Characterization and Chemical Synthesis of Cm39 (α-KTx 4.8): A Scorpion Toxin That Inhibits Voltage-Gated K+ Channel KV1.2 and Small- and Intermediate-Conductance Ca2+-Activated K+ Channels KCa2.2 and KCa3.1

A novel peptide, Cm39, was identified in the venom of the scorpion Centruroides margaritatus. Its primary structure was determined. It consists of 37 amino acid residues with a MW of 3980.2 Da. The full chemical synthesis and proper folding of Cm39 was obtained. Based on amino acid sequence alignment with different K+ channel inhibitor scorpion toxin (KTx) families and phylogenetic analysis, Cm39 belongs to the α-KTx 4 family and was registered with the systematic number of α-KTx 4.8. Synthetic Cm39 inhibits the voltage-gated K+ channel hKV1.2 with high affinity (Kd = 65 nM). The conductance–voltage relationship of KV1.2 was not altered in the presence of Cm39, and the analysis of the toxin binding kinetics was consistent with a bimolecular interaction between the peptide and the channel; therefore, the pore blocking mechanism is proposed for the toxin–channel interaction. Cm39 also inhibits the Ca2+-activated KCa2.2 and KCa3.1 channels, with Kd = 502 nM, and Kd = 58 nM, respectively. However, the peptide does not inhibit hKV1.1, hKV1.3, hKV1.4, hKV1.5, hKV1.6, hKV11.1, mKCa1.1 K+ channels or the hNaV1.5 and hNaV1.4 Na+ channels at 1 μM concentrations. Understanding the unusual selectivity profile of Cm39 motivates further experiments to reveal novel interactions with the vestibule of toxin-sensitive channels.


Introduction
Potassium ion (K + ) channels regulate several vital physiological processes such as membrane potential, cell volume control, calcium signaling, cell proliferation and action potential firing in both excitable and non-excitable cells [1,2]. Voltage-gated potassium (K V ) channels constitute the largest family of K + channels with other smaller families of K + channels that are stimulated either by calcium ions or those that are constitutively active [3,4]. K V channels are assembled by four homologous subunits, with each subunit consisting of six transmembrane segments (TM); four TM segments are part of the voltage sensing

Purification of Cm39 and Primary Sequence Determination
A number of peptide toxins were isolated from the venom of C. margaritatus following a three-step purification scheme. A comprehensive description about the purification and proteomic analysis of these peptides was reported in our earlier publication [29]. The first step was size-exclusion chromatography, resulting in three fractions FI, FII and FIII. Fraction FII, which typically contains the toxic components of the venom, was subjected to ion-exchange chromatography (IEC) using a carboxy-methylcellulose (CMC) column. The resulting 10 subfractions (FII.1-10) were individually separated by HPLC using a C 18 column and molecular weights (MW) were analyzed for the principal peaks by ESI-MS as shown in the previous publication [29]. A peptide of 3980.2 Da MW was found in sub-fraction FII.8 ( Figure 1A) and named "Cm39" after the scorpion C. margaritatus and its MW. Cm39 toxin eluted at a retention time (R T ) of 25.2 min from C 18 column as indicated in HPLC chromatogram ( Figure 1A). The complete sequence of Cm39 was obtained by direct automatic Edman degradation. It has 39 residues with six cysteines and three putative disulfide bridges ( Figure 1B).
Toxins 2023, 15, 41 3 of 21 of its pharmacological activity on ten different K + channels and two Na + channels by single-cell electrophysiology (patch-clamp) assay. Comparison of the amino acid sequence of Cm39 with other known KTxs and phylogenetic analysis showed great resemblance with α-KTx 4 members.

Purification of Cm39 and Primary Sequence Determination
A number of peptide toxins were isolated from the venom of C. margaritatus following a three-step purification scheme. A comprehensive description about the purification and proteomic analysis of these peptides was reported in our earlier publication [29]. The first step was size-exclusion chromatography, resulting in three fractions FI, FII and FIII. Fraction FII, which typically contains the toxic components of the venom, was subjected to ion-exchange chromatography (IEC) using a carboxymethylcellulose (CMC) column. The resulting 10 subfractions (FII.1-10) were individually separated by HPLC using a C18 column and molecular weights (MW) were analyzed for the principal peaks by ESI-MS as shown in the previous publication [29]. A peptide of 3980.2 Da MW was found in sub-fraction FII.8 ( Figure 1A) and named "Cm39" after the scorpion C. margaritatus and its MW. Cm39 toxin eluted at a retention time (RT) of 25.2 min from C18 column as indicated in HPLC chromatogram ( Figure 1A). The complete sequence of Cm39 was obtained by direct automatic Edman degradation. It has 39 residues with six cysteines and three putative disulfide bridges ( Figure 1B).  Peptides were eluted with a linear gradient of solution A (0.12% TFA in water) to 60% solution B (0.1% TFA in acetonitrile) over 60 min (black dashed line). Cm39 was identified in the peak that eluted at 25.2 min as indicated with red arrow. (B) Full length amino acid sequence of Cm39 determined by automatic Edman degradation. Cysteine residues are bold.

Chemical Synthesis
In order to obtain a suitable quantity of the Cm39 peptide for functional assay we synthesized it using the solid phase method from Merrifield [31]. The peptide was purified using HPLC, and the main component was analyzed ( Figure 2). The MW of synthetic Cm39 (sCm39) obtained was 3980.68 Da and the amino acid sequence determined by automatic Edman degradation showed that the peptide contained the expected sequence.

