AgTx2-GFP, Fluorescent Blocker Targeting Pharmacologically Important Kv1.x (x = 1, 3, 6) Channels

The growing interest in potassium channels as pharmacological targets has stimulated the development of their fluorescent ligands (including genetically encoded peptide toxins fused with fluorescent proteins) for analytical and imaging applications. We report on the properties of agitoxin 2 C-terminally fused with enhanced GFP (AgTx2-GFP) as one of the most active genetically encoded fluorescent ligands of potassium voltage-gated Kv1.x (x = 1, 3, 6) channels. AgTx2-GFP possesses subnanomolar affinities for hybrid KcsA-Kv1.x (x = 3, 6) channels and a low nanomolar affinity to KcsA-Kv1.1 with moderate dependence on pH in the 7.0–8.0 range. Electrophysiological studies on oocytes showed a pore-blocking activity of AgTx2-GFP at low nanomolar concentrations for Kv1.x (x = 1, 3, 6) channels and at micromolar concentrations for Kv1.2. AgTx2-GFP bound to Kv1.3 at the membranes of mammalian cells with a dissociation constant of 3.4 ± 0.8 nM, providing fluorescent imaging of the channel membranous distribution, and this binding depended weakly on the channel state (open or closed). AgTx2-GFP can be used in combination with hybrid KcsA-Kv1.x (x = 1, 3, 6) channels on the membranes of E. coli spheroplasts or with Kv1.3 channels on the membranes of mammalian cells for the search and study of nonlabeled peptide pore blockers, including measurement of their affinity.


Introduction
Many peptide toxins from scorpion venoms are potent and specific blockers of potassium channels [1,2], which are tetrameric transmembrane proteins with a central ionconducting pore [3]. The toxins bind at the outer vestibule of a channel, occluding the pore with a lysine residue. The main targets of scorpion peptide toxins are voltage-gated K v 1.x channels, calcium-activated K + channels (K Ca ), and K v 11.x channels [1,2,4]. The affinity profile of each peptide toxin usually comprises a limited set of K + -channel isoforms, suggesting that scorpion toxins are able to discriminate a few targets from dozens of different K + channels [1,5]. For example, the OSK1 toxin from the scorpion Orthochirus scrobiculosus 2. Results and Discussion 2.1. Interactions of AgTx2-GFP with KcsA-Kv1.x (x = 1, 3, 6) Channels AgTx2-GFP was obtained by fusing GFP with the C-terminus of AgTx2 by means of a flexible 20 amino acid linker (GSGGSGGSGGTGGAGGATST) [46].
To characterize quantitatively the previously revealed interactions of AgTx2-GFP with the pore blocker binding sites of K v 1.x (x = 1, 3, 6) channels, the concentration dependences of AgTx2-GFP binding to the hybrid potassium channels KcsA-K v 1.x (x = 1, 3, 6) expressed in the inner membranes of E. coli cells were measured ( Figure 1). KcsA-K v 1.x (x = 1, 3, 6) channels comprised the pore blocker outer-binding sites of corresponding eukaryotic channels, which were transferred to the bacterial KcsA channel (Figure 1c) [44,50,51]. As shown earlier, the affinities of peptide pore blockers to KcsA-K v 1.x (x = 1, 3, 6) channels were similar to their affinities to corresponding eukaryotic K v 1 channels [52].
The confocal microscopy analysis of AgTx2-GFP interactions with KcsA-K v 1.
