Characterization of a novel radiolabeled peptide selective for a subpopulation of voltage-gated potassium channels in mammalian brain.

BgK, a 37-amino acid voltage-gated potassium (Kv) 1 channel blocker isolated from the sea anemone Bunodosoma granulifera, can be modified at certain positions to alter its pharmacological profile (Alessandri-Haber, N., Lecoq, A., Gasparini, S., Grangier-Macmath, G., Jacquet, G., Harvey, A. L., de Medeiros, C., Rowan, E. G., Gola, M., Ménez, A., and Crest, M. (1999) J. Biol. Chem. 274, 35653-35661). In the present study, we report the design of two BgK analogs that have been radiolabeled with (125)INa. Whereas BgK(W5Y/Y26F) and its radiolabeled derivative, (125)I-BgK(W5Y/Y26F), bind to Kv1.1, Kv1.2, and Kv1.6 channels with potencies similar to those for the parent peptide, BgK, BgK(W5Y/F6A/Y26F) and its monoiodo-tyrosine derivative, (125)I-BgK(W5Y/F6A/Y26F), display a distinctive and unique pharmacological profile; they bind with high affinity to homomultimeric Kv1.1 and Kv1.6 channels, but not to Kv1.2 channels. Interaction of BgK(W5Y/F6A/Y26F) with potassium channels depends on the nature of a residue in the mouth of the channel, at a position that determines channel sensitivity to external tetraethylammonium. In native brain tissue, (125)I-BgK(W5Y/F6A/Y26F) binds to a population of Kv1 channels that appear to consist of at least two sensitive (Kv1.1 and/or Kv1.6) subunits, in adjacent position. Given its unique pharmacological properties, (125)I-BgK(W5Y/F6A/Y26F) represents a new tool for studying subpopulations of Kv1 channels in native tissues.

Voltage-gated potassium (Kv) 1 channels regulate numerous cellular processes by controlling plasma membrane potential and electrical excitability (1). The existence of a large number of pore-forming subunit genes contributes to the large diversity of potassium channels found in native tissues (2,3). Because potassium channels are tetrameric structures (4) made up by the association of four identical or closely related subunits (5)(6)(7)(8)(9)(10)(11), it is difficult to determine the subunit composition of a given channel in vivo based on the biophysical properties of the resultant proteins.
A number of peptides isolated from scorpion, snake, and sea anemone venoms have been characterized and shown to block Kv1 channels by binding to residues located in the external vestibule of the channel (12). Despite their limited selectivity for a given Kv1 subtype, these peptides represent unique pharmacological tools for studying the structure-function relationship of Kv1 channels and have proved to be important for: 1) defining the physiological role that channels play in native tissue, 2) purifying channels from native tissues and determining their subunit composition, and 3) developing the pharmacology of potassium channels (4,(12)(13)(14)(15)(16)(17)(18)(19)(20). Design of peptides with new pharmacological profiles should help to develop our understanding of both structure and function of Kv channels.

EXPERIMENTAL PROCEDURES
Materials-Plasmids were amplified in Escherichia coli XL1Blue or DH5␣ and purified using the plasmid purification kit from Qiagen (maxi protocol). cDNA encoding hKv1.2 in pGEMA was provided by Prof. Stephan Grissmer (Department of Applied Physiology, University of Ulm, Ulm, Germany). After introduction of two cloning sites by PCR (5Ј BamHI and 3Ј XbaI), the cDNA was cloned into the mammalian expression vector pcDNA3.1/HisC (Invitrogen). The resulting construct, whose integrity was assessed by nucleotide sequencing, encodes a channel with an N-terminal polyhistidine tag/epitope. To construct a cDNA coding for a Kv1.1-Kv1.2 dimer, human Kv1.1 and Kv1.2 DNAs were amplified from linearized plasmids with the primers CGATAGCTAG-CACGCCACCATGACGGTGATGTCTGGGGAAG/CGATGGATCCCTG-TTGCTGTTGCTGTTGCTGTTGCTGTTGAACATCGGTCAGTAGCTT-GCTCTTATTAAC (hKv1.1), and GCATGGATCCATGACAGTGGCCA-CCGGAGACC/GCATCTCGAGTCAGACATCAGTTAACATTTTGG (hKv1.2), respectively. The resulting PCR fragment for Kv1.1 had a unique restriction site 5Ј to ATG. The stop codon at the 3Ј end was replaced by a glutamine codon, followed by a stretch of nine additional glutamine codons. After the last glutamine codon, a unique BamHI site was introduced, which added two additional amino acids (Gly, Ser) to the linker between Kv1.1 and Kv1.2. The PCR fragment for Kv1.2 had a BamHI site immediately before the ATG codon and a unique restric-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Rat Brain Synaptosomal Membranes-Rat brain synaptosomal membranes were prepared as described in Ref. 25. Protein concentration was determined according to Lowry, using bovine serum albumin as a standard.