Chemical Synthesis
In order to obtain a suitable quantity of the Cm39 peptide for functional assay we synthesized it using the solid phase method from Merrifield [31]. The peptide was purified using HPLC, and the main component was analyzed ( Figure 2). The MW of synthetic Cm39 (sCm39) obtained was 3980.68 Da and the amino acid sequence determined by automatic Edman degradation showed that the peptide contained the expected sequence (data not shown).

Figure 2.
Purification of the synthetic Cm39. After re-folding, sCm39 was purified using C18 preparative column (238TP1022 Vydac) with a linear gradient of solution A (0.12% TFA in water) to 40% solution B (0.1% TFA in acetonitrile) over 80 min (black dashed line) and at a flow rate of 5mL/min. Inset shows the re-purification of the main fraction (RT 54.8, indicated with red asterisk) in a C18 analytical column (218TP54 Vydac) with a linear gradient from 0 to 60% B solution in 60 min (black dashed line).

Pharmacological Properties of Cm39
The primary sequence of Cm39 has a significant resemblance with several scorpion toxins that are highly potent inhibitors of voltage-gated K + channels. We, therefore, aimed at testing the pharmacological activity of sCm39 on a battery of various potassium and sodium ion channels. Among the tested channels were six members of human voltagegated K + (hKV) channels from Shaker family, hKV1.1-hKV1.6 ( Figure 3A-F) and three members of Ca 2+ -activated potassium channel; hKCa2.2 (SK2, Figure 3I), the small-conductance Ca 2+ -activated channel; hKCa3.1 (IKCa1, SK4. Figure 3J), the intermediate-conductance Ca 2+activated channel expressed in T lymphocytes; and mKCa1.1 (BK, Slo1, MaxiK, Figure 3H), the large-conductance voltage-and Ca 2+ -activated channel of mice. In addition, we also screened the effect of Cm39 on hKV11.1 (hERG1, Figure 3G), a voltage-gated cardiac K + channel, and two human voltage-gated sodium (hNaV) channels, hNaV1.4 ( Figure 3K) and hNaV1.5 ( Figure 3L), expressed in skeletal and cardiac muscles, respectively. These ion Figure 2. Purification of the synthetic Cm39. After re-folding, sCm39 was purified using C 18 preparative column (238TP1022 Vydac) with a linear gradient of solution A (0.12% TFA in water) to 40% solution B (0.1% TFA in acetonitrile) over 80 min (black dashed line) and at a flow rate of 5mL/min. Inset shows the re-purification of the main fraction (R T 54.8, indicated with red asterisk) in a C 18 analytical column (218TP54 Vydac) with a linear gradient from 0 to 60% B solution in 60 min (black dashed line).

Pharmacological Properties of Cm39
The primary sequence of Cm39 has a significant resemblance with several scorpion toxins that are highly potent inhibitors of voltage-gated K + channels. We, therefore, aimed at testing the pharmacological activity of sCm39 on a battery of various potassium and sodium ion channels. Among the tested channels were six members of human voltage-gated K + (hK V ) channels from Shaker family, hK V 1.1-hK V 1.6 ( Figure 3A-F) and three members of Ca 2+ -activated potassium channel; hK Ca 2.2 (SK2, Figure 3I), the small-conductance Ca 2+activated channel; hK Ca 3.1 (IK Ca 1, SK4. Figure 3J), the intermediate-conductance Ca 2+activated channel expressed in T lymphocytes; and mK Ca 1.1 (BK, Slo1, MaxiK, Figure 3H), the large-conductance voltage-and Ca 2+ -activated channel of mice. In addition, we also screened the effect of Cm39 on hK V 11.1 (hERG1, Figure 3G), a voltage-gated cardiac K + channel, and two human voltage-gated sodium (hNa V ) channels, hNa V 1.4 ( Figure 3K) and hNa V 1.5 ( Figure 3L), expressed in skeletal and cardiac muscles, respectively. These ion channels were heterologously expressed in the CHO or HEK cell lines except for K V 1.3. Human peripheral T lymphocytes. They were stimulated with Phytohemagglutinin A (PHA) to increase the K V 1.3 channel expression, and Ca 2+ -free intracellular solution was used to avoid K Ca 3.1 channel activation [30]. Appropriate depolarization protocols were used to record ionic currents in voltage-clamped cells (see Figure 3 insets and Material and Methods for details). Freshly dissolved sCm39 in extracellular solution was applied at 1 µM concentration on the cell using a custom-built micro-perfusion system at~200 µL/min flow rate. The complete exchange of solution in the recording chamber was ensured by frequently using either fully reversible blockers, such as TEA + for K V 1.1, K V 1.3, K V 1.6 and mK Ca 1.1; Charybdotoxin (ChTx) for K V 1.2; and apamin for K Ca 2.2 or specific solutions that allow testing the solution exchange: HK-150 solution for K V 1.4 and K V 1.5; and Na + -free solution for Na V 1.4 and Na V 1.5 as positive controls. The decrease in the current upon perfusion of the cell with these controls confirmed the nature of ion channel expressed, especially when specific inhibitors were used, and that full solution exchange was achieved.  with high extracellular (150 mM) K + to reduce the K + driving force or ChTx: charybdotoxin, TEA + : tetraethylammonium chloride, Apamin, Tram34 and E4031, known inhibitors of the appropriate channels as indicated). Voltage protocols are displayed above the current traces in each panel. For extracellular and intracellular solution composition see Materials and Methods section.
(M) Remaining current fraction (RCF, I/I 0 ) values were calculated as the ratio of the peak currents in the presence (I) or absence (I 0 ) of 1 µM sCm39 at steady-state block (for K V 1.2, K Ca 2.2 and K Ca 3.1) or after 3−5 min of application of toxin. In the case of K Ca 2.2 and K Ca 3.1, the peak currents were measured at +48 mV (time point 148 ms) of the ramp. Bars with individual data points (triangles) represent RCF values determined on individual cells, bars in red color highlight ion channels that are potential targets of Cm39. Error bars indicate the mean ± SEM (n = 3−5).
We found that sCm39 at 1 µM concentration did not exhibit any significant effect on 9 of the 12 tested channels as shown in panels A-L of Figure 3. The summarized selectivity data in Figure 3M indicates that the three channels which were affected by sCm39 were K V 1.2, K Ca 2.2 and K Ca 3.1. At equilibrium block, 1 µM of Cm39 reduced 95% of K V 1.2, 55% of K Ca 2.2 and 86% of K Ca 3.1 currents ( Figure 3M) whereas decrease in the current of other channels was <3%. The selectivity profile indicates that sCm39 is selective for K V 1.2, K Ca 2.2 and K Ca 3.1 over the other channels investigated in this study.

