Observation of σ-pore currents in mutant hKv1.2_V370C potassium channels

Current through the σ-pore was first detected in hKv1.3_V388C channels, where the V388C mutation in hKv1.3 channels opened a new pathway (σ-pore) behind the central α-pore. Typical for this mutant channel was inward current at potentials more negative than -100 mV when the central α-pore was closed. The α-pore blockers such as TEA+ and peptide toxins (CTX, MTX) could not reduce current through the σ-pore of hKv1.3_V388C channels. This new pathway would proceed in parallel to the α-pore in the S6-S6 interface gap. To see whether this phenomenon is restricted to hKv1.3 channels we mutated hKv1.2 at the homologue position (hKv1.2_V370C). By overexpression of hKv1.2_V370C mutant channels in COS-7 cells we could show typical σ-currents. The electrophysiological properties of the σ-pore in hKv1.3_V388C and hKv1.2_V370C mutant channels were similar. The σ-pore of hKv1.2_V370C channels was most permeable to Na+ and Li+ whereas Cl- and protons did not influence current through the σ-pore. Tetraethylammonium (TEA+), charybdotoxin (CTX) and maurotoxin (MTX), known α-pore blockers, could not reduce current through the σ-pore of hKv1.2_V370C channels. Taken together we conclude that the observation of σ-pore currents is not restricted to Kv1.3 potassium channels but can also be observed in a closely related potassium channel. This finding could have implications in the treatment of different ion channel diseases linked to mutations of the respective channels in regions close to homologue position investigated by us.


Introduction
Earlier studies showed that mutation in voltage-gated and potassium channels could open other pathways besides the central α-pore through the complex channel molecules. These pathways could be described as alternative pores and were initially observed with mutations in the voltage-sensing domain (S1-S4) of the channels. For example by exchanging a positively charged arginine at position 362 in R1 S4 of the Shaker channel by cysteine or serine, an alternative pore (ω-pore) could be opened [1]. The ω-pore produce leak current conducting monovalent cations and is most permeable to K + . In addition, alternatives ω-pores in sodium channels could be observed with mutations in the voltage sensor S4 of Nav1.2 and Nav1.7 [2,3].
Kv1.2 and Kv1.3 channels are voltage-activated channels that open with depolarizations. Both channel proteins consist of four subunits. The N-and C-terminal regions of the channels are located at the intracellular side [4]. Each subunit of these channels contain six membrane- spanning regions (S1-S6) with a P-region between S5 and S6. This S5-P-S6 forms, together with three other subunits, the central, potassium selective α-pore. Segments S1-S4 form the voltage-sensing domain (VSD). This VSD controls the gates and is located around the pore domain [1,5]. Mutations in the VSD of the voltage-gated sodium channel Nav1.2 or the voltage-gated Shaker potassium channel can open another ion permeation pathway through the channel molecule [1][2][3][4][6][7]. This new pathway through the VSD was described as ω-current, was selective for monovalent cations and was open at potentials when the α-pore was closed [1].
Yet another pathway, the σ-pore through a mutant potassium channel was described [8] in a valine to cysteine mutant channel at position 388 in hKv1.3 (Shaker position 438, for a sequence alignment please see Table 1). This mutant hKv1.3_V388C channel showed an additional inward current at membrane potentials more negative than -100 mV. This σ-current showed similarities to the ω-current that flows through the voltage-sensing domain of the R1C/S mutated Shaker channel described above: First, ω-and σ-currents can only be observed at potentials more negative than -100 mV, a potential range where the central α-pore is normally closed; second, ω-and σ-currents can be carried by different monovalent cations like Li + and Cs + ; third, extracellularly applied α-pore blockers reduced current through the α-pore, however, had no effect on the ω-or σ-current. Since the ω-current was carried best by K + and the σ-current carried best by Na + , the authors concluded that the pathway of the ω-current was distinct from the pathway of the σ-current. Moreover, the hKv1.3_V388C mutant channel not only showed a sustained current at potentials more negative than -100 mV in external solutions containing in mM [160 Na + + 4.5 K + ] o but displayed normal current behavior in [164.5 K + ] o compared with the hKv1.3_wt channel. Based on this normal current behavior in the hKv1.3_V388C mutant channel in high potassium outside the authors concluded that the V388C mutation in hKv1.3 generated a channel with two ion-conducting pathways. One, the central α-pore allowing K + permeation in the presence of extracellular K + and another pathway, the σ-pore, functionally similar but physically distinct from the ω-pathway.
According to the model of the mutant hKv1.3_V388C channel, the exchange of the valine by cysteine, removing the two methyl groups of the valine at position 388 enlarged the space in between Y395 and W384 and may now allow the passage of ions [8]. The σ-pore was located behind the central α-pore at the back of the selectivity filter and proceeded parallel to the central α-pore. The entry of the σ-pore was located between the Tyr-395 of the GYG motif and the Trp-384 of the pore helix [8].
To find out whether the σ-pore is restricted to hKv1.3 channels we mutated hKv1.2, a very closely related voltage-gated potassium channel, at the homologue position (hKv1.2_V370C) and observed current behavior identical to current through the σ-pore.

