Leucine 18, a hydrophobic residue essential for high affinity binding of anthopleurin B to the voltage-sensitive sodium channel.

Anthopleurin B is a potent anemone toxin that binds with nanomolar affinity to the cardiac and neuronal isoforms of the voltage-gated sodium channel. A cationic cluster that includes Arg-12, Arg-14 and Lys-49 has been shown previously to be important in this interaction. In this study, we have used site-directed mutagenesis to determine the contribution to activity of two aliphatic residues, Leu-18 and Ile-43, that have previously been experimentally inaccessible. Leu-18, a residue proximal to the cationic cluster, plays a critical role in defining the high affinity of the toxin. In ion flux studies, this is exemplified by the several hundredfold loss in affinity (231-672-fold) observed for both L18A and L18V toxins on either isoform of the sodium channel. When analyzed electrophysiologically, L18A, the most severely compromised mutant, also displays a substantial loss in affinity (34-fold and 328-fold) for the neuronal and cardiac isoforms. This difference in affinities may reflect an increased preference of the L18A mutant for the closed state of the neuronal channel. In contrast, Ile-43, a residue distal to the cationic cluster, plays at most a very modest role in affinity toward both isoforms of the sodium channel. Only conservative substitutions are tolerated at this position, implying that it may contribute to an important structural component. Our results indicate that Leu-18 is the most significant single contributor to the high affinity of Anthopleurin B identified to date. These results have extended the binding site beyond the cationic cluster to include Leu-18 and broadened our emphasis from the basic residues to include the crucial role of hydrophobic residues in toxin-receptor interactions.

Voltage-sensitive sodium channels play a crucial role in the transmission of electrical signals in excitable cells (1,2). Several laboratories have used toxins that alter channel function to advance our understanding of sodium channel biology (3,4). These toxins have been classified by Catterall (4) as binding to one of five different receptor sites on the channel with consequent effects on either activation, inactivation, or ion conductance. Of particular interest, primarily because of their potential as cardiotonic agents, are anemone toxins and ␣-scorpion toxins that delay sodium channel inactivation, resulting in enhanced sodium influx and ultimately giving rise to an increase in the force of contraction (5)(6)(7)(8). The classical drugs used to treat heart failure are the cardiac glycosides ouabain and digoxin, which are potent inhibitors of the Na ϩ /K ϩ -ATPase, and the ␤-adrenoreceptor agonists dopamine and dobutamine (9 -11). The life threatening toxicity associated with digoxin and the decreased effectiveness of the ␤-agonists as heart failure progresses underscore the importance of identifying novel agents displaying cardiotonic activity (11).
Sea anemones are a rich source of biologically active polypeptides with diverse pharmacological activities. While most anemone toxins were isolated as neurotoxins, Anthopleurin A and B (ApA and ApB) 1 obtained from the sea anemone Anthopleura xanthogrammica were originally isolated based on their displaying cardiac stimulatory activity (8,12). In isolated cardiac muscle, ApA is more potent than digoxin, being effective at nanomolar (nM) concentrations (13). More importantly, studies using anesthetized dogs demonstrate that this activity is not associated with adverse effects on heart rate or blood pressure (13,14). ApB is even more potent than ApA, displaying its maximal inotropic activity at 0.3 nM (8). While their antigenicity and lack of oral activity in animals preclude the use of these toxins as drugs in their naturally occurring forms, understanding the molecular interactions between these toxins and the sodium channel could form the basis for rational design of new drugs displaying enhanced cardiotonic activity. It is essential that the approach taken combine available structural information with functional studies in order to define regions of the molecule that contribute either to high affinity or selectivity for the cardiac channel.
ApA and ApB are naturally occurring homologs that differ in only 7 out of 49 residues and are cross-linked by three disulfide bonds (8,10). Ion flux studies have established that ApB exhibits nanomolar affinity for both the cardiac and neuronal isoforms of the sodium channel, while ApA binds much less tightly to the latter (15). Under voltage clamp conditions, both ApA and ApB bind preferentially to the cardiac channel, and ApB binding affinity is increased 100-fold in comparison with the values obtained by ion flux (16). In contrast, ApB binding to the neuronal channel differs by only about 4-fold in these two assays. These results indicate that ApB binds both to the open and closed conformations of the channel and greatly prefers binding to the closed channel over other states. This laboratory has previously cloned a synthetic gene for * These studies were aided by Grants HL-41543 (to K. M. B.) and HL-PO1-20592 (to D. A. H.) from the National Institutes of Health. 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.
