Advertisement

Artificial pore blocker acts specifically on voltage-gated potassium channel isoform KV1.6

Open AccessPublished:September 07, 2022DOI:https://doi.org/10.1016/j.jbc.2022.102467
      Among voltage-gated potassium channel (KV) isoforms, KV1.6 is one of the most widespread in the nervous system. However, there are little data concerning its physiological significance, in part due to the scarcity of specific ligands. The known high-affinity ligands of KV1.6 lack selectivity, and conversely, its selective ligands show low affinity. Here, we present a designer peptide with both high affinity and selectivity to KV1.6. Previously, we have demonstrated that KV isoform-selective peptides can be constructed based on the simplistic α-hairpinin scaffold, and we obtained a number of artificial Tk-hefu peptides showing selective blockage of KV1.3 in the submicromolar range. We have now proposed amino acid substitutions to enhance their activity. As a result, we have been able to produce Tk-hefu-11 that shows an EC50 of ≈70 nM against KV1.3. Quite surprisingly, Tk-hefu-11 turns out to block KV1.6 with even higher potency, presenting an EC50 of ≈10 nM. Furthermore, we have solved the peptide structure and used molecular dynamics to investigate the determinants of selective interactions between artificial α-hairpinins and KV channels to explain the dramatic increase in KV1.6 affinity. Since KV1.3 is not highly expressed in the nervous system, we hope that Tk-hefu-11 will be useful in studies of KV1.6 and its functions.

      Keywords

      Abbreviations:

      MD (molecular dynamics), MS (mass spectrometry), PDB (Protein Data Bank)
      Ion channels play a crucial role in physiology, underlying signal transduction in excitable cells, muscle contraction, regulation of cell volume, release of hormones and neurotransmitters, etc. Potassium (K+) channels are the most abundant in humans with ∼80 genes encoding the pore-forming α-subunits. Voltage-gated K+ channels (KV) have six transmembrane segments (S1–S6) in their α-subunits, and functional channels contain four identical or different α-subunits. The first four transmembrane segments (S1–S4) of each α-subunit form a voltage-sensing domain, whereas the fifth and sixth transmembrane segments (S5 and S6) from all four subunits come together to form the centrally located pore domain. These transmembrane segments are joined by a so-called re-entrant P-loop, which contains a short pore (P) helix and the selectivity filter region. The major function of KV channels is to provide the repolarization stage of the action potential (
      • Hille B.
      Ion Channels of Excitable Membranes.
      ).
      KV1.6 is a member of the Shaker-related subfamily of voltage-gated K+ channels (KV1) encoded by the KCNA6 gene in humans or Kcna6 in mice or rats. This isoform is widely expressed in the nervous system representing one of the major K+ channels found in the brain and peripheral neurons (
      • Zeisel A.
      • Hochgerner H.
      • Lönnerberg P.
      • Johnsson A.
      • Memic F.
      • van der Zwan J.
      • et al.
      Molecular architecture of the mouse nervous system.
      ). KV1.6 is also expressed in muscles and nonelectroexcitable cells and tissues including the ophthalmic artery (
      • Manicam C.
      • Staubitz J.
      • Brochhausen C.
      • Grus F.H.
      • Pfeiffer N.
      • Gericke A.
      The gatekeepers in the mouse ophthalmic artery: endothelium-dependent mechanisms of cholinergic vasodilation.
      ) and nephrons (
      • Carrisoza-Gaytán R.
      • Salvador C.
      • Diaz-Bello B.
      • Escobar L.I.
      Differential expression of the Kv1 voltage-gated potassium channel family in the rat nephron.
      ). Additionally, there is evidence that this isoform contributes to forming heteromeric K+ channels with other KV1 subunits (
      • Shamotienko O.G.
      • Parcej D.N.
      • Dolly J.O.
      Subunit combinations defined for K+ channel Kv1 subtypes in synaptic membranes from bovine brain.
      ,
      • Dodson P.D.
      • Barker M.C.
      • Forsythe I.D.
      Two heteromeric Kv1 potassium channels differentially regulate action potential firing.
      ). Changes in KCNA6 expression are linked to some pathologies, for instance, amyotrophic lateral sclerosis (
      • Gunasekaran R.
      • Narayani R.S.
      • Vijayalakshmi K.
      • Alladi P.A.
      • Shobha K.
      • Nalini A.
      • et al.
      Exposure to cerebrospinal fluid of sporadic amyotrophic lateral sclerosis patients alters Nav1.6 and Kv1.6 channel expression in rat spinal motor neurons.
      ). However, KV1.6 is not well studied compared to some other members of the KV1 subfamily such as KV1.1–1.3. Ligands with high affinity and selectivity are needed to reveal the blind spots of KV1.6 physiological functions and its role in diseases. Unfortunately, among the available repertoire, ligands with high (nanomolar) affinity show little selectivity; and conversely, ligands with higher selectivity are active at high concentrations (hundreds of nanomoles or micromoles) (
      • Orts D.J.B.
      • Peigneur S.
      • Madio B.
      • Cassoli J.S.
      • Montandon G.G.
      • Pimenta A.M.C.
      • et al.
      Biochemical and electrophysiological characterization of two sea anemone type 1 potassium toxins from a geographically distant population of bunodosoma caissarum.
      ,
      • Orts D.J.B.
      • Peigneur S.
      • Silva-Gonçalves L.C.
      • Arcisio-Miranda M.
      • Bicudo J.E.P.W.
      • Tytgat J.
      AbeTx1 is a novel sea anemone toxin with a dual mechanism of action on shaker-type K+ channels activation.
      ,
      • ElFessi-Magouri R.
      • Peigneur S.
      • Othman H.
      • Srairi-Abid N.
      • ElAyeb M.
      • Tytgat J.
      • et al.
      Characterization of Kbot21 reveals novel side chain interactions of scorpion toxins inhibiting voltage-gated potassium channels.
      ,
      • Imperial J.S.
      • Chen P.
      • Sporning A.
      • Terlau H.
      • Daly N.L.
      • Craik D.J.
      • et al.
      Tyrosine-rich conopeptides affect voltage-gated K+ channels.
      ,
      • Vandendriessche T.
      • Kopljar I.
      • Jenkins D.P.
      • Diego-Garcia E.
      • Abdel-Mottaleb Y.
      • Vermassen E.
      • et al.
      Purification, molecular cloning and functional characterization of HelaTx1 (Heterometrus laoticus): the first member of a new κ-KTX subfamily.
      ,
      • Imperial J.S.
      • Bansal P.S.
      • Alewood P.F.
      • Daly N.L.
      • Craik D.J.
      • Sporning A.
      • et al.
      A novel conotoxin inhibitor of Kv1.6 channel and nAChR subtypes defines a new superfamily of conotoxins.
      ).
      Previously, we have developed a computational approach to the design of selective peptides that block KV channels (
      • Gigolaev A.M.
      • Kuzmenkov A.I.
      • Peigneur S.
      • Tabakmakher V.M.
      • Pinheiro-Junior E.L.
      • Chugunov A.O.
      • et al.
      Tuning scorpion toxin selectivity: switching from KV1.1 to KV1.3.
      ,
      • Tabakmakher V.M.
      • Kuzmenkov A.I.
      • Gigolaev A.M.
      • Pinheiro-Junior E.L.
      • Peigneur S.
      • Efremov R.G.
      • et al.
      Artificial peptide ligand of potassium channel KV1.1 with high selectivity.
      ,
      • Tabakmakher V.M.
      • Gigolaev A.M.
      • Peigneur S.
      • Krylov N.A.
      • Tytgat J.
      • Chugunov A.O.
      • et al.
      Potassium channel blocker crafted by α-hairpinin scaffold engineering.
      ). In particular, we noticed that plant α-hairpinins are structurally similar to some α-helical cone snail and scorpion toxins. The similarity of Tk-APM-X2 from the wheat Triticum kiharae to κ-hefutoxin-1 from the scorpion Heterometrus fulvipes allowed us to design and obtain Tk-hefu, an artificial blocker of KV1.3 (
      • Berkut A.A.
      • Usmanova D.R.
      • Peigneur S.
      • Oparin P.B.
      • Mineev K.S.
      • Odintsova T.I.
      • et al.
      Structural similarity between defense peptide from wheat and scorpion neurotoxin permits rational functional design.
      ). We then built computer models of Tk-hefu with KV1.3 and other channels, applied protein surface topology to assess the binding interfaces and molecular dynamics (MD) to analyze the interactions, and proposed amino acid replacements to enhance the affinity (
      • Berkut A.A.
      • Chugunov A.O.
      • Mineev K.S.
      • Peigneur S.
      • Tabakmakher V.M.
      • Krylov N.A.
      • et al.
      Protein surface topography as a tool to enhance the selective activity of a potassium channel blocker.
      ). As a result, we produced Tk-hefu-10, which is selective to and active against KV1.3 in the submicromolar range (
      • Tabakmakher V.M.
      • Gigolaev A.M.
      • Peigneur S.
      • Krylov N.A.
      • Tytgat J.
      • Chugunov A.O.
      • et al.
      Potassium channel blocker crafted by α-hairpinin scaffold engineering.
      ).
      Based on our previous experiments, here we analyzed the relations between amino acid substitutions and the affinity of Tk-hefu derivatives to KV1.3. We chose to retract some substitutions that we had introduced in Tk-hefu-7 and 9 to further increase the activity of the new peptide Tk-hefu-11 to KV1.3. As expected, we obtained a 2-fold increase in the affinity to that isoform. Surprisingly, we also observed a dramatic increase in affinity toward KV1.6. With an EC50 of ≈10 nM on KV1.6, Tk-hefu-11 is >500-fold more active on this isoform compared to any other Tk-hefu peptide. This new KV1.6 ligand combines both high activity and high selectivity, outperforming all other known ligands to date, and is therefore a valuable tool for neurobiology.