Molecular cloning and site directed mutagenesis
The hKv1.2_wt template cDNA was a generous gift from Prof. Dr. O. Pongs (Institute for Neural Signal Processing, Center for Molecular Neurobiology, Hamburg Germany) and was cloned

Cell culture
For the expression of channels the adherent cell line COS-7 (passage 6 and 12, DSMZ no. ACC 60, Braunschweig, Germany) was used. The COS-7 cells were grown according to the standard protocol in DMEM high glucose with 10% FBS. The cells were incubated at 37˚C, 5% CO 2 with saturated humidity. Cells were grown to 95% confluence and transfected with 2 μg of total hKv1.2_V370C DNA plus 0.2 μg of pEGFP-C1 (CLONTECH) DNA using FuGENE 6 (Roche Molecular Biochemicals). The cells were replated the day after transfection on poly-Llysine-coated coverslips, and EGFP-positive cells were patch clamped 36-48 h after transfection, as described below.

Electrophysiology
The patch-clamp measurements were performed as described earlier [8]. Briefly, measurements were performed at room temperature 19-22˚C in the whole-cell configuration [9][10]. Cells were visualized with an inverted microscope Axiovert 25 (Carl Zeiss AG, Jena, Germany) installed on a vibration-isolation table (Newport Corporation, Irvine, USA) equipped with a xenon lamp and fluorescence detection unit. The amplifier EPC-9 (HEKA Elektronik GmbH, Lambrecht, Germany) was connected to a Dell computer running Patchmaster 2.0 data acquisition software. Currents were filtered through a 2.9 kHz Bessel Filter and capacitative and leakage currents were not subtracted. All voltage ramp protocols were preceded by a 100-ms prepulse to the starting potential to avoid complications associated with the slow "activation" of the σ-current. The analysis of the data was performed with the programs Fitmaster v2.15 (HEKA Elektronik GmbH) and Igor Pro 3.1.2 (Wave Metrics Inc., Lake Oswego, Oregon).

Solution and chemicals
The measurements were performed in different external bath solutions. . Osmolarity of the bath solutions was 300-310 mOsm. The internal pipette solution contained 145 mM KF, 2 mM MgCl 2 , 10 mM HEPES, 10 mM EGTA and was adjusted with KOH to pH 7.2 and the osmolarity was 310 mOsm. Charybdotoxin, CTX (Bachem, Bubendorf, Switzerland) and maurotoxin, MTX (Sigma-Aldrich, Saint Louis, USA) were dissolved in bath solution with 0.1% BSA.

Modeling
The model of the σ-pore in hKv1.2_V370C was created as described earlier for the hKv1.3_V388C mutant channel [8]. Briefly, we mutated V370C in the hKv1.2 wt (2A79) monomer with the help of the Deep Viewer software (Swiss PDB viewer, Expasy Server) followed by the creation of the hKv1.2_V370C homotetramer. The σ-pore was simulated with CAVER software (Loschmidt Laboratories, http://www.caver.cz) and visualized with PyMOL viewer (DeLano Scientific LLC, Schrödinger).

