The asparagine 533 residue in the outer pore loop region of the mouse PKD2L1 channel is essential for its voltage‐dependent inactivation

Voltage‐dependent inactivation of ion channels contributes to the regulation of the membrane potential of excitable cells. Mouse polycystic kidney disease 2‐like 1 (PKD2L1) forms voltage‐dependent nonselective cation channels, which are activated but subsequently inactivated in response to membrane depolarization. Here, we found that the mutation of an asparagine 533 residue (N533Q) in the outer pore loop region of PKD2L1 caused a marked increase in outward currents induced by depolarization. In addition, the tail current analysis demonstrated that the N533Q mutants are activated during depolarization but the subsequent inactivation does not occur. Interestingly, the N533Q mutants lacked the channel activation triggered by the removal of stimuli such as extracellular alkalization and heating. Our findings suggest that the N533 residue in the outer pore loop region of PKD2L1 has a key role in the voltage‐dependent channel inactivation.

Voltage-dependent inactivation of ion channels contributes to the regulation of the membrane potential of excitable cells. Mouse polycystic kidney disease 2-like 1 (PKD2L1) forms voltage-dependent nonselective cation channels, which are activated but subsequently inactivated in response to membrane depolarization. Here, we found that the mutation of an asparagine 533 residue (N533Q) in the outer pore loop region of PKD2L1 caused a marked increase in outward currents induced by depolarization. In addition, the tail current analysis demonstrated that the N533Q mutants are activated during depolarization but the subsequent inactivation does not occur. Interestingly, the N533Q mutants lacked the channel activation triggered by the removal of stimuli such as extracellular alkalization and heating. Our findings suggest that the N533 residue in the outer pore loop region of PKD2L1 has a key role in the voltage-dependent channel inactivation.
Transient receptor potential (TRP) proteins constitute a superfamily of nonselective cation channels. The TRP protein subunit has six transmembrane domains with intracellular N and C termini and a putative pore region between transmembrane segments 5 and 6 [1,2]. Polycystic kidney disease 2-like 1 (PKD2L1) (also called TRP polycystin 3 [TRPP3]) is a member of the TRP superfamily [3]. We have so far demonstrated that mouse PKD2L1 functions as a voltage-dependent nonselective cation channel regulated by extracellular pH and ambient temperature [4][5][6][7]. Interestingly, the PKD2L1 channel exhibits unique gating kinetics. Although the cells expressing the PKD2L1 channels show quite small depolarization-triggered outward currents, large tail currents are observed during the subsequent repolarization. The current response indicates that the PKD2L1 channels can enter an inactivated state. In the gating model, depolarization first activates but then immediately inactivates the PKD2L1 channels, therefore showing small outward currents upon depolarization. The tail currents triggered by the following repolarization represent quick transition of the PKD2L1 channel from inactivated to open states. Thus, the voltage-dependent inactivation is essential for robust activation of the PKD2L1 channels upon repolarization. The structural basis for the channel inactivation during depolarization, however, is poorly understood.
It is well known that some voltage-dependent K + channels exhibit two types of voltage-dependent inactivation, which are referred to as N-and C-type inactivation [8][9][10][11]. In N-type inactivation, the intracellular N-terminal region of the channel occludes the inner pore. In contrast, C-type inactivation is associated with structural changes around the selectivity filter of the channel. Among voltage-dependent K + channels, human ether-a-go-go-related gene (HERG) K + channels display a characteristic voltage-dependent gating [12][13][14]. In response to depolarization, the HERG K + channels first activate but subsequently inactivate. Then, upon repolarization, the inactivated channels are rapidly reactivated. The voltage-dependent inactivation in the HERG K + channel is considered to be due to C-type inactivation, because the mutation of amino acid residues adjacent to the selectivity filter of the channel, N629 and S631, suppresses the inactivation [12,15,16]. As the voltage-dependent gating of the PKD2L1 channel has some resemblance to that of the HERG K + channel, in the present study, we focused on amino acid residues behind the selectivity filter of the PKD2L1 channel, N531 and N533, to investigate the mechanism of the PKD2L1 channel inactivation.