x (x = 1, 3, 6) channels on the membranes of spheroplasts prepared from E. coli cells revealed saturable binding that occurred in subnanomolar (KcsA-Kv1.3) and low nanomolar (KcsA-Kv1.1 and KcsA-Kv1.6) concentration ranges (Figure 1d-f). This AgTx2-GFP binding was reversible, and peptide blockers of Kv1.x (x = 1, 3, 6) channels displaced AgTx2-GFP from spheroplasts bearing KcsA-Kv1.x (x = 1, 3, 6) channels [46]. In addition, AgTx2-GFP binding to the E. coli membrane itself, as well as to the KcsA channel embedded in the E. coli membrane, was negligible compared to its binding to spheroplasts bearing KcsA-Kv1.x (x = 1, 3, 6) channels [46]. These findings allowed us to validly fit experimental data using Equation (1) (Figure 1d-f) and to determine the dissociation constants Kd of AgTx2-GFP complexes with KcsA-Kv1.x (x = 1, 3, 6) channels (Table 1). The affinity of AgTx2-GFP to the hybrid channels reduced in the manner of KcsA-Kv1.3 > KcsA-Kv1.6 >> KcsA-Kv1.1 and was found to be moderately sensitive to pH in the neutral to slightly basic pH range ( Table 1). The affinity of AgTx2-GFP to KcsA-Kv1.1 was the lowest at pH 7.0 and increased at higher pH values ( Table 1). The affinities of AgTx2-GFP to KcsA-Kv1.3 and KcsA-Kv1.6 were highest at pH 7.5 and decreased at both pH 7.0 and pH 8.0 (Table 1). Since the pKa values of most amino acids are far from the pH range of 7-8 and cysteines of AgTx2 are in the form of cystines, the histidine residues with pKa = 6.3 were the most likely candidates whose protonation-deprotonation equilibrium  The affinity of AgTx2-GFP to the hybrid channels reduced in the manner of KcsA-K v 1.3 > KcsA-K v 1.6 >> KcsA-K v 1.1 and was found to be moderately sensitive to pH in the neutral to slightly basic pH range ( Table 1). The affinity of AgTx2-GFP to KcsA-K v 1.1 was the lowest at pH 7.0 and increased at higher pH values ( Table 1). The affinities of AgTx2-GFP to KcsA-K v 1.3 and KcsA-K v 1.6 were highest at pH 7.5 and decreased at both pH 7.0 and pH 8.0 (Table 1). Since the pK a values of most amino acids are far from the pH range of 7-8 and cysteines of AgTx2 are in the form of cystines, the histidine residues with pK a = 6.3 were the most likely candidates whose protonation-deprotonation equilibrium affected the K d of the studied complexes. H34 of AgTx2 and H58 of the binding site of KcsA-K v 1.1 could be responsible for the revealed pH dependence, while the absence of histidine residues in the binding sites of KcsA-K v 1.3 and KcsA-K v 1.6 could be a reason for the variation in the character of pH dependences of these hybrid channels compared to KcsA-K v 1.1 (Table 1). Additionally, a decrease in the protonation state of the N-terminal amino group of AgTx2, which occurred at the increase in pH from 7.0 to 8.0, could affect peptide binding.
As compared to previously studied FP-Txs, AgTx2-GFP was inferior to RFP-AgTx2 and GFP-OSK1 in its affinity to KcsA-K v 1.1 (Figure 2a) [7]. At the same time, AgTx2-GFP surpassed all the studied FP-Txs in its reported affinities to KcsA-K v 1.3 ( Figure 2a) and KcsA-K v 1.6 [7]. It should be noted that K d were measured for AgTx2-GFP at an ionic strength that was noticeably higher than in the measurements of K d values for other FP-Txs. A decrease in the ionic strength of the solution can increase the affinities of peptide blockers to the binding sites of K v 1 channels due to the considerable role of electrostatic interactions in the formation of a peptide-channel complex [50,51].
interactions in the formation of a peptide-channel complex [50,51].
