INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

VOLTAGE-GATED K⁺ channels represent a structurally and functionally diverse group of membrane proteins. These channels establish the resting membrane potential and modulate the frequency and duration of action potentials in nerve and muscle (18). In addition, K⁺ channels are involved in processes not usually associated with electrically excitable membranes, such as T lymphocyte activation, cell volume regulation, and pancreatic -cell function (5, 22). Multiple Shaker-like K⁺ channel -subunit genes have been cloned from mammalian brain, heart, skeletal muscle, pancreas, and smooth muscle and functionally expressed in heterologous systems (6). The recent discovery of -subunits, some of which convert delayed rectifiers into rapidly inactivating channels, has revealed additional mechanisms of voltage-gated K⁺ channel modulation (9, 12-14, 17, 26, 28, 32, 33). Functional analysis has shown that four of these -subunits, Kv1.1, Kv1.2, Kv1.3, and Kv3.1, confer varying degrees of rapid inactivation on certain members of the Kv1 family of delayed rectifiers (12, 13, 16, 17, 26, 28, 32, 33). In addition, Kv1.2, Kv1.3, and Kv2.1 modify the voltage dependence of Kv1.5 channel opening (8-20 mV hyperpolarizing shift in the midpoint of activation) (12, 13, 40). However, it has been suggested that the apparent shift in the Kv1.2 activation curve following coexpression with Kv1.2 is completely derived from the -subunit-induced N-type or fast inactivation (10). In this respect, it is important to note that Kv2.1, which does not induce fast inactivation, does shift the activation midpoint and only enhances slow or C-type inactivation of the Kv1.5 delayed rectifier (40). Each Kv family (Kv1, Kv2, and Kv3) appears to derive from a separate gene, and additional variability in the Kv1 family results from alternative splicing in the amino-terminal region, thus yielding the Kv1.1, Kv1.2, and Kv1.3 subunits (12, 27). The variable amino-terminal domains are responsible for the functional differences (16, 29, 32), whereas the conserved carboxy-terminal domain most likely governs assembly with the -subunit (16, 30, 34, 43, 45).

The rapid inactivation induced by Kv1.1 when coexpressed with the delayed rectifier, Kv1.1, is thought to occur by the same "ball-and-chain" mechanism reported for the Shaker channel (20, 32). For example, exposure of the cytoplasmic face of Kv1.1 to a peptide corresponding to the 24 amino-terminal amino acids of Kv1.1 causes inactivation (32). One interpretation of this finding is that the peptide induces inactivation by binding within the open pore. Alternatively, the peptide could bind to a site removed from the pore and allosterically modulate ion permeation. Indeed, Morales and co-workers (29) have presented evidence that Kv1.2 confers inactivation on a delayed rectifier by enhancing primarily C-type inactivation (29), suggesting that Kv1.2-induced inactivation occurs by an allosteric mechanism. Thus no a priori assumptions should be made about the inactivation mechanism of other Kv subunits.

The Kv1.3 subunit converts Kv1.5 from a delayed rectifier with a modest degree of slow inactivation to a channel with both fast and slow components of inactivation. The rapid inactivation has the characteristics of open channel block and is similar to block of Kv1.5 by antiarrhythmic agents such as quinidine. Mutagenesis of - and -subunits followed by kinetic analysis of channel gating was used in the present study to examine the mechanism by which Kv1.3 modulates Kv1.5 function. The data indicate that Kv1.3-mediated fast inactivation likely occurs in part by an open channel block mechanism that is sensitive to both membrane potential and external K⁺. Both the -subunit-enhanced slow inactivation and the modulation of -subunit-induced fast inactivation by an external, but not an internal, pore mutation reveal the importance of allosteric mechanisms in K⁺ channel - interactions. Results presented here clearly indicate that the Kv1.3 subunit directly alters -subunit activation, for Kv1.3 mutants that produce no inactivation still induce a hyperpolarizing shift in the voltage dependence of Kv1.5 activation.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. Enzymes and buffers were from New England Biolabs (Beverly, MA), Boehringer Mannheim (Indianapolis, IN), and Promega (Madison, WI). The Sequenase 2.0 kit was purchased from United States Biochemical (Cleveland, OH). The origins of other materials are specified below.