To quantify expression levels of the PKD2L1 mutants, we introduced a sequence encoding an Xpress epitope [amplified from the pcDNA4/His expression vector (Invitrogen, Carlsbad, CA, USA)] into each mouse PKD2L1 expression vector. The generated vectors were sequenced to confirm the appropriate location of the Xpress epitope.

Cell culture and transfection
Human embryonic kidney HEK293T cells were grown in Dulbecco's modified Eagle's medium (Sigma-Aldrich Japan, Tokyo, Japan) supplemented with 10% fetal bovine serum, 100 unitÁmL À1 penicillin, and 100 lgÁmL À1 streptomycin at 37°C in a humidity-controlled incubator with 5% CO 2 . Each expression vector was transiently transfected into HEK293T cells using Lipofectamine 2000 reagent (Invitrogen). Transfected HEK293T cells were detached from the plastic substrate and cultured on cover slips (Matsunami Glass, Osaka, Japan) before electrophysiological experiments. Patch-clamp experiments were carried out with GFP-positive cells 48 h after transfection.

Electrophysiological experiments
Whole-cell patch-clamp recordings were performed using an Axopatch 200B patch-clamp amplifier (Molecular Devices, Sunnyvale, CA, USA). CLAMPEX 9.2 software (Molecular Devices) was used for command pulse control and data acquisition. Currents were filtered at 1 kHz and digitized at 10 kHz. The data were analyzed with CLAMPFIT 9.2 (Molecular Devices) and WINASCD software (kindly provided by G. Droogmans). Patch electrodes had a resistance of 2-4 MΩ when filled with pipette solution. The access resistance was electrically compensated by 70% to minimize voltage errors. To investigate the voltage dependence of PKD2L1 channels, step pulses were applied from À100 to +200 mV in 20-mV increments with a postpulse to À100 mV. PKD2L1 channel activity (NPo, where N is the number of channels and Po is the open probability) was obtained by dividing each tail current by the corresponding single-channel current amplitude recorded at À100 mV. The voltage dependence of the channel activity was fitted to the Boltzmann equation.
To analyze single-channel properties of the PKD2L1 channel, whole-cell patch membranes were clamped at À60 mV. All-point amplitude histograms were created from continuous current traces of 30 s. The single-channel current amplitude was measured as the peak-to-peak distance in Gaussian fits of the histograms. The channel activity (NPo) was calculated by dividing the mean current amplitude of each current trace by the corresponding singlechannel amplitude, as previously described [4][5][6].

Statistics
Data are presented as means AE SEM of n observations. Statistical differences in the data were evaluated by unpaired Student's t-test or one-way ANOVA with Tukey's post hoc tests and were considered significant at P < 0.05.

Membrane expression of mouse PKD2L1 mutants
Western blotting was performed using membrane fractions prepared from the HEK293T cells overexpressing Xpress-tagged PKD2L1 proteins. As shown in Fig. 1B, Xpress-tagged PKD2L1 proteins were detected as a band at expected size (~90 kDa). The expression level of each PKD2L1 mutant was similar to that of the wild-type (WT) channel (Fig. 1C).