ApB and produced the recombinant protein using a bacterial expression system (17,18). Using site-directed mutagenesis, we were able to specifically target residues and determine their contribution to activity by measuring sodium uptake in tissue culture cells. These studies have emphasized cationic residues that are either unique to ApB or conserved among anemone toxins (15,19). Studies on the unique cationic residues in ApB indicated that the polar side chain of Arg-12 is important in activity (15). In contrast, mutating the conserved positively charged residues Arg-14 and Lys-48 establishes that these residues play a smaller role in activity (19). A set of three mutant toxins containing pairwise replacements of the cationic side chains were not completely inactivated although their apparent affinities were significantly diminished raising the possibility of compensatory effects upon replacement of single cationic sites (16). Although our previous studies have identified residues which contribute 1-2 kcal/mol of binding energy, it is clear that additional sites of interaction remain unidentified (15).
The solution structures of both ApA and ApB have recently been solved by multidimensional NMR and reveal a fourstranded anti-parallel ␤-sheet structure common to both polypeptides (20,21). In a recently published model, the cationic side chains of Arg-12, Arg-14, and Lys-49 of ApB are located close together, and the solution structure, while not unambiguous, is generally consistent with this model (16,21). The three-dimensional structures of homologous scorpion toxins that interact with the sodium channel and affect either activation or inactivation are highly similar to each other although unrelated to those of anemone toxins (22)(23)(24). A notable conserved structural similarity among scorpion toxins is a surface-exposed hydrophobic region (22,23,25). Based on sequence analogies and chemical modification studies, Fontecilla-Camps et al. (26) proposed that this hydrophobic region is either directly involved in channel binding or helps align other residues that are important for both binding and specificity. Although ApB lacks an analogous surface hydrophobic face, a number of its hydrophobic side chains are at least partially exposed, including a subset of those found in proximity to the cationic cluster mentioned above (8,16,21). The present study was initiated in order to ascertain the extent to which these exposed hydrophobic residues in ApB might participate in the toxin-channel interaction. Sequence comparisons among homologous toxins and three-dimensional structural information lead us to target Ile-43 and Leu-18 to determine their role in toxin activity. Our results indicate that only highly conservative substitutions, resulting in modest changes in affinity, are tolerated for Ile-43, a residue which is distal to the proposed cationic cluster. In striking contrast, Leu-18, a proximal residue, contributes significantly to the high affinity of ApB for both isoforms of the sodium channel, thus further delineating the binding surface presented by these toxins. This study represents the first direct demonstration that hydrophobic residues play an essential role in sodium channel neurotoxin function.

Reagents and Enzymes
The highest grade chemicals commercially available were used in all experiments. Restriction endonucleases and modification enzymes were obtained from Life Technologies, Inc. The radiochemicals [ 32 P]dATP and 22 NaCl were purchased from DuPont NEN. The Sequenase kit from U. S. Biochemical Corp. was used to sequence all mutant constructs. Staphylococcal protease (V8 protease) was acquired from ICN, and veratridine, ouabain, and penicillin-streptomycin from Sigma. Tissue culture media was obtained from JRH Scientific and Fisher Scientific while the fetal calf serum was from Hyclone and United Biochemicals Inc.

