Inhibitory Effects of 2-Aminoethoxydiphenyl Borate (2-APB) on Three KV1 Channel Currents

2-Aminoethoxydiphenyl borate (2-APB), a boron-containing compound, is a multitarget compound with potential as a drug precursor and exerts various effects in systems of the human body. Ion channels are among the reported targets of 2-APB. The effects of 2-APB on voltage-gated potassium channels (KV) have been reported, but the types of KV channels that 2-APB inhibits and the inhibitory mechanism remain unknown. In this paper, we discovered that 2-APB acted as an inhibitor of three representative human KV1 channels. 2-APB significantly blocked A-type Kv channel KV1.4 in a concentration-dependent manner, with an IC50 of 67.3 μM, while it inhibited the delayed outward rectifier channels KV1.2 and KV1.3, with IC50s of 310.4 μM and 454.9 μM, respectively. Further studies on KV1.4 showed that V549, T551, A553, and L554 at the cavity region and N-terminal played significant roles in 2-APB’s effects on the KV1.4 channel. The results also indicated the importance of fast inactivation gating in determining the different effects of 2-APB on three channels. Interestingly, a current facilitation phenomenon by a short prepulse after 2-APB application was discovered for the first time. The docked modeling revealed that 2-APB could form hydrogen bonds with different sites in the cavity region of three channels, and the inhibition constants showed a similar trend to the experimental results. These findings revealed new molecular targets of 2-APB and demonstrated that 2-APB’s effects on KV1 channels might be part of the reason for the diverse bioactivities of 2-APB in the human body and in animal models of human disease.


Introduction
Boron, often as boric acid, is a trace and necessary element for bone metabolism, inflammation attenuation, wound healing, modulating hormone levels, enhancing magnesium absorption, attenuating oxidative stress, and ameliorating cisplatin-induced peripheral neuropathy. 2-Aminoethoxydiphenyl borate (2-APB), a boron-containing compound, exerts effects on neurons, adaptive and innate immunity, and muscle cells, and it provides a cytoprotective effect against reactive oxygen. The molecular pharmacology of 2-APB reveals its complex actions through multiple targets including ion channels, transporters, and enzymes. 2-APB has the potential to act as both as a drug and as a precursor of drugs. Interest in 2-APB and its derivatives regarding the design of bioactive molecules keeps increasing in recent years [1,2].
The ion channel is the third-largest family of signaling molecules and the secondlargest druggable target group. 2-APB has been shown to act as either a blocker or an
The concentration-response curves of K V 1.2 and K V 1.3 did not seem sigmoidal due to the responses not reaching the maximum yet.

Characteristics of the Inhibition of 2-APB on Kv1.4
To understand the potential inhibition mechanisms of 2-APB, we further investigated the 2-APB effect on KV1.4 since 2-APB was the most potent in inhibiting Kv1.4 among three KV1 channels. A two-pulse protocol ( Figure 4A) was set to examine the recovery of the inactivation. Currents were evoked by the same amplitude, a 40 mV prepulse, and a test pulse from a holding potential at −80 mV. The interval between these two pulses was varied, and both peak currents evoked from the test pulse (I2) and the prepulse (I1) were