AgTx2-GFP could be displaced from complexes with KcsA-Kv1.x (x = 1, 3, 6) not only by pure, specific peptide pore blockers [46], but also by complex mixtures of natural peptides, such as crude scorpion venoms containing pore blockers (Figure 2b). This result confirms that AgTx2-GFP in combination with KcsA-Kv1.x (x = 1, 3, 6) channels could be used to search for Kv1 channel pore blockers among individual compounds and in complex natural extracts and venoms. As demonstrated earlier [51], the targeted detection of specific activity using a combination of a fluorescent ligand and KcsA-Kv1.x (x = 1, 3, 6) channels facilitated significantly the targeted isolation of individual peptide blockers of Kv1 channels from venoms. x (x = 1, 3) complexes with AgTx2-GFP and with previously studied FP-tagged pore blockers. * [7], # [47], & [46], ^ [45]. (b) Displacement of AgTx2-GFP (5 nM) AgTx2-GFP could be displaced from complexes with KcsA-K v 1.x (x = 1, 3, 6) not only by pure, specific peptide pore blockers [46], but also by complex mixtures of natural peptides, such as crude scorpion venoms containing pore blockers (Figure 2b). This result confirms that AgTx2-GFP in combination with KcsA-K v 1.x (x = 1, 3, 6) channels could be used to search for K v 1 channel pore blockers among individual compounds and in complex natural extracts and venoms. As demonstrated earlier [51], the targeted detection of specific activity using a combination of a fluorescent ligand and KcsA-K v 1.x (x = 1, 3, 6) channels facilitated significantly the targeted isolation of individual peptide blockers of K v 1 channels from venoms.
Displacement of AgTx2-GFP from the complexes with KcsA-K v 1.
x (x = 1, 3, 6) channels by the nonlabeled peptide blocker occurred due to the competitive binding to the K v 1 channel binding site. A quantitative analysis of such displacement by the studied peptide blocker for the particular channels (Figure 2c-e) allowed the estimation of the peptide concentration, which displaced 50% of AgTx2-GFP from the complexes (DC 50 ), and the calculation of the apparent dissociation constant K ap of the studied peptide blocker (Equations (2) and (3)) [44]. The K ap values determined for KTx1 and AgTx2 using the AgTx2-GFP/KcsA-K v 1.x (x = 1, 3, 6) bioengineering analytical systems (Table 2) allowed us to conclude that AgTx2-GFP was applicable as a component of these systems for the study of peptide blockers and their affinities to K v 1 channels. Table 2. Apparent dissociation constants K ap (nM) of AgTx2 and KTx1 complexes with KcsA-K v 1.x (x = 1, 3, 6) channels.

Toxin
KcsA The affinities of AgTx2 to KcsA-K v 1.3 and KcsA-K v 1.6 estimated with AgTx2-GFP (Table 2) are in general agreement with the data on AgTx2 activity (IC 50 , blocker concentration inhibiting current through the channel by 50%) reported in electrophysiological studies: 4 pM [10], 50 pM [34], 200 pM [53], and 170 pM ( Figure 3, Table 3) for K v 1.3 and 37 pM [10] and 600 pM ( Figure 3, Table 3) for K v 1.6. The measured affinity of AgTx2 to KcsA-K v 1.1 ( Table 2) is close to the previously reported K ap value determined with another fluorescent ligand [51], as well as to recent IC 50 values of 2 nM [7] and 4.2 nM ( Figure 3, Table 3) obtained in electrophysiological studies of the K v 1.1 channel. At the same time, earlier electrophysiological studies have reported much lower IC 50 values (44 pM [10] and 130 pM [34]) for inhibition of the K v 1.1 channel by AgTx2, and the reason for these differences is unclear.
crude venoms of Mesobuthus eupeus and Orthochirus scrobiculosus scorpions. (c-e) Displacement of AgTx2-GFP from complexes with KcsA-Kv1.1 (c), KcsA-Kv1.3 (d), and KcsA-Kv1.6 (e) by different concentrations B of KTx1 and AgTx2. Ia-relative fluorescence intensity of AgTx2-GFP bound to spheroplasts bearing KcsA-Kv1.x (x = 1, 3, 6) channels. Concentrations of AgTx2-GFP in the experiments with KcsA-Kv1.1, KcsA-Kv1.3, and KcsA-Kv1.6 were 80, 8, and 2 nM, respectively. Ia values are presented as means ± SD. Displacement of AgTx2-GFP from the complexes with KcsA-Kv1.x (x = 1, 3, 6) channels by the nonlabeled peptide blocker occurred due to the competitive binding to the Kv1 channel binding site. A quantitative analysis of such displacement by the studied peptide blocker for the particular channels (Figure 2c-e) allowed the estimation of the peptide concentration, which displaced 50% of AgTx2-GFP from the complexes (DC50), and the calculation of the apparent dissociation constant Kap of the studied peptide blocker (Equations (2) and (3)) [44]. The Kap values determined for KTx1 and AgTx2 using the AgTx2-GFP/KcsA-Kv1.x (x = 1, 3, 6) bioengineering analytical systems ( Table 2) allowed us to conclude that AgTx2-GFP was applicable as a component of these systems for the study of peptide blockers and their affinities to Kv1 channels.