DNA Methodology
Well established protocols for bacterial transformation, plasmid isolation, and cloning techniques were used (27). The expression vector pMG2, which encodes the synthetic gene for ApB fused to the 3Ј end of the gene for the bacteriophage gene-9 protein under control of the T7 promoter, was used to express high levels of mutant and wild-type proteins (18). The plasmid pMG7, designed to encode a polyglutamic acid sequence at the 5Ј end of the gene for ApB, was constructed using PCR technology. The sense strand primer contained the polyglutamate coding sequence, 5Ј-TCCTCTGAATTCGAGGAGGAAGAGGAGGGGG-3Ј, and was paired with an antisense primer recognizing the 3Ј end of the gene. The resulting PCR product was then cloned into pMG2 (18) to generate the plasmid pMG7. This vector was further engineered to introduce a polylinker at the downstream end by converting the EcoRI site of pMG7 to an SstI site by PCR amplification using the antisense primer 5Ј-CGTCTTCAAGAGCTCTTATTACTT-3Ј and a sense strand primer that recognized the 3Ј end of gene-9. The product was digested with EcoRI and SstI and cloned into pER9 (15) to generate the plasmid pKB13. Mutations were introduced by PCR amplification of the ApB gene encoded in pKB13 using primers BD-21 and KB-56 for the Leu-18 mutations and BD-13, BD-15, BD-16, BD-17, and KB-56 for the Ile-43 replacements (Table I). Each reaction included 30 repetitive cycles: an initial melting step (94°C, 1 min), an annealing step (2 min), and elongation (72°C, 2 min), with annealing temperatures based on the melting temperatures of the primers. PCR products encoding the Leu-18 or Ile-43 mutants were digested with appropriate restriction enzymes and gel-purified prior to ligation into pKB13. Ligation reactions were transformed into JM109 or XL1-Blue cells to obtain the desired clones pBK17 (L18A/L18V/L18F/L18I), pBK11 (I43G), and pBK12 (I43L/I43V/I43A/I43F). These bacterial strains were used routinely for plasmid propagation. All constructs were sequenced by dideoxy chain termination (28) using the methodology outlined by U. S. Biochemical Corp.

Expression and Purification of Mutant Proteins
The expression plasmid pKB13, which encodes a synthetic gene for ApB fused to the 3Ј end of the bacteriophage gene-9 protein, was used to express large quantities of fusion protein (18). The mutant plasmids pBK17 (L18A/L18V), pBK11 (I43G), and pBK12 (I43L/I43V/I43A/I43F) were transformed into the Escherichia coli expression strain BL21(DE3). Mutant fusion proteins were purified using anion exchange chromatography, reoxidized in the presence of glutathione (GSSH: GSSG, 1 mM:0.2 mM), and cleaved with staphylococcal protease to obtain intact toxin. Final purification to homogeneity was achieved by reverse-phase HPLC on a C4 column.

Cell Cultures
Cell lines expressing either the neuronal or cardiac isoform of the sodium channel were used for functional characterization of mutant toxins. The murine neuroblastoma cell line (N1E-115) was a generous gift from Dr. Marshall Nirenberg (NHLBI, National Institutes of Health). Cells were grown for 3 days in 24-well tissue culture dishes in 90% Dulbecco's modified Eagle's medium containing 10% heat-inactivated fetal calf serum and 110 units/ml penicillin-streptomycin in a humid atmosphere containing 10% CO 2 at 37°C. After the cells reached confluency, they were maintained in medium containing 1.5% fetal calf serum and 1.5% dimethyl sulfoxide for 48 h to enhance adherence to the plates prior to uptake assays. Dr. Laurie Donahue (Health Sciences Center, Texas Tech University) kindly provided us with a rat peripheral neurotumor cell line (RT4-B) which expresses predominantly the tetrodotoxin-resistant cardiac/denervated skeletal muscle isoform of the sodium channel (29,30). These cells were grown under similar conditions, but since they adhered readily to the plates, maintenance in dimethyl sulfoxide/low serum was not required and the cells were assayed when confluent.

Analytical Methods
Prior to amino acid analyses, samples of the mutant proteins were hydrolyzed with constant boiling HCl in vacuo for 22 h at 110°C. Hydrolysates were derivatized with phenylisothiocyanate and analyzed on a Picotag column. The secondary structures of the proteins were analyzed by circular dichroism spectropolarimetry using a Jasco J-710 spectropolarimeter. Samples were prepared in 5 mM sodium phosphate buffer (pH 6.8), and their far UV spectra were compared by a least squares fit to a composite structure based on the proteins of known secondary structure: myoglobin, lysozyme, ribonuclease A, papain, cytochrome c, hemoglobin, ␣-chymotrypsinogen, trypsin, and horse liver alcohol dehydrogenase.