      Results

      Design of new Tk-hefu derivative

      Our previous works (
      • Tabakmakher V.M.
      • Gigolaev A.M.
      • Peigneur S.
      • Krylov N.A.
      • Tytgat J.
      • Chugunov A.O.
      • et al.
      Potassium channel blocker crafted by α-hairpinin scaffold engineering.
      ,
      • Berkut A.A.
      • Usmanova D.R.
      • Peigneur S.
      • Oparin P.B.
      • Mineev K.S.
      • Odintsova T.I.
      • et al.
      Structural similarity between defense peptide from wheat and scorpion neurotoxin permits rational functional design.
      ,
      • Berkut A.A.
      • Chugunov A.O.
      • Mineev K.S.
      • Peigneur S.
      • Tabakmakher V.M.
      • Krylov N.A.
      • et al.
      Protein surface topography as a tool to enhance the selective activity of a potassium channel blocker.
      ) aimed at crafting a high-affinity KV1.3 pore blocker and culminated in Tk-hefu-10 production. On the first stage, we created Tk-hefu-1 based on 3D structure similarity between α-hairpinins and κ-hefutoxin-1 (
      • Srinivasan K.N.
      • Sivaraja V.
      • Huys I.
      • Sasaki T.
      • Cheng B.
      • Kumar T.K.S.
      • et al.
      κ-Hefutoxin1, a novel toxin from the scorpion Heterometrus fulvipes with unique structure and function.
      ) (Table 1) (
      • Berkut A.A.
      • Usmanova D.R.
      • Peigneur S.
      • Oparin P.B.
      • Mineev K.S.
      • Odintsova T.I.
      • et al.
      Structural similarity between defense peptide from wheat and scorpion neurotoxin permits rational functional design.
      ). Next, we introduced one amino acid replacement and produced Tk-hefu-2 with ten times greater potency (
      • Berkut A.A.
      • Chugunov A.O.
      • Mineev K.S.
      • Peigneur S.
      • Tabakmakher V.M.
      • Krylov N.A.
      • et al.
      Protein surface topography as a tool to enhance the selective activity of a potassium channel blocker.
      ). Finally, we mutated the peptide sequentially to attain Tk-hefu-6 to 10 (
      • Tabakmakher V.M.
      • Gigolaev A.M.
      • Peigneur S.
      • Krylov N.A.
      • Tytgat J.
      • Chugunov A.O.
      • et al.
      Potassium channel blocker crafted by α-hairpinin scaffold engineering.
      ). Although Tk-hefu-10 features a dramatically improved affinity to KV1.3 (IC50 ≈ 150 nM), close inspection of the activity of all intermediate Tk-hefu derivatives (Table 1) indicates that some amino acid substitutions actually had a neutral or even negative effect. In particular, when moving from Tk-hefu-6 to 7, we observed a decrease in activity. And when we went on from Tk-hefu-8 to 9, the activity was not affected. Apparently, the D2Q, Y6K, and R7Q mutations were either neutral or reduced the affinity. Our idea was then to reverse those substitutions and produce a new Tk-hefu derivative with enhanced activity. This new derivative was named Tk-hefu-11.
      Table 1Amino acid sequences of Tk-hefu derivatives and their activity against KV channels
      a Gray shading highlights cysteine residues that form disulfide bonds; amino acid substitutions introduced in Tk-AMP-X2 to attain Tk-hefu peptides are in bold.
      b -, no activity at 20 μM.
      c KD values as reported in
      • Srinivasan K.N.
      • Sivaraja V.
      • Huys I.
      • Sasaki T.
      • Cheng B.
      • Kumar T.K.S.
      • et al.
      κ-Hefutoxin1, a novel toxin from the scorpion Heterometrus fulvipes with unique structure and function.
      .
      d Empty box means that the activity was not tested.
      e “X/Y” means that at concentration of X nM, Y percent of block was observed.
      f X ± Y represent IC50 values in nM.
      g These data represent EC50 values in nM.

      Production of Tk-hefu-11

      To obtain Tk-hefu-11 (Table 1), we used the Escherichia coli SHuffle T7 Express strain as an expression system. A synthetic gene encoding the derivative was cloned into the pET-32b expression vector, and thioredoxin (Trx) was used as the fusion partner to ensure a high yield of the disulfide-containing peptide. The target peptide was produced as a result of fusion protein cleavage by enteropeptidase followed by separation using reversed-phase HPLC and identification by MALDI mass spectrometry (MS) (Fig. 1). The final yield of Tk-hefu-11 was ∼6 mg per 1 l of bacterial culture.
      Figure thumbnail gr1
      Figure 1Purification of Tk-hefu-11. A, reversed-phase HPLC separation profile of fusion protein Trx-Tk-hefu-11 cleaved by enteropeptidase. B, MALDI-TOF mass spectrum of purified Tk-hefu-11. C, zoomed region of the spectrum. The measured [M + H]+ monoisotopic molecular mass is 3629.47 Da matching the calculated value of 3629.75 Da.

      Tk-hefu-11 activity

      Tk-hefu-11 was tested against four KV isoforms. The calculated EC50 values on hKV1.3 and rKV1.6 were 70.4 ± 2.9 nM and 10.0 ± 0.5 nM, respectively, while 1 μM Tk-hefu-11 inhibited only 41.2 ± 2.9% of hKV1.1 and 50.1 ± 0.8% of hKV1.2-mediated currents. At concentrations higher than 2 μM, lysis of oocytes was detected. These data indicate a stronger inhibition of KV1.6 compared to other KV isoforms tested, making Tk-hefu-11 a unique peptide with an ability to inhibit this channel in the nanomolar range and a selectivity factor of >7 (Fig. 2, AC and Table 1).
      Figure thumbnail gr2
      Figure 2Electrophysiological profile of Tk-hefu-11. A and B, representative current traces recorded from X. laevis oocytes. Shown are recordings on (A) KV1.1 and (B) KV1.2 in control conditions (solid lines) and after the application of the peptide. The dashed lines represent steady-state current traces after the application of 1 μM Tk-hefu-11, the dotted lines represent current after the application of 2 μM Tk-hefu-11, and the dash-dotted lines represent zero current level. C, dose-response curves for Tk-hefu-11 inhibition of KV1.3 and 1.6, obtained by plotting the percentage of blocked current as a function of increasing peptide concentrations. DG, representative currents traces recorded for (D) KV1.3, (E) KV1.3-mut1, (F) KV1.3-mut2, and (G) KV1.3-mut3 in control (solid lines) and after application of 80 nM Tk-hefu-11 (dashed lines). The dash-dotted lines represent zero current level. H, comparison of Tk-hefu-11 potency against different channels. The percentage of block is shown as mean ± SD, and individual values are presented; ∗∗∗∗p < 0.0001; ∗∗∗p < 0.001 (one-way ANOVA, F value = 185.2 and p < 0.0001). I, conductance-voltage relation (g/gmax-V) for KV1.6 in control (closed symbols) and in the presence of 10 nM Tk-hefu-11 (open symbols).
      To determine if Tk-hefu-11 inhibits KV1 channels as a pore blocker or gating modifier, we used two approaches. In one, Tk-hefu-11 was tested on mutants of KV1.3 previously described by us (
      • Kuzmenkov A.I.
      • Nekrasova O.V.
      • Peigneur S.
      • Tabakmakher V.M.
      • Gigolaev A.M.
      • Fradkov A.F.
      • et al.
      KV1.2 channel-specific blocker from Mesobuthus eupeus scorpion venom: structural basis of selectivity.
      ) (Fig. 2, DG). The constructs named mut1 (423DPTSGFS429423ERDSQFP429), mut2 (451HPV453451VPT453), and mut3 (425TSG427425DSQ427) represent hKV1.3 channels that harbor residues from hKV1.2 in the S5-P and P-S6 segments (residue numbering is according to UniProt accession number P22001). Significant differences in the blocking potency of Tk-hefu-11 were observed for all constructs (Fig. 2H), suggesting that the peptide interacts with the channel pore. Of note, mut3 comprises a small fragment of mut1 (underlined in its sequence) with only two replacements. A slightly higher level of current inhibition by Tk-hefu-11 for mut3 compared to wild type KV1.3 evidences that these replacements play a minor but beneficial role in the binding. Conversely, the dramatically reduced inhibitory effect observed for mut1 suggests the importance of D423, P424, or S429 for the interaction of the peptide with KV1.3 compared to the corresponding residues in KV1.2. In the other approach, we evaluated the conductance-voltage relationship for KV1.6 (Fig. 2I). No significant difference in the V1/2 values was observed: 7.7 ± 0.8 mV and 12.9 ± 0.9 mV for control and in the presence of 10 nM Tk-hefu-11, respectively. For a gating modifier, we would expect a profound shift in the voltage dependence of activation (
      • Swartz K.J.
      • MacKinnon R.
      Hanatoxin modifies the gating of a voltage-dependent K+ channel through multiple binding sites.
      ,
      • Tao H.
      • Chen J.J.
      • Xiao Y.C.
      • Wu Y.Y.
      • Su H.B.
      • Li D.
      • et al.
      Analysis of the interaction of tarantula toxin Jingzhaotoxin-III (β-TRTX-Cj1α) with the voltage sensor of Kv2.1 uncovers the molecular basis for cross-activities on Kv2.1 and Nav1.5 channels.
      ). Together, these data indicate that Tk-hefu-11 interacts with the pore of KV channels.
      Because we observed oocyte lysis at higher Tk-hefu-11 concentrations, we decided to test its antimicrobial activity, which is a hallmark of cytolytic peptides (
      • Kozlov S.A.
      • Vassilevski A.A.
      • Feofanov A.V.
      • Surovoy A.Y.
      • Karpunin D.V.
      • Grishin E.V.
      Latarcins, antimicrobial and cytolytic peptides from the venom of the spider Lachesana tarabaevi (Zodariidae) that exemplify biomolecular diversity.
      ). We did not detect any activity at concentrations up to 20 μM against neither Gram-positive nor Gram-negative bacteria. In addition, the peptide did not exhibit any hemolytic or cytolytic activity at the studied concentrations up to 20 μM.

