Separable Kvβ Subunit Domains Alter Expression and Gating of Potassium Channels*

Kvβ subunits have been shown to affect kinetic properties of voltage-gated K+ channel Kv1α subunits and increase the number of cell surface dendrotoxin-binding sites when coexpressed with Kv1.2. Here, we show that Kvβ1.2 alters both current expression and gating of Kvα1 channels and that each effect is mediated by a distinct Kvβ1.2 domain. The Kvβ1.2 N terminus or Kvα1-blocking domain introduced steady state current block, an apparent negative shift in steady state activation, and a slowing of deactivation along with a dramatic reduction in single channel open probability. N-terminal deletions of Kvβ1.2 no longer altered channel kinetics but promoted dramatic increases in Kv1.2 current. The conserved Kvβ1 C terminus or Kvα1 expression domain alone was sufficient to increase the number of functional channels. The same effect was observed with the normally noninactivating subunit, Kvβ2. By contrast, Kv1.5 currents were reduced when coexpressed with either the Kvβ1 C terminus or Kvβ2, indicating that the Kvα1 expression domain has Kvα1 isoform-specific effects. Our results demonstrate that Kvβ subunits consist of two domains that are separable on the basis of both primary structure and functional modulation of voltage-gated K+ channels.

Kv␤ subunits have dual effects on Kv␣1 subunit currents. The three Kv␤1 isoforms, as well as Kv␤3, introduce inactivation into Kv␣1 subunit currents but with variable potency (2,3,9). The Kv␤1 N termini are thought to produce inactivation by acting like a ball peptide to mimic N-type inactivation (2,11). Coexpression of Kv␤1.2 with Kv1.5 has also been shown to shift the activation curve to more negative voltages and to increase the time constant of deactivation (6), although these effects may be secondary to open channel block (11). In contrast, Kv␤2 subunits do not alter the inactivation or deactivation properties of coexpressed Kv␣1 subunits but do shift activation kinetics of Kv1.5 to more negative voltages (2,12).
More recently, Kv␤ subunits were shown to increase the surface expression of Kv1.2 as measured by an increase in dendrotoxin-binding sites at the plasma membrane of transfected cells (13). The functional consequences of this increase as well as the Kv␤ subunit domain responsible for this effect have not been determined. The Drosophila Kv␤ subunit encoded by the hyperkinetic locus has also been shown to alter both the kinetics and current amplitude of Shaker Kv␣ subunits (14), but it is not known whether the increase results from alterations in intrinsic inactivation or functional expression. Interestingly, a recent report found that Kv␤2 did not increase the surface expression of Shaker channels, although the two proteins were found to associate in the endoplasmic reticulum (15).
We have examined the functional consequences of Kv␤1 assembly with Kv␣1 channels and determined the Kv␤1 domains responsible. We show that the Kv␤1.2 C terminus (Kv␤1-C) as well as Kv␤2 produce dramatic increases in Kv1.2 currents when coexpressed in Xenopus oocytes. This increase in functional expression is due to an increase in the number of functional channels and not by alterations in single channel properties. In contrast, Kv1.5 currents show a dramatic decrease upon coexpression with Kv␤1-C, indicating that the nature of the expression effect is Kv␣1 isoform-dependent. All kinetic effects of Kv␤1.2 can be explained by open channel block of Kv1.2 by the Kv␤ N terminus. Thus, our studies of noninactivating Kv␣1 channels have demonstrated that Kv␤1.2 consists of two functional domains: a Kv␣1-blocking domain located in the first 20 amino acids of the N terminus and a Kv␣1 expression domain consisting of the entire C terminus. scription-PCR 1 from rat brain poly(A) ϩ RNA. Protein-protein interactions were monitored with the yeast GAL-4 Matchmaker Two-hybrid System from CLONTECH. In-frame EcoRI and SalI sites were incorporated into the 5Ј and 3Ј ends, respectively, of the coding sequences of Kv␤1.2 (amino acids 1-408), Kv␤2 (amino acids 1-367), Kv␤1.2-N⌬20 (amino acids 21-408), and Kv␤1.2-C terminus (amino acids 81-408) by PCR for cloning into the yeast shuttle vectors, pGBT9 and pGAD424. Constructs generating fusion proteins with the N terminus of Kv1.2 (Kv1.2-N; amino acids 1-124) were also made. The PCR-amplified constructs were sequenced to confirm the correct reading frame for the yeast fusions and check that no unwanted PCR mutations were introduced.
Protein-protein interactions were tested in the yeast host strain Y190 by cotransformation with pairs of pGBT9 and pGAD424 fusion constructs. Cotransformants were selected on medium lacking tryptophan (trpϪ) and leucine (leu Ϫ ) after growth for 2-3 days at 30°C. Representative colonies from these transformations were replated on trpϪ/leuϪ medium to allow direct comparison of individual colonies. Transcription of the reporter gene, lacZ, was tested by a ␤-galactosidase filter assay.
In Vitro Transcription of cRNAs and Oocyte Injection-Kv1.2, Kv1.5, and Kv1.4-⌬N2-146 cRNAs were prepared as described previously (9). Kv␤1.2 and Kv␤2 were subcloned into pSP64 (Promega) for cRNA synthesis. The Kv␤1.2-C terminus was subcloned into a modified version of pCRII, A ϩ -pCRII, which we had previously altered to include a poly(A) ϩ tail (16). For oocyte expression, the Kv␤1.2-N⌬20 construct was removed from the yeast vector, pGBT9, by digestion with EcoRI and SalI and subcloned into a modified version of pSP64. The modified pSP64 vector was constructed as follows: an adaptor containing a Kozak concensus sequence for initiation of translation followed by an ATG start codon and an EcoRI site was ligated into the HindIII and PstI sites of the multiple cloning region of pSP64. The HindIII site was destroyed in this ligation. In addition, the EcoRI site following the poly(A) ϩ tail in pSP64 was deleted, and a HindIII site was inserted by ligation of an adaptor. The EcoRI/SalI insert from the yeast vector was then subcloned into the new EcoRI site and the downstream SalI site. As a result, the Kv␤1.2-N⌬20 construct for oocyte expression has three amino acids, Met, Glu, and Phe, preceding amino acid number 21 of Kv␤1.2 sequence.
Constructs in the pSP64 and A ϩ -pCRII vectors were linearized with EcoRI and BamHI, respectively, and transcribed into cRNA with the mMESSAGE mMACHINE in vitro transcription kit (Ambion) using SP6 or T7 polymerase as required. cRNA concentrations were estimated on denaturing agarose gels stained with ethidium bromide by comparison with RNA standards. RNA was injected into Xenopus oocytes as described previously (4).
Electrophysiology-Measurement of Xenopus oocyte currents were performed using the standard two microelectrode voltage clamp technique. For comparison of whole cell current amplitudes, all measurements were done 3 days post-injection. 2-4 days after injection, single channel recordings were performed in the cell-attached configuration using pipettes pulled from hard borosilicate glass with resistances of 5-10 M⍀. Pipettes were coated with Sylgard and fire-polished immediately before use. The seal resistance was 50 -500 G⍀. Xenopus oocytes were always patched on the dark hemisphere.
Data Analysis Single Channels-Data were low pass filtered at 2 kHz (Ϫ3dB, 4-pole Bessel filter) before digitalization at 10 kHz. PClamp software (Axon Instruments) was used for generation of the voltagepulse protocols and for data acquisition. All single channel measurements were leak subtracted in Fetchan and analyzed using TRANSIT software (17). This produced histograms for the open time, ensemble current traces, and values of open probability. The maximum likelihood method is utilized by TRANSIT to determine time constants from open time histograms. The data for open time from TRANSIT are presented on a log scale as described previously (17). Single channel parameters were given as the means Ϯ S.D.
After forming a giga-seal, single channel I-V protocols were performed to ensure that the investigated channel had the expected acti-vation threshold and reversal potential (9). Measurements with only one channel in the patch were used for further investigation. From a holding potential of Ϫ80 mV, we measured 1000 test pulses to a potential of 70 mV at a frequency of 2 per second. Leak subtraction was done by averaging 10 null records and subtracting it from each trace as described previously (18).