Functional Characterization of ApB Mutants by Ion Flux
Assessment of toxin function is done by measuring veratridine-dependent sodium uptake in both the N1E-115 and RT4-B cell lines under conditions which have been previously described in detail (15,31). RT4-B cells were assayed when confluent (day 4), whereas the N1E-115 cells were differentiated when confluent and assayed 48 h later. Cells were preincubated in sodium-free binding solutions containing 20 M veratridine and different concentrations of the toxin. 22 Na uptake rates were measured by incubating the cells for 1 min in a binding solution of identical composition, except for containing 1 Ci/ml 22 Na ([NaCl] ϭ 10 mM). After uptake was terminated by washing in sodium-free solution, the cells were solubilized and total protein was estimated by the Bradford method. Absolute uptake is reported in terms of nmol/min/mg. The maximal rates of uptake for the mutants were normalized to those at saturating levels of ApB (500 nM). The ratio of the apparent affinity of the mutant to that of wild type toxin in the N1E-115 cells to that in the RT4-B line is described as the discrimination index and denotes the preference of the mutant for the cardiac channel. Data are fitted to a hyperbolic function as described by Cleland to obtain the kinetic constants V max and K 0.5 (32). The data shown were obtained from multiple experiments at each concentration and are corrected for uptake due to veratridine alone.

Functional Characterization by Electrophysiological Methods
Solutions-For both cell types, the control bath solution consisted of 70 mM NaCl, 70 mM CsCl, 1 mM CaCl 2 , 1 mM MgCl 2 , and 10 mM HEPES (pH ϭ 7.4). The pipette solution consisted of 130 mM CsF, 10 mM CsCl, and 10 mM HEPES (pH ϭ 7.4). Toxin concentrations tested were in the range of 0.1-0.5 times K D .
Recording Protocol-Recordings were made using an Axopatch 200 amplifier (Axon Instruments, Foster City, CA). Voltage protocols were executed on a 486DX2-50 computer running CLAMPEX 6.0.1 (Axon) as described previously (16) except that data were filtered at 5 kHz and pipette resistances ranged from 700 kiloohms to 1.1 megaohms. Suction was used to gain electrical access to the cell, which was then voltageclamped to Ϫ140 (RT4-B) or Ϫ120 mV (N1E-115). A current-voltage relationship was recorded, and a steady state inactivation protocol was performed. In every case, the holding potential used was sufficiently hyperpolarized to ensure complete availability of I Na . Voltage control was determined by examination of the time to peak of the whole-cell current and the slope of the conductance transform. In order to measure the time course of modification, cells were depolarized for 10 ms every 1-2 s (depending on the rate of modification). In fitting the data, the first two points of a 1-Hz record train and the first point of a 0.5 Hz record train were excluded. Separate current-voltage and steady state inactivation protocols were taken after the modification in order to monitor continued voltage control. In order to observe dissociation of the toxin, 10-ms depolarizations were begun again at a rate of 1 Hz or 0.5 Hz, and the cell moved back to the control bath.
Analysis-Modification rates were determined by taking an average of the current between 7.6 ms and 8 ms. For both channel isoforms, all the unmodified current had decayed before this window. Thus, by averaging the current 7.6 -8.0 ms after depolarization, the measured current was assured to be directly proportional to the number of modified channels in the cell. The averages were plotted against time and a single exponential curve fit to the data using a least squares minimization routine included with the Origin 3.5.2 software (Microcal Soft-ware, Northampton, MA). Unmodification time constants were determined similarly. Because unmodification (toxin dissociation) is a first order process, k off was simply the inverse of the unmodification time constant. The on-rate (k on ) was determined through the relationship: where k mod was the inverse of the fitted of modification time course and [T] the toxin concentration.

Identification and Characterization of the ApB Mutants-
Using PCR methodology we have generated a set of mutants at Leu-18 and Ile-43. Primers BD-21 and KB-56 were used to obtain the Leu-18 mutants, and cassette mutagenesis was used to replace the wild-type ApB sequence in ApaI-HindIII cleaved pKB13 with the PCR product containing the Leu-18 mutations, generating plasmids pBK17. Several replacements, including L18A/L18V/L18F/L18I, were identified by sequencing. In a similar manner we created the plasmids pBK11 (I43G) and pBK12 (I43L/I43A/I43F/I43V) that encoded replacements for Ile-43. Dideoxy sequencing confirmed the presence of the desired substitutions in the resulting plasmids pBK17, pBK11, and pBK12 and the absence of spontaneous mutations. The plasmids pBK17 (L18A/L18V), pBK11 (I43G), and pBK12 (I43L/I43V/I43A/I43F) were used in subsequent investigations to ascertain the importance of these sites in anemone toxin activity.