Mutagenesis. The series of amino-terminal deletions in Kv1.3 shown in Fig. 1 was created by PCR mutagenesis. The 5' oligonucleotide primer replaced amino acids 10, 37, 68 and 91, respectively, with an Xba I restriction site and a consensus translation initiation site (GGTCTAGAATG. . .). The 3' oligonucleotide annealed downstream from a unique endogenous Kpn I (N10 and N37) or Pst I (N68 and N91) restriction site. The 5' sequence of the previously described Kv1.3 construct in the modified pSP64T vector (12) was excised with the appropriate restriction enzymes, and the amplified mutant PCR products were ligated in its place. Two additional mutants were made in which the first 19 or 87 amino acids of Kv1.3 were linked to the amino terminus of Kv1.5. The 5' region of Kv1.3 was PCR amplified to contain an Sph I site at the 3' end in the same reading frame as an endogenous Sph I site in the 5' untranslated region of Kv1.5. Use of this site introduced seven amino acids (His-Ala-Leu-Cys-Ser-Arg-Ala) that are not part of either wild-type subunit. The tandem constructs were then ligated into a modified pSP64T vector for oocyte expression. These constructs will be referred to as tandems even though only portions of the amino terminus of Kv1.3 are appended to the full-length Kv1.5 (Fig. 1). All PCR-generated mutants were verified by double-stranded sequencing. The point mutations in the Kv1.5 -subunit have been described previously (44). These mutants were subcloned into the pSP64T vector for oocyte expression. It was necessary to add ~710 base pairs of the 3' untranslated sequence of Kv1.5 to constructs containing the -subunit to achieve expression levels in excess of 1.5 µA.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Kv1.3 mutations. PCR-generated deletion and tandem mutant constructs are shown. Solid bar, carboxy-terminal 329 amino acids of Kv1.3; shaded bar, variable amino-terminal domain of 91 amino acids. Nomenclature for deletions uses  to indicate a deletion and N to indicate that deletion is amino-terminal, and number indicates amino acid that was replaced with the initiating methionine. Tandem constructs link coding sequence of Kv1.3 amino terminus with coding sequence of complete Kv1.5. Nomenclature for tandems uses N to indicate that amino terminus of Kv1.3 was added to amino terminus of Kv1.5, and number indicates number of Kv1.3 amino acids added, as indicated by shaded bar. Circle, 7 amino acids (His-Ala-Leu-Cys-Ser-Arg-Ala) inserted between - and -subunit domains; shaded boxes, transmembrane domains within -subunit.

Expression in Xenopus oocytes and electrophysiological recording. Templates were linearized with EcoR I before in vitro cRNA synthesis using the SP6 mMessage mMachine kit (Ambion) according to the manufacturer's instructions. Defolliculated Xenopus oocytes were prepared as described previously (12, 13) and injected with ~40 nl (4-20 ng) of in vitro-transcribed cRNA. The cRNA was diluted to yield peak currents of 1-10 µA.