Characteristics of the Inhibition of 2-APB on Kv1.4
To understand the potential inhibition mechanisms of 2-APB, we further investigated the 2-APB effect on K V 1.4 since 2-APB was the most potent in inhibiting Kv1.4 among three K V 1 channels. A two-pulse protocol ( Figure 4A) was set to examine the recovery of the inactivation. Currents were evoked by the same amplitude, a 40 mV prepulse, and a test pulse from a holding potential at −80 mV. The interval between these two pulses was varied, and both peak currents evoked from the test pulse (I 2 ) and the prepulse (I 1 ) were recorded. The ratio of I 2 /I 1 was plotted against the intervals and fitted by a single exponential. The time constant was 683.0 ± 80.5 ms and 624.2 ± 78.8 ms before and after 2-APB (100 µM) application (n = 5, p > 0.05) ( Figure 4B). These data indicate that 2-APB had no effects on the recovery of the inactivation.  The effects on the onset of inactivation by 2-APB were also investigated. A two-pulse protocol ( Figure 4A) was used. Currents were evoked by the same amplitude, a 40 mV prepulse, and a test pulse from a holding potential at −80 mV. The duration of the prepulse was varied, while the interval between these two pulses was fixed at 20 ms. Both peak currents evoked from the test pulse (I2) and the prepulse (I1) were recorded. The ratio of I2/I1 was plotted against the intervals and fitted by a single exponential. The onset of slow inactivation was measured by a prepulse duration from 32 ms to 2048 ms. The time constant was 74.5 ± 11.5 ms and 67.0 ± 8.2 ms before and after 2-APB (100 μM) application (n = 4, p > 0.05) ( Figure 4D). The onset of fast inactivation was measured by a prepulse dura-  The effects on the onset of inactivation by 2-APB were also investigated. A two-pulse protocol ( Figure 4A) was used. Currents were evoked by the same amplitude, a 40 mV prepulse, and a test pulse from a holding potential at −80 mV. The duration of the prepulse was varied, while the interval between these two pulses was fixed at 20 ms. Both peak currents evoked from the test pulse (I2) and the prepulse (I1) were recorded. The ratio of I2/I1 was plotted against the intervals and fitted by a single exponential. The onset of slow inactivation was measured by a prepulse duration from 32 ms to 2048 ms. The time constant was 74.5 ± 11.5 ms and 67.0 ± 8.2 ms before and after 2-APB (100 μM) application (n = 4, p > 0.05) ( Figure 4D). The onset of fast inactivation was measured by a prepulse duration from 2 ms to 120 ms. However, the fit started at 8 ms. The time constancy was 36.2 ± points to the current amplitudes facilitated at a 2 ms prepulse length. (D). The onset of the slow inactivation of K V 1.4 channels in the absence or presence of 2-APB. The prepulse length varied from 32 ms to 1024 ms.
The effects on the onset of inactivation by 2-APB were also investigated. A two-pulse protocol ( Figure 4A) was used. Currents were evoked by the same amplitude, a 40 mV prepulse, and a test pulse from a holding potential at −80 mV. The duration of the prepulse was varied, while the interval between these two pulses was fixed at 20 ms. Both peak currents evoked from the test pulse (I 2 ) and the prepulse (I 1 ) were recorded. The ratio of I 2 /I 1 was plotted against the intervals and fitted by a single exponential. The onset of slow inactivation was measured by a prepulse duration from 32 ms to 2048 ms. The time constant was 74.5 ± 11.5 ms and 67.0 ± 8.2 ms before and after 2-APB (100 µM) application (n = 4, p > 0.05) ( Figure 4D). The onset of fast inactivation was measured by a prepulse duration from 2 ms to 120 ms. However, the fit started at 8 ms. The time constancy was 36.2 ± 2.1 ms before and was 32.9 ± 0.9 ms after 2-APB (100 µM) application (n = 5, p > 0.05) ( Figure 4C). These findings show that 2-APB had no effects on the onset of inactivation. Figure 4C illustrates that the 2 ms length prepulse facilitated currents evoked by the test pulse. The facilitation of the amplitudes of currents (I 2 /I 1 ) was 1.21 ± 0.03 vs. 2.80 ± 0.32 before and after 2-APB application (n = 5). It seemed like the facilitation was more than doubled after 2-APB application. However, it was actually because 2-APB suppressed the currents of the prepulse (I 1 ) more than those of the test pulse (I 2 , Figure 5A,B). At this stage, the kinetics of the activation of the current were well fitted by the Boltzmann equation rather than a single exponential ( Figure 5C). The average time to activate half of the maximal current was delayed by 0.6 ± 0.15 ms by 30 µM 2-APB (p < 0.01, n = 10) ( Figure 5D). Meanwhile, the slope, which measured the speed of the activation, was also significantly slowed (p < 0.01, n = 10). It was 0.34 ± 0.12 pA/ms with 30 µM 2-APB, compared to 0.24 ± 0.02 pA/ms without 2-APB ( Figure 5D). The data suggested that 2-APB delayed the K V 1.4 activation at a 2 ms length of the prepulse.
Molecules 2023, 28, x FOR PEER REVIEW 7 of 18 doubled after 2-APB application. However, it was actually because 2-APB suppressed the currents of the prepulse (I1) more than those of the test pulse (I2, Figure 5A,B). At this stage, the kinetics of the activation of the current were well fitted by the Boltzmann equation rather than a single exponential ( Figure 5C). The average time to activate half of the maximal current was delayed by 0.6 ± 0.15 ms by 30 μM 2-APB (p < 0.01, n = 10) ( Figure  5D). Meanwhile, the slope, which measured the speed of the activation, was also significantly slowed (p < 0.01, n = 10). It was 0.34 ± 0.12 pA/ms with 30 μM 2-APB, compared to 0.24 ± 0.02 pA/ms without 2-APB ( Figure 5D). The data suggested that 2-APB delayed the KV1.4 activation at a 2 ms length of the prepulse. Considering that 2-APB exerts a much higher potency against KV1.4 channel currents, . Enlarged example prepulse current traces with (Black) and without 2-APB (Red). (C). Averages of current amplitudes at different times were analyzed by the Boltzman equation. represents data after 2-APB application. represents data before 2-APB application. Current amplitudes were normalized to the maximal peak currents in prepulses without 2-APB. represents the normalized current amplitudes with 2-APB, which were normalized to the maximal peak currents in the prepulse with 2-APB. (D). Bar graphs show the half-current activation time (left), which is relative to that without 2-APB treatment, and slopes (right) with and without 2-APB. * represents p < 0.01.