Binding Assays-All binding assays were carried out at room temperature in a medium consisting of 20 mM Tris-HCl, pH 7.4, 100 mM NaCl, 5 mM KCl, 0.1% bovine serum albumin. Incubations with 125 I-␣DTX and 125 I-BgK(W5Y/F6A/Y26F) were carried out for 2 h. At the end of the incubation period, samples (0.25-8 ml) were filtered through Whatman GF/C glass-fiber filters presoaked with 0.5% (w/v) polyethylenimine (Sigma). Filters were rinsed three time with 3 ml of ice-cold buffer (20 mM Tris-HCl, pH 7.4, 150 mM NaCl). Duplicate samples were run for each experimental point, and the data were averaged. Incubation of 125 I-HgTX 1 (A19Y/Y37F) with either Kv1.3 or Kv1.3(H399Y) channels was carried for 20 h, in a total volume of 6 ml. Nonspecific binding was determined in the presence of 2 nM margatoxin. At the end of the incubation period, samples were diluted with 4 ml of ice-cold 100 mM NaCl, 20 mM Tris-HCl, pH 7.4, filtered through Whatman GF/C glass-fiber filters presoaked with 0.5% polyethylenimine and rinsed twice with 4 ml of ice-cold buffer. Triplicate samples were run for each experimental point, and the data were averaged. Standard deviation of the mean was typically less than 5%. Radioactivity associated with filters was determined in a ␥ counter.
Data Analysis-Data from saturation experiments were analyzed according to Equation 1, where B eq is the amount of ligand bound at equilibrium, B max the maximum receptor concentration, K d the ligand dissociation constant, and L* the free ligand concentration.
Competition experiments were analyzed according to the Hill equation in Equation 2, where B eq is the amount of ligand bound at equilibrium, B max the amount of radioligand bound in the absence of inhibitor, I the inhibitor concentration, and n H the Hill coefficient.
K i values were calculated from IC 50 values using the Cheng and Prusoff relationship (29).
Data from Fig. 4C were analyzed using either the Hill equation for a single-site model or the following equation for a two-site model, where IC 50 1 ϭ K i ⅐(1 ϩ (L*/K d )), and K i ϭ 229 pM, L* ϭ 0.77 pM, and K d ϭ 4.5 pM.

Synthesis of BgK
Analogs-To develop new pharmacological tools for studying Kv1 channels, two analogs of BgK, a Kv1 channel blocker isolated from the sea anemone, B. granulifera (22,23), have been radiolabeled and used to characterize Kv1 channels. Direct iodination of BgK leads to a peptide that lacks biological activity, as inferred from its failure to bind to rat brain synaptosomal membranes. This finding is consistent with the observation that the only Tyr residue of BgK, Tyr 26 , is critical for conferring high affinity interaction of the peptide to either homomultimeric or native Kv1 channels (21,24,30). Substitution of Phe at Tyr 26 and Tyr at Trp 5 led to a peptide, (BgK(W5Y/Y26F)), that can be radiolabeled without loss of biological activity. Moreover, a substitution that has been shown previously to modify the pharmacological profile of BgK for homomultimeric Kv1 channels (21,30) was introduced to yield BgK(W5Y/F6A/Y26F). The two BgK analogs were radiolabeled with 125 INa and their monoiodo-derivatives used in radioligand binding studies using either heterologous expressed channels or channels present in native tissues.