Pulse protocols and data analysis. Activation curves were obtained from the tail current amplitude 6 ms after repolarization to 40 mV. The voltage dependence of channel opening ("activation curve") was fit with the Boltzmann equation I/I_max = {1 + exp[(E_m  E_h)/k]}¹, in which I is current, I_max is maximum current, E_m is membrane potential, E_h is the voltage at which 50% of the channels are open, and k is the slope factor. The impact of inactivation on the generation of activation curves is reviewed in DISCUSSION. The time course of deactivating tail currents was fit with a single exponential by a nonlinear least squares algorithm. Goodness of the fit was judged by visual inspection for nonrandom trends in the residuals of the fit. For steady-state current values, raw data points were averaged over a small time window (2-5 ms).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

The present study examined the mechanism by which Kv1.3 alters Kv1.5 function and determined regions within each subunit that affect their functional interactions. In the presence of the Kv1.3 subunit, the Kv1.5 current phenotype was changed from a delayed rectifier to one that displayed a rapid, but partial, inactivation (time constant = 5.76 ± 0.14 ms at +50 mV, n = 17; Table 1) that was apparent only at membrane potentials more positive than ~0 mV (see Fig. 2). A slower rate of channel deactivation was also observed in the presence of Kv1.3 relative to Kv1.5 alone (32.6 ± 1.2 ms vs. 13.7 ± 0.5 ms at 40 mV; Table 1). Additionally, the threshold for Kv1.5 channel activation occurred at more negative membrane potentials, with a parallel 13-mV hyperpolarizing shift in the voltage dependence of activation (Table 1).

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Fast inactivation localizes to amino terminus of Kv1.3. Whole cell current was recorded from Xenopus oocytes injected with Kv1.5 (A), Kv1.5 + Kv1.3 (B), Kv1.5 + Kv1.3(N10) (C), and tandem(N87) (D). All currents were elicited by 200-ms depolarizations from a holding potential of 80 mV to test potentials between 60 and +70 mV (in 10-mV increments). Test potential was followed by a repolarizing step to 40 mV to record deactivating currents.

Role of the Kv1.3 amino-terminal sequence in fast inactivation. The various Kv1 subfamily members confer different functional effects on Kv1.5 (12, 13, 16, 26, 41). The amino termini of these subunits vary considerably, whereas the 329 carboxy-terminal amino acids are identical. Therefore, we constructed several mutations within the Kv1.3 amino terminus and examined the effect of these alterations on Kv1.3-induced inactivation. Mutant Kv1.3 subunits lacking the first 10, 37, 68 and 91 amino acids of the variable amino terminus were constructed. Two tandem mutants were also made in which the first 19 or 87 amino acids of the amino terminus of Kv1.3 were linked to the amino terminus of Kv1.5 (Fig. 1). Representative outward currents recorded from oocytes expressing Kv1.5, Kv1.5 + Kv1.3, Kv1.5 + Kv1.3(N10), and tandem(N87) are shown in Fig. 2. Kv1.5 displayed a delayed rectifier phenotype when expressed alone (Fig. 2A). Coexpression of the Kv1.5 subunit with wild-type Kv1.3 resulted in currents that exhibit fast inactivation at membrane potentials positive to +10 mV but failed to show any apparent inactivation at more negative potentials (Fig. 2B). At +70 mV, the extent of inactivation at 200 ms, defined as the decline from the peak current level, averaged 54.4 ± 1.1% (n = 16; Table 1). The time constant of Kv1.3-mediated inactivation at +70 mV was 5.7 ± 0.1 (n = 17; Table 1) and that for recovery from fast inactivation at 80 mV was 5.1 ± 0.1 ms (n = 7). Removal of as few as 10 amino acids from the amino terminus of Kv1.3 prevented fast inactivation (Fig. 2C). The other more extensive Kv1.3 amino-terminal deletions also did not induce fast inactivation (data not shown). In contrast to the currents observed with Kv1.2 and amino-terminal-truncated Kv1.2 (1), no significant increase in outward current was observed following the removal of Kv1.3-induced inactivation. Therefore, the ratio of the peak to steady-state current shown in Fig. 2B is likely to represent a reasonable approximation of the percentage of Kv1.5 channels inactivated by Kv1.3. Increases in the ratio of - to -subunit cRNA failed to further enhance inactivation and often suppressed current. This suppression occurred with both the wild-type and mutant -subunits and is in agreement with the data of Accili et. al. (1) showing that overexpression of the conserved Kv1 core can decrease Kv1.5 current.