Mechanism of K V 1.2 Block
For comprehensive characterization of Cm39 activity on K V 1.2, we determined the concentration-response of current inhibition by Cm39, the kinetic parameters of its binding and assessed the effect on voltage dependence of steady-state activation. The whole-cell K V 1.2 currents were recorded in CHO cells using depolarization pulses to +50mV from −120mV holding potential (V h ) every 15 s. Since the activation kinetics of K V 1.2 are highly variable [32], the duration of pulses was set on a cell-by-cell basis between 15-300 ms to allocate sufficient time for reaching maximum currents at the test potentials. Figure 4A shows K V 1.2 current traces recorded sequentially in the same cell in the absence of toxin (control trace) and after perfusing the cell with different concentrations of sCm39 until the equilibrium block. sCm39 reduces the K V 1.2 currents in a concentration-dependent manner, i.e., 5, 20, 50, 100, 250 and 500 nM inhibited approximately 5, 22, 43, 62, 78 and 87% of current at equilibrium block. The remaining current fractions (RCF = I/I 0 , where I 0 is the peak current in the absence of the toxin, and I is the peak current at equilibrium block in the presence of sCm39 at a given concentration) were calculated and plotted as a function of toxin concentration. The Hill equation was fit to the data points of the concentration-response relationship (see Materials and Methods), which resulted in a dissociation constant (K d ) of 65 nM with a Hill coefficient of 0.96 ( Figure 4B).
The development and recovery from the block at various concentrations of sCm39 are displayed in Figure 4C, where normalized peak currents were plotted against the time. The block of K V 1.2 by various concentrations of sCm39 was fully reversible and kinetics of both toxin association and dissociation were very fast. Figure 4D shows the analysis of the kinetic parameters of K V 1.2 current inhibition by sCm39. Single-exponential decay functions were fitted to the normalized peak currents in the presence of each sCm39 concentration ( Figure 4C, data points in red shades) to determine the time constant for the onset of block (τ on , wash-in or association time constant). The dissociation time constant for the relief from block (τ off , wash-out time constant) was obtained by fitting a singleexponential rising function to the normalized peak currents in the wash-out procedure ( Figure 4C, data point in black empty circles). Assuming a simple bimolecular interaction between the toxin and the channel, time constants for the onset and recovery from the block can be stated as follows: where k on represents the second-order association rate constant, k off indicates the first-order dissociation rate constant and [toxin] is the sCm39 concentration. The 1/τ on and 1/τ off values were plotted as a function of toxin concentration where 1/τ on rises linearly with the Cm39 concentration; however, the dissociation rate (1/τ off ) remains constant with a k off value of 0.019 ± 0.0013 s −1 ( Figure 4D), as also described for the ChTx binding to a Shaker channel [33] and for the sVmKTx interaction with K V 1.3 [34]. The second-order rate constant of association (k on ) was determined by fitting the 1/τ on data points using linear regression with k off as the y-intercept. The slope of the regression line corresponds to the k on of 2.15 × 10 −4 ± 2.19 × 10 −5 nM −1 s −1 (r 2 = 0.97; Figure 4D). The dissociation constants (K d = k off /k on ) calculated from the block kinetic parameters resulted in 87.8 nM, which is comparable to the K d value determined from equilibrium block ( Figure 4B). Mostly, scorpion toxins inhibit the K V channels by physically occluding the pore region of the channel thereby preventing the flow of K + ions. There are some other peptide toxins which bind to the voltage-sensor domains of Kv channels and significantly shift the voltagedependence of steady-state activation towards depolarized potentials which results in decreased K + current [35,36]. To confirm the blocking mechanism of Cm39 we determined the conductance-voltage (G-V) relationship for K V 1.2 in the absence and presence of the toxin. Whole-cell currents of K V 1.2 expressing CHO cells were recorded by applying the 300 ms long depolarization pulses ranging from −70 to +80 in 10 mV steps from V h of −120 mV. Due to the highly variable activation properties of K V 1.2 [32] we restricted the analysis to the current records which showed similar gating mode. The conductance values were calculated for each voltage step, normalized to the maximum conductance and plotted as a function of membrane potential (E m ) as shown in Figure 4E. The fitting of averaged data points with the Boltzmann sigmoidal function resulted in the superimposed solid lines ( Figure 4E), demonstrating that the presence of 65 nM of sCm9 did not affect the voltage dependence of steady-state activation of K V 1.2 channels. The midpoint voltage (V 50 ) of the G-V relationship in the control solution (−1.0 ± 4.8, n = 4) was statistically similar to the V 50 value at equilibrium block with 65 nM sCm39 (1.6 ± 4.0, n = 4) as shown in Figure 4F. These results suggest that Cm39 is not modifying the gating of K V 1.2, rather it binds to the pore region of the channel.