The affinities of AgTx2 to KcsA-Kv1.3 and KcsA-Kv1.6 estimated with AgTx2-GFP ( Table 2) are in general agreement with the data on AgTx2 activity (IC50, blocker concentration inhibiting current through the channel by 50%) reported in electrophysiological studies: 4 pM [10], 50 pM [34], 200 pM [53], and 170 pM ( Figure 3, Table 3) for Kv1.3 and 37 pM [10] and 600 pM ( Figure 3, Table 3) for Kv1.6. The measured affinity of AgTx2 to KcsA-Kv1.1 ( Table 2) is close to the previously reported Kap value determined with another fluorescent ligand [51], as well as to recent IC50 values of 2 nM [7] and 4.2 nM ( Figure 3,  [34]. To our best knowledge, there are no quantitative electrophysiological data on the inhibition of Kv1.6 by KTx1, but only qualitative ones [54].  The K ap values of KTx1 complexes with KcsA-K v 1.1 and KcsA-K v 1.3 (Table 2) correlated to the IC 50 values reported for KTx1 complexes with K v 1.1 (1.1 nM) and K v 1.3 (0.1 nM) [34]. To our best knowledge, there are no quantitative electrophysiological data on the inhibition of K v 1.6 by KTx1, but only qualitative ones [54]. Table 3. Concentrations of AgTx2 and AgTx2-GFP that induced 50% inhibition of the current through K v 1.x (x = 1-3,6) channels in oocytes (IC 50 , nM).

Electrophysiological Studies of AgTx2-GFP
According to the data on electrophysiological measurements on oocytes, AgTx2-GFP was a pore blocker of K v 1.x (x = 1-3, 6) channels ( Figure 3). Its affinity to the channels decreased in the following manner: Table 3). C-terminal attachment of GFP did not affect the affinity of AgTx2 to K v 1.1 but decreased its activity by two, ten, and more than thirty times to K v 1.6, K v 1.3, and K v 1.2, respectively, thus changing noticeably the natural pharmacological profile of AgTx2 (Table 3). A significant decrease in the affinity of AgTx2-GFP to the Kv1.2 channel could be associated with the disturbance of strong interactions between the channel and amino acid residues near the C-terminus of the peptide by a volumetric fluorescent protein. According to 1H-NMR studies, these could be the Lys32 and His34 of AgTx2 forming a salt bridge with Asp355 and hydrophobic contacts with the Gly378, Asp379, and Val381 of the channel, respectively [55].
Finally, K v 1.6, K v 1.3, and K v .1.1 differed in affinity to AgTx2-GFP by less than three times, which made AgTx2-GFP a high-affinity fluorescent marker of Kv1 channels of a wide spectrum of action.
The affinities of AgTx2-GFP to hybrid KcsA-K v 1.
x (x = 1, 3, 6) channels (Table 1) and corresponding K v 1.x (x = 1, 3, 6) channels (Table 3) differed noticeably, reflecting the influences of buffer composition and the structural differences of hybrid and natural channels on the binding of the bulky AgTx2-GFP ligand. Apparently, the FP (in our case, GFP) could participate in interactions with amino acid residues of the channel surrounding or beyond the natural binding site of the peptide blocker and, thus, modulate the overall affinity of the ligand to the channel. Differences in the amino acid compositions of KcsA-K v 1.x and the corresponding K v 1.x channels beyond the binding site led to differences in interactions with GFP moiety and, finally, to differences in the stability of the complexes.