Mutations at the Pore Region
Considering that 2-APB exerts a much higher potency against K V 1.4 channel currents, we further studied its inhibitory mechanism. Most known small-molecule inhibitors of K V channels bind a water-filled cavity below the selectivity filter that is formed by residues located at the base of the selectivity filter and by pore-lining amino acids of the inner (S6) helices [13]. We hypothesized that 2-APB might have a specific interaction site(s) on the K V 1.4 channel and that the 2-APB interaction site(s) with the K V 1.4 channel might locate in the channel cavity region. To identify the site(s), we conducted alanine scanning by making a series of site mutations, except for the A553 site that was mutated to A553V, in the channel cavity region (G548A, V549A, L550A, T551A, I552A, A553V, L554A, P555A, V556A, P557A, V558A, I559A, V560A). These mutations were transiently expressed in CHO cell lines. Despite some of these mutations having no currents or too few currents elicited for analysis, 100 µM 2-APB was used for studying the mutations of measurable currents. We found that four mutations (V549A, T551A, A553V, and L554A) significantly attenuated the 2-APB inhibition of the K V 1.4 channel currents (Figure 6), whereas the other mutations had no significant effects (Table 1). This indicated that 2-APB might regulate K V 1.4 channel activity through interactions with the four residues.

N-Terminal Truncation
K V 1.4, as a typical A-type channel, has the fastest kinetic inactivation among K V 1s, and 2-APB also has the strongest inhibition on K V 1.4. It is interesting to investigate whether there is an impact on the inhibition of 2-APB if the fast inactivation is removed. Because the N-terminal of K V 1.4 serves as the "ball" to deliver the fast inactivation by blocking the path of ions at the bundle cross of the S6 region [14], 1-61 amino acids at the N-terminal were removed to make a truncation mutation. Compared with the other mutations, the Nterminal-removed mutation was the least inhibited by 2-APB. Furthermore, 10, 30, 100, 300, and 1000 µM 2-APB inhibited peak amplitudes of K V 1.4 currents by 1.5 ± 1.7%, 9.0 ± 0.9%, 23.1 ± 2.5%, 49.1 ± 2.9%, and 77.1 ± 3.8%, respectively (n = 5-6) (Figure 7). The IC 50 value was 310.4 ± 16.6 µM (n = 6).
V556A, P557A, V558A, I559A, V560A). These mutations were transiently expressed in CHO cell lines. Despite some of these mutations having no currents or too few currents elicited for analysis, 100 μM 2-APB was used for studying the mutations of measurable currents. We found that four mutations (V549A, T551A, A553V, and L554A) significantly attenuated the 2-APB inhibition of the KV1.4 channel currents (Figure 6), whereas the other mutations had no significant effects (Table 1). This indicated that 2-APB might regulate KV1.4 channel activity through interactions with the four residues.

