Insights into R-Κ Toxin Specificity for K+ Channels Revealed through Mutations in Noxiustoxin†

Noxiustoxin (NxTX) displays an extraordinary ability to discriminate between large conductance, calcium-activated potassium (maxi-K) channels and voltage-gated potassium (Kv1.3) channels. To identify features that contribute to this specificity, we constructed several NxTX mutants and examined their effects on whole cell current through Kv1.3 channels and on current through single maxi-K channels. Recombinant NxTX and the site-specific mutants (P10S, S14W, A25R, A25∆) all inhibited Kv1.3 channels with Kd values of 6, 30, 0.6, 112, and 166 nM, respectively. In contrast, these same NxTX mutants had no effect on maxi-K channel activity with estimated Kd values exceeding 1 mM. To examine the role of the R-carbon backbone in binding specificity, we constructed four NxTX chimeras, which altered the backbone length and the R/â turn. For each of these chimeras, six amino acids comprising the R/â turn in iberiotoxin (IbTX) replaced the corresponding seven amino acids in NxTX (NxTX-YGSSAGA21-27-FGVDRG21-26). The chimeras differed in length of Nand C-terminal residues and in critical contact residues. In contrast to NxTX and its site-directed mutants, all of these chimeras inhibited single maxi-K channels. Under low ionic strength conditions, Kd values ranged from 0.4 to 6 μM, association rate constant values from 3 × 107 to 3 × 108 M-1 s-1, and time constants for block from 5 to 20 ms. The rapid blocked times suggest that key microscopic interactions at the toxin-maxi-K channel interface may be absent. Under physiologic external ionic strength conditions, these chimera inhibited Kv1.3 channels with Kd values from 30 to 10 000 nM. These results suggest that the extraordinary specificity of NxTX for Kv1.3 over maxi-K channels is controlled, in part, by the toxin R-carbon backbone. These differences in the R-carbon backbone are likely to reflect fundamental structural differences in the external vestibules of these two channels. The R-K channel toxin (R-KTx)1 peptides inhibit the flow of K+ ions through the K+ channel pore by simply binding to and occluding the extra cellular pore. This simple, bimolecular plugging mechanism is based on detailed kinetic studies of toxin block of single large-conductance, calciumactivated potassium (maxi-K) channels (1, 2) and of macroscopic current through a voltage-gated potassium (Kv) channel (3). The location of the R-KTx binding site to the external vestibule has revealed amino acids that form the potassium channel pore (4, 5), and it has yielded a lowresolution image of the maxi-K and Kv (6-11) channel vestibules. Thus, the R-KTx peptides provide invaluable molecular insight into protein structures that underlie gating and permeation in the K channel pore. The R-ΚΤx peptides can be grouped into different subfamilies (R-KTx 1.x, R-KTx 2.x, and R-KTx 3.x) on the basis of differences in their amino acid sequences (12, 13). These R-KTx subfamilies display an extraordinary ability to distinguish between the large family of voltage-gated potassium (Kv) channels and the large-conductance calciumactivated potassium (maxi-K) channel (12). The peptides from the R-KTx 1.x subfamily display high-affinity interactions for the maxi-K channel. However, only two of these toxins, iberiotoxin (IbTX or R-KTx 1.3) and limbatus toxin (LbTX or R-KTx 1.4), appear to be highly selective for the maxi-K channel (14, 15). Other toxins from this subfamily show high-affinity interactions with some Kv channels. For instance, charybdotoxin (ChTX or R-KTx 1.1) and LQ2 (RKTx 1.2) inhibit the Kv1.3 (16) and Shaker Kv (17) channels with high affinity, respectively. Indeed, with the exception of IbTX, many of the R-KTx peptides from different subfamilies interact with the Kv1.3 channel with high affinity. In stark contrast, the R-KTx 2.x and R-KTx 3.x subfamily of toxins that block Kv channels with high affinity do not block the maxi-K channel with high affinity (12). Thus, the Kv1.3 channel is rather promiscuous in its † This work was supported in part by NIH Grant GM52179. * Corresponding author. E-mail: [email protected]. Phone: 215-707-8170. Fax: 215-707-7536. ‡ Temple University School of Medicine. § Astra-Zeneca Pharmaceuticals. 