Expression and Isolation of Mutant Toxins-The L18A/L18V and the I43L/I43V/I43G/I43A/I43F mutants were all cloned into the expression plasmid pKB13 in which recombinant ApB protein is expressed as a fusion protein with the bacteriophage T7 gene-9 protein (15,16). Typical yields for wild-type ApB range from 3-4 mg per 500 mg of fusion protein. Consistently higher yields of 5-6 mg per 500 mg of fusion protein were obtained for the L18A mutation, while the yields of L18V were lower and varied between different protein preparations. We attribute this decreased yield to the lower specific activity of the batch of staphylococcal protease used for L18V. For the conservative substitutions I43L and I43V, we were able to isolate normal amounts of intact toxin. However, further truncation of the side chain to I43A or I43G resulted in the absence of an identifiable ApB-like HPLC peak after staphylococcal protease digestion, despite isolation of normal quantities of fusion protein. Several modifications of the reoxidation protocol including using dithiothreitol as a reducing agent early in the isolation procedure, diluting the protein to 0.1 mg/ml prior to oxidation or maintaining the sample under nitrogen prior to oxidation failed to yield the I43A toxin. While a small split peak migrating at the same position as wild-type ApB on the HPLC

I43G
NcoI BD-15 Sense  (Table II), although low levels of glutamic acid (Յ1 residue/mol) were observed for some samples (see "Discussion"). Secondary structures and thermal stabilities of the mutant proteins were assessed by circular dichroism. The far UV (190 -250 nm) data collected for the mutants demonstrate that their secondary structures are comparable to wild-type ApB, with ␤-sheet contents ranging between 52 and 60% (data not shown). Thermal denaturation studies were carried out to determine whether the mutants retained structural stabilities similar to that of ApB. As with wild-type ApB, all mutant toxins retained their ␤-sheet structure in the presence of 1.5 M guanidine chloride at temperatures up to 80°C (data not shown).
Functional Characterization of the ApB Mutants by Ion Flux-The ability of the ApB mutants to stimulate 22 Na uptake in N1E-115 and RT4-B cells was measured using an established ion flux assay (15,16,31) in cell lines expressing either the tetrodotoxin-sensitive neuronal (N1E-115) or the tetrodotoxin-resistant cardiac (RT4-B) isoform of the sodium channel. ApB alone does not induce sodium uptake in these cells, and, therefore, subsaturating quantities of veratridine are added to induce the conducting state of the channel. Addition of ApB, which increases the open probability of the channel by causing delayed inactivation, thus results in a dose-dependent increase in sodium uptake. Representative dose-response curves and the derived kinetic constants are shown in Figs. 1 and 2. The data obtained for all the mutants in both cell lines are contrasted in the bar graph depicted in Fig. 3.
The apparent binding affinities of wild-type ApB for the neuronal and cardiac isoforms of the sodium channel, as estimated by ion flux measurements, are 22 nM and 9 nM, respectively (15,18,19). Thus, the apparent neuronal K 0.5 for the L18A mutant (7.9 M) represents at least a 359-fold reduction in affinity. Similarly, a 231-fold reduction is observed for L18V, displaying a neuronal K 0.5 of 5.1 M. The same trend is evident in RT4-B cells, with 672-and 320-fold decreases in apparent affinity for the L18A and L18V toxins, respectively. Comparison of affinities in the two cell types yields a discrimination index for L18A of 0.5, indicating that its ability to preferentially bind to the cardiac channel is compromised. L18V also has a reduced discrimination index of 0.7. The highest velocity observed for the L18A/L18V mutants is between 40 and 70% that of wild-type ApB in both cell lines and appears not to represent a V max . These results are consistent with the essential nature of the contact made between Leu-18 and both the cardiac and neuronal isoforms of the sodium channel.