Cm39 Inhibits Small-Conductance/Intermediate Conductance Ca 2+ -Activated Channels with Nanomolar Affinity
Cm39 also targets Ca 2+ -activated potassium channels K Ca 2.2 and K Ca 3.1 as described in the selectivity profile experiments above; therefore, we studied the effect of Cm39 in concentration-dependent manner on these channels. The whole-cell K + currents of CHO cells transiently expressing K Ca 2.2 or K Ca 3.1 channels were recorded sequentially by applying the 150 ms long voltage ramp to +50 mV from V h of −120 mV every 10 s in the presence of control solution or upon perfusing the cell with different concentrations of sCm39. Normalized peak currents were calculated and plotted as function of time as shown in Figure 5A for K Ca 2.2 and in Figure 5C for K Ca 3.1 in the presence of the toxin at various concentrations and upon wash-out. Figure 5A,C show that block of both channels by sCm39 was fully reversible, and toxin association and dissociation kinetics were very fast. The development and almost full recovery from the equilibrium block took place in 1-3 depolarization pulses separated by 10 s. RCF values were determined at equilibrium block in the presence of various concentrations of sCm39 and fitted with the Hill equation (see Materials and Methods for details). The resulting dissociation constant (K d ) values and Hill coefficients (H) were K d = 502 nM, H = 0.7 for K Ca 2.2 ( Figure 5B) and K d = 57.7 nM, H = 0.68 for K Ca 3.1 ( Figure 5D). The affinity of Cm39 for K Ca 3.1 is~10 times higher than for K Ca 2.2.   were very fast. The development and almost full recovery from the equilibrium block took place in 1-3 depolarization pulses separated by 10 s. RCF values were determined at equilibrium block in the presence of various concentrations of sCm39 and fitted with the Hill equation (see Materials and Methods for details). The resulting dissociation constant (Kd) values and Hill coefficients (H) were Kd = 502 nM, H = 0.7 for KCa2.2 ( Figure 5B) and Kd = 57.7 nM, H = 0.68 for KCa3.1 ( Figure 5D). The affinity of Cm39 for KCa3.1 is ~10 times higher than for KCa2.2.

Comparative Sequence and Phylogenetic Analyses
Amino acid sequence alignment of the different KTx families with Cm39 showed that Cm39 is most closely related to the α-KTx family. Within the α-KTx family, the subfamilies with the highest average identity percentages were: α-KTx 4 with 60.8%, α-KTx 2 with 51.4%, α-KTx 23 with 44.9%, α-KTx 1 with 43.9%, α-KTx 3 and 19 with 43.8%, and α-KTx 12 with 43.5% with the most conserved region being the C-terminus of the peptides. Table 1 shows the alignment between Cm39, the α-KTx 4 subfamily and a representative sequence of the other α-KTxs subfamilies. As can be seen, except for α-KTx 4.2, the other members of subfamily 4 show more than 60% identity to Cm39. Amino acid sequence alignment of the different KTx families with Cm39 showed that Cm39 is most closely related to the α-KTx family. Within the α-KTx family, the subfamilies with the highest average identity percentages were: α-KTx 4 with 60.8%, α-KTx 2 with 51.4%, α-KTx 23 with 44.9%, α-KTx 1 with 43.9%, α-KTx 3 and 19 with 43.8%, and α-KTx 12 with 43.5% with the most conserved region being the C-terminus of the peptides. Table  1 shows the alignment between Cm39, the α-KTx 4 subfamily and a representative sequence of the other α-KTxs subfamilies. As can be seen, except for α-KTx 4.2, the other members of subfamily 4 show more than 60% identity to Cm39.  Phylogenetic analysis was performed by comparing the amino acid sequence of Cm39 with representative sequences of the most related α-KTxs subfamilies. Cm39 was clustered with all other α-KTx 4 toxins, indicating that Cm39 belongs to this subfamily and that α-KTx 4 toxins are the ancestor of subfamilies 19 and 23 ( Figure 6).
Amino acid sequence alignment of the different KTx families with Cm39 showed that Cm39 is most closely related to the α-KTx family. Within the α-KTx family, the subfamilies with the highest average identity percentages were: α-KTx 4 with 60.8%, α-KTx 2 with 51.4%, α-KTx 23 with 44.9%, α-KTx 1 with 43.9%, α-KTx 3 and 19 with 43.8%, and α-KTx 12 with 43.5% with the most conserved region being the C-terminus of the peptides. Table  1 shows the alignment between Cm39, the α-KTx 4 subfamily and a representative sequence of the other α-KTxs subfamilies. As can be seen, except for α-KTx 4.2, the other members of subfamily 4 show more than 60% identity to Cm39.  Phylogenetic analysis was performed by comparing the amino acid sequence of Cm39 with representative sequences of the most related α-KTxs subfamilies. Cm39 was clustered with all other α-KTx 4 toxins, indicating that Cm39 belongs to this subfamily and that α-KTx 4 toxins are the ancestor of subfamilies 19 and 23 ( Figure 6).
Amino acid sequence alignment of the different KTx families with Cm39 showed that Cm39 is most closely related to the α-KTx family. Within the α-KTx family, the subfamilies with the highest average identity percentages were: α-KTx 4 with 60.8%, α-KTx 2 with 51.4%, α-KTx 23 with 44.9%, α-KTx 1 with 43.9%, α-KTx 3 and 19 with 43.8%, and α-KTx 12 with 43.5% with the most conserved region being the C-terminus of the peptides. Table  1 shows the alignment between Cm39, the α-KTx 4 subfamily and a representative sequence of the other α-KTxs subfamilies. As can be seen, except for α-KTx 4.2, the other members of subfamily 4 show more than 60% identity to Cm39.  Phylogenetic analysis was performed by comparing the amino acid sequence of Cm39 with representative sequences of the most related α-KTxs subfamilies. Cm39 was clustered with all other α-KTx 4 toxins, indicating that Cm39 belongs to this subfamily and that α-KTx 4 toxins are the ancestor of subfamilies 19 and 23 ( Figure 6).
Amino acid sequence alignment of the different KTx families with Cm39 showed that Cm39 is most closely related to the α-KTx family. Within the α-KTx family, the subfamilies with the highest average identity percentages were: α-KTx 4 with 60.8%, α-KTx 2 with 51.4%, α-KTx 23 with 44.9%, α-KTx 1 with 43.9%, α-KTx 3 and 19 with 43.8%, and α-KTx 12 with 43.5% with the most conserved region being the C-terminus of the peptides. Table  1 shows the alignment between Cm39, the α-KTx 4 subfamily and a representative sequence of the other α-KTxs subfamilies. As can be seen, except for α-KTx 4.2, the other members of subfamily 4 show more than 60% identity to Cm39.