1 Abbreviations: R-ΚΤx, potassium channel toxin; R-ΚΤx 1.1, charybdotoxin; R-ΚΤx 1.2, LQ2; R-ΚΤx 1.3, iberiotoxin; R-ΚΤx 2.1, noxiustoxin; R-ΚTx 3.2, agitoxin 2; AgTX2, agitoxin 2; ChTX, charybdotoxin; IbTX, iberiotoxin; NxTX, noxiustoxin; Kd, equilibrium dissociation constant; koff, dissociation rate constant; kon, second-order association rate constant; Kv channel, voltage-gated potassium channel; maxi-K channel, large-conductance calcium-activated potassium channel; Po, single-channel open probability; Tblock, time constant for toxin block. 10987 Biochemistry 2001, 40, 10987-10997 10.1021/bi010227m CCC: $20.00 © 2001 American Chemical Society Published on Web 08/14/2001 interactions with R-KTx peptides while the maxi-K channel exhibits extraordinary specificity among the R-KTx subfamilies. These differences in specificity likely point to fundamental differences in the structure of the maxi-K and Kv channel vestibules. The three-dimensional, solution NMR structures of the R-ΚΤx peptides provide insight into the determinants for specificity (12). All of these peptides show a classic motif with three antiparallel â-strands forming a â-sheet face on one side of the molecule and a helix on the other. This rigid structure is maintained by the presence of three disulfide bonds that are conserved among the R-KTx peptides (12). Thus, all of the R-KTx peptides share a similar 3D structure. However, among the R-KTx subfamilies, there are differences in the R-carbon backbone that may contribute to differences in binding specificity. The most striking differences occur between the R-KTx 2.x and R-KTx 1.x subfamilies; see Figure 1. The R-KTx 2.x subfamily contains two extra amino acids that extend the Nand C-termini, each by one residue relative to the R-KTx 1.x toxins (12). In addition, the R/â turns in R-KTx 1.x and R-KTx 2.x show structural differences (12). The structural differences between these two subfamilies may contribute to their specificity for maxi-K versus Kv channels. Insight into the determinants for this specificity also comes from knowledge of toxin residues that form an interaction surface with the maxi-K channel. Identification of toxin residues critical for a high-affinity interaction with the maxi-K channel comes from site-directed mutagenesis studies with ChTX (11). This work showed that eight residues, protruding from the ChTX â-sheet face, are critical for a high-affinity interaction with the maxi-K channel (Figure 1). Moreover, two of these residues, R25 and W14, are unique to the R-KTx 1.x subfamily (12). However, these maxi-K channel contact residues are not well conserved among the R-KTx 2.x and R-KTx 3.x subfamilies. Thus, the absence of this repertoire of critical contact residues among the R-KTx 2.x and R-KTx 3.x subfamilies may contribute to their discrimination against the maxi-K channel. Noxiustoxin (NxTX or R-ΚΤx 2.1) is an exquisite discriminator of maxi-K versus Kv1.3 channels. In this work, we show that the difference in NxTX binding free energy for the maxi-K and Kv channels exceeds 7 kcal/mol. To test whether this specificity derives from the absence of critical contact residues or from differences in the R-carbon backbone, we constructed various site-specific and chimeric NxTX mutants and examined their functional interaction with Kv1.3 and maxi-K channels. Remarkably, we found that altering the toxin R-carbon backbone had profound effects on toxin binding specificity for the maxi-K and Kv channels. MATERIALS AND METHODS Materials. Sarcolemmal membranes from bovine aortic smooth muscle were purified as described (18). The plasmid construct pG9-M was a generous gift from Dr. Maria Garcia (Merck Research Laboratories). Polystyrene cuvettes for bilayer experiments contained either a 100 or 150 μm aperture and were purchased from Warner Instruments, Inc. (Hamden, CT). 