Interaction of AgTx2-GFP with K v 1.3 on Mammalian Cell Membranes
Recently, we created a plasmid encoding a natural (M1-D52)-truncated variant of the K v 1.3 channel [56] fused with red fluorescent protein mKate2 (mKate2-K v 1.3), which is effectively expressed in the plasma membranes of mammalian cells and retains voltagegated K + conductivity and the ability to bind peptide pore blockers [42]. Our studies show that AgTx2-GFP bound to the membrane of Neuro 2A and HEK293 cells expressing the mKate2-K v 1.3 channel (Figures 4 and S1). The K v 1.3 channel was obviously a target of AgTx2-GFP in these cells since no binding of AgTx2-GFP to cells was observed in the absence of mKate2-K v 1.3 channels ( Figure S2). Moreover, AgTx2 displaced AgTx2-GFP from the surfaces of cells expressing the mKate2-K v 1.3 channel, indicating competition of these ligands for the pore blocker binding site of K v 1.3 (Figures 4 and S1).
To characterize the affinity of AgTx2-GFP to K v 1.3 on the membranes of mammalian cells, we measured the concentration dependence of AgTx2-GFP binding to mKate2-K v 1.3expressing Neuro 2A cells using confocal microscopy and calculated the ratios (R av ) of the fluorescent intensity of bound AgTx2-GFP to the fluorescent intensity of membranous mKate2-K v 1.3 at different concentrations of added AgTx2-GFP (Figure 5a). Previously, it was shown that the analysis of such a dependence using Equation (5) allowed one to calculate the K d of the formed complexes [42]. The measurements were carried out after a 30 min incubation of the cells with AgTx2-GFP when the complexation process reached equilibrium ( Figure S3). The calculated K d of AgTx2-GFP complexes with K v 1.3 on the Neuro 2A cell membranes was 3.4 ± 0.8 nM, which was consistent with the data of the electrophysiological measurements (Table 3). To characterize the affinity of AgTx2-GFP to Kv1.3 on the membranes of mammalian cells, we measured the concentration dependence of AgTx2-GFP binding to mKate2-Kv1.3-expressing Neuro 2A cells using confocal microscopy and calculated the ratios (Rav) of the fluorescent intensity of bound AgTx2-GFP to the fluorescent intensity of membranous mKate2-Kv1.3 at different concentrations of added AgTx2-GFP (Figure 5a). Previously, it was shown that the analysis of such a dependence using Equation (5) allowed one to calculate the Kd of the formed complexes [42]. The measurements were carried out after a 30 min incubation of the cells with AgTx2-GFP when the complexation process reached equilibrium ( Figure S3). The calculated Kd of AgTx2-GFP complexes with Kv1.3 on the Neuro 2A cell membranes was 3.4 ± 0.8 nM, which was consistent with the data of the electrophysiological measurements (Table 3). It is worth mentioning that the measurements of the affinities of AgTx2-GFP to the Kv1.3 channel with the fluorescent cell-based assay (Figure 5a) and electrophysiological   (Figure 5a). Pre ously, it was shown that the analysis of such a dependence using Equation (5) allow one to calculate the Kd of the formed complexes [42]. The measurements were carried after a 30 min incubation of the cells with AgTx2-GFP when the complexation proc reached equilibrium ( Figure S3). The calculated Kd of AgTx2-GFP complexes with Kv on the Neuro 2A cell membranes was 3.4 ± 0.8 nM, which was consistent with the data the electrophysiological measurements (Table 3).  It is worth mentioning that the measurements of the affinities of AgTx2-GFP to the K v 1.3 channel with the fluorescent cell-based assay (Figure 5a) and electrophysiological approach (Figure 3) were performed at different states of the channel-closed and open, respectively. Similar values of K d and IC 50 obtained in these measurements allowed us to assume that the affinity of AgTx2-GFP to K v 1.3 depended weakly, if at all, on the state of the channel (open or closed). To verify this assumption using the same technique, we compared changes in AgTx2-GFP binding to K v 1.3 on the cell membranes in buffers containing either 150 mM NaCl or 150 mM KCl (Figure 5b). The presence of NaCl in the outer medium maintained the predominantly closed state of the channels, while KCl induced depolarization of the plasma membrane and the opening of the channels. The calculated K d values of the AgTx2-GFP complexes with open and closed K v 1.3 channels were 3 ± 1 and 4.7 ± 0.8 nM, respectively. The difference between these K d values was insignificant, thus confirming the assumption made.