Binding of BgK and Its Analogs to Kv1.3(H399Y)-It has recently been shown that substitution of Ala at position 6 alters the affinity of BgK for a Kv1.1 channel in which Tyr 379 , the residue that confers high sensitivity to inhibition by extracellular tetraethylammonium ion, is replaced by His, the corresponding residue present in Kv1.3 (30). These data suggest that the nature of the residue at that position could be a major determinant in the interaction of BgK(W5Y/F6A/Y26F) with Kv1 channels. To examine this in further detail, we determined the ability of BgK and its analogs to inhibit 125 I-HgTX 1 (A19Y/ Y37F) binding to a Kv1.3 mutant, Kv1.3(H399Y), in which His was replaced by the corresponding residue, Tyr, present in Kv1.1. Results of these experiments are presented in Table I and Fig. 2 (A and B). Modification of this single residue in Kv1.3 is sufficient for enhancing the affinity of both BgK and BgK(W5Y/Y26F) by 30 -40-fold (K i values of 23.5 Ϯ 1.9 and 44 Ϯ 4 pM, respectively). These values are similar to those of Kv1.1 channels (Table I)   are not independent; the effect of F6A substitution in BgK (W5Y/Y26F) depends on the nature of the residue at the position that determines sensitivity of the channel to external tetraethylammonium. These results are consistent with data that indicate that the affinity of BgK(W5Y/F6A/Y26F) for Kv1.6, which also possesses a Tyr residue at the equivalent position, is only 10-fold lower than that of BgK(W5Y/Y26F), but is much more reduced for Kv1.2, which possesses a valine residue at that position (Table I).
Binding of BgK Analogs to Heteromeric Channels-Given that BgK(W5Y/Y26F) and BgK(W5Y/F6A/Y26F) differ in their pharmacological profile toward homomultimeric Kv1 channel subtypes, we then evaluated the effects of these peptides on heteromeric channels present in rat brain synaptic membranes. 125 I-BgK(W5Y/Y26F) binds to a single class of sites with a K d of 4.5 Ϯ 1 pM and a B max of 2.7 Ϯ 0.3 pmol/mg of protein (n ϭ 4) (Fig. 3A). Using the same membrane preparation, and under the same experimental conditions, 125 I-␣DTX binds to a single class of sites (B max value of 2.6 Ϯ 0.7 pmol/mg of protein) with a K d of 3.5 Ϯ 1.7 pM (n ϭ 5), a value virtually identical to that published previously (25). Binding of both 125 I-BgK(W5Y/Y26F) and 125 I-␣DTX to brain membranes is sensitive to inhibition by ␣DTX, BgK, BgK(W5Y/Y26F), and BgK(W5Y/F6A/Y26F), and K i values for these peptides are similar regardless of the radioligand used (Table II). These data indicate that the concentration of binding sites labeled by both 125 I-BgK(W5Y/Y26F) and 125 I-␣DTX, as well as the affinities of these sites for ␣DTX, BgK, BgK(W5Y/Y26F), and BgK(W5Y/F6A/Y26F) are similar and suggest that both radioligands bind to the same receptor population in rat brain membranes. DTX-K inhibits binding of 125 I-BgK(W5Y/Y26F) (Fig.  4A) and 125 I-␣DTX (data not shown) with Hill coefficients lower than 1, indicating that this peptide distinguishes between several subpopulations of these receptors. These data are consistent with previous observations suggesting that 125 I-␣DTX receptors do not form a homogeneous population (14).
b Data were fitted using the Hill equation for one-site model. c Data were fitted using a two-site model in which the high affinity represents 20% of the total sites and has a K i of 229 pM (see text and ments whether or not BgK(W5Y/F6A/Y26F) distinguishes subpopulations of 125 I-BgK(W5Y/Y26F) receptors. These findings illustrate the importance of using 125 I-BgK(W5Y/F6A/Y26F), instead of its unlabeled analog, in binding studies. Although 125 I-BgK(W5Y/F6A/Y26F) binds to a subpopulation of brain 125 I-BgK(W5Y/Y26F) receptors, these sites do not appear to form an homogeneous population, as indicated by the fact that DTX-K inhibits binding of 125 I-BgK(W5Y/F6A/Y26F) with Hill coefficient lower than 1 (Fig. 4B).