1-Palmitoyl-2oleoylphosphatidylethanolamine (POPE) and 1-palmitoyl2-oleoylphosphatidylcholine (POPC) were from Avanti Polar Lipids, Inc. (Birmingham, AL). Decane from Fisher Scientific, Inc. (Springfield, NJ), was 99.9% mole purity. All other reagents were of the highest purity commercially available. Construction of the NxTX Plasmid and Mutants. The plasmid pG9-NxTX, encoding six histidine residues between the T7 gene 9 fusion protein and the factor Xa cleavage site, and the wild-type NxTX sequence were constructed as described (19). NxTX mutations were generated using a twostep polymerase chain reaction point mutagenesis strategy. Plasmids containing the NxTX gene and site-directed mutants were propagated using the Escherichia coli strain DH5R. Τhe identity of all DNA constructs was verified using dideoxy sequencing (20). Expression and Purification of Recombinant Noxiustoxin Mutants. The E. coli strain BL21(DE3) harboring the pG9NxTX plasmids was cultured and induced with isopropyl 1-thio-â-galactopyranoside (IPTG), and the T7 gene 9 toxin fusion protein was purified by DEAE ion-exchange chromatography as described (19) with the following modifications. After folding, the fusion protein was cleaved from the toxin with TPCK-treated trypsin as described (21). This reaction was inhibited with L-1-chloro-3-(4-tosylamido)-7amino-2-heptanone hydrochloride (TLCK) and immediately loaded onto a 300 mL SP-Sephadex column equilibrated at pH 9.0 with 20 mM sodium borate and then eluted with a linear gradient of NaCl (0 to 0.75 M in 0.5 h) at a flow rate of 5 mL/min. Final purification of the peptides was achieved by reverse-phase HPLC as described (19). The purity of each peptide was >99% as judged by HPLC chromatography and mass spectrometric analysis. The identity of each peptide was confirmed by MALDI-MS mass spectrometric analysis (Protein Chemistry Laboratory, University of Pennsylvania). The predicted and measured masses for each peptide, respectively, were NxTX (4202, 4202), P10S (4192, 4197), S14W (4301, 4319), A25R (4287, 4288), A25∆ (4131, 4131), NxTX-IbTX I (4241, 4241), NxTX-IbTX II (4124, 4129), NxTX-IbTX III (4247, 4350), and NxTX-IbTX IV (4283, 4382). With the exception of NxTX-IbTX III and NxTX-IbTX IV, the measured masses were well within experimental error of the MALDI-MS instrumentation (Dr. William Moore, Protein Chemistry Laboratory, University of Pennsylvania, personal communication). The masses of these two mutants were further confirmed by amino acid analysis. Amino acid analyses performed on rNxTX, NxTXS14W, NxTX-IbTX I, NxTX-IbTX II, NxTX-IbTX III, and NxTX-IbTX IV (Commonwealth Biotechnologies, Inc.) were in excellent agreement with the predicted mass of each peptide. Correct folding and disulfide bond formation for NxTX-IbTX I and NxTX-IbTX II is inferred from their solution NMR structures (37) and from the fact that all NxTX mutants elute with similar (∼13%) organic solvent by HPLC. The extinction coefficient for wild type, 12.82 cm-1 (mg/ mL)-1 at 214 nm, was used for all mutants except NxTXS14W, 13.94 cm-1 (mg/mL)-1 at 214 nm. The NxTX extinction coefficient was based on the extinction coefficient for margatoxin (MgTX or R-KTx 2.2), a peptide that displays 82% sequence identity with NxTX (19) and a nearly identical 3D structure (22). The extinction coefficient for NxTX-S14W was calculated by determining the amino acid composition of 1 nmol of toxin. Protein content was then correlated with absorbance monitored at 214 nm. 10988 Biochemistry, Vol. 40, No. 37, 2001 Mullmann et al.