Considering a probable application of AgTx2-GFP as a fluorescent ligand for the study of interactions of various peptide blockers with K v 1.3 channels on mammalian cells, we measured the concentration dependence of the competitive binding of AgTx2 to K v 1.3 at a constant concentration (8 nM) of AgTx2-GFP added to cells following an earlier developed procedure [42]. An analysis of the displacement of AgTx2-GFP from complexes with mKate2-K v 1.3 by increasing concentrations of AgTx2 allowed one to calculate the AgTx2 concentration displacing 50% of the AgTx2-GFP from the complexes (DC 50 , Equation (6)) and the K ap of the complexes between AgTx2 and K v 1.3 (Equation (7)). The calculated K ap was 0.7 ± 0.2 nM, which is in a good agreement with the electrophysiological data presented in Table 3 and published earlier [10,34,53].
The data obtained allowed one to conclude that AgTx2-GFP was suitable for the fluorescent imaging of K v 1.3 channels in mammalian cells, as well as for recognizing K v 1.3 blockers among the studied peptide toxins and assessing their affinity for K v 1.3, when AgTx2-GFP was used in combination with mammalian cells expressing fluorescent (FPtagged) K v 1.3 channels. Such studies are also useful at an earlier stage of the selection of a candidate drug among peptides. In vitro cell experiments are a mandatory part of preclinical studies for any candidate drug. The initial selection and study of peptide blockers is often performed in buffer solutions by means of electrophysiological techniques on Xenopus oocytes or radioligand analyses on isolated membranes. These results need to be verified in a more complex and natural environment, where background proteolytic activity and various nonspecific interactions with the components of a mammalian cell growth medium can affect the blocker-channel interactions.
AgTx2-GFP is a high-affinity fluorescent ligand of the K v 1.1, K v 1.3, and K v 1.6 channels whose increased expression and activity are associated with a number of diseases. The K v 1.3 channel is involved in the development of autoimmune or neuroimmune inflammation through the induction of pro-inflammatory response and the activation of T lymphocytes or microglial cells, respectively [57,58]. CD4+ effector memory T cells mediate many autoimmune diseases, including multiple sclerosis, type-I diabetes mellitus, and rheumatoid arthritis. The selective blockage of K v 1.3 channels in these cells inhibits their proliferation and cytokine production, thereby contributing to a reduction in inflammation and providing a therapeutic effect [23]. The pathogeneses of oncological diseases, such as breast and prostate cancers, pancreatic ductal adenocarcinoma, or B-type chronic lymphocytic leukemia, are associated with the upregulation of K v 1.3 channel expression [59]. The inhibition of K v 1.3 channels by either small organic blockers or peptide toxins from animal venoms results in a significant therapeutic effect without major side effects [60]. Inherited loss-of-function mutations of the K v 1.1 channel in humans are associated with episodic ataxia type I [61], while some other deletions or mutations of the K v 1.1 and K v 1.2 channels provoke epilepsy [62]. Although the blockage of K v 1.1 and K v 1.2 channels has no therapeutic value in cases of increased excitability of neurons, the inhibition of K v 1.1 channels in inhibitory (demyelinated) axons is known to improve neuronal conduction in multiple sclerosis [63]. While the physiological role of the K v 1.6 channel, which is widely expressed in the central nervous system and peripheral neurons, is still poorly determined, the localization of K v 1.6 in microglial cells in the striatum suggests its potential role in neuroinflammation and striatal disorders, such as Parkinson's disease [64].