      Spatial structure of Tk-hefu-11

      The 3D structure of the peptide was solved in water (Fig. 3). For this, a 15N-labeled analog of the peptide (15N-Tk-hefu-11) was prepared by using a culture medium containing only 15N as the source of nitrogen. Hundred structures were calculated using the torsion angle restraints, upper and lower NOE-based distance restraints, as well as hydrogen bond and disulfide bond restraints (data are shown in Table S2). The resulting set of ten NMR structures is characterized by a low RMSD value for backbone atoms (0.39 ± 0.15) and insignificant restraint violations, suggesting that the structure is well defined by the experimental data. Chemical shifts and coordinates are deposited to the Protein Data Bank (PDB) and Biological Magnetic Resonance Data Bank (BMRB) databases under the accession codes 7QXJ and 34703, respectively.
      Figure thumbnail gr3
      Figure 3Spatial structure of Tk-hefu-11. The amino acid sequence is presented above, and the color code is the same as in the 3D structure. The N-terminal α-helix is shown in light pink and the C-terminal, pink. Disulfide bonds are shown as yellow sticks; side chains of positively charged (at neutral pH) amino acids are shown as blue lines; acidic, red; hydrophobic, orange; and hydrophilic uncharged, purple.
      The solution structure of Tk-hefu-11 consists of two antiparallel α-helices (residues K3–Q10 and R15–G24) joined by a short loop (Figs. 3 and S1), corresponding to the α-hairpinin fold that we and others described in many plant peptides (
      • Nolde S.B.
      • Vassilevski A.A.
      • Rogozhin E.A.
      • Barinov N.A.
      • Balashova T.A.
      • Samsonova O.V.
      • et al.
      Disulfide-stabilized helical hairpin structure and activity of a novel antifungal peptide EcAMP1 from seeds of barnyard grass (Echinochloa crus-galli).
      • Oparin P.B.
      • Mineev K.S.
      • Dunaevsky Y.E.
      • Arseniev A.S.
      • Belozersky M.A.
      • Grishin E.V.
      • et al.
      Buckwheat trypsin inhibitor with helical hairpin structure belongs to a new family of plant defence peptides.
      ). The structure is stabilized by two disulfide bridges (C5–C25 and C9–C21) and nine hydrogen bonds formed according to α-helical conformation (Fig. S2). Detailed analysis of the obtained structure reveals that Y6 is likely to form a cation-π contact with K3 and/or K22. We notice that Tk-hefu-11 has a small hydrophobic core formed by the disulfide bonds surrounded by M8 and Y12 on one side and A1, Y6, and Y27 on the other side. Tk-hefu-11 also reveals an anisotropic distribution of electrostatic parameters; a pronounced positively charged face is formed by arginine and lysine residues, and an uncharged and relatively apolar face is formed by the side chains of cystine, methionine, and tyrosine residues.

      Molecular modeling explains the selectivity of Tk-hefu-11

      To uncover the molecular determinants underlying the differences in Tk-hefu-10 and 11 activities on KV channels, we performed a computational study of the molecular complexes of these peptides with hKV1.3 and rKV1.6 (Fig. 4). We took advantage of the solved NMR structure of Tk-hefu-11 and built a homology model of Tk-hefu-10 based on it. As mentioned in Experimental procedures, we aligned K/Y6 and K22 of Tk-hefu-10/11 onto the classical dyad residues Y36 and K27 in ChTx complexed with KV1.2 pore (
      • Banerjee A.
      • Lee A.
      • Campbell E.
      • MacKinnon R.
      Structure of a pore-blocking toxin in complex with a eukaryotic voltage-dependent K+ channel.
      ) to get a starting conformation. MD trajectories were calculated, and computational analysis of contact surfaces and energy contributions of residues to complex formation was carried out.
      Figure thumbnail gr4
      Figure 4Modeled structure of complex between KV1.6 and Tk-hefu-11. The structure is shown after 501 ns of molecular dynamics simulation inside a hydrated lipid bilayer. KV1.6 subunits are shown in a cartoon representation (colored gray, brown, cyan, and blue); the pore domain helices of the channel subunit in the foreground and the voltage-sensing domain of the adjacent subunit, as well as extended extracellular loops of the voltage-sensing domains are omitted for clarity. Lipids are in a semitransparent space-filling representation, and some are omitted for clarity. Atoms are colored: oxygen, red; phosphorus, orange; nitrogen, blue; hydrogen of amino and hydroxyl group, white; carbon of POPC, light-yellow; carbon of POPE, yellow; and carbon of cholesterol, beige. Tk-hefu-11 is in pink; K22 (plugs the channel pore) and disulfide bridges (yellow) are shown as sticks.
      In line with the findings of our recent study of Tk-hefu derivatives (
      • Tabakmakher V.M.
      • Gigolaev A.M.
      • Peigneur S.
      • Krylov N.A.
      • Tytgat J.
      • Chugunov A.O.
      • et al.
      Potassium channel blocker crafted by α-hairpinin scaffold engineering.
      ), the analysis of residual contributions to intermolecular interaction energy yielded expectable results (Fig. 5). We note that the calculated absolute energy values have little physical meaning and should only be considered in comparison with other calculated values. In going from Tk-hefu-10 to 11, the substitution of a neutral residue by a negatively charged one (Q2D), as well as the substitution of a positively charged residue by a neutral one (K6Y) provides a negative contribution to the channel binding (rise in complex energy values). On the contrary, the Q7R substitution results in a positive contribution (drop in energy values). These tendencies are observed in complexes of both KV1.3 and 1.6.
      Figure thumbnail gr5
      Figure 5Interaction energy profiles for Tk-hefu-10 and 11 complexes with hKV1.3 and rKV1.6. Bar charts show residual contributions to the interaction energy averaged over molecular dynamics simulations. Error bars indicate SDs. Letters below each bar group indicate amino acid residues in Tk-hefu-10/11; identical residues are shown in light gray.
      Analysis of MD trajectories revealed that in the complexes rKV1.6–Tk-hefu-11, rKV1.6–Tk-hefu-10, and hKV1.3–Tk-hefu-11, the peptides change their positions from the initial state considerably, while the position of Tk-hefu-10 in complex with hKV1.3 changes less significantly (Fig. 6). In the complexes with KV1.6, the flexibility of the N terminus of the peptides is restrained by hydrogen bonding and salt bridge formation between their N-terminal amino group and the D404 side chain carboxyl group of the channel. The positively charged side chain of R11 tends to approach the rKV1.6 channel-specific negatively charged motif 399EADDVD404 located in the S5-P loop (Fig. 7, AC). Thus, Tk-hefu-11 rotates slightly around the KV1.6 pore axis during MD, which results in multiple polar contacts: R11 forms salt bridges with E399 and D404, hydrogen bonds with S405 and Y430, and a cation-π contact with F407. Moreover, the reorientation of Tk-hefu-11 allows R7 to reach D412 and D428 side chains (and form two salt bridges), as well as the side chains of W415 and Y430 of a neighboring channel subunit (and form two cation-π contacts), which stabilizes the complex greatly (Fig. 7B and Table S3). A shorter and uncharged side chain of Q7 in Tk-hefu-10 cannot reach those channel residues, so its position in the complex is less stabilized, and R11 forms only one salt bridge with D404 in KV1.6 (Fig. 7C).
      Figure thumbnail gr6
      Figure 6RMSD of atomic positions of Tk-hefu-10 and 11 complexed to KV1.3 and 1.6.
      Figure thumbnail gr7
      Figure 7Tk-hefu-10 and 11 interactions with KV1.3 and 1.6 predicted by molecular dynamics. A, amino acid sequence alignment of the extracellular pore region of hKV1.3 and rKV1.6 channels. Residue numbering is above each sequence (UniProt accession numbers P22001 and P17659); identical residues are shaded gray. Functional segments of the channels are marked above the alignment. BE, orientation of peptides and intermolecular contacts in the complexes rKV1.6–Tk-hefu-11 (B), rKV1.6–Tk-hefu-10 (C), hKV1.3–Tk-hefu-11 (D), and hKV1.3–Tk-hefu-10 (E). Orientations of the peptides in the channel pore vestibule are shown in the up-left corner of each panel; the channel-specific motifs 399EADDVD404/420EADDPT425 of the S5-P loop in KV1.6/KV1.3 are shown in red, Tk-hefu-10 or 11 are shown in pink; the initial orientation of the peptides is shown in sand yellow. Interacting amino acid residues of the peptides and channels are shown as sticks; residues not involved in the interactions are shown semitransparently. Intermolecular contacts (hydrogen bonds and salt bridges) are shown as dashed yellow lines.
      Notably, in the complex with KV1.6, residue D2 of Tk-hefu-11 tends to repulse from D404 (Fig. 7B). However, that same residue D2 also repulses the C-terminal carboxyl group from the N terminus of the peptide. In Tk-hefu-11, the C terminus interacts with K23 forming an intramolecular salt bridge that stabilizes the secondary structure of the peptide. In Tk-hefu-10 that has Q2 instead of D2, the N- and C-terminal groups interact. This causes peptide secondary structure distortion and brings the C-terminal carboxyl group in close proximity to D404 of KV1.6, destabilizing the complex (Fig. 7, B and C).
      The channel-specific negatively charged motif 420EADDPT425 in the S5-P loop of hKV1.3 has a lower formal charge compared to the aforementioned 399EADDVD404 in rKV1.6 because D404 is substituted by T425 in the former (Fig. 7A). Besides, the side chain of T425 is not long enough to reach and fix the position of the N-terminal peptide residue in a manner described previously. Therefore, the electrostatic attraction from the S5-P loop affects the peptide position to a lesser degree in case of KV1.3. Tk-hefu-11 in the complex with KV1.3 moves slightly during MD, and R7 forms a salt bridge with D433, a hydrogen bond with G427, and a cation-π contact with F428 (Fig. 7D). Such a shift does not happen in the complex with Tk-hefu-10 because Q7 is not charged. Instead, it forms a single hydrogen bond with D449 of a neighboring channel subunit. Together with two hydrogen bonds formed by Q2 with D433 and D449, this interaction forces Tk-hefu-10 to disrupt its secondary structure (Fig. 7E).
      Analysis of the interactions during MD has shown that Tk-hefu-11 in the complexes with KV1.3 and 1.6 forms more contacts with the channels than Tk-hefu-10 (Tables 2 and S3), which is in agreement with the results of electrophysiological recordings. Despite Tk-hefu-10 being involved in a decent number of medium-lived and short-lived contacts (19 and 11, respectively), it forms only 16 long-lived specific interactions with KV1.6. This indicates that Tk-hefu-10 forms a weak complex with KV1.6, which is consistent with our experimental data.
      Table 2Intermolecular contacts observed in complexes of Tk-hefu-10 and 11 with KV1.3 and 1.6 during MD simulations
      Contacts
      Lifetime of each contact is counted as part of MD trajectory time (450 ns in total, the first 50 ns were not taken into account to get representative data). Weak contacts: lifetime is equal to or greater than 7% and less than 10%; short-lived contacts: lifetime is is equal to or greater than 10% but less than 20%; medium-lived contacts: lifetime is is equal to or greater than 20% but less than 50%; long-lived contacts: lifetime is is equal to or greater than 50%.
      Number of contacts in complexes
      KV1.3KV1.6
      Tk-hefu-10Tk-hefu-11Tk-hefu-10Tk-hefu-11
      All specific interactions
      Hydrogen bonds, salt bridges, stacking/π-π interactions, and cation-π interactions.
       Long-lived20271625
       Medium-lived991920
       Short-lived168118
       Weak5266
      Hydrogen bonds
       Long-lived1011711
       Medium-lived65109
       Short-lived9596
       Weak3132
      Salt bridges
       Long-lived4756
       Medium-lived--36
       Short-lived52--
       Weak--22
      Stacking/π-π interactions
       Long-lived3322
       Medium-lived1312
       Short-lived-112
       Weak-1--
      Cation-π interactions
       Long-lived3626
       Medium-lived2153
       Short-lived2-1-
       Weak2-12
      All nonspecific (hydrophobic) interactions
       Long-lived175163130135
       Medium-lived77718285
       Short-lived52394353
       Weak24182131
      a Lifetime of each contact is counted as part of MD trajectory time (450 ns in total, the first 50 ns were not taken into account to get representative data). Weak contacts: lifetime is equal to or greater than 7% and less than 10%; short-lived contacts: lifetime is is equal to or greater than 10% but less than 20%; medium-lived contacts: lifetime is is equal to or greater than 20% but less than 50%; long-lived contacts: lifetime is is equal to or greater than 50%.
      b Hydrogen bonds, salt bridges, stacking/π-π interactions, and cation-π interactions.
      Our analysis of the intermolecular contacts during MD shows that the substitutions introduced in Tk-hefu-11 affect the binding to the channels in a complex manner. On the one hand, when moving from Tk-hefu-10 to 11, the Q2D substitution results in a loss of interactions with KV1.3 due to electrostatic repulsion from the side chains of D433 and D449 (Fig. 7, D and E). However, neither Q2 nor D2 in Tk-hefu-10 or 11 forms specific contacts (H-bonds, salt bridges, cation-π, or stacking) with KV1.6. On the other hand, Tk-hefu-10 residue K6 is involved in a salt bridge, three hydrogen bonds, and two cation-π contacts in the complex with KV1.3, while Tk-hefu-11 residue Y6 forms just one hydrogen bond and one stacking interaction. In the complex with KV1.6, that same residue K6 in Tk-hefu-10 is involved in two hydrogen bonds and two cation-π contacts, while Y6 of Tk-hefu-11 forms just one hydrogen bond (Table S2). The substitution Q7R affects contact distribution in the complexes even more prominently. According to our analysis, together with the N terminus and R11, R7 provides the high stability of the Tk-hefu-11 complex with KV1.6 and 1.3.