N-Terminal Truncation
KV1.4, as a typical A-type channel, has the fastest kinetic inactivation among KV1s, and 2-APB also has the strongest inhibition on KV1.4. It is interesting to investigate whether there is an impact on the inhibition of 2-APB if the fast inactivation is removed. Because the N-terminal of KV1.4 serves as the "ball" to deliver the fast inactivation by blocking the path of ions at the bundle cross of the S6 region [14], 1-61 amino acids at the N-terminal were removed to make a truncation mutation. Compared with the other mutations, the N-terminal-removed mutation was the least inhibited by 2-APB. Furthermore, 10, 30, 100, 300, and 1000 μM 2-APB inhibited peak amplitudes of KV1.4 currents by 1.5 ± 1.7%, 9.0 ± 0.9%, 23.1 ± 2.5%, 49.1 ± 2.9%, and 77.1 ± 3.8%, respectively (n = 5-6) (Figure 7). The IC50 value was 310.4 ± 16.6 μM (n = 6).  As the K V 1.4 channels do not inactivate yet at 2 ms ( Figure 5A,B), the activation delay is more probably correlated with the fast inactivation. Thus, the onset of the inactivation of the N-terminal truncation mutation was investigated. Since the N-terminal truncation mutation has lost inactivation, there was no clear inactivation detected in the first 32 ms ( Figure 8A). Clear facilitation was discovered at at least 2, 4, and 8 ms with or without 2-APB. The facilitation of current amplitudes (I 2 /I 1 ) at a 2 ms length prepulse was 2.23 ± 0.32 vs. 4.68 ± 0.69 before and after 30 µM 2-APB application (p < 0.05, n = 4) ( Figure 8B). Similar to the effect on WT, the facilitation was more than doubled after 2-APB application. The average time needed to activate half of the maximal current was delayed by 0.4 ± 0.09 ms by 30 µM 2-APB (p < 0.05, n = 4). Meanwhile, the slope was 0.33 ± 0.02 pA/ms with 30 µM 2-APB, compared to 0.42 ± 0.07 pA/ms without 2-APB (n = 4, p > 0.05) (Figure 9). mutation of KV1.4 channel currents inhibited by 2-APB. The percentage inhibitions against t APB concentration were analyzed by the Hill equation.
As the KV1.4 channels do not inactivate yet at 2 ms ( Figure 5A,B), the activation d is more probably correlated with the fast inactivation. Thus, the onset of the inactiva of the N-terminal truncation mutation was investigated. Since the N-terminal trunca mutation has lost inactivation, there was no clear inactivation detected in the first 3 ( Figure 8A). Clear facilitation was discovered at at least 2, 4, and 8 ms with or witho APB. The facilitation of current amplitudes (I2/I1) at a 2 ms length prepulse was 2.23 ± vs. 4.68 ± 0.69 before and after 30 μM 2-APB application (p < 0.05, n = 4) ( Figure 8B). Sim to the effect on WT, the facilitation was more than doubled after 2-APB application average time needed to activate half of the maximal current was delayed by 0.4 ± 0.0 by 30 μM 2-APB (p < 0.05, n = 4). Meanwhile, the slope was 0.33 ± 0.02 pA/ms with 30 2-APB, compared to 0.42 ± 0.07 pA/ms without 2-APB (n = 4, p > 0.05) (Figure 9).  The data suggested that the truncation of N-terminal reduced the 2-APB inhibition on K V 1.4 channels but does not affect the short prepulse facilitation.