Discussion
In this study, we described the purification, determination of primary structure, chemical synthesis and characterization of pharmacological activities of a new peptide toxin, Cm39, from the venom of Centruroides margaritatus. Cm39 consists of 37 residues with six cysteines. In electrophysiological evaluation, it inhibited human KV1.2 and intermediate-conductance Ca 2+ -activated (KCa3.1) channel with nanomolar affinity (Kd = 65 nM

Discussion
In this study, we described the purification, determination of primary structure, chemical synthesis and characterization of pharmacological activities of a new peptide toxin, Cm39, from the venom of Centruroides margaritatus. Cm39 consists of 37 residues with six cysteines. In electrophysiological evaluation, it inhibited human K V 1.2 and intermediateconductance Ca 2+ -activated (K Ca 3.1) channel with nanomolar affinity (K d = 65 nM and 58 nM, respectively). Additionally, it also showed a blocking effect on a small-conductance Ca 2+ -activated channel (K Ca 2.2) with comparatively low affinity (K d = 502 nM). However, Cm39 did not show any effect on several other ion channels tested in this study including five other subtypes of voltage-gated K + channels (K V 1.1, K V 1.3, K V 1.4, K V 1.5, K V 1.6 and K V 11.1), two subtypes of Na V channels (Na V 1.4 and Na V 1.5) and the large-conductance Ca 2+ -activated channel (mK Ca 1.1).
When the amino acid sequence was compared, Cm39 showed high similarity to α-KTxs. Within the α-KTxs family, peptides of subfamily 4 showed a higher percentage of identity with Cm39 (~60%). Most of the peptides in subfamily 4 are very similar (percentage identity ranging from 78 to 97%) and share common pharmacological targets, including K V 1.3, K V 1.2 and K Ca channels. Like the subfamily 4 peptides, Cm39 also showed an inhibitory effect on K V 1.2, K Ca 2.2 and K Ca 3.1 channels, but not on the K V 1.3 channel. Not all subfamily 4 toxins have been tested on the K V 1.3 channel, but for those that were tested (α-KTxs 4.1, 4.4 and 4.6), very similar K d values were found (19.8 nM, 16 nM and 10.7 nM, respectively) [37][38][39]. It has been reported that there are several features in the primary structure of toxins that confer selectivity for the K V 1.3 or K V 1.2 channel. A higher overall positive charge, a greater number of basic residues at the C-terminus, a Met residue two positions downstream of the dyad Lys and a residue other than Tyr in the dyad confer selectivity for the K V 1.3 channel. Contrastingly, a lower overall charge, a lower number of basic amino acids, a Tyr residue in the dyad and an Ile residue two positions downstream of the dyad Lys, confer selectivity for the K V 1.2 channel [39]. Cm39 fully satisfies some of these characteristics. First, Cm39 has an overall charge of 3.6, which is lower compared to the other toxins mentioned (> 4.8). Second, while the other toxins have two Lys residues at the C-terminus, Cm39 has a change from Lys to Thr (T34). Third, the Met residue in Cm39 has been replaced by Ile (I31), and the Tyr residue is located in the essential dyad. All these together could explain the selectivity of Cm39 towards the K V 1.2 channel over the K V 1.3 channel. On the other hand, the effect of subfamily 4 peptides on the K Ca 2.2 channel has been demonstrated only for the α-KTxs 4.2. The molecular mechanism of interaction with the K Ca 2 (small-conductance K + channel, SK) was attributed to the conserved motif "RXCQ" found in toxins such as Lei-I and PO5 [40], in addition to apamin, which has a blocking effect in the picomolar range [41]. However, this motif is not present in α-KTxs 4.2 or Cm39, thus the mechanism of interaction is likely to be different. Mutagenesis studies on α-KTx 4.2 have shown that Arg6 and, to a lesser extent, Arg9 are important residues for the interaction of the toxin with the SK channels [42]. In Cm39, Arg6 is replaced by Lys, which also results in a positive charge, and Arg9 is replaced by Ser, although a Lys is present at position 8. These changes could explain the difference in K d value between Cm39 and α-KTxs 4.2 (502 nM and 80 nM, respectively). In addition, these amino acid changes (K6 and K8) are conserved in the other subfamily 4 toxins, and further investigation may help to clarify the molecular mechanism of the interaction of subfamily 4 toxins with the K Ca 2.2 channel. None of the subfamily 4 toxins have been reported to block the K Ca 3.1 channel. However, based on sequence similarity with toxin Cm39, it is possible that other members of the α-KTx 4 family also inhibit the K Ca 3.1 channel. There are two other peptides that have been reported as high affinity blockers of the K Ca 3.1 channel, charybdotoxin (ChTx, α-KTx-1.1) and maurotoxin (MTX, α-KTx-6.2), whose K d values are 5 nM and 1 nM, respectively [43]. The K d value of these toxins is not far from that found for Cm39 (58 nM); however, none of the toxins is selective for this channel (like Cm39). ChTx shows effects on K V 1.2 and K V 1.3 channels [44] and MTX inhibits the Kv1.2 channel [45]. A new toxin such as Cm39 could help to identify the motifs involved in selectivity toward the K Ca 3.1 channel by expanding the repertoire of peptides with which comparative analysis can be performed to find amino acids likely involved in this interaction. Thus, amino acid sequence and phylogenetic analyses, where clearly Cm39 is clustered within subfamily 4, and the pharmacological properties of Cm39 indicate that it is a novel member of the α-KTx 4 subfamily. Therefore, the Cm39 toxin was registered with the systematic number of α-KTx 4.8, and its primary amino acid sequence will appear in the UniProt Knowledgebase under the accession number C0HM65.