      Discussion

      Quite unexpectedly, Tk-hefu-11, which was designed to target KV1.3, showed pronounced affinity and selectivity to KV1.6. Our molecular modeling suggests that the observed Tk-hefu-11 activity is due to (i) direct contacts between the channel and those residues that differ from other Tk-hefu peptides and (ii) an indirect effect of the substitutions.
      The 3D structure of Tk-hefu-11 established here (Fig. 3) diverged from the initial structure of Tk-hefu-1 reported previously (PDB ID: 5LM0) (
      • Berkut A.A.
      • Chugunov A.O.
      • Mineev K.S.
      • Peigneur S.
      • Tabakmakher V.M.
      • Krylov N.A.
      • et al.
      Protein surface topography as a tool to enhance the selective activity of a potassium channel blocker.
      ). The major differences are the angle between the α-helices and the positioning of Y6. While in Tk-hefu-1 the α-helices are more X-crossed with the interhelix angle of ≈160°, Tk-hefu-11 is more planar with this angle being ≈−170°, that is, the α-helices twist by 30°. This twist is explained by a cation-π interaction between Y6 and K3 found in Tk-hefu-11. In Tk-hefu-1, on the other hand, D3 cannot engage in such contact, and Y6 flips to interact with Q10 and K18. These unanticipated changes lead to a different mode of peptide interaction with the channels and should be taken into account in molecular modeling.
      Previously, we used a rather straightforward approach to suggest amino acid replacements in Tk-hefu derivatives. We analyzed the interaction energy profiles of the models of the complexes and sought to substitute just those residues that made a negative contribution. Here, in case of Tk-hefu-11, we observe that sometimes such negative contribution to the complex formation may actually lead to a conformational reorganization of the peptide, which in turn affects the binding capacity positively. Thus, we note that in the complex with KV1.6, although residue D2 of Tk-hefu-11 is repulsed from D404, it also repulses the C-terminal carboxyl group of the peptide. This latter repulsion allows Tk-hefu-11 to assume an optimal position in the pore vestibule of the channel. Conversely, residue Q2 of Tk-hefu-10 is not repulsed from D404, but it does not repulse the C terminus, and this peptide does not fit into the channel vestibule optimally. This finding may be utilized in further attempts to improve ligand binding to ion channels.
      With an EC50 of ≈10 nM, Tk-hefu-11 is one of the most potent KV1.6 ligands. Table 3 compares Tk-hefu-11 with other known peptides showing some selectivity to KV1.6 isoform. As mentioned previously and seen from the table, other ligands have either high affinity or high selectivity toward KV1.6, while our peptide combines both properties. The closest competitor is conopeptide Y-PI1 (
      • Imperial J.S.
      • Chen P.
      • Sporning A.
      • Terlau H.
      • Daly N.L.
      • Craik D.J.
      • et al.
      Tyrosine-rich conopeptides affect voltage-gated K+ channels.
      ), but it is less potent (IC50 ≈ 170 nM) and the activity against KV1.1 is not known.
      Table 3Comparison of known peptide toxins presenting selective activity to KV1.6
      LigandKV1.1KV1.2KV1.3KV1.6Reference
      Tk-hefu-111000/41
      “X/Y” means that at concentration of X nM, Y percent of block was observed.
      1000/5070
      EC50 values in nM.
      10
      EC50 values in nM.
      This work
      Scorpion toxins
       Kbot211000/1001000/1001000/10076
      IC50 values in nM.
      (
      • ElFessi-Magouri R.
      • Peigneur S.
      • Othman H.
      • Srairi-Abid N.
      • ElAyeb M.
      • Tytgat J.
      • et al.
      Characterization of Kbot21 reveals novel side chain interactions of scorpion toxins inhibiting voltage-gated potassium channels.
      )
       HelaTx19900
      EC50 values in nM.
      30,000/530,000/2030,000/60(
      • Vandendriessche T.
      • Kopljar I.
      • Jenkins D.P.
      • Diego-Garcia E.
      • Abdel-Mottaleb Y.
      • Vermassen E.
      • et al.
      Purification, molecular cloning and functional characterization of HelaTx1 (Heterometrus laoticus): the first member of a new κ-KTX subfamily.
      )
      Sea anemone toxins
      AbeTx1672
      IC50 values in nM.
      167
      IC50 values in nM.
      3000/20116
      IC50 values in nM.
      (
      • Orts D.J.B.
      • Peigneur S.
      • Silva-Gonçalves L.C.
      • Arcisio-Miranda M.
      • Bicudo J.E.P.W.
      • Tytgat J.
      AbeTx1 is a novel sea anemone toxin with a dual mechanism of action on shaker-type K+ channels activation.
      )
       κ-Actitoxin-Bcs3b14
      IC50 values in nM.
      80
      IC50 values in nM.
      13
      IC50 values in nM.
      8
      IC50 values in nM.
      (
      • Orts D.J.B.
      • Peigneur S.
      • Madio B.
      • Cassoli J.S.
      • Montandon G.G.
      • Pimenta A.M.C.
      • et al.
      Biochemical and electrophysiological characterization of two sea anemone type 1 potassium toxins from a geographically distant population of bunodosoma caissarum.
      )
      Conotoxins
       Conopeptide Y-Fe1
      Empty box, no data.
      >30,000>50,0008800
      IC50 values in nM.
      (
      • Imperial J.S.
      • Chen P.
      • Sporning A.
      • Terlau H.
      • Daly N.L.
      • Craik D.J.
      • et al.
      Tyrosine-rich conopeptides affect voltage-gated K+ channels.
      )
       Conopeptide Y-Pl12000
      IC50 values in nM.
      >50,000170
      IC50 values in nM.
      (
      • Imperial J.S.
      • Chen P.
      • Sporning A.
      • Terlau H.
      • Daly N.L.
      • Craik D.J.
      • et al.
      Tyrosine-rich conopeptides affect voltage-gated K+ channels.
      )
       α/κ-Conotoxin pl14a>1000>1000>10001590
      IC50 values in nM.
      • Imperial J.S.
      • Bansal P.S.
      • Alewood P.F.
      • Daly N.L.
      • Craik D.J.
      • Sporning A.
      • et al.
      A novel conotoxin inhibitor of Kv1.6 channel and nAChR subtypes defines a new superfamily of conotoxins.
      a “X/Y” means that at concentration of X nM, Y percent of block was observed.
      b EC50 values in nM.
      c IC50 values in nM.
      d Empty box, no data.
      Previously, using the α-hairpinin fold as template, we designed only KV1.3-targeting peptides. Tk-hefu-11 exemplifies the applicability of this fold to obtain selective blockers of other KV channels. Interestingly, Tk-hefu-11 outperforms not only all previously known artificial α-hairpinins (Table 1) but also natural toxins with the same type of fold (called cysteine-stabilized α/α fold or CSα/α) such as κ-hefutoxin-1 (
      • Srinivasan K.N.
      • Sivaraja V.
      • Huys I.
      • Sasaki T.
      • Cheng B.
      • Kumar T.K.S.
      • et al.
      κ-Hefutoxin1, a novel toxin from the scorpion Heterometrus fulvipes with unique structure and function.
      ) or HelaTx1 (
      • Kasheverov I.E.
      • Oparin P.B.
      • Zhmak M.N.
      • Egorova N.S.
      • Ivanov I.A.
      • Gigolaev A.M.
      • et al.
      Scorpion toxins interact with nicotinic acetylcholine receptors.
      ). In conclusion, we hope that Tk-hefu-11 will be useful as a molecular tool to study the function of KV1.6 isoform.