Docking of the Interactions between 2-APB and Three Channels
To further examine the possible interaction at the atomic level between the ligand 2-APB and hK V 1.3 and hK V 1.4 and the template structure of rK V 1.2 channels, homology modeling of hK V 1.3 and hK V 1.4 and molecular docking of protein-ligands complexes were carried out. The sequence of the central pore of hK V 1.3 and hK V 1.4 was aligned with that of rK V 1.2 according to the previous studies [15][16][17]. The hK V 1.3, hK V 1.4, and rK V 1.2 share a~90% sequence identity in the pore domain. Hence, using a Swiss-model workspace, we constructed a reliable 3D structure of hK V 1.3 and hK V 1.4 composed of S5, S6, and P-loop according to the crystal structure of rK V 1.2. The quality of the refined models was then evaluated by the Ramachandran plot, which suggests that the Phi/Psi angles of most residues (~90%) are within the reasonable ranges. The flexible docking protocol was carried out to explore the molecular basis of the interaction between rK V 1.2, hK V 1.3, and hK V 1.4 and 2-APB. Based on the docking analysis, the phenolic hydroxyl group of 2-APB showed an interaction with T401 (K V 1.2 (2A79)), A473 (K V 1.3 model), and V549 (K V 1.4 model), respectively (Figures 10-12). The best binding poses were chosen based on the binding energy score, the inhibition constant, the VDW_HB desolv_energy, and the ligand efficiency, and the surrounding residues of ligands may have an essential impact on the binding affinity between hK V 1.4 and 2-APB. According to the computational docking results, the superposition of the binding structures for the inspected ligand is shown in Figures 10-12  ○ represents data after 2-APB application, and • represents data before 2-APB application. Current amplitudes were normalized to the maximal peak currents in prepulses without 2-APB. □ represents the normalized current amplitudes with 2-APB, which were normalized to the maximal peak currents in a prepulse with 2-APB. (D). Bar graphs show the half-current activation time (left), relative to the one without 2-APB, and a slope (right) with and without 2-APB. * represents p < 0.05; "n.s." means "not significant".
The data suggested that the truncation of N-terminal reduced the 2-APB inhibition on KV1.4 channels but does not affect the short prepulse facilitation.

Docking of the Interactions between 2-APB and Three Channels
To further examine the possible interaction at the atomic level between the ligand 2-APB and hKV1.3 and hKV1.4 and the template structure of rKV1.2 channels, homology . Enlarged example prepulse current traces with (Black) and without 2-APB (Red). (C). Average current amplitude at different times analyzed by the Bolzman equation. represents data after 2-APB application, and • represents data before 2-APB application. Current amplitudes were normalized to the maximal peak currents in prepulses without 2-APB. represents the normalized current amplitudes with 2-APB, which were normalized to the maximal peak currents in a prepulse with 2-APB. (D). Bar graphs show the half-current activation time (left), relative to the one without 2-APB, and a slope (right) with and without 2-APB. * represents p < 0.05; "n.s." means "not significant". ergy score, the inhibition constant, the VDW_HB desolv_energy, and the ligand efficiency, and the surrounding residues of ligands may have an essential impact on the binding affinity between hKV1.4 and 2-APB. According to the computational docking results, the superposition of the binding structures for the inspected ligand is shown in Figures 10-12. The inhibition constants (Ki) predicted were 218.87 μM, 350.52 μM, and 118.52 μM for rKV1.2, hKV1.3, and hKV1.4, respectively (Table 2). This shows a similar trend with the experimental data.