The outer pore loop region of the mouse PKD2L1 channel regulates the voltage-dependent inactivation We have previously reported that mouse PKD2L1 channels exhibit voltage-dependent inactivation [4]. Although depolarization-triggered outward currents of the PKD2L1 channels are small, large tail currents are observed during the subsequent repolarization (see Fig. 2A), suggesting that PKD2L1 channels activate but then immediately inactivate upon depolarization. Here, we investigated whether the outer pore loop region of the PKD2L1 channel is involved in the depolarization-induced inactivation, based on the functional role of the corresponding region in the HERG K + channel. As we previously had a variety of values for half-maximal activation voltage (V half ) in the voltage-dependent activation curve of PKD2L1 channels due to small tail currents [4][5][6], in the present experiments, we applied stronger depolarizing pulses up to +200 mV to precisely analyze the voltage-dependent gating of the PKD2L1 channel ( Fig. 2A, inset). HEK293T cells overexpressing the WT channels showed typical PKD2L1 currents; that is, the outward currents triggered by depolarization were small, but larger tail currents were observed upon repolarization to À100 mV ( Fig. 2A-C). In comparison with the WT channels, the N531A/N533A double mutants, in which the length of the side chains becomes shorter, exhibited larger outward currents at potentials greater than +160 mV ( Fig. 2A,B). The decay of tail currents was accelerated in the cells expressing the N531A/N533A mutants ( Fig. 2A). In addition, the current-voltage relationship for the tail currents indicates that the N531A/N533A mutants had a different voltage dependence compared to the WT channels (Fig. 2C). As we could measure the single-channel PKD2L1 current at the end of the tail currents as shown in Fig. 2A inset, we calculated the PKD2L1 channel activity (NPo), which was obtained by dividing the tail currents by the corresponding single-channel current amplitude. Figure 2D shows the activation curves for WT channel activity (NPo) and for N531A/N533A mutant activity (NPo). The double mutants showed a rightward shift and steeper slope of the activation curve compared with the WT channels (Fig. 2D). The V half was significantly shifted from 163.1 AE 7.3 mV in the WT channels to 204.1 AE 6.7 mV in the N531A/N533A mutants (P = 0.01). The slope factor was statistically different between the WT channels (29.2 AE 1.7 mV) and the N531A/N533A mutants (18.5 AE 0.9 mV, P = 0.02). On the other hand, the single-channel current amplitude in the N531A/N533A mutants (À19.6 AE 0.6 pA, P = 0.47) was similar to that in the WT channels (À20.6 AE 0.6 pA).
We further investigated which of the two asparagine residues, N531 or N533, affects the voltage-dependent inactivation of the PKD2L1 channel. To generate the single PKD2L1 mutants, each asparagine residue was replaced with glutamine instead of an alanine, because the side chain of the glutamine is longer than that of asparagine. Thereby, we tried to examine an appropriate length of the side chain at positions 531 and 533 in the voltage-dependent inactivation of the PID2L1 channels. The N531Q mutants exhibited PKD2L1 currents similar to the WT channels ( Fig. 3A-C). The V half and the slope factor in the activation curve of the N531Q mutants were 169.7 AE 10.7 mV (P = 0.62) and 30.3 AE 3.3 mV (P = 0.46), respectively (Fig. 3D). On the other hand, the N533Q mutants surprisingly showed larger outward currents upon depolarization and a faster decay of tail currents compared with the WT channels (Fig. 3A,B). As with the double mutants, voltage dependence of the tail currents of the N533Q mutants was different from that of the WT channels (Fig. 3C). The N533Q mutation shifted the activation curve to more positive potentials and decreased its slope factor (Fig. 3D). The activation curve of the N533Q mutant activity had a V half of 204.6 AE 16.5 mV (P = 0.02) and a slope factor of 18.6 AE 2.1 mV (P = 0.04). The single-channel current amplitudes in the N531Q and N533Q mutants were À19.8 AE 0.4 pA (P = 0.40) and À19.9 AE 0.8 pA (P = 0.50), respectively. These results suggest that the asparagine residue at position 533 of the PKD2L1 channel regulates the channel inactivation upon depolarization.

The N533 mutant of the mouse PKD2L1 channel lacks an increase in the channel activity after the reversal of alkalization and heating
We have previously demonstrated that heating and a shift to alkaline pH at the extracellular side regulate the voltage-dependent inactivation of the PKD2L1 channel, thereby causing a marked increase in the channel currents during a subsequent decrease in pH and temperature [5,6]. We therefore investigated whether the mutations of the PKD2L1 channel affect its alkali and temperature sensitivity.
Single-channel PKD2L1 currents can be observed at a holding potential of À60 mV in whole-cell patchclamp recordings of PKD2L1-expressing cells [4][5][6]. As shown in Fig. 4A, the WT channels exhibited a constitutive activity, which increases and then decreases as extracellular pH is increased. Consistent with our previous findings [5], subsequent neutralization to pH 7.4 caused transient activation of the PKD2L1 channels (Fig. 4A). The current change is dependent on the shift of the voltage-dependent activation curve. The N531Q and N533Q mutants had a similar basal activity to the WT channels (Fig. 4A). The singlechannel conductance and channel activity (NPo) of these mutants were comparable to those of the WT channels (Fig. 4B,C). On the other hand, the N531Q mutants were activated by a pH drop after alkalization, whereas the N533Q mutants were not (Fig. 4A). Single-channel analysis demonstrated that the N533Q mutation influenced the channel activity (NPo), but not the single-channel conductance (Fig. 4B,C).