DISCUSSION
This study concerns the development of new pharmacological tools for studying the composition of potassium channels expressed in native tissues. Two analogs, derived from BgK, a high affinity blocker of certain Kv1 channels (22)(23)(24), have been characterized in radioligand binding studies using homomultimeric Kv1 channels expressed in mammalian cells. Furthermore, the two BgK analogs were radiolabeled and used to study the molecular composition of native Kv1 channels present in brain tissue.
To obtain a radiolabeled derivative of BgK that retains biological activity, two substitutions, Trp 5 to Tyr and Tyr 26 to Phe, had to be introduced in the molecule to: 1) prevent conformational changes caused by the incorporation of iodine in the functionally important residue Tyr 26 (21,24,30), and 2) provide a iodination site at a position, Trp 5 , that is not critical for the interaction of the peptide with its target channels (21,24,30). Both BgK(W5Y/Y26F) and its radiolabeled derivative bind to Kv1 channels with similar affinities as BgK. It has been shown previously that the selectivity of BgK for Kv1 channels can be altered by substitution of certain residues in the peptide (21). Among them, Phe 6 to Ala substitution is of interest because it reduces the affinity of BgK for homomultimeric Kv1.2 and Kv1.3 channels, while having no effect on homomultimeric Kv1.1 channels (21). In this study, we show that this effect depends on the nature of a residue in the mouth of the channel, at a position that determines channel sensitivity to external tetraethylammonium. The substitution F6A does not affect the affinity of BgK for channels bearing a tyrosine at that position, such as Kv1.1, Kv1.6, or Kv1.3(H399Y), but it significantly decreases affinity for channels containing a His residue such as Kv1.3 or a valine residue such as Kv1.2. Consistent with this, BgK(W5Y/F6A/Y26F) and BgK(W5Y/Y26F) display similar affinities for Kv1.1 and Kv1.6 channels, but the affinity of BgK(W5Y/F6A/Y26F) for Kv1.2 and Kv1.3 channels is much lower than that of BgK(W5Y/Y26F).
Results obtained with channels expressing a Kv1.1-Kv1.2 tandem dimer deserve some comments. Immunoprecipitations and competition experiments indicate that the resulting channels are heteromers containing both types of subunits. However, as suggested previously (36), dimers can assemble to yield different tetrameric structures; some channels will have the identical subunits in opposing geometry, and other channels will have them adjacent. The affinities of 125 I-BgK(W5Y/Y26F) for these Kv1.1-Kv1.2 channels and for native brain channels are similar (Table II). 125 I-BgK(W5Y/F6A/Y26F) binds with high affinity to a subpopulation of 125 I-BgK(W5Y/Y26F) receptors formed by Kv1.1-Kv1.2 channels. This suggests that the geometry of the tetrameric channels resulting from the association of two Kv1.1-Kv1.2 subunits is important for conferring high affinity 125 I-BgK(W5Y/F6A/Y26F) binding. Recently, a model describing the interaction of BgK with homomultimeric Kv1 channels has been proposed (30), using distance constraints derived from double-mutant cycle analysis, and the structure of the bacterial potassium channel KcsA as a template (35). In this model, BgK lies on two adjacent subunits. Thus, BgK(W5Y/F6A/Y26F) could bind with low affinity to channels containing two sensitive Kv1.1 subunits in a diagonal orientation and with high affinity to those containing adjacent Kv1.1 subunits. Expression of constructs containing the four subunits in a tandem, so that the geometry of the tetramer is fixed, will be necessary to validate this hypothesis.
Modification of BgK has led to a new radioligand, 125 I-BgK(W5Y/F6A/Y26F), that recognizes a restricted population of Kv1 channels in mammalian brain. All results, taken together, suggest that native 125 I-BgK(W5Y/F6A/Y26F) receptors are formed by the association of at least two toxin-sensitive Kv1.1 and/or Kv1.6 subunits in adjacent position. Therefore, 125 I-BgK(W5Y/F6A/Y26F) represents a novel tool for studying structure-function relationships of native potassium channels.