      Experimental procedures

      Recombinant peptide production

      Tk-hefu-11 was produced by a standard protocol that we used in previous work (
      • Berkut A.A.
      • Usmanova D.R.
      • Peigneur S.
      • Oparin P.B.
      • Mineev K.S.
      • Odintsova T.I.
      • et al.
      Structural similarity between defense peptide from wheat and scorpion neurotoxin permits rational functional design.
      ,
      • Berkut A.A.
      • Chugunov A.O.
      • Mineev K.S.
      • Peigneur S.
      • Tabakmakher V.M.
      • Krylov N.A.
      • et al.
      Protein surface topography as a tool to enhance the selective activity of a potassium channel blocker.
      ). Shortly, a bacterial expression system was used to produce peptides as fusion proteins with the carrier protein Trx (
      • McCoy J.
      • LaVallie E.
      Expression and purification of thioredoxin fusion proteins.
      ), containing a site of cleavage by human enteropeptidase light chain (
      • Gasparian M.E.
      • Ostapchenko V.G.
      • Schulga A.A.
      • Dolgikh D.A.
      • Kirpichnikov M.P.
      Expression, purification, and characterization of human enteropeptidase catalytic subunit in Escherichia coli.
      ), and a His-tag for affinity chromatography purification.

      Expression vector construction

      DNA sequence encoding Tk-hefu-11 was obtained in two steps by PCR using synthetic oligonucleotides as primers (Table S1). On the first step, all four primers were used for five PCR cycles. Then, the reaction mixture was diluted 1000 times and used as a matrix on the second step with flanking primers (F1 and R1). The resulting PCR fragment was cloned into the expression vector pET-32b (Novagen) using the KpnI and BamHI restriction enzymes to produce pET-32b-Tk-hefu-11.

      Fusion protein expression and purification

      E. coli SHuffle T7 Express cells (New England Biolabs) were transformed using the expression vector pET-32b-Tk-hefu-11 and cultured at 37 °C in LB medium to the mid-log phase. Expression was then induced by 0.2 mM IPTG. Cells were cultured at room temperature (RT) (24 °C) overnight (16 h) and harvested by centrifugation. The cell pellet was resuspended in 30 ml of 300 mM NaCl, 50 mM Tris–HCl buffer (pH 8.0), and ultrasonicated. The lysate was applied to a HisPur Cobalt Resin (Thermo Fisher Scientific) and the fusion protein Trx-Tk-hefu-11 was purified according to the manufacturer’s protocol.
      To produce 15N-labeled Tk-hefu-11, M9 minimal medium with the addition of ISOGRO and 15NH4Cl (both at 1 g/l; Sigma–Aldrich) was used instead of LB medium. E. coli was first cultured in LB, harvested by centrifugation, and resuspended to A600 ≈ 0.1 in M9 medium with ISOGRO and 15NH4Cl. The bacterial culture was then grown at 37 °C for ∼4 h to the mid-log phase (A600 ≈ 0.5) using the New Brunswick BioFlo/CelliGen 115 bioreactor (Eppendorf) with intense aeration and agitation. Transgene expression was induced as aforementioned, and after the induction, the culture was incubated at RT (24 °C) overnight (16 h). 15N-Trx-Tk-hefu-11 was purified as the unlabeled protein aforementioned.

      Fusion protein cleavage and purification of recombinant peptides

      Fusion proteins were dissolved in 50 mM Tris–HCl (pH 8.0) to a concentration of 1 mg/ml. Protein cleavage with human enteropeptidase light chain (1 U of enzyme per 1 mg of substrate) was performed at 37 °C overnight (16 h). Recombinant peptides were purified by reversed-phase HPLC on a Jupiter C5 column (4.6 × 250 mm; Phenomenex) in a linear gradient of acetonitrile concentration (0%–60% in 60 min, followed by a quick step to 80%) in the presence of 0.1% TFA. The purity of the target peptides was checked by MALDI MS and analytical chromatography on a Vydac C18 column (4.6 × 250 mm; Separations Group) in the same acetonitrile gradient.

      MS

      MALDI MS was performed on an Ultraflex TOF-TOF (Bruker Daltonik) spectrometer as described previously (
      • Kuzmenkov A.I.
      • Sachkova M.Y.
      • Kovalchuk S.I.
      • Grishin E.V.
      • Vassilevski A.A.
      Lachesana tarabaevi, an expert in membrane-active toxins.
      ). 2,5-Dihydroxybenzoic acid (Sigma–Aldrich) was used as a matrix. Measurements were performed in the reflectron mode with a mass accuracy error not exceeding 100 ppm. Mass spectra were analyzed with the Data Analysis 4.3 and Data Analysis Viewer 4.3 software (Bruker).

      Electrophysiology

      Expression of KV channels in Xenopus laevis oocytes

      Voltage-gated potassium channels (hKV1.1, hKV1.2, hKV1.3, and rKV1.6) were expressed in X. laevis oocytes. Frogs were kept in compliance with the regulations of the European Union concerning the welfare of laboratory animals as declared in Directive 2010/63/EU. The use of X. laevis oocytes was approved by the Animal Ethics Committee of KU Leuven with the license number P186/2019. Mature female animals were purchased from Nasco (Fort Atkinson) and housed in the Aquatic Facility at KU Leuven. Stage V–VI oocytes were collected from anaesthetized frogs as described previously (
      • Boldrini-França J.
      • Pinheiro-Junior E.L.
      • Peigneur S.
      • Pucca M.B.
      • Cerni F.A.
      • Borges R.J.
      • et al.
      Beyond hemostasis: a snake venom serine protease with potassium channel blocking and potential antitumor activities.
      ). Human KCNA1 (GenBank accession number: NM_000217) and KCNA2 (NM_004974) genes were cloned in pcDNA3.1(+) vector, which was linearized using EcoRV. Human KCNA3 (NM_002232) was in pCI-neo, which was linearized by NotI. Rat Kcna6 (X17621) was in pGEM-HE, which was linearized by NdeI. The linearized plasmids were transcribed using the mMESSAGE mMACHINE T7 transcription kit (Ambion). mRNA was injected into oocytes using a microinjector (Drummond Scientific), with a programmed RNA injection volume of 4 to 50 nl depending on channel subtype. The oocytes were incubated in ND96 solution (96 mM NaCl, 2 mM KCl, 1.8 mM CaCl2, 2 mM MgCl2, and 5 mM Hepes, pH 7.4), supplemented with 50 mg/l gentamicin sulfate.

      Electrophysiological recordings

      Electrophysiological measurements were performed at RT (18–22 °C) using the two-electrode voltage-clamp technique. Data were obtained using a GeneClamp 500 amplifier and Clampex 9 software (Molecular Devices). Micropipettes were produced using borosilicate glass capillaries (1B120-6) and drawn in a manual stretcher (World Precision Instruments). The bath and perfusion solutions were ND96.
      Whole cell currents were recorded 1 to 2 days after RNA injection. Current and voltage electrodes were filled with 3 M KCl and their resistance was adjusted to 0.7–2.0 MΩ. Currents were sampled at 2 kHz and filtered at 0.5 kHz using a four-pole Bessel low-pass filter. The holding potential was set at −90 mV. Leak subtraction was performed using a P/4 protocol. KV currents were evoked by a 500 ms depolarization to 0 mV followed by a 500 ms pulse to −50 mV. For conductance-voltage relationship studies, currents were evoked by 5 mV depolarization steps. Potassium conductance was calculated using Ohm’s law:
      gK=IK|VmVrev|,


      where IK is the maximal current at the test potential Vm, and Vrev is the reversal potential. The conductance-voltage data were fitted using the Boltzmann equation:
      gKgmax=[1+eV1/2Vmk]1,


      where gmax represents maximal gK, V1/2 is the voltage corresponding to half-maximal conductance, and k is the slope factor.
      Different concentrations of the peptides diluted in ND96 were used for the concentration–response assays. These solutions were added to the bath containing the oocyte and mixed immediately, thereby obtaining the desired final concentrations. KV currents were then recorded in the presence of the tested peptide, and the data were fitted with the Hill equation:
      y=1001+(EC50[peptide])h


      where y is the amplitude of the compound-induced effect in percent (i.e., the percent of maximal current inhibition for the given compound), [peptide] is the peptide concentration, EC50 is the half-maximal effective concentration, and h is the Hill coefficient.
      All data were obtained in at least three independent repeats (n ≥ 3) using different batches of X. laevis oocytes. Statistical significance was determined using one-way ANOVA with Dunnett’s post-test. Experimental values are presented as mean ± SEM. Bar charts present all individual data points and the SD to represent variation.

      Antimicrobial assay

      Determination of antimicrobial activity was performed following a previously described protocol (
      • Vassilevski A.A.
      • Kozlov S.A.
      • Egorov T.A.
      • Grishin E.V.
      Purification and characterization of biologically active peptides from spider venoms.
      ). Bacteria (Enterococcus faecalis (Andrewes and Horder) Schleifer and Kilpper-Balz (ATCC 29212), E. coli (Migula) Castellani and Chalmers (ATCC 25922), Pseudomonas aeruginosa (Schroeter) Migula (ATCC 27853), and Staphylococcus aureus subsp. aureus Rosenbach (ATCC 29213)) were cultured overnight in LB medium at 37 °C. Determination of the minimal inhibitory concentrations for the peptide was performed using a 2-fold microtiter broth dilution assay in 96-well sterile plates at a final volume of 100 μl. Mid-log phase cultures were diluted to a final concentration of 105 colony-forming units/ml. Dried peptide was dissolved in 10 μl of water and added to 90 μl of the bacterium dilution. The peptide, a nontreated control, and a sterility control were tested in triplicate. The microtiter plates were incubated for 24 h at 37 °C; growth inhibition was determined by measuring the absorbance at 620 nm. Minimal inhibitory concentrations are expressed as the lowest concentration of peptides that caused 100% growth inhibition.