The hyperpolarizing shift in the midpoint of activation does not require an intact amino terminus. Figure 3 shows activation curves constructed from normalized deactivating tail current magnitudes as described in MATERIALS AND METHODS. Coexpression of Kv1.5 with the wild-type Kv1.3 subunit produced the greatest leftward shift compared with Kv1.5 expressed alone. Coexpression of Kv1.5 with Kv1.3(N10) also shifted the curve leftward but to a slightly lesser extent. These data demonstrate the functional independence of the Kv1.3-mediated effects on inactivation and the change in the midpoint of activation. As shown in Fig. 3, the 68-amino acid deletion still caused a shift in the activation curve. The Kv1.3(N91) construct did not induce a shift in the voltage dependence of activation (data not shown). Neither tandem construct (N19 or N87) shifted the midpoint of activation relative to wild-type Kv1.5 (Table 1). All of these data suggest that the -subunit domain involved in inactivation is distinct from that governing the shift in the voltage dependence of activation.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Hyperpolarizing shift in midpoint of activation does not require an intact amino terminus. Activation curves obtained from oocytes expressing Kv1.5 in absence or presence of Kv1.3, Kv1.3(N10), or Kv1.3(N68) are compared with that obtained from expression of tandem(N87). Curves were created from Boltzmann equations fit to normalized deactivating tail current magnitudes as described in MATERIALS AND METHODS.

Prolonged depolarization revealed a slow component of inactivation. Fast inactivation occurred with a time constant of 5.7 ms, suggesting that this process is essentially at steady state within 25 ms (Table 1 and Fig. 4B, inset). However, the current continued to decay at a slower rate after this time, indicating a slow component of inactivation. To examine the slow component of inactivation, oocytes expressing Kv1.5 or Kv1.5 + Kv1.3 were depolarized for 5 s. Unscaled currents resulting from 5-s depolarizations to +70, +50, and +30 mV are shown in Fig. 4, A and B, for Kv1.5 and Kv1.5 + Kv1.3, respectively. The magnitude of current measured at 1 s increases with membrane potential in the absence of Kv1.3 (Fig. 4A). However, in the presence of Kv1.3, the K⁺ current measured at 1 s decreases with depolarizations positive to +30 mV (Fig. 4B). Because fast inactivation is complete within 50 ms (Fig. 4B, inset), normalization to the current magnitude at 100 ms allows qualitative comparison of slow inactivation in the absence and presence of Kv1.3 (Fig. 4, C and D). The normalized tracings in the absence of Kv1.3 superimposed (Fig. 4C), indicating that the extent and rate of slow inactivation of Kv1.5 was independent of membrane potential, i.e., the tracings reach a plateau at 10 mV (36). In contrast, the tracings at each membrane potential in the presence of Kv1.3 did not superimpose (Fig. 4D), indicating that the extent of slow inactivation depends on the membrane potential.

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Kv1.3 enhances slow inactivation through a mechanism that requires a functional fast inactivation domain. Oocytes expressing Kv1.5 (A and C) or Kv1.5 + Kv1.3 (B and D) were depolarized for 5 s to +70, +50, and +30 mV. To emphasize differences in slow decay rates, currents were normalized to amplitude at 100 ms (indicated by arrows in A and B) and superimposed as shown in C and D, respectively. E: experimental pulse protocol used to quantitate slow component of inactivation and a typical current tracing from an oocyte expressing Kv1.5 + Kv1.3. Slow inactivation (S) was defined as amount of current that did not recover within 50 ms at 80 mV, fast inactivation (F) as amount that recovered during 50-ms pulse, and open (O) as noninactivated current. F: data obtained from a minimum of 4 oocytes expressing Kv1.5, Kv1.5 + Kv1.3, Kv1.5+ Kv1.3(N10), and tandem(N87) were averaged and plotted as fraction of slow inactivation vs. duration of 1st depolarizing pulse. Error bars, SE. Average values were fit with a single exponential. P1 and P2, peak currents of 1st and 2nd pulses, respectively.

