzmann fit to the data with V_1/2 = –8.3 ± 0.9 mV and Z = 2.1 ± 0.14 (n = 5) and V_1/2 = 18.5 ± 0.8 mV and Z = 1.3 ± 0.08 (n = 5), respectively. (E) Exemplary Kv1.3 current records obtained using a double-pulse protocol as indicated on top of current traces. Currents were elicited from a holding potential of –80 mV in absence (black trace) and presence of 0.5 µM por3 (red trace). Bars indicate Kv1.3 current amplitude and test pulse duration. (F) Dependence of Kv1.3 channel block on por3 concentration. Fractional inhibition data were obtained measuring tail current amplitudes after a preceding test potential to –20 mV (n = 5). A Hill equation (solid line) was fit to the data with a Hill coefficient of 3.9 and K_d = 0.13 µM.

Figure 3. Dependence of por3 block on Kv channel state. (A) Exemplary Kv1.3 current traces (I_control) elicited from a holding potential of –80 mV to test potentials of ≥ 0 mV as indicated. Dotted line is base line recorded at –80 mV. Bars indicate current amplitude and test pulse duration. (B) Exemplary current traces as in (A), but recorded in the presence of 0.2 µM por3 (I_por3) (C). Plot of I_por3/I_control against test pulse duration to calculate τ_U at different test potentials. Dashed line indicates base line. (D) Plot of τ_U against voltage of test pulse. For further details see Results and Equation 1.

Figure 4. Por3 affects Kv1.3 gating charge movement. (A) Kv1.3 gating currents elicited by depolarization to –20 mV and subsequent repolarization to the holding potential of –80 mV. Black and red trace correspond to gating-current of control and, respectively, after application of 0.5 µM por3. Bars indicate current amplitude and test pulse duration. (B) Normalized Q_off–V relation obtained in absence (black trace) and presence of 0.5 µM por3 (red trace). A Boltzmann function was fit to the data with V_{1/2, Qoff} = –22.1 ± 1.1 mV (n = 4) for gating charge movement in absence (black line) and V_{1/2, Qoff} = –10 ± 0.5 (n = 4) in presence of 0.5 µM por3 (red line). (C) Exemplary wash-in and washout kinetics of por3 block of gating charge movement mV as shown in (A). Red bar indicates duration of 0.5 µM por3 application by bath perfusion. Q_Off (black dots) was recorded every 20 s at –20 mV. Kinetics of por3 block were fit with τ_on,Qoff = 26.3 s and those of por3 unblock with τ_off,Qoff = 26.6 s (dashed curve). (D) Probabibility (ρ₄) of por3 block of Kv1.3 gating charge movement at different por3 concentration. Data (n = 3) were evaluated as described by Equation 2 in Results (K_d = 45 nM; solid curve), i.e., cooperative por3 binding at four binding sites is needed to effect block. Dotted line is obtained with K_d = 45 nM, if por3 binding to 1 of the 4 sites suffices to effect block. Dashed line corresponds to a single binding site with K_d = 240 nM effecting por3 block of Kv1.3 gating charge movement at –20 mV.

Figure 5. Por3 blocks constitutively opened Kv1.3P427D channel. (A) Exemplary Kv1.3 P427D currents elicited by depolarization to voltages between –80 mV and +80 mV. Bath solution contained 100 mM K⁺. Holding potential was at 0 mV. Control currents are in black, currents in presence of 0.5 µM por3 are shown in red. (B) Current-voltage relation of Kv1.3 P427D currents (n = 5) in absence (black dots) and presence of 0.5 µM por3 (red dots) as shown in (A). (C) Bar diagram of fractional inhibition of Kv1.3 P427D current obtained by 0.5 µM por3 application at different test voltages (n = 5). (D) Exemplary wash-in and washout kinetics of por3 block of Kv1.3 P427D. Red bar indicates duration of 0.5 µM por3 application by bath perfusion. Kv1.3 P427D current (open circles) was recorded every 20 s at +80 mV. Kinetics of por3 block were fit with τ_on = 31.3 s and those of por3 unblock with τ_off,Qoff = 285.0 s (dashed curve). (E) Bar diagram of fractional inhibition of Kv1.5 and Kv1.5R487Y current in comparison to Kv1.1 and Kv1.3 block by por3 (n for each bar = 5). (F) Dependency of Kv1.1 (squares) and Kv1.3 channel block (circles) on por3 concentration. Fractional inhibition data were obtained measuring tail current amplitudes as in Figure 1. An equation assuming cooperative binding sites (K_d1) and an additional independent single binding site (K_d2) was fit to the data (solid curves) with a Hill coefficient of 4.5 and K_d1 = 0.16 µM and K_d2 = 0.20 µM in the case of Kv1.1 and with a Hill coefficient of 4.5 and K_d1 = 0.18 µM and K_d2 = 0.23 µM in the case of Kv1.3.

Figure 6. Study of por3-lipid interactions by MAS NMR. (A) Structures of DMPC with atom nomenclature and (B) of por3. (C) NOESY spectrum of DPMC:por3 (10:1, mol:mol) obtained for a ¹H-¹H mixing time of 250 ms under MAS. Intermolecular cross peaks between lipid and por3 resonances were labeled by dashed boxes. (D) Normalized per ¹H cross relaxation rates computed for the porphyrin core of por3 to sp.ins of the DPMC head group. (E) Model for the binding mode of por3 to the lipid bilayer as seen from ¹H-¹H cross relaxation rates.