      Cytolytic assay

      Human capillary blood was collected in a tube with heparin (10 units/ml), diluted to (1.0 ± 0.1) × 107 cells/ml with RPMI-1640 medium (PanEco) containing 10% fetal bovine serum (HyClone), and incubated with the peptides (0.6–20 μM, 2-fold dilutions) for 3 h at 37 °C with gentle shaking. Hemoglobin release was measured as described previously (
      • Vorontsova O.V.
      • Egorova N.S.
      • Arseniev A.S.
      • Feofanov A.V.
      Haemolytic and cytotoxic action of latarcin Ltc2a.
      ).
      Human lung adenocarcinoma A549 cells were cultured in Dulbecco's modified Eagle's medium with addition of 2 mM L-glutamine and 10% fetal bovine serum (complete medium) at 37 °C in humidified atmosphere with 5% CO2. Cell reseeding was performed twice a week. To study the cytolytic activity of the peptides, cells were seeded in 96-well plates (seeding density of 5 × 103 cells per well) 1 day before the experiment. Peptides were added to cells (2.5–20 μM, 2-fold dilutions). The cytotoxicity was estimated after incubation of the cells with peptides for 3 h at 37 °C in humidified atmosphere with 5% CO2. Cell survival was analyzed by staining cell nuclei with Hoechst 33342 (stains all cells) and propidium iodide (stains dead cells) and examining them with an inverted fluorescence microscope Axio Observer (Zeiss) as described previously (
      • Efremenko A.V.
      • Ignatova A.A.
      • Grin M.A.
      • Sivaev I.B.
      • Mironov A.F.
      • Bregadze V.I.
      • et al.
      Chlorin e6 fused with a cobalt-bis(dicarbollide) nanoparticle provides efficient boron delivery and photoinduced cytotoxicity in cancer cells.
      ). At least three independent experiments were carried out (n ≥ 3).

      NMR spectroscopy

      All NMR experiments were performed using the Avance 700 MHz spectrometer (Bruker Biospin) at 30 °C. 15N-Tk-hefu-11 was dissolved in H2O/D2O (19:1) and pH was adjusted to 5.5. 1H, 15N chemical shift assignments were obtained by the standard procedure based on 2D TOCSY, 2D NOESY (mixing time of 80 ms), 3D NOESY, 15N-heteronuclear single quantum coherence (HSQC), and 13C-HSQC spectra. After recording the set of spectra, the peptide sample was freeze dried and redissolved in 100% D2O (Acros Organics) to measure the proton-deuterium exchange rates and record the 2D NOESY and DQF correlated spectroscopy spectra.
      Spatial structure calculation was performed using the simulated annealing/MD protocol as implemented in the CYANA software package version 3.98.13 (L.A. Systems) (
      • Herrmann T.
      • Güntert P.
      • Wüthrich K.
      Protein NMR structure determination with automated NOE-identification in the NOESY spectra using the new software ATNOS.
      ). Upper interproton distance restraints were obtained by 1/r6 calibration of NOESY crosspeak intensities. Torsion angle restraints and stereospecific assignments were obtained based on the J-couplings and NOE intensities. Hydrogen bonds were added at the final stage of the structure calculation, if they were formed in at least 70% of the obtained conformers. The disulfide linkages were introduced based on the previously published data for α-hairpinins and their derivatives (
      • Berkut A.A.
      • Usmanova D.R.
      • Peigneur S.
      • Oparin P.B.
      • Mineev K.S.
      • Odintsova T.I.
      • et al.
      Structural similarity between defense peptide from wheat and scorpion neurotoxin permits rational functional design.
      ,
      • Berkut A.A.
      • Chugunov A.O.
      • Mineev K.S.
      • Peigneur S.
      • Tabakmakher V.M.
      • Krylov N.A.
      • et al.
      Protein surface topography as a tool to enhance the selective activity of a potassium channel blocker.
      ) and were additionally tested when calculating the spatial structure. Visual analysis of the calculated structures and figure drawings were performed using PyMOL (Schrodinger, LLC) and MOLMOL (
      • Koradi R.
      • Billeter M.
      • Wüthrich K.
      MOLMOL: a program for display and analysis of macromolecular structures.
      ).

      Molecular modeling

      Structural model of Tk-hefu-10 was generated in PyMOL using the in silico mutagenesis option based on the NMR solution structure of Tk-hefu-11 in a similar way to the procedure described elsewhere (
      • Berkut A.A.
      • Chugunov A.O.
      • Mineev K.S.
      • Peigneur S.
      • Tabakmakher V.M.
      • Krylov N.A.
      • et al.
      Protein surface topography as a tool to enhance the selective activity of a potassium channel blocker.
      ). rKV1.6 model was generated analogously to the model of hKV1.3 (
      • Gigolaev A.M.
      • Kuzmenkov A.I.
      • Peigneur S.
      • Tabakmakher V.M.
      • Pinheiro-Junior E.L.
      • Chugunov A.O.
      • et al.
      Tuning scorpion toxin selectivity: switching from KV1.1 to KV1.3.
      ,
      • Berkut A.A.
      • Chugunov A.O.
      • Mineev K.S.
      • Peigneur S.
      • Tabakmakher V.M.
      • Krylov N.A.
      • et al.
      Protein surface topography as a tool to enhance the selective activity of a potassium channel blocker.
      ,
      • Kuzmenkov A.I.
      • Nekrasova O.V.
      • Peigneur S.
      • Tabakmakher V.M.
      • Gigolaev A.M.
      • Fradkov A.F.
      • et al.
      KV1.2 channel-specific blocker from Mesobuthus eupeus scorpion venom: structural basis of selectivity.
      ) in MODELLER 9.19 (
      • Webb B.
      • Sali A.
      Comparative protein structure modeling using MODELLER.
      ) using the rKV1.2 structure (PDB ID: 3LUT) (
      • Chen X.
      • Wang Q.
      • Ni F.
      • Ma J.
      Structure of the full-length Shaker potassium channel Kv1.2 by normal-mode-based X-ray crystallographic refinement.
      ) as a template. Complexes of Tk-hefu-10 and 11 with KV channels were modeled based on the crystal structure of the KV1.2/2.1 paddle chimera in complex with charybdotoxin (ChTx; PDB ID: 4JTA) (
      • Banerjee A.
      • Lee A.
      • Campbell E.
      • MacKinnon R.
      Structure of a pore-blocking toxin in complex with a eukaryotic voltage-dependent K+ channel.
      ) analogously to the procedures described in (
      • Webb B.
      • Sali A.
      Comparative protein structure modeling using MODELLER.
      ,
      • Kudryavtsev D.S.
      • Tabakmakher V.
      • Budylin G.S.
      • Egorova N.S.
      • Efremov R.G.
      • Ivanov I.A.
      • et al.
      Complex approach for analysis of snake venom α-neurotoxins binding to HAP, the high-affinity peptide.
      ). Briefly, the model of each channel was structurally superimposed onto KV1.2/2.1 in the complex with ChTx to get a model of ChTx complex with that channel. Then, ChTx was replaced in the complex with Tk-hefu-10 or 11 by structural alignment of the appropriate dyads: K/Y6 and K22 of the peptide onto the classical Y36 and K27 in ChTx with minor manual adjustment.

      MD simulations

      The resulting complexes of Tk-hefu-10 and 11 with KV channels were placed inside a lipid bilayer mimicking neuronal membrane in terms of lipid composition. We used a pre-equilibrated fragment of bilayer (7.0 × 7.0 × 13.5 nm3; 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine/1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine/cholesterol, POPC:POPE:Ch = 100:50:50 molecules, solvated with 14,172 water molecules) that has been described in detail in our previous works (
      • Berkut A.A.
      • Chugunov A.O.
      • Mineev K.S.
      • Peigneur S.
      • Tabakmakher V.M.
      • Krylov N.A.
      • et al.
      Protein surface topography as a tool to enhance the selective activity of a potassium channel blocker.
      ,
      • Lyukmanova E.N.
      • Shenkarev Z.O.
      • Shulepko M.A.
      • Paramonov A.S.
      • Chugunov A.O.
      • Janickova H.
      • et al.
      Structural insight into specificity of interactions between nonconventional three-finger weak toxin from Naja kaouthia (WTX) and muscarinic acetylcholine receptors.
      ,
      • Chugunov A.O.
      • Volynsky P.E.
      • Krylov N.A.
      • Nolde D.E.
      • Efremov R.G.
      Temperature-sensitive gating of TRPV1 channel as probed by atomistic simulations of its trans- and juxtamembrane domains.
      ); some phospholipid and cholesterol molecules were removed to provide room for the protein. The TIP3P model water (
      • Jorgensen W.L.
      • Chandrasekhar J.
      • Madura J.D.
      • Impey R.W.
      • Klein M.L.
      Comparison of simple potential functions for simulating liquid water.
      ) and the required number of Na+ ions (to maintain electroneutrality) were used for resolvation.
      All systems were equilibrated (heated) during 100 ps of MD simulation. Positions of the channel Cα-atoms of residues not involved in the channel pore vestibule, as well as the Nε atom of K22 in Tk-hefu-10 and 11 were restrained during the equilibration to prevent initial complex destabilization. Systems were then subjected to 501 ns of MD. All simulations were performed with the GROMACS software (https://www.gromacs.org/) (
      • Abraham M.J.
      • Murtola T.
      • Schulz R.
      • Páll S.
      • Smith J.C.
      • Hess B.
      • et al.
      Gromacs: high performance molecular simulations through multi-level parallelism from laptops to supercomputers.
      ) (versions 2018/2020) using the AMBER99SB-ILDN parameters set (
      • Lindorff-Larsen K.
      • Piana S.
      • Palmo K.
      • Maragakis P.
      • Klepeis J.L.
      • Dror R.O.
      • et al.
      Improved side-chain torsion potentials for the Amber ff99SB protein force field.
      ). Simulations were carried out at 37 °C with a time step of 1 fs for the first 1 ns and 2 fs for the rest of MD calculation and imposing 3D periodic boundary conditions, in the isothermal-isobaric (NPT) ensemble with a semi-isotropic pressure of 1 bar, using the Berendsen pressure coupling algorithm (
      • Berendsen H.J.C.
      • Postma J.P.M.
      • Van Gunsteren W.F.
      • Dinola A.
      • Haak J.R.
      Molecular dynamics with coupling to an external bath.
      ) and V-rescale thermostat (
      • Bussi G.
      • Donadio D.
      • Parrinello M.
      Canonical sampling through velocity rescaling.
      ). Van der Waals interactions were truncated using a 1.4 nm spherical cutoff function. Electrostatic interactions were treated with the PME algorithm.
      The mobility of the peptides in the complexes was assessed by calculating the RMSD of the atomic coordinates. The calculations were performed by using the rms utility of GROMACS software (
      • Abraham M.J.
      • Murtola T.
      • Schulz R.
      • Páll S.
      • Smith J.C.
      • Hess B.
      • et al.
      Gromacs: high performance molecular simulations through multi-level parallelism from laptops to supercomputers.
      ), and the results were visualized using the Grace plotting tool (http://plasma-gate.weizmann.ac.il/Grace/).