Characterization of Kv1.3 interaction with internal and external Kv1.5 pore mutations. One possible mechanism for the fast component of inactivation is that the amino terminus of Kv1.3 acts as an open channel blocker that binds within the open pore. Because the antiarrhythmic agent quinidine probably interacts by such a mechanism (44), we tested whether a Kv1.5 point mutant that affects quinidine block (V512A) would also affect Kv1.3-mediated inactivation. The Kv1.5(V512A) mutation is located in the internal mouth of the ion-conducting pore and has three major effects on Kv1.5 function (44): a shift in the midpoint of activation (Table 1), a slowing of deactivation, and a fourfold increase in the affinity for quinidine. If the binding site for the inactivation particle of Kv1.3 overlaps that of quinidine, then the V512A mutation may also show an enhanced affinity for the -subunit blocking particle (and hence display more inactivation). Currents elicited from oocytes expressing the Kv1.5(V512A) mutant -subunit in the absence and presence of Kv1.3 are shown in Fig. 5, A and B. The internal pore mutation (V512A) did not affect either the Kv1.3-mediated inactivation or the hyperpolarizing shift in the midpoint of activation (Fig. 5 and Table 1). Another internal pore mutation of Kv1.5, T505I, which enhances quinidine block ~10-fold (44), also did not affect Kv1.3 function (data not shown).

View larger version (31K): [in this window] [in a new window]   Fig. 5.   An external pore mutation alters fast inactivation. Current from oocytes expressing Kv1.5 subunit point mutants V512A (A and B) and R485Y (C and D) are shown in absence (A and C) and presence (B and D) of Kv1.3. Currents in A and B were elicited by 200-ms depolarizations from a holding potential of 100 mV to potentials between 90 and +70 mV in 10-mV increments. Currents in C and D were elicited by 200-ms depolarizations from a holding potential of 80 mV to potentials between 60 and +70 mV in 10-mV increments. Test potentials were followed by a repolarizing step to 40 mV to record deactivating currents.

Effect of high external K⁺ on inactivation. Sensitivity of block to concentrations of the conducting ion has been used to support an open pore model of block by both pore-blocking drugs and the inactivation particle (11). Increasing the external K⁺ concentration also reduces the slow component of inactivation (2, 24, 31). To test whether Kv1.3-mediated inactivation was sensitive to external K⁺ concentration, we compared the Kv1.3-mediated fast and slow inactivation of Kv1.5 and Kv1.5(R485Y) in the absence and presence of 96 mM extracellular K⁺ (Fig. 6). High external K⁺ concentrations reduced the extent of both fast and slow inactivation. Figure 6, A and B, shows representative current tracings during depolarizations to +70 mV in oocytes expressing Kv1.5 + Kv1.3 and Kv1.5 (R485Y) + Kv1.3, either in normal K⁺, in high K⁺, or after return to normal K⁺ (wash). Currents were normalized to the peak to emphasize the difference in the extent of inactivation. For Kv1.3 with either the wild-type or R485Y -subunit, the presence of high external K⁺ reduced the extent of inactivation at 200 ms (Table 1). To compare the time course of inactivation, currents from Fig. 6, A and B, were further normalized to the current level at 1 s as shown in Fig. 6, C and D. For the wild-type subunit, increasing external K⁺ eliminated the slow component of inactivation without affecting the rate of fast inactivation (see also Table 1). The slow component of inactivation returned on washout of external K⁺. Figure 6D illustrates that the rate of slow inactivation of Kv1.5(R485Y) in the presence of Kv1.3 was not affected by the presence of 96 mM K⁺.