Results and discussion
The single point mutation V370C (Shaker position 438) in the hKv1.2 background channel showed an inward current at potentials more negative than -100 mV similar to the σ-current found in the homologous hKv1.3_V388C mutant channel [8] and different from the ω-current [1,6,7]. Below we characterized the electrophysiological and pharmacological properties of the inward current in hKv1.2_V370C mutant channels and compared it with the known properties of the σ-current found in the hKv1.3_V388C mutant channels. To confirm this assumption we performed the experiments shown in Fig 2. In the hKv1.3_V388C mutant channel in [160 Na + + 4.5 K + ] o (Fig 2B) we observed an outward current at +40 mV through the α-pore that inactivated much faster than the wild type ( Fig 2B) together with an inward current at -180 mV. In comparison, the hKv1.3 mutant channel in [164.5 K + ] o (Fig 2D) showed slightly slower inactivation at +40 mV compared with that in [160 Na + + 4.5 K + ] o (Fig 2B). At -180 mV in [164.5 K + ] o we could observe a current that deactivated slower compared with hKv1.3_wt (Fig 2C), however, with a smaller sustained inward current as seen in [160 Na + + 4.5 K + ] o . These observations are in agreement with earlier findings [8]. Similar observations regarding current through the α-and σ-pore as described above for hKv1.3_V388C mutant channels in normal and high external potassium solutions can be made for currents through hKv1.2_V370C mutant channels: At +40 mV in [160 Na + + 4.5 K + ] o an outward current through the α-pore of the hKv1.2_V370C mutant channels (Fig 2F) can be seen that inactivated much faster than in the wild type hKv1.2 channel (Fig 2E) together with an inward current at -180 mV that increased during the 100-ms hyperpolarization in most of our experiments (21 out of 24) using this protocol. In a minority of these experiments (3 out of 24) the increase in σ-current amplitude at -180 mV was followed by a slight decrease during this 100-ms hyperpolarization. In [164.5 K + ] o the current at -180 mV deactivated slower ( Fig 2H) compared with hKv1.2_wt (Fig 2G) with a smaller sustained inward current compared to Fig 2F. σ-currents were not inhibited by CTX and MTX, known α-pore-blocking peptide toxins acting at the external mouth of the channel Fig 3A clearly shows that application of 700 nM CTX in a bathing solution containing 4.5 mM [K + ] o cannot block current through hKv1.2_V370C mutant channels at potentials more negative than -60 mV indicating that CTX is unable to block current through the σ-pore while still able to block current through the central α-pore as can be seen in Fig 3B in a bathing  reduce outward current in this record. The answer to this phenomenon is similar to what has been reported earlier [8]: the time course of inactivation of the mutant hKv1.2_V370C channel, even in high external potassium, shown in Fig 2H, is so fast that during the first 300 ms of the 400-ms voltage ramp (showing the inward current) the channel did completely inactivate. Therefore, the outward current in the ramp current shown in Fig 3B cannot go through the mutant hKv1.2_V370C channel. We conclude that the outward current in Fig 3B is either a nonspecific leak current or flows through some other endogenous channels in the cell, for example through chloride channels.
In additional experiments we compared the application of 1 and 18 nM MTX on currents through hKv1.2_wt (Fig 3C, left) and hKv1.2_V370C mutant channels (Fig 3C, right). Through both channels the outward current through the α-pore at a potential of +40 mV during depolarization was similarly reduced. For example in the wild type hKv1.2 channel 1 nM MTX reduced current to about one third of the control current and in the hKv1.2_V370C mutant channel the same concentration reduced peak current to about one half. These current reductions indicate minor changes in the ability of MTX to block current through the α-pore of hKv1.2_wt and hKv1.2_V370C mutant channels. More importantly, amplitude and kinetic properties of the inward current through the σ-pore of the hKv1.2_V370C mutant channel at -180 mV did not change (Fig 3C, right) at any of the applied MTX concentrations.

Ion selectivity of the σ-current
To further characterize the inward current in the hKv1.2_V370C mutant channel we determined which ions could pass through the σ-pore. Replacing extracellular Clby aspartate as shown in Fig 4A did not change the inward current suggesting that the inward current was not selective for Cl -.
Is the inward current through hKv1.2_V370C mutant channels insensitive to protons similar to the situation in the hKv1.3_V388C mutant channel [8]? To answer this question we tested external bathing solutions [160 Na + + 4.5 K + ] O with different pH O . Decreasing pH to 5.5 or increasing pH to 8.0 did not influence σ-current (blue and red traces, Fig 4B) through the hKv1.2 V370C mutant channels similar to what was described for current through the hKv1.3_V388C mutant channel [8]. In both cases, the σ-current was not carried by H + .
To elucidate which ions could generate σ-currents we replaced the major cations in the external bathing solution. Extracellular Rb + and K + generated very small inward currents through hKv1.2_V370C mutant channels whereas extracellular Cs + , NH 4 + , Na + or Li + could carry larger inward currents at potentials more negative than -100 mV. From the amplitudes of the ramp currents (I x + ) at -180 mV we calculated the ratios (I x + /I Na + ) as measure of ion conductance. The measurement resulted in an ion permeation efficiency in the following order: Li + (1.1) >Na + (1) >NH 4 + (0.7) >Cs + (0.3) > K + (0.18) >Rb + (0.12) similar to what was described for currents through the σ-pore of hKv1.3_V388C mutant channels [8].