Figs. 1 and 2 also show the considerably smaller changes in apparent binding affinity observed for the Ile-43 mutants. K 0.5
a Not corrected for partial degradation of serine and threonine during acid hydrolysis; numbers in parentheses represent summation of the known sequence of ApB.
b Determined as cysteic acid after performic acid oxidation. c ND, not determined. d Determined spectrophotometrically.
values for I43L were only 2-fold different from that of wild type ApB in both cell types assayed. A comparable reduction in apparent affinity of 2.8-fold was obtained for the I43V substitution on the cardiac channel. Surprisingly, while the maximal uptake values obtained for I43V in both cell types and for I43L in the neuronal line are comparable to ApB, the I43L mutant displays a reduced relative V max of 0.6 in the RT4-B line (Fig. 2). Functional Characterization of the L18A Mutant by Electrophysiology-Electrophysiologic determinations of toxin K D were qualitatively similar to those determined by ion flux, although the affinity of ApB for the cardiac channel is substan-tially higher in this system, consistent with our previous observations (16). ApB L18A thus exhibits a 328-fold decreased affinity for the cardiac channel and a 34-fold decreased affinity for the neuronal isoform when compared with the wild-type toxin. The major factor contributing to the reduced affinity displayed by the mutant toxin for both isoforms is an increase in off-rates. This increase is greater for the cardiac isoform, bringing the difference in K D between channel isoforms to 6-fold, as compared to wild type toxin, which prefers the cardiac isoform by 57-fold (Table III). Thus, in both assay systems, L18A displays a lesser preference for the cardiac channel than does wild-type toxin. DISCUSSION This study targets hydrophobic residues in the toxin ApB by site-directed mutagenesis to establish their role in activity. Previous experiments demonstrated the importance of a cationic cluster including Arg-12, Arg-14, and Lys-49 to toxin activity (15,16,19). Using a model structure of ApB described by Khera et al. (16), we have identified residues based on proximity to this basic region and assessed their contribution to biological activity. In this model, the CD1 carbon of Leu-18 is within 10 Å of the NH1 and NH 2 of Arg-14 while Ile-43 is clearly on the opposite face of the molecule. Selecting Ile-43, distal to the cluster, and Leu-18, a residue proximal to the cationic region, has enabled us to to determine the nature of the potential intermolecular contacts made by two distinct surfaces of the toxin.
Using a PCR-based approach with a wobble containing primer, we generated two panels of mutants, of which a subset was expressed and characterized. Only highly conservative substitutions to leucine or valine were tolerated at position 43. We interpret these results as supporting the hypothesis that Ile-43 is involved in ApB folding, consistent with our model suggesting that it packs tightly against the Phe-24 and Tyr-25 side chains (16). We suggest that the I43A and I43G mutants disrupt the resulting hydrophobic region, preventing the protein from folding to a form allowing correct pairing of the three disulfide bonds. Since the ApB coordinates are not accessible in the Brookhaven data base, we are presently unable to verify this prediction.
Structural characterization of the mutants includes amino acid analysis and circular dichroism studies. While the amino acid compositions are overall in good agreement with that of wild-type ApB, the glutamic acid contents unexpectedly ranged from 0.6 -1.1 residues per mol. Previously, we attributed this to a system artifact (16). Recently, however, characterization by mass spectrometry and N-terminal sequencing of ApB expressed from pKB13 revealed the presence of two forms, one having a residue of glutamate at the amino terminus. 2 This extra glutamate is found only in ApB expressed from pKB13, in which the toxin sequence is preceded by five consecutive glutamates. Glu-ApB, like an N-terminally extended form we characterized earlier (18), is functionally identical with the wild-type toxin. Furthermore, all the mutants retained as their predominant secondary structural motif the ␤-sheet as assessed by circular dichroism. Thermal denaturation profiles for all mutants were essentially unchanged, confirming the structural stability of the toxins.
We have made two replacements to evaluate the contribution of the hydrophobic residue Leu-18. Substitution with alanine represents a side chain truncation in which all interactions made by atoms beyond the ␤-carbon are removed. Replacement with valine should restore some of the hydrophobicity. Characterization of both mutants by ion flux demonstrates a pivotal 2 R. A. Byrd, personal communication.

FIG. 3. Comparison of toxin-channel interactions by ion flux and voltage clamp.
Affinities were estimated in N1E-115 (N) and RT4-B (C) cell lines as described under "Experimental Procedures," and the results (Fig. 1, Fig. 2, and Table III) were compared on a log scale Ϯ S.E. of the estimate. role for Leu-18 in toxin activity as exemplified by the several hundredfold (231-672-fold) loss in activity observed in both model systems analyzed. By comparison, double neutralization mutants in the cationic cluster reduce apparent affinities by a maximum of 72-fold (16).