Comparing the prospects of the best organic fluorophores and FPs for the fluorescent labeling of peptide ligands of K v 1 channels, it can be stated that organic fluorophores are usually brighter and more photostable, but the differences are gradually being leveled due to the extensive development of new FP variants [65][66][67]. For many laboratories that are familiar with the technology of recombinant proteins, FP-Tx production is simpler than the synthesis of fluorophore-labeled peptide blockers, especially because the recombinant production of FP-Txs usually provides a high yield of a fully active ligands with properly formed disulfide bonds without additional refolding of the product [46]. It should be noted that even small organic fluorophores can unpredictably modulate the affinities and pharmacological profiles of peptide blockers of K v 1 channels [43], not only FPs [46].

Conclusions
AgTx2-GFP was shown to have subnanomolar affinities to hybrid KcsA-K v 1.3 and KcsA-K v 1.6 channels and a nanomolar affinity to the KcsA-K v 1.1 channel (Table 1). Since these affinities were slightly sensitive to pH in the 7.0-8.0 range, a constant pH value should be maintained in studies to increase the reproducibility of the results. AgTx2-GFP could be used as a component of the analytical system based on hybrid Kcsa-K v 1.x (x = 1, 3, 6) channels expressed in the membranes of E.coli cells [52] for the study of peptide blockers and the evaluation of their affinities to K v 1 channels (Table 2). Data from electrophysiological measurements on oocytes indicated that AgTx2-GFP was a pore blocker of K v 1.x (x = 1, 3, 6) channels possessing nanomolar activity (Table 3). AgTx2-GFP could be equally used as a high-affinity fluorescent marker of the K v 1.6, K v 1.3, and K v .1.1 channels. Nanomolar activity of AgTx2-GFP was preserved for both open and closed K v 1.3 channels expressed in the membranes of mammalian cells ( Figure 5). Fluorescent imaging of K v 1.3 channels and studies of the activities of K v 1.3 peptide blockers are potential applications of AgTx2-GFP in mammalian cells.
Data on the properties of AgTx2-GFP and previously studied FP-Txs [7,[45][46][47] have shown that FP-Txs can be considered as a reliable alternative to peptide blockers labeled with organic fluorophores for the fluorescent imaging of K v 1 channels, as well as for a number of analytical applications. FPs can modulate the pharmacological profiles of peptide blockers of K v 1 channels, maintaining a high (nanomolar) affinity to some of the target channels or even enhancing considerably the ligand selectivity for a particular channel. A wide variety of peptide blockers differing in the repertoire of amino acid residues interacting with K v 1 channels, together with a wide variety of FPs having not only different colors but also distinct compositions and distributions of surface-charged residues, can provide multiple combinations of FP-Txs with valuable properties. The design and study of such FP-Txs is still at an early stage.

Reagents
AgTx2-GFP was produced as described previously [46]. Its concentration in an aqueous solution was determined using the molar extinction coefficient of 55,000 M −1 cm −1 at 489 nm. Recombinant peptides of KTx1 and AgTx2 were obtained as described earlier [68]. Concentrations of KTx1 and AgTx2 were measured in an aqueous solution containing 20% acetonitrile and 0.1% trifluoroacetic acid using molar extinction coefficients (at 214 nm) of 49,300 and 49,200 M −1 cm −1 , respectively.
The plasmid pmKate2-KCNA3-del was obtained as described earlier [42]. Crude venoms of M. eupeus and O. scrobiculosus scorpions were a kind gift from Dr. Alexander Vassilevski.
K d values of AgTx2-GFP complexes with KcsA-K v 1.
x (x = 1, 3, 6) were calculated from the measured I a dependencies on the concentration L of added AgTx2-GFP using the following equation: where I s is the I a value at the saturation of binding. Data on the displacement of AgTx2-GFP from complexes with KcsA-K v 1.
x (x = 1, 3, 6) by a nonlabeled peptide blocker were fitted with the following equation: where B is the concentration of the added nonlabeled peptide, DC 50 is the concentration of the nonlabeled peptide displacing 50% of AgTx2-GFP from the complex with the KcsA-K v 1.x channel, and I 0 is the I a at B = 0. The K ap value of the nonlabeled peptide blocker was calculated using the Cheng-Prusoff equation: where L is the constant concentration of AgTx2-GFP. The K d and K ap values were averaged over three independent experiments and presented as means ± SEM.