      Interaction energy and intermolecular contacts

      To assess the potency of Tk-hefu derivatives to interact with KV1.3 and 1.6, we identified the intermolecular contacts during MD and estimated the residual contributions to the interaction energy based on the MD trajectory using the IMPULSE software (https://www.ibch.ru/en/structure/groups/lbm), analogously to the procedures described in our previous studies (
      • Gigolaev A.M.
      • Kuzmenkov A.I.
      • Peigneur S.
      • Tabakmakher V.M.
      • Pinheiro-Junior E.L.
      • Chugunov A.O.
      • et al.
      Tuning scorpion toxin selectivity: switching from KV1.1 to KV1.3.
      ,
      • Tabakmakher V.M.
      • Gigolaev A.M.
      • Peigneur S.
      • Krylov N.A.
      • Tytgat J.
      • Chugunov A.O.
      • et al.
      Potassium channel blocker crafted by α-hairpinin scaffold engineering.
      ,
      • Berkut A.A.
      • Chugunov A.O.
      • Mineev K.S.
      • Peigneur S.
      • Tabakmakher V.M.
      • Krylov N.A.
      • et al.
      Protein surface topography as a tool to enhance the selective activity of a potassium channel blocker.
      ,
      • Kudryavtsev D.S.
      • Tabakmakher V.
      • Budylin G.S.
      • Egorova N.S.
      • Efremov R.G.
      • Ivanov I.A.
      • et al.
      Complex approach for analysis of snake venom α-neurotoxins binding to HAP, the high-affinity peptide.
      ). Briefly, H-bonds were assigned using parameters set from the h-bond utility of GROMACS (
      • Abraham M.J.
      • Murtola T.
      • Schulz R.
      • Páll S.
      • Smith J.C.
      • Hess B.
      • et al.
      Gromacs: high performance molecular simulations through multi-level parallelism from laptops to supercomputers.
      ) (the distance D—A ≤ 0.35 nm and the angle D—H—A ≥ 150° for the hydrogen bond D—H···A); salt bridges, cation-π, stacking, and hydrophobic contacts were calculated using the algorithms described in our previous works (
      • Pyrkov T.V.
      • Efremov R.G.
      A fragment-based scoring function to re-rank ATP docking results.
      ,
      • Pyrkov T.V.
      • Chugunov A.O.
      • Krylov N.A.
      • Nolde D.E.
      • Efremov R.G.
      PLATINUM: a web tool for analysis of hydrophobic/hydrophilic organization of biomolecular complexes.
      ). The AMBER99SB-ILDN parameters set (
      • Lindorff-Larsen K.
      • Piana S.
      • Palmo K.
      • Maragakis P.
      • Klepeis J.L.
      • Dror R.O.
      • et al.
      Improved side-chain torsion potentials for the Amber ff99SB protein force field.
      ) and 1.5 nm cutoff distance for Lennard-Jones or electrostatic interactions were used during the intermolecular short-range nonbonded interaction energy estimation. The latter is the sum of the Lennard-Jones and electrostatic terms. All drawings of 3D structures were prepared with PyMOL. Graphical representation of interaction energy profiles was performed using Python built-in libraries and NumPy package.

      Data availability

      All data are contained within the article and supporting information.

      Supporting information

      This article contains supporting information.

      Conflict of interest

      The authors declare that they have no conflicts of interest with the contents of this article.

      Acknowledgments

      We thank Ilya Yu. Toropygin (V.I. Orekhovich Research Institute of Biomedical Chemistry, Moscow, Russia) for molecular mass measurements and Daniil M. Pavlenko (Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences, Moscow, Russia) for guidance on bioreactor usage. We are grateful to M.L. Garcia for sharing the hKV1.3 clone and to O. Pongs for the rKV1.6 clone. MD simulations were carried out with the use of computational facilities of the Supercomputer Center “Polytechnical” at the St Petersburg Polytechnic University and IACP FEB RAS Shared Resource Center “Far Eastern Computing Resource” equipment (https://cc.dvo.ru).

      Author contributions

      Conceptualization: A.A.V.; Data curation: V.M.T., A.A.V.; Formal Analysis: V.M.T., A.V.F., R.G.E., K.S.M., J.T., A.A.V.; Funding acquisition: S.P., J.T., A.A.V.; Investigation: A.M.G., V.A.L., E.L.P.-J., V.M.T., S.P., A.A.I.; Methodology: V.M.T., A.V.F., R.G.E., K.S.M., J.T., A.A.V.; Project administration: J.T., A.A.V.; Resources: A.V.F., RGE, KSM, JT, AAV. Software: VMT, AVF, RGE, KSM, JT, AAV. Supervision: AVF, R.G.E., K.S.M., JT, AAV. Validation: A.V.F., R.G.E., K.S.M., J.T., AAV. Visualization: A.MG., E.L.P.-J., V.M.T., A.A.I., K.S.M.; Writing – original draft: AMG, ELP-J, VMT, AAI, KSM. Writing – review & editing: R.G.E., A.A.V.

      Funding and additional information

      This work was supported by the Russian Science Foundation (grant no. 20-44-01015 to A. A. V.), FWO-Vlaanderen (grants GOA4919N , GOE7120N and GOC2319N to J. T. and 12W7822N to S. P.), and KU Leuven ( PDM/19/164 to S. P.).