We next examined the temperature sensitivity of the PKD2L1 channel. As previously reported [6], although heating to 40°C increased and decreased a single-channel conductance and channel activity (NPo) of the WT channels, respectively (data not shown), rapid decrease in the temperature induced an obvious activation of the WT channels (Fig. 5A). The N531Q mutants showed activation by a temperature drop similar to that of the WT channels (Fig. 5A). On the other hand, the N533Q mutants did not induce such a current activation (Fig. 5A). In contrast to the WT channels and N531Q mutants, the N533Q mutants lacked the increase in the channel activity (NPo) during cooling (Fig. 5C), although the single-channel conductance was not altered by the mutations (Fig. 5B). These results suggest that the N533 residue of the PKD2L1 channel also regulates its sensitivity to alkalization and heating.

Discussion
Voltage-dependent inactivation of ion channels is crucial in controlling the membrane excitability of excitable cells. In the present study, we focused on the outer pore loop region of the PKD2L1 channel to investigate its inactivation mechanism, because it is well known that the region modulates C-type inactivation of voltage-dependent K + channels such as the HERG K + channel [8][9][10]12,21,22]. We here identified a key amino acid residue for voltage-dependent inactivation of the mouse PKD2L1 channel. While the WT PKD2L1 channels exhibited small outward currents upon depolarization, the outward currents of the N533Q mutants were much larger (Fig. 3A,B). The increase in outward currents in the N533Q mutants is due to lack of the transition from open to inactivated states during depolarization, because we could not observe similar tail currents in the N533Q mutants to that in the WT channels. Decay of the tail currents in the N533Q mutants was much faster than that in the WT channels (Fig. 3A). In addition, the voltage dependence of the tail currents in the N533Q mutants was clearly different from that in the WT channels (Fig. 3C). The N533Q mutants had a rightward shift and a steeper slope of the activation curve compared with the WT channels (Fig. 3D). Importantly, the tail currents of N533Q mutants were observed only upon repolarization from depolarized potentials that showed the large outward currents (Fig. 3B,C). These results suggest that the N533Q mutants activate, but do not subsequently inactivate, upon depolarization and that the following repolarization causes its deactivation (the transition process from open to closed states). On the other hand, although the single-channel conductance and the basal channel activity (NPo) of the N533Q mutants were similar to those of the WT channels (Figs 4 and 5), the gating property of the N533Q mutants seems to be different from that of the WT channels. The WT channels showed long opening events in the decaying process of the tail currents, whereas the N533Q mutants displayed flickering events (Fig. 3A inset). Further studies are awaited to investigate the gating difference between the WT PKD2L1 channel and the N533Q mutant.
C-type inactivation in voltage-dependent K + channels is considered to involve structural changes around the selectivity filter [9,10,22]. Actually, the 'open-inactivated' structure of the Streptomyces lividans K + channel, KcsA, has been crystallized and proposed as a model of C-type inactivation [11]. The structure revealed that conformation changes in the selectivity filter mainly restrict ion occupancy in the filter, although the backbone of the outer pore loop region shows only a modest conformational rearrangement. In addition, it has been recently reported that the transition between the inactivated and conductive states of the KcsA channel is accompanied by changes in local outer vestibule dynamics without great conformational changes [23]. These results suggest that the local structural dynamics of the outer pore loop region are involved in C-type inactivation of the KcsA channel. In the present study, we demonstrated that the asparagine residue at position 533 (N533) modulates the voltage-dependent inactivation of the PKD2L1 channel. The outer pore loop region of the mouse PKD2L1 channel may contribute to the voltage-dependent inactivation, in consistent with C-type inactivation of these voltage-dependent K + channels.