Expressions of rat K v 1.1, K v 1.2, and K v 1.6 and human K v 1.3 channels in X. Laevis oocytes were performed as described previously [71]. Transcription of linearized plasmids bearing K v 1 genes was performed using either T7 or SP6 mMESSAGE mMACHINE transcription kits (Invitrogen, CA, USA).
Two-electrode voltage clamp recording was performed at 18-22 • C using a Geneclamp 500 amplifier (Molecular Devices, CA, USA) and an Axon pClamp data acquisition system (Molecular Devices, CA, USA), as described earlier [72]. Whole-cell currents from oocytes were recorded 1-4 days after injection of 50 nl of cRNA (1 g/L). Current and voltage electrodes were filled with 3 M KCl, and their resistances were kept between 0.7 and 1.5 MΩ. The bath solution was ND96. The induced currents were filtered at 0.5 kHz and sampled at 2 kHz using a four-pole low-pass Bessel filter. Leak was subtracted using a-P/4 protocol. Currents through K v 1.x (x = 1-3, 6) channels were induced by 250 ms depolarization to 0 mV followed by a 250 ms pulse to −50 mV from a holding potential of −90 mV.
Concentration dependences of the inhibition of currents through K v 1.
x (x = 1-3, 6) channels by AgTx2-GFP and AgTx2 were measured and fitted with the Hill equation: where y is the relative decrease (%) in current amplitude at the concentration C of a pore blocker, IC 50 is the pore blocker concentration inducing a 50% decrease in the current amplitude, and h is the Hill coefficient. The Hill coefficients for all the measured dependences were found to be close to 1 (not shown). Concentrations of AgTx2-GFP and AgTx2 varied from 1 nM to 1 µM. Experimental data obtained in several independent experiments are presented as means ± SEM (n ≥ 3). Fluorescent images of cells were recorded with a laser scanning confocal microscope Leica-SP2 (Leica Microsystems GmbH, Wetzlar, Germany) equipped with a water immersion 63×/1.2 NA HCX PL APO objective (0.2 µm lateral and 0.6 µm axial resolutions). Fluorescence of AgTx2-GFP was excited at the 488 nm wavelength and recorded in the 495-535 nm range. Fluorescence of mKate2-K v 1.3 was excited at the 561 nm wavelength and recorded in the 650-700 nm range. Sequential scanning was used.

Experiments with
Fluorescent images were treated using ImageJ software (National Institutes of Health, Bethesda, MD, USA). Average fluorescence intensities of membrane-bound AgTx2-GFP and mKate2-K v 1.3 were determined for each examined cell and corrected for background signal, as described earlier [42], and their ratios R i were calculated. The R i values of 20-25 cells measured in the same conditions were averaged, producing the R av values (±SD). The R av values were plotted as a function of the concentration L of added AgTx2-GFP ( Figure 5a) and fitted with the following equation: where R m is R av at the saturation of AgTx2-GFP binding. The R av values were plotted as a function of the concentration B of added AgTx2 at a fixed concentration L of AgTx2-GFP ( Figure 5c) and fitted with the following equation: where DC 50 is the concentration of AgTx2 displacing 50% of AgTx2-GFP from the complex with the K v 1.3 channel, and R av0 is R av at B = 0. The K ap value of AgTx2 was calculated using the Cheng-Prusoff equation: K ap = DC 50 /(1 + L/K d ) All the measurements were repeated in two independent experiments, and data on the K d and K av were averaged and presented as means ± SEM.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/toxins15030229/s1, Figure S1: Confocal imaging of AgTx2-GFP interactions with mKate2-tagged K v 1.3 channels transiently expressed in HEK293 cells. Figure S2: Confocal imaging of an interaction of AgTx2-GFP with Neuro 2A cells that did not express mKate2-K v 1.3. Figure

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to local regulations.