      Supporting information

      References

        • Hille B.
        Ion Channels of Excitable Membranes.
        3rd Ed. Sinauer Associates, Inc, Sunderland, MA2001
        • Zeisel A.
        • Hochgerner H.
        • Lönnerberg P.
        • Johnsson A.
        • Memic F.
        • van der Zwan J.
        • et al.
        Molecular architecture of the mouse nervous system.
        Cell. 2018; 174: 999-1014.e22
        • Manicam C.
        • Staubitz J.
        • Brochhausen C.
        • Grus F.H.
        • Pfeiffer N.
        • Gericke A.
        The gatekeepers in the mouse ophthalmic artery: endothelium-dependent mechanisms of cholinergic vasodilation.
        Sci. Rep. 2016; 6: 1-13
        • Carrisoza-Gaytán R.
        • Salvador C.
        • Diaz-Bello B.
        • Escobar L.I.
        Differential expression of the Kv1 voltage-gated potassium channel family in the rat nephron.
        J. Mol. Histol. 2014; 45: 583-597
        • Shamotienko O.G.
        • Parcej D.N.
        • Dolly J.O.
        Subunit combinations defined for K+ channel Kv1 subtypes in synaptic membranes from bovine brain.
        Biochemistry. 1997; 36: 8195-8201
        • Dodson P.D.
        • Barker M.C.
        • Forsythe I.D.
        Two heteromeric Kv1 potassium channels differentially regulate action potential firing.
        J. Neurosci. 2002; 22: 6953-6961
        • Gunasekaran R.
        • Narayani R.S.
        • Vijayalakshmi K.
        • Alladi P.A.
        • Shobha K.
        • Nalini A.
        • et al.
        Exposure to cerebrospinal fluid of sporadic amyotrophic lateral sclerosis patients alters Nav1.6 and Kv1.6 channel expression in rat spinal motor neurons.
        Brain Res. 2009; 1255: 170-179
        • Orts D.J.B.
        • Peigneur S.
        • Madio B.
        • Cassoli J.S.
        • Montandon G.G.
        • Pimenta A.M.C.
        • et al.
        Biochemical and electrophysiological characterization of two sea anemone type 1 potassium toxins from a geographically distant population of bunodosoma caissarum.
        Mar. Drugs. 2013; 11: 655-679
        • Orts D.J.B.
        • Peigneur S.
        • Silva-Gonçalves L.C.
        • Arcisio-Miranda M.
        • Bicudo J.E.P.W.
        • Tytgat J.
        AbeTx1 is a novel sea anemone toxin with a dual mechanism of action on shaker-type K+ channels activation.
        Mar. Drugs. 2018; 16: 360
        • ElFessi-Magouri R.
        • Peigneur S.
        • Othman H.
        • Srairi-Abid N.
        • ElAyeb M.
        • Tytgat J.
        • et al.
        Characterization of Kbot21 reveals novel side chain interactions of scorpion toxins inhibiting voltage-gated potassium channels.
        PLoS One. 2015; 10e0137611
        • Imperial J.S.
        • Chen P.
        • Sporning A.
        • Terlau H.
        • Daly N.L.
        • Craik D.J.
        • et al.
        Tyrosine-rich conopeptides affect voltage-gated K+ channels.
        J. Biol. Chem. 2008; 283: 23026-23032
        • Vandendriessche T.
        • Kopljar I.
        • Jenkins D.P.
        • Diego-Garcia E.
        • Abdel-Mottaleb Y.
        • Vermassen E.
        • et al.
        Purification, molecular cloning and functional characterization of HelaTx1 (Heterometrus laoticus): the first member of a new κ-KTX subfamily.
        Biochem. Pharmacol. 2012; 83: 1307-1317
        • Imperial J.S.
        • Bansal P.S.
        • Alewood P.F.
        • Daly N.L.
        • Craik D.J.
        • Sporning A.
        • et al.
        A novel conotoxin inhibitor of Kv1.6 channel and nAChR subtypes defines a new superfamily of conotoxins.
        Biochemistry. 2006; 45: 8331-8340
        • Gigolaev A.M.
        • Kuzmenkov A.I.
        • Peigneur S.
        • Tabakmakher V.M.
        • Pinheiro-Junior E.L.
        • Chugunov A.O.
        • et al.
        Tuning scorpion toxin selectivity: switching from KV1.1 to KV1.3.
        Front. Pharmacol. 2020; 11: 1010
        • Tabakmakher V.M.
        • Kuzmenkov A.I.
        • Gigolaev A.M.
        • Pinheiro-Junior E.L.
        • Peigneur S.
        • Efremov R.G.
        • et al.
        Artificial peptide ligand of potassium channel KV1.1 with high selectivity.
        J. Evol. Biochem. Physiol. 2021; 57: 386-403
        • Tabakmakher V.M.
        • Gigolaev A.M.
        • Peigneur S.
        • Krylov N.A.
        • Tytgat J.
        • Chugunov A.O.
        • et al.
        Potassium channel blocker crafted by α-hairpinin scaffold engineering.
        Biophys. J. 2021; 120: 2471-2481
        • Berkut A.A.
        • Usmanova D.R.
        • Peigneur S.
        • Oparin P.B.
        • Mineev K.S.
        • Odintsova T.I.
        • et al.
        Structural similarity between defense peptide from wheat and scorpion neurotoxin permits rational functional design.
        J. Biol. Chem. 2014; 289: 14331-14340
        • Berkut A.A.
        • Chugunov A.O.
        • Mineev K.S.
        • Peigneur S.
        • Tabakmakher V.M.
        • Krylov N.A.
        • et al.
        Protein surface topography as a tool to enhance the selective activity of a potassium channel blocker.
        J. Biol. Chem. 2019; 294: 18349-18359
        • Srinivasan K.N.
        • Sivaraja V.
        • Huys I.
        • Sasaki T.
        • Cheng B.
        • Kumar T.K.S.
        • et al.
        κ-Hefutoxin1, a novel toxin from the scorpion Heterometrus fulvipes with unique structure and function.
        J. Biol. Chem. 2002; 277: 30040-30047
        • Kuzmenkov A.I.
        • Nekrasova O.V.
        • Peigneur S.
        • Tabakmakher V.M.
        • Gigolaev A.M.
        • Fradkov A.F.
        • et al.
        KV1.2 channel-specific blocker from Mesobuthus eupeus scorpion venom: structural basis of selectivity.
        Neuropharmacology. 2018; 143: 228-238
        • Swartz K.J.
        • MacKinnon R.
        Hanatoxin modifies the gating of a voltage-dependent K+ channel through multiple binding sites.
        Neuron. 1997; 18: 665-673
        • Tao H.
        • Chen J.J.
        • Xiao Y.C.
        • Wu Y.Y.
        • Su H.B.
        • Li D.
        • et al.
        Analysis of the interaction of tarantula toxin Jingzhaotoxin-III (β-TRTX-Cj1α) with the voltage sensor of Kv2.1 uncovers the molecular basis for cross-activities on Kv2.1 and Nav1.5 channels.
        Biochemistry. 2013; 52: 7439-7448
        • Kozlov S.A.
        • Vassilevski A.A.
        • Feofanov A.V.
        • Surovoy A.Y.
        • Karpunin D.V.
        • Grishin E.V.
        Latarcins, antimicrobial and cytolytic peptides from the venom of the spider Lachesana tarabaevi (Zodariidae) that exemplify biomolecular diversity.
        J. Biol. Chem. 2006; 281: 20983-20992
        • Nolde S.B.
        • Vassilevski A.A.
        • Rogozhin E.A.
        • Barinov N.A.
        • Balashova T.A.
        • Samsonova O.V.
        • et al.
        Disulfide-stabilized helical hairpin structure and activity of a novel antifungal peptide EcAMP1 from seeds of barnyard grass (Echinochloa crus-galli).
        J. Biol. Chem. 2011; 286: 25145-25153
        • Oparin P.B.
        • Mineev K.S.
        • Dunaevsky Y.E.
        • Arseniev A.S.
        • Belozersky M.A.
        • Grishin E.V.
        • et al.
        Buckwheat trypsin inhibitor with helical hairpin structure belongs to a new family of plant defence peptides.
        Biochem. J. 2012; 446: 69-77
        • Banerjee A.
        • Lee A.
        • Campbell E.
        • MacKinnon R.
        Structure of a pore-blocking toxin in complex with a eukaryotic voltage-dependent K+ channel.
        Elife. 2013; 2e00594
        • Kasheverov I.E.
        • Oparin P.B.
        • Zhmak M.N.
        • Egorova N.S.
        • Ivanov I.A.
        • Gigolaev A.M.
        • et al.
        Scorpion toxins interact with nicotinic acetylcholine receptors.
        FEBS Lett. 2019; 593: 2779-2789
        • McCoy J.
        • LaVallie E.
        Expression and purification of thioredoxin fusion proteins.
        Curr. Protoc. Mol. Biol. 2001; (Chapter 16:Unit16.8)
        • Gasparian M.E.
        • Ostapchenko V.G.
        • Schulga A.A.
        • Dolgikh D.A.
        • Kirpichnikov M.P.
        Expression, purification, and characterization of human enteropeptidase catalytic subunit in Escherichia coli.
        Protein Expr. Purif. 2003; 31: 133-139
        • Kuzmenkov A.I.
        • Sachkova M.Y.
        • Kovalchuk S.I.
        • Grishin E.V.
        • Vassilevski A.A.
        Lachesana tarabaevi, an expert in membrane-active toxins.
        Biochem. J. 2016; 473: 2495-2506
        • Boldrini-França J.
        • Pinheiro-Junior E.L.
        • Peigneur S.
        • Pucca M.B.
        • Cerni F.A.
        • Borges R.J.
        • et al.
        Beyond hemostasis: a snake venom serine protease with potassium channel blocking and potential antitumor activities.
        Sci. Rep. 2020; 10: 4476
        • Vassilevski A.A.
        • Kozlov S.A.
        • Egorov T.A.
        • Grishin E.V.
        Purification and characterization of biologically active peptides from spider venoms.
        Methods Mol. Biol. 2010; 615: 87-100
        • Vorontsova O.V.
        • Egorova N.S.
        • Arseniev A.S.
        • Feofanov A.V.
        Haemolytic and cytotoxic action of latarcin Ltc2a.
        Biochimie. 2011; 93: 227-241
        • Efremenko A.V.
        • Ignatova A.A.
        • Grin M.A.
        • Sivaev I.B.
        • Mironov A.F.
        • Bregadze V.I.
        • et al.
        Chlorin e6 fused with a cobalt-bis(dicarbollide) nanoparticle provides efficient boron delivery and photoinduced cytotoxicity in cancer cells.
        Photochem. Photobiol. Sci. 2014; 13: 92-102
        • Herrmann T.
        • Güntert P.
        • Wüthrich K.
        Protein NMR structure determination with automated NOE-identification in the NOESY spectra using the new software ATNOS.
        J. Biomol. NMR. 2002; 24: 171-189
        • Koradi R.
        • Billeter M.
        • Wüthrich K.
        MOLMOL: a program for display and analysis of macromolecular structures.
        J. Mol. Graph. 1996; 14: 51-55
        • Webb B.
        • Sali A.
        Comparative protein structure modeling using MODELLER.
        Curr. Protoc. Bioinformatics. 2016; 54: 5.6.1-5.6.37
        • Chen X.
        • Wang Q.
        • Ni F.
        • Ma J.
        Structure of the full-length Shaker potassium channel Kv1.2 by normal-mode-based X-ray crystallographic refinement.
        Proc. Natl. Acad. Sci. U. S. A. 2010; 107: 11352-11357
        • Kudryavtsev D.S.
        • Tabakmakher V.
        • Budylin G.S.
        • Egorova N.S.
        • Efremov R.G.
        • Ivanov I.A.
        • et al.
        Complex approach for analysis of snake venom α-neurotoxins binding to HAP, the high-affinity peptide.
        Sci. Rep. 2020; 10: 3861
        • Lyukmanova E.N.
        • Shenkarev Z.O.
        • Shulepko M.A.
        • Paramonov A.S.
        • Chugunov A.O.
        • Janickova H.
        • et al.
        Structural insight into specificity of interactions between nonconventional three-finger weak toxin from Naja kaouthia (WTX) and muscarinic acetylcholine receptors.
        J. Biol. Chem. 2015; 290: 23616-23630
        • Chugunov A.O.
        • Volynsky P.E.
        • Krylov N.A.
        • Nolde D.E.
        • Efremov R.G.
        Temperature-sensitive gating of TRPV1 channel as probed by atomistic simulations of its trans- and juxtamembrane domains.
        Sci. Rep. 2016; 6: 33112
        • Jorgensen W.L.
        • Chandrasekhar J.
        • Madura J.D.
        • Impey R.W.
        • Klein M.L.
        Comparison of simple potential functions for simulating liquid water.
        J. Chem. Phys. 1983; 79: 926
        • Abraham M.J.
        • Murtola T.
        • Schulz R.
        • Páll S.
        • Smith J.C.
        • Hess B.
        • et al.
        Gromacs: high performance molecular simulations through multi-level parallelism from laptops to supercomputers.
        SoftwareX. 2015; 1–2: 19-25
        • Lindorff-Larsen K.
        • Piana S.
        • Palmo K.
        • Maragakis P.
        • Klepeis J.L.
        • Dror R.O.
        • et al.
        Improved side-chain torsion potentials for the Amber ff99SB protein force field.
        Proteins. 2010; 78: 1950-1958
        • Berendsen H.J.C.
        • Postma J.P.M.
        • Van Gunsteren W.F.
        • Dinola A.
        • Haak J.R.
        Molecular dynamics with coupling to an external bath.
        J. Chem. Phys. 1984; 81: 3684
        • Bussi G.
        • Donadio D.
        • Parrinello M.
        Canonical sampling through velocity rescaling.
        J. Chem. Phys. 2007; 126: 014101
        • Pyrkov T.V.
        • Efremov R.G.
        A fragment-based scoring function to re-rank ATP docking results.
        Int. J. Mol. Sci. 2007; 10: 1083-1094
        • Pyrkov T.V.
        • Chugunov A.O.
        • Krylov N.A.
        • Nolde D.E.
        • Efremov R.G.
        PLATINUM: a web tool for analysis of hydrophobic/hydrophilic organization of biomolecular complexes.
        Bioinformatics. 2009; 25: 1201-1202