Although the outer pore loop region is considered to be involved in C-type inactivation of voltage-dependent K + channels, the mechanism of inactivation remains obscure. A model in which the selectivity filter is dilated on the extracellular side during C-type inactivation, however, has been recently proposed [22]. The dilation changes K + occupancy at the outer side of the selectivity filter, and results in C-type inactivation. At present, the detailed gating mechanism of the PKD2L1 channel is unknown. Quite recently, the three-dimensional structure of human PKD2, which is an isoform of PKD2L1, has been revealed [24][25][26]. In the structural analysis, the PKD2 channel has a tetrameric architecture similar to many voltage-gated K + channels [27,28] and TRP channels [29][30][31]. However, the large extracellular domain between transmembrane segments 1 and 2 (termed polycystin domain [24] or TOP domain [25,26]) is unique to the PKD2 channel. The extracellular domain covers the pore without blocking the ion-conducting pathway. An aspartate residue (D523) within the selectivity filter in the human PKD2L1 channel has been previously demonstrated to be responsible for Ca 2+ -dependent inactivation of the channel [32]. Consistently, the PKD2 structure in the multi-ion mode revealed that Ca 2+ ion is bound to a high-affinity site formed by the conserved aspartate residue (N643) in the selectivity filter of PKD2, suggesting that the binding of Ca 2+ ion into the entrance of the selectivity filter might contribute to Ca 2+ -dependent inactivation [26]. On the other hand, we have here demonstrated that mutation of an asparagine residue (N533Q) in the outer pore loop region of the PKD2L1 channel modified its voltage-dependent inactivation. The extracellular domain (between transmembrane segments 1 and 2) of PKD2 directly interacts with its outer pore loop region and the interaction is considered to regulate the channel gating. Interestingly, the asparagine residue (N653) in the outer pore loop region of PKD2 corresponding to the N533 residue of PKD2L1 [see Fig. 2 in 24, Fig. S1 in 25] is demonstrated to be important for the formation of a tight tripartite complex with the extracellular domain (between transmembrane segments 1 and 2) and membrane lipids in single-ion mode of the PKD2 structure [26]. Therefore, this interaction through the N533 residue of PKD2L1 might regulate the voltage-dependent inactivation. In the present study, replacement of the asparagine residue (N533) in PKD2L1 with alanine or glutamine lacked the voltage-dependent inactivation (Figs 2 and 3). Judging from properties of the side chains, proper length rather than functional similarity of the side chains is suggested to be essential for the voltage-dependent inactivation. The length difference in the side chains might trigger deformation of the tripartite complex in the PKD2L1 channel. It is intriguing that the width of the selectivity filter of PKD2 is changed in the state-dependent manner [26]. Similar to C-type inactivation of the voltage-dependent K + channels, it seems possible that a small dilation of the selectivity filter by changing the interaction of the extracellular domain (between transmembrane segments 1 and 2) with the outer pore loop region might cause the voltage-dependent inactivation of the PKD2L1 channel.
We also demonstrated that the N533Q mutants, in contrast to the N531Q mutants, showed no pH dropor temperature drop-induced increase in channel activity (Figs 4 and 5). We consider that no changes in the channel activity in the N533Q mutants are due to lack of the voltage-dependent inactivation. It is currently unknown how extracellular alkalization and heating affect the voltage-dependent inactivation of the PKD2L1 channels. However, it could be possible that structural changes in the large extracellular domain (between transmembrane segments 1 and 2) triggered by extracellular alkalization and heating allosterically regulate the gating of the PKD2L1 channels through the interaction between the extracellular domain and the outer pore loop region.
The present study suggests that local structural changes in the outer pore loop region of the PKD2L1 channel regulate its voltage-dependent inactivation. This is the first report showing that the N533 residue in the outer pore loop region contributes to the voltage-dependent inactivation of PKD2L1, a member of the TRP channel family. Most recently, an important role for the PKD2L1 channel in spinal cerebrospinal fluid contacting (CSF-c) neurons of the mouse brainstem has been reported. The activity of PKD2L1 channels expressed in CSF-c neurons regulates their excitability, suggesting that the PKD2L1 channel acts as a spike generator [33]. In addition, the CSF-c neurons have a sensory function being capable of sensing the movement and composition of CSF. Intriguingly, extracellular alkalization has been demonstrated to increase the firing frequency of the CSF-c neurons, in which firing is mediated by the activity of the PKD2L1 channel [33,34]. Therefore, it would be interesting to investigate how the voltage-dependent inactivation of the PKD2L1 channels contributes to the sensory functions of the CSF-c neurons.