The estimated maximal level of uptake obtained for the Leu-18 mutants is only 40 -70% of that seen at saturating levels of ApB. Either we have not saturated ApB binding sites even at concentrations of 10 -25 M mutant toxin, or the ability of the bound mutants to stabilize the open conformation is compromised. Based on the results depicted in Figs. 1 and 2, we favor the former explanation, which suggests that the K 0.5 values are even higher than those presented here. Because the range of concentrations we are able to assay is restricted by the amount of protein we can produce, we are unable to confirm this prediction. Nonetheless, the key role of Leu-18 is amply supported by the data presented.
In order to more closely assess the role of Leu-18 in activity, we assayed the most severely impaired mutant, L18A, by whole cell patch clamp. There are important differences in the two assay systems used. Under ion flux conditions, the channel is maintained in the open conformation due to the presence of subsaturating concentrations of the alkaloid veratridine. In contrast, in the electrophysiological assays, the cells are clamped at a potential sufficient to ensure the maintenance of channels in the closed conformation during toxin binding. Substantial losses (34-and 328-fold) in binding affinity are seen for the L18A mutant in both cell types although the loss of affinity in N1E-115 cells is less dramatic (Table III). This disparity between the effects of L18A on electrophysiological versus ion flux parameters in N1E-115 cells led us to a more general consideration of the role of electrostatic interactions in toxin affinity. Our previous characterization of an R12S/K49Q charge neutralization mutant pointed to electrostatics as a more important determinant of binding to neuronal than cardiac channels (16). In order to test this hypothesis, we first increased the extracellular calcium concentration from 0.5 to 2 mM. Under these conditions, the K D for ApB is unaffected in RT4B cells (n ϭ 2) but increases 9-fold from 5 to 45 nM (n ϭ 5) in the N1E-115 line. Reducing the ionic strength by one-half (with sucrose replacement) has no effect on the K D in RT4B cells (n ϭ 4) but increases the affinity for ApB in N1E-115 cells to 27 nM (n ϭ 3). These results are consistent with the hypothesis that electrostatics are more important for binding of ApB to the neuronal than to the cardiac isoform of the sodium channel and underscore the contribution of hydrophobic interactions in the latter system. It is important to emphasize that the reductions in apparent affinity estimated for L18A by either method represent the greatest changes detected to date for either single or double site substitutions. This study has therefore extended the limits of the binding site to include Leu-18 as a residue making a vital contribution to apparent binding affinity. Changes observed for the Ile-43 mutants are more modest, ranging from 2-8-fold, and are consistent with the nonessentiality of Ile-43.
Hydrophobic forces are generally considered to be important in protein-protein interactions (33). Computational studies by Miyamoto and Kollman (34,35) emphasize the importance of hydrophobic interactions between ligand and protein. In an elegant study on the human growth hormone (hGH):receptor (hGHR) interface, Clackson and Wells (36) used alanine scanning mutagenesis to identify functionally important residues on the hGHR. While the greatest losses in binding affinity (4.5 kcal/mol) were observed when tryptophan residues were replaced, large effects (1.5-3.5 kcal/mol) were also detected when the hydrophobic groups Ile and Pro were substituted. In contrast, generally smaller effects (1-2 kcal/mol) were seen with charged residues (36). They concluded that a central hydrophobic region, surrounded by less important contact sites that are frequently hydrophilic in nature, forms a functional epitope accounting for three-quarters of the binding free energy (36). We propose that in our system the hydrophobic residue Leu-18 is of primary importance in binding, while the cationic residues identified previously may be analogous to the less important contacts (16).
Identification of Leu-18, as a critical determinant of high affinity channel binding, represents a potential starting point in drug design. We have successfully extended the currently available functional map beyond the cationic cluster to include Leu-18. We have also learned that while Ile-43 has minor effects on binding it is involved in folding and is probably an important structural determinant. Beyond the scope of this study, but an interesting direction for the future, will be the identification of complementary residues on the sodium channel with which defined sites of ApB interact. Off-rates were measured by least-squares fitting of a single exponential curve to the mean current between 7.6 ms and 8 ms after the depolarization. On-rates were determined through the equation k mod ϭ k on [T] ϩ k off . k mod was determined through the same method as k off . Numbers are shown Ϯ S.E. Dissociation constants were calculated as described under "Experimental Procedures." Wild-type ApB values were taken from Ref. 16 (n ϭ 6) (n ϭ 6) (n ϭ 4)