View larger version (29K): [in this window] [in a new window]   Fig. 6.   High external K+ reduces extent of fast inactivation. Oocytes expressing Kv1.5 + Kv1.3 (A and C) or Kv1.5 (R485Y) + Kv1.3 (B and D) were depolarized to +70 mV for 1 s, and currents were recorded in normal K+ (2 mM KCl), high K+ (96 mM KCl) and after washing back to normal K+. A and B: currents were normalized to peak to compare extent of inactivation. For each tracing, slow inactivation continued at same rate for duration of 1-s pulse (only 1st 300 ms are shown). C and D: currents were further scaled to steady-state current at 1 s to compare inactivation rates under different conditions. Again, 1st 300 ms after peak are shown.

Voltage dependence of the inactivation induced by Kv1.3. Two lines of evidence suggested that Kv1.3-mediated inactivation was voltage dependent. First, macroscopic inactivation was apparent only at large depolarizations. Second, the steady-state current level from depolarizations more positive than +20 mV does not increase with a corresponding increase in the membrane potential (12), which would occur if the open channel probability were decreasing as the driving force increased. To quantitate this voltage dependence, we used a two-pulse protocol in which the membrane potential of the first pulse was varied as shown in Fig. 7A. Figure 7B shows normalized data plotted vs. the membrane potential of the 25-ms first pulse. Because fast inactivation is essentially at steady state and slow inactivation is negligible at 25 ms, this approach isolates the fast component of inactivation. Data were obtained in both the absence and presence of 96 mM extracellular K⁺. The fractional inactivation increased rapidly in the voltage range corresponding to channel activation, indicating that Kv1.3-mediated inactivation required the channel to enter the open state. Note that the 10 mV tracing in Fig. 7A shows no apparent inactivation in the first pulse, whereas there was an obvious reduction in the peak current of the second pulse, i.e., channels inactivate at 10 mV even though inactivation is not apparent in the macroscopic current. Positive to potentials at which the channels are maximally activated (+10 mV), the fractional inactivation continued to increase with membrane potential, indicating that Kv1.3-mediated fast inactivation was voltage dependent (Fig. 7B). The Woodhull model (42) was used to quantitate the voltage dependence, and the derived z ·  values are indicated in Fig. 7. Similar curves and z ·  values were obtained with tandem(N87) and the -subunit point mutants (Fig. 7C). These data suggest that the voltage dependence of Kv1.3-mediated block is independent of individual -subunit properties, consistent with the localization of the voltage-dependent inactivating particle to the amino terminus of Kv1.3 (Fig. 2). The fits are extrapolated to negative membrane potentials to predict the fractional block of open channels at those voltages. Note that elevating external K⁺ to 96 mM reduces the extent of fast inactivation but does not alter its voltage dependence (z · ).

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Fast inactivation is mediated by external K+-sensitive, voltage-dependent, open channel block. Voltage dependence of inactivation was quantitated using a 2-pulse protocol as depicted in A. In 1st pulse, oocytes were depolarized to varying potentials for either 25 ms (A-C) or 1 s (D) and stepped to +70 mV for 200 ms in 2nd pulse. A: tracings from test pulses of 70, 10, and +70 mV. Time point of peak current of 2nd pulse was determined from a 1st pulse potential of 70 mV and is represented by an arrow. Current amplitudes of subsequent pulses were determined at this time point, normalized with respect to maximum, and plotted vs. potential of 1st depolarizing pulse (B). Similar plots were made for tandem(N87) and -subunit point mutants (C). A Boltzmann equation was fit to data from voltages exceeding that of maximal activation (greater than +20 mV). Fits (solid lines) are extended into negative voltages to predict fractional inactivation of open channels at those voltages. Using the Woodhull model (42), z ·  values were determined from derived slope factor (z represents effective charge of blocking particle, and  represents fraction of transmembrane potential at blocking site). Data obtained in high external K+ (96 mM) are indicated (B and D). Data are from 5 oocytes.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Separation of Kv1.3-mediated functional effects. Other investigators have suggested that the hyperpolarizing shift in the activation curve observed following coexpression of Kv1.5 with Kv1.2 is specifically linked to inactivation and that the -subunit does not directly alter activation properties (10). Precise quantitation of the voltage dependence of activation is difficult in the presence of the voltage-dependent inactivation observed with open channel block. Once current values are scaled, an artifactual shift in the activation curve can be generated. Because analysis of peak current levels is especially prone to artifact, activation curves were obtained from the tail current amplitude 6 ms after repolarization to 40 mV. Because the time constant for recovery from fast inactivation at 80 mV was 5.1 ± 0.1 ms, these tail current measurements are not perfect. However, even before the experiments reported here with the Kv1.3 mutants, it seemed likely that, at least in the case of Kv1.5 and Kv1.3, the -subunit-induced activation shift was independent of inactivation. Indeed, current was observed at membrane potentials as low as 30 mV, a potential generating no current in the absence of the -subunit. In addition, the Kv2.1 subunit shifts the Kv1.5 activation curve without inducing fast inactivation (16, 40).