Discussion
Our study showed that 2-APB could significantly block three representative human K V 1 channels. The most potent inhibition against A-type K V 1.4 currents and a relatively weaker inhibition against two delayed outward rectifier channels were discovered. The mutations V549A, T551A, A553V, L554A, and N-terminal removal significantly attenuated 2-APB's effects on K V 1.4 channel currents. 2-APB did not affect the onset and recovery of K V 1.4 inactivation but significantly slowed the initial activation kinetics of the channels. The importance of fast inactivation gating in determining the different 2-APB effects on the two types of channels was revealed. Interestingly, a current facilitation phenomenon by short prepulses was discovered for the first time.
Functional K V 1 channels not only exist throughout both central and peripheral nervous systems but can also be expressed in peripheral tissues, including the cardiovascular and the immune system. Among them, K V 1.2 is mostly in the central nervous system, and K V 1.3 is more prominent in peripheral tissues. K V 1.4, an A-type K V channel, is widely distributed in excitable cells of mammalian tissues and exists in the cardiac ventricular endocardium [18,19]. The inhibitory effects of 2-APB on three channels can be used to explain the effects of 2-APB on the human body and animal models of human disease.
Inhibition on K V 1.3 ought to be a reason for modulating adaptive and innate immunity by 2-APB, because K V 1.3, as a drug target for autoimmune diseases, widely exists in immune cells. Inhibition on K V 1.2 and K V 1.4 should play a role in 2-APB's effects on neurons, smooth muscle cells, and cardiomyocytes [2]. Inhibition on both the transient and sustained voltage-activated potassium current of Limulus ventral photoreceptors comprising a delayed outward rectifier K + current and a rapidly inactivating conductance could also result from 2-APB's effects on Kv1 channels [7].
2-APB showed the most potent effects on K V 1.4 among three channels. Previous studies showed that 2-APB's inhibitory effects on K V channels of guinea pig arteriole cells were more marked on the fast component than they were on the slow component [8]. Our results could provide an explanation.
Among the antagonists of K V 1.4 channels, 2-APB, with an IC 50 of 67.3 µM, is comparable to the effects of fluoxetine and La 3+ , and its inhibition is 4 and 100 times more potent than 4-aminopyridine and TEA, respectively [18]. The data indicated that the inhibitory potential of 2-APB is quite considerable. 2-APB has been widely used as a tool in many studies. It is worth noticing that 2-APB has also been used in several systems (e.g., rat dorsal root ganglia) that endogenously express K V 1 as their major potassium background channels [20]. We think that any application of 2-APB over 10 µM at cellular levels should take into consideration the potential inhibition of K V 1.4.
The V549A, T551A, A553V, and L554A mutations significantly attenuate 2-APB's effects on K V 1.4, and these four sites were conserved with K V 1.2 (V399, T401, A403, L404) and K V 1.3 (V469, T471, A473, L474) channels. It is reasonable to speculate that these sites might also play a key role in the inhibition of K V 1.2 and K V 1.3 by 2-APB. The variable IC 50 s in K V 1.2 and K V 1.3 indicated that other diverse regions or sites of these channels might be involved in the interaction between 2-APB and channels.
Fast inactivation is the characteristic of K V 1.4 that distinguishes it from other K V 1.x channels [21]. N-terminal removal will disrupt the fast inactivation of the channel. Our data of the N-terminal deleted channel showed the weakest inhibition by 2-APB, suggesting that the fast inactivation plays a vital role in the inhibition of 2-APB. Interestingly, N-terminal truncation caused the IC 50 on K V 1.4 to be nearly the same as that of K V 1.2. Moreover, A553V mutation also disrupted the fast inactivation kinetics ( Figure 6C) and mostly significantly attenuated 2-APB inhibition, which is consistent with the N-terminal truncation data. These findings provide additional evidence on the key role of the fast inactivation gate in the 2-APB effects on K V 1.4. Considering that fast inactivation is the major difference between K V 1.4 and the other K V 1 channels, the fast inactivation characteristic of K V 1.4 might determine the different inhibitory potencies of 2-APB against K V 1.4 and the other two K V 1channels.
Our data showed that 2-APB delayed the kinetic activation of K V 1.4 when the prepulse length was 2 ms. The N-terminal truncated mutation altered neither this effect nor the amount of time delayed. We do not know whether other K V 1.4 inhibitors have the same effect, or, in other words, whether this effect is 2-APB-specific. We suspect that 2-APB might have a direct impact on the channel opening, but further evidence is needed.
In the docking analysis, 2-APB had an interaction with V549 and did not bind with the other three key residues in the K V 1.4 channel. 2-APB could also interact with T401 in Kv1.2 and A473 in K V 1.3, which are conserved with T551 and A553 of the K V 1.4 channel. The docking predictions were consistent with the inhibitory effects on three channels by 2-APB.

Site-Directed and N-Terminal Truncation Mutations of K V 1.4
Mutation PCR was performed with primers synthesized by Sangon biotech (Shanghai, China) using the KOD DNA polymerase (TOYOBO, Osaka, Japan). PCR products were digested by the DpnI restriction enzyme (TAKARA, Dalian, China), purified, and transformed into DH5α E. coli competent cells. Additionally, 1-61 amino acids were deleted for N-terminal truncation mutation. After sequencing, plasmids with the desired mutations were extracted and transiently transfected into Chinese hamster ovary (CHO) cells using Lipofectamine™ 2000.