Generally, scorpion toxins interact at two different regions of K V channels to modify their function. Typical functional dyad (lysine and tyrosine) bearing toxins bind to the extracellular vestibule in such a way that lysine goes deep into selectivity filter, thereby plugging the pore of the channel [33,[46][47][48]. Other toxins interact with the voltage-sensing domain (VSD) causing the significant change in gating of the channel [35,36]. We found that the voltage dependence of steady-state-activation of K V 1.2 channel was completely insensitive to the presence of Cm39, thereby clarifying that it does not interfere with the VSD (Figure 4E and F). On the other hand, we propose that Cm39 could bind to the pore region and Lys27 is responsible for obstructing the pore of K V 1.2 channel. In addition, the binding kinetics analysis showed a simple bimolecular interaction between Cm39 and K V 1.2 channel as described previously for typical pore blocker toxins such as charybdotoxin and margatoxin [33]. An apparent first-order association rate showed direct correlation with the Cm39 concentration, being faster at higher toxin concentration; however, the firstorder dissociation rate was not affected by the change in toxin concentration ( Figure 4D). The dissociation constant (K d ) for K V 1.2 calculated by the k off /k on ratio (87.8 nM) is comparable to the K d value determined by fitting the Hill equation to the concentration dependence of current inhibition (65 nM). Moreover, we observed that the block of all Cm39-sensitive channels (K V 1.2, K Ca 2.2 and K Ca 3.1) was fully reversible and association and dissociation rates of Cm39 were fast. Although Cm39 quickly binds to the channel and blocks the conductance presumably by protruding Lys27 into the selectivity filter, unfavorable interactions between residues of toxin and the external vestibule of the channel destabilizes the toxin binding, thereby drastically shorting the residence time of Cm39 on the channel. Goldstein and Miller have showed this by generating several mutants of ChTx with very high off rates [33]. Similarly, the double substitution in AnTx[N17A/F32T] resulted in unfavorable interaction with the pore of the K V 1.3 channel and exhibited a significantly fast dissociation rate as compared to the wild-type toxin [49].
To date, a number of scorpion-derived peptide toxins are known to block K + channels with great affinity. However, these native toxins target more than one channel which compromises their therapeutic use in the management of channelopathies [23]. For example, Urotoxin (α-KTx-6.21) [50], MgTx (α-KTx-2.2) [51] and ChTx (α-KTx-1.1) [52] inhibit several K V 1.x channel subtypes and Ca 2+ -activated channels with high affinity ranging from picomolar to nanomolar. Vm24 (α-KTx-23.1) is a high affinity blocker of the K V 1.3 channel and its effect is >1500-fold selective for K V 1.3 over the K V 1.1, K V 1.2 and K Ca 3.1 channels [53]. The selectivity of these potential peptides can be enhanced through protein engineering. For instance, by substituting two residues (N17A/F32T) in Anuroctoxin (α-KTx 6.12) it gained 16,000-fold selectivity for K V 1.3 over K V 1.2 while retaining its native affinity for K V 1.3 [49]. Leiurotoxin I (α-KTx-5.1, Scyllatoxin, isolated from Leiurus quinquestriatus hebraeus) is a potent inhibitor of small-conductance Ca 2+ -activated potassium channel K Ca 2.x subtypes. The Lei-Dab 7 mutant, a less potent but highly selective blocker for K Ca 2.2, was designed by replacing single residue with unnatural amino acid diaminobutonoic acid [40]. Recently, our group developed an analog of Vm24 by mutating a single amino acid (K32E) which is still a high affinity blocker of K V 1.3 while gaining~9000-fold selectivity over K V 1.2, while it was insensitive to K V 1.1 and K Ca 3.1 at a 2.5 µM concentration [34].
As discussed, Cm39 blocks K V 1.2 and K Ca 3.1 with similar affinities and shows comparatively less affinity for the K Ca 2.2 channel (Figures 4B and 5). Cm39 at a 1µM concentration did not block several other channels included in this study (Figure 3). This pattern of ion channel inhibition influences the potential biological application of Cm39 in its native form.
For example, K V 1.2 is a typical ion channel in the central nervous system which is protected by the blood-brain barrier and thus, restricts the access of peptide blockers to the channels. On the other hand, conjugation of the therapeutic peptides with BBB shuttle peptides is known to increase the applicability of these peptides [54,55] For example, inhibition of K V 1.2 may have beneficial effects in epilepsy associated with gain of function mutations of K V 1.2 [14]. Inhibition of K Ca 3.1, especially in combination with targeting K Ca 2.2, can be interesting for influencing atrial fibrillation (AF). It was shown recently that inhibition if K Ca 3.1 or K Ca 2.2 in isolation may reduce AF [56][57][58]. It may be an interesting scenario to target both ion channels simultaneously, and based on our results Cm39 would be a suitable candidate for a dual-action drug candidate. In addition, selectivity of Cm39 for a particular ion channel over the other channels can be improved through peptide engineering after revealing the candidate residues which interact with the channel vestibule.