Deactivation is affected by fast inactivation. Prolonged deactivation has been observed in the presence of open pore-blocking drugs and N-type inactivation (11, 35, 37). Blockade of the pore by either the drug or inactivation particle is thought to prevent deactivation (11, 35). Thus the drug or inactivating particle must dissociate before the channel can close. The fact that the open channel is subject to reblock further complicates the deactivating current. An open channel will conduct current until it deactivates or is reblocked. If the channel becomes reblocked, deactivation is again prevented, and the particle must dissociate before the channel can deactivate. Therefore, reblock by the drug or inactivating particle can significantly contribute to prolonging the deactivating current (11). Shifting the midpoint of activation also results in altered kinetics of deactivation (39). The tandem(N87) construct, which conferred fast inactivation without altering the midpoint of activation, and the Kv1.3(N10) truncation, which hyperpolarized the midpoint of activation without inducing fast inactivation, demonstrated deactivation time constants intermediate between those observed with Kv1.5 alone and Kv1.5 plus the wild-type Kv1.3. These data indicate that the prolonged deactivation observed in the wild-type subunit may be the result of the combined effects of a shift in the midpoint of activation and reentry into the inactivated state.

External, not internal, Kv1.5 pore mutations alter Kv1.3-induced inactivation. It was conceivable that the inactivation particle of Kv1.3 interacts with the -subunit at the same cytoplasmic binding site as quinidine, since internal TEA derivatives can compete with N-type inactivation in Shaker (7). However, the two -subunit point mutants, V512A and T505I, which increase quinidine block (44), failed to alter Kv1.3-induced inactivation. Thus the molecular determinants in Kv1.5 for Kv1.3-induced fast inactivation and open channel drug block appear to be distinct. However, it is still possible that there is overlap in the -subunit domains involved in quinidine block and Kv1.3-induced inactivation.

Inactivation is dependent on both external K⁺ and voltage. A model introduced by Baukrowitz and Yellen (2-4) for the Shaker channel proposes that occupancy of an external K⁺ binding site inhibits slow inactivation. This model proposes that the enhancement of slow inactivation in the presence of internal blockade (drug or inactivation ball) is a direct result of depletion of K⁺ from this external site. Increasing the external K⁺ concentration allows occupancy of this site, despite internal blockade, returning the extent of slow inactivation to the level observed in the absence of the blocking agent. In agreement with this model, elevating external K⁺ inhibited the slow component of inactivation, as seen in Fig. 6C. However, it also clearly inhibited the extent of fast inactivation, as shown in Figs. 6B and 7B. Because the extent of slow inactivation appears to be coupled to the extent of fast inactivation (Fig. 4F), we cannot discern whether the reduction in slow inactivation is the result of reduced fast inactivation or a direct effect of K⁺ on slow inactivation.