Model of the σ-pore
Prütting et al. [8] modelled the σ-pore of the hKv1.3_V388C mutant channel and according to their model postulated that the entrance of the σ-pore from the outside should be located before and after extracellular application of CTX. Ramp currents were elicited as described in the legend to Fig 1. (C), effect of 1 and 18 nM MTX on currents through the α-pore of hKv1.2_wt channels (left) and through the α-and σ-pores of hKv1.2_V370C mutant channels in [160 Na + + 4.5 K + ] o , elicited with 100-ms depolarizing pulses from the holding potential of -120 mV to +40 mV followed by a 100-ms hyperpolarizing pulse to -180 mV every 30 s.  between Y395 (Shaker position 445) on the backside of the central α-pore and W384 (Shaker position 434) of the channel. Since the S5-P-S6 region of hKv1.3 is very similar to hKv1.2 we modelled the σ-pore in hKv1.2_V370C similar to what was described for the hKv1.3_V388C mutant channel [8] i.e. using the Caver program, visualizing the pore with PyMOL 1 and verifying it with PoreWalker as shown in Fig 5. For the hKv1.2_V370C mutant channel the entry of the σ-pore is located on the extracellular side of the channel between Y377 (Shaker position 445) on the back surface of the α-pore and W366 (Shaker position 434), it runs parallel to the GYG motif of the selectivity filter in the S6-S6 interface gap and ends between S5 and S6 at the intracellular side of one α-subunit. The ending of this pathway might be responsible for the fact that σ-current can only occur in a potential range where the α-pore is closed, i.e. the voltage sensor S4 is in its resting position. The position of S4 seems to be important for the opening or closing of the σ-pore. During hyperpolarization or at the resting potential of a cell, the gap between S5 and S6 is larger (see Fig 7 of [11]). During depolarizations of the channel the voltage sensors S4 move towards the extracellular side leading to a concerted movement of S5 and S6 via the S4-S5 linker. This results in a structural change of the channel narrowing the gap between S5 and S6 [11][12][13]. The gap between S5 and S6 could then be too narrow to allow the flux of Na + through the σpore. Therefore to open the σ-pore the voltage sensor S4 must be in its resting position. One Observation of σ-pore currents in mutant hKv1.2_V370C potassium channels could speculate that S4 moves even further towards the intracellular side or even tilts towards the side at strong hyperpolarized potentials to widen the σ-pore thereby increasing current amplitude towards more negative potentials. Such a movement would be slow (>30 ms) as can be judged by the time course of activation of the σ-current compared to the classical gating charge movements observed when opening or closing the α-pore (<3 ms).

Conclusion
The newly described permeation pathway of the mutant hKv1.2_V370C channel is likely to be similar to the σ-pore described in hKv1.3_V388C mutant channels [8]. In both channels, αpore blockers were unable to block current through the σ-pore. In addition, σ-pore current had a similar potential range of activation (more negative than -100 mV) and had the same ion selectivity. We conclude that the V370C mutation in hKv1.2_V370C channels opens up a similar pathway like in the hKv1.3_V388C mutant channel suggesting that the observation of a σpore is not restricted to Kv1.3 channels but may be a common structural element of a variety of voltage-gated ion channels. Therefore this finding could have implications for the interpretation of the cause and the treatment of different ion channel diseases associated with mutations in the pore-region of the respective channels reviewed in [14]. In such a scenario the observation of Na + currents leading to long depolarizations resulting in arrhythmias [15][16] or migraine [17] could be interpreted as a result of a current similar to the σ-pore current and treatment would then require the development of a selective σ-pore blocker.