Data Recording
CHO cells with K V 1.2, 1.3, or 1.4 stably expressed were routinely cultured in the solution with 10% fetal bovine serum (FBS), 1% P/S (100U Penicillin and 0.1 mg/mL streptomycin), and 100 µg/mL G418. Cells not connected with other cells were voltageclamped using the PC505B (Warner Instrument Corporation) patch clamp amplifier in the whole-cell configuration. Electrodes ranged from 2 to 3 MΩ in resistance. The voltage clamp data were filtered at 2 kHz and digitized at 100 or 150 ms/point. Voltage protocols were generated and analyzed by Clampex and Clampfit patch clamp software (Version 10.4, Axon Instruments). The recordings from cells were carried out at room temperature (25 ± 1 • C) [22].

Statistical Analysis
Electrophysiological data were analyzed by Clampfit software (Version 10.4, Axon Instruments) and Origin 8.0 (OriginLab Corporation, Northampton, MA, USA). All data were expressed as the mean ± S.E.M using Student's t-tests with statistical significance (p < 0.05). The IC 50 value was obtained by fitting the concentration-dependent data to the following Hill equation: I (%) = 1/{1 + (IC 50 /[D]) n }. In the equation, I% is the percentage inhibition of current amplitudes; IC 50 is the concentration of the half-maximal inhibition; [D] is the concentration of a compound; and n is the Hill coefficient.

Homology Modeling
The pore region sequences of hK V 1.3 and hK V 1.4 were retrieved from the Uniprot database (accession number: hK V 1.3-P22001, hK V 1.4-P22459), and multiple sequence alignment was carried out with the sequence of rat K V 1.2. Since the crystal structure of the human voltage-gated potassium channels K V 1.3 and K V 1. 4 has not yet been determined, the homology model of both channels was constructed by the Swiss-model based on the crystal structure of the rat K V 1.2 channel (PDB entry: 2A79) [14,23]. The residues Met288 to Thr421, forming the S5 pore helix and S6 in rat K V 1.2, correspond to the region from Met358 to Thr491 in hK V 1.3, and Met438 to Thr571 in hK V 1.4 was selected as the template for homology modeling. The sequence identity of these regions between the rat K V 1.2, hK V 1.3, and hK V 1.4 was~90%, which can enable us to construct a reliable homology model based on the high-resolution crystal structure of rat K V 1.2. As the crystal structure of rat K V 1.2 contains only one subunit, the transformation matrices of the structural coordinate file were employed to generate the missing subunits of rat K V 1.2, creating the fourfold symmetry required to build the hK V 1.3 and hK V 1.4 channel homo-tetramer. The Ramachandran plot further assessed the quality of the refined model.

Induced Fit Docking
The compound 2-APB, the modeled target proteins hK V 1.3 and hK V 1.4, and the template structure of rat K V 1.2 were used in the molecular docking analysis by the AutoDock tools (ADT) with the MGL Tools v1.5.6rc3 program [24]. All three target protein molecules were prepared by the Python Molecule Viewer (PMV); after repairing the missing atoms, the polar hydrogen atoms and Gasteiger charges were added into the protein structures. The active binding sites were selected at the central cavity of the pore region of hK V 1.3 and hK V 1.4 in each of the four subunits and were considered as flexible according to the previous study [25]; the rest of the target protein sites were treated as a rigid body. The parameter library was loaded for the ligand molecule 2-APB, and the atoms were also considered flexible. The grid map was constructed based on the ligand atom types with a default grid spacing of 0.375 A • in the box size of 90 × 90 × 90. After the grid map calculation, the target proteins and the ligand molecule were induced for docking analysis. The default genetic algorithm parameters with the addition of AutoDock 4.2 parameters were used for building 27,000 generations with a population size of 150 individuals. The docking conformations were calculated by the Lamarckian genetic algorithm (LGA) [26]. Finally, the least scored conformation was evaluated and picked for the ligand-receptor docking analysis; the interaction images were developed by PyMOL.

Conclusions
In conclusion, this study demonstrated that 2-APB could significantly block three representative human K V 1 channels. Moreover, the potential inhibitory mechanism was also investigated. Therefore, 2-APB's effects on K V 1 channels might be part of the reason for the diverse bioactivities of 2-APB in the human body and in animal models of human disease. In other words, our study revealed new targets of 2-APB at the molecular level. More comprehensive research is needed to understand the inhibitory mechanism in more detail and more 2-APB effects on other channels.