Conclusions
In summary, we characterized a new member of the α-KTx 4 family from the venom of Colombian scorpion Centruroides margaritatus. The synthetic Cm39 peptide blocks K V 1.2 and Ca 2+ -activated potassium channels (K Ca 2.2 and K Ca 3.1) with nanomolar affinities while it does not affect several other K + and Na + channels. Cm39 targets two distinct families of potassium channel (voltage-gated and Ca 2+ -activated), and it bears >15-fold selectivity for K V 1.2 over the K V 1.3 channel. These properties of Cm39 make it a good candidate for designing a specific inhibitor of a certain potassium channel to use as a therapeutic drug or as a tool for physiological studies.

Isolation and Amino Acid Sequence Determination of Native Cm39
A detailed procedure of venom preparation and purification of peptide toxins including Cm39 (MW 3980 Da) from C. margaritatus venom was described previously [29]. Concisely, the soluble venom was first fractionized using the Sephadex G-50 column in 20 mM ammonium acetate buffer of pH 4.7 at a flow rate of 2 mL/min. Then, a toxic peptide containing fraction was subjected to ion exchange chromatography (IEC) using a carboxy-methylcellulose (CMC) column and components were eluted with a linear gradient 0-100% of 500 mM ammonium acetate buffer in 200 min at 2 mL/min flow rate. Lastly, IEC fractions were further purified by high-performance liquid chromatography (HPLC) using analytical C 18 reverse-phase column (Vydac). To elute the peptides from the column, a linear gradient of 0-60% of solution B (0.1% TFA in acetonitrile) in solution A (0.12% trifluoroacetic acid in water) was run for 60 min at a 1 mL/min flow rate. HPLC fractions were collected manually, and after vacuum drying stored at −20 • C until further analysis. A sample from a single peak of HPLC was analyzed in the LCQ Fleet mass spectrometer coupled with an electrospray ionization (Thermo Fisher Scientific Inc., San Jose, CA, USA) to determine the molecular mass of pure peptide. The amino acid sequence of the peptide was revealed by automated Edman degradation using Biotech PPSQ-31A Protein Sequencer equipment (Shimadzu Scientific Instruments, Inc., Columbia, MD, USA) according to the procedure as described for other components from the same venom [29,30]. First, pure native peptide was loaded directly for sequencing. Then, a reduced and alkylated sample of the same peptide was sequenced to identify the cysteine residues.

Chemical Synthesis and Folding of sCm39
The peptide Cm39 was chemically synthesized by using the Merrifield technique [31]. The synthesis was prepared manually using the standard Fmoc-based solid phase technique on NovaSyn TGA resin (0.25 mmol/g resin). HBTU and HOBT were used as coupling reagents. A three-fold excess of Fmoc amino acids was added during each coupling cycle. The Fmoc group was removed with 20% piperidine in DMF. Unreacted or deblocked free amines were monitored through the ninhydrin test in each cycle of the peptide synthesis. After cleavage and deprotection for 1.5 h at room temperature with reagent K (TFA 84%, pH 7.35. In the HK-150 and Na + -free bath solution bath all Na + was substituted by K + or Choline-Cl, respectively, other ingredients remained unchanged. In the various TEA + -containing solutions, Na + was substituted by tetraethylammonium-Cl in equimolar concentration ( Figure 3). The bath solution for K Ca 2.2 and K Ca 3.1 consisted of (in mM) 145 L-Aspartic Na + salt, 5 KCl, 2.5 CaCl 2 , 1 MgCl 2 , 5.5 glucose and 10 HEPES, pH 7.4 and for K V 11.1 (in mM) 140 Choline-Cl, 5 KCl, 2 MgCl 2 , 2 CaCl 2 , 10 HEPES, 20 glucose and 0.1 CdCl 2 , pH 7.35. The osmolarity of bath solutions was ranging between 302 to 308 mOsm/L. A total of 0.1 mg/mL bovine serum albumin (BSA, Sigma-Aldrich, Budapest, Hungary) was added to all the bath solutions before the patch-clamp assay to avoid toxin adsorption to the plastic surfaces of the perfusion system. Internal or pipette solution for recording K V 1.x and mK Ca 1.1 currents consisted of (in mM) 140 KF, 2 MgCl 2 , 1 CaCl 2 , 11 EGTA and 10 HEPES, pH 7.22. The composition of the internal solution for sodium channels was (in mM) 10 NaCl, 105 CsF, 10 HEPES and 10 EGTA, pH 7.2; for K V 11.1 (in mM) 140 KCl, 2 MgCl 2 , 10 HEPES and 10 EGTA, pH 7.3; and for K Ca channels (in mM) 150 K-Aspartate, 5 HEPES, 8.5 CaCl 2 , 2 MgCl 2 and 10 EGTA, pH 7.22. The osmolarity of internal solutions was~295 mOsm/L.