Interaction of Kv1.2 N Terminus with Kv␤ Subunit
Fragments-To investigate the functional role of distinct Kv␤ subunit domains on Kv1 channels, we constructed noninactivating forms of Kv␤1.2 by deleting either 20 or 79 amino acids from the N terminus. In the deletion lacking the first 20 amino acids (Kv␤1.2-N⌬20), one of two stretches in the N terminus suggested previously to contain a "ball" peptide sequence was deleted (9). Deletion of 79 amino acids removed the entire N terminus, leaving the conserved Kv␤1-C terminus. These deletions did not affect the ability of Kv␤1.2 to bind to the Kv1.2 N terminus as assessed with the yeast two hybrid system (Fig. 1). Kv␤2, which has a C-terminal domain highly homologous to Kv␤1, also interacted with the N terminus of Kv1.2 in the yeast two-hybrid assay. Because the C terminus of Kv␤ subunits confers the ability to interact with the N terminus of Kv␣1 subunits, we examined the functional consequences of the interaction between truncated Kv␤ subunits and Kv1.2 channels.

Kv␤1 C Terminus Alters Functional Expression of Kv1.2-
The properties of whole cell Kv1.2 currents were examined upon coexpression with full-length Kv␤1.2 and N-terminal deletion mutants in Xenopus oocytes. Kv1.2 alone produced noninactivating currents ( Fig. 2A). Co-expression of Kv1.2 cRNA with saturating amounts of Kv␤1.2 cRNA resulted in currents exhibiting partial inactivation (Fig. 2B) as described previously (9). As predicted, coexpression of Kv1.2 with saturating amounts of Kv␤1.2-N⌬20 produced noninactivating currents (Fig. 2C). In addition, we observed a dramatic increase in the amplitude of Kv1.2 currents with the truncated Kv␤ subunit. When Kv1.2 was coexpressed with Kv␤1.2-N⌬20, the current amplitude increased by a factor of 6.4 Ϯ 2.6 (n ϭ 4 -11 eggs in each group in five injection series) compared with Kv1.2 alone (Fig. 2C). Fig. 2D shows the summary of steady state current amplitudes obtained in one series of measurements (n ϭ 5-11 eggs in each group). Because the amount of Kv1.2 cRNA injected remained constant in each case, it appeared that the truncated Kv␤ subunit increased the functional expression of Kv1.2 channels. This effect was largely masked with the wildtype Kv␤1.2 subunit because of the open channel block conferred by the Kv␤1.2 N terminus.
To determine whether this property could be further localized on the Kv␤ subunit, we tested the effects of the Kv␤1 C 1 The abbreviation used is: PCR, polymerase chain reaction. terminus on Kv1.2 currents. The Kv␤1 C terminus did not alter the inactivation properties of Kv1.2 (not shown) but did result in a significant increase (about 5-fold) in the amplitude of Kv1.2 currents when cRNAs