Patch-Clamp Recording Conditions
Whole-cell currents were recorded in voltage-clamped cells following standard protocols [60] by using a Multiclamp 700B amplifier connected to a computer with Axon Digidata1440 digitizer and for data acquisition, Clampex 10.7 software was used (Molecular Devices). Micropipettes were pulled from GC150F-7.5 borosilicate capillaries (Harvard Apparatus) resulting in electrodes having 3-5 MΩ resistance in the bath solution. Current traces were lowpass-filtered through the built-in analog 4-pole Bessel filters of the amplifiers and sampling frequency was set at 4-50 kHz, at least twice the filter cutoff frequency. Current records were discarded when the leak current at V h was >10% of peak current at the depolarization potential. Recordings were carried out at room temperature (20-25 • C). The cell was perfused with control and test solutions by using a gravity-flow perfusion system and excess bath solution was removed constantly by vacuum suction. Voltage protocols were not corrected for changes in the liquid-junction potentials (typically < 5 mV) upon perfusion of the recording chamber with different K + or Na + concentration solutions (see above), as these controls were merely used to confirm solution exchange.
In general, the V h was set at −120 mV and the depolarization pulses were applied every 15 s except when indicated. For recording the K + currents from K V 1.x, 15-300 ms long voltage pulses to +50 mV were applied. To record the K V 1.2 currents for the G-V relationship, CHO cells were depolarized to voltages ranging from -60mV to +100 mV in steps of 10 mV every 15 s. For K V 11.1 channels, currents were evoked with voltage steps to +20 mV for 1.25 s from a V h of -80 mV followed by a step to -40 mV for 2 s, during which peak currents were recorded and pulses were delivered every 30 s. mK Ca 1.1 currents were recorded by applying voltage steps to +100 mV for 600 ms from a V h of -100 mV. To record the K Ca 2.2 and K Ca 3.1 channel currents, a 150 ms long voltage ramp to +50 mV from −120 mV was applied every 10 s. Na + currents through Na V 1.x were recorded with 15 ms long voltage steps to 0 mV every 10 s.

Patch-Clamp Data Analyses
Patch-clamp data was analyzed using the pClamp 10.7 software package (Molecular Devices). Generally, before analysis current records were corrected for ohmic leakage and digitally filtered with 3-point boxcar smoothing. Each data point in concentrationdependent dose curves represents the mean of ≥3 individual records and these data points were fitted with the Hill equation: where RCF is the remaining current fraction (RCF = I/I 0 , where I 0 is the peak current in the absence of the toxin, and I is the peak current at equilibrium block at a given concentration of toxin), [Tx] is the concentration of the toxin, K d is the dissociation constant and H is the Hill coefficient. When the voltage ramp protocol was used to record the Ca 2+ -activated K + currents, the peak currents were measured at +48 mV (time point 148 ms) of the ramp. Moreover, in the case of K Ca 2.2 the RCF values at more than 1µM concentration of Cm39 were normalized to the RCF value obtained at 100 nM of Apamin (Smartox Biotechnology, France) to eliminate the influence of any contaminating current. To determine the voltage dependance of steady-state activation of K V 1.2 current, peak conductance (G) for each test potential was calculated from peak current (I 0 ) at a test potential (E m ) and K + reversal potential (E K ) by using the chord-conductance equation G = I 0 /(E m − E K ). The G values were normalized to the maximum value and plotted as a function of test potential and data points were fitted with Boltzmann sigmoidal equation: where G norm is the normalized conductance, V is the test potential, V 50 is the midpoint voltage and k is the slope factor of the function.
To study the Cm39 binding kinetics, the association and dissociation rate constants (k on , k off ) for the K V 1.2 channel were determined following the procedure as described previously [30,34]. Normalized peak currents (I norm = I t /I 0 , where I t is peak current in the presence of the toxin at time t and I 0 is peak current in the absence of toxin) were plotted against the time and data points during the wash-in and wash-out procedures and were fitted with single-exponential function as shown below, to determine the time constants for association (τ on ) and dissociation (τ off ) of toxin.
Then, k on and k off values were calculated from these time constants assuming a simple bimolecular interaction between the channel and toxin, and by using equations shown below, also demonstrated previously in detail [30,33]:

Statistics
For statistical analyses and graph plotting, GraphPad Prism software (version 8.0.1) was used. Data points were always given as mean ± SEM. Student's t test with Mann-Whitney rank sum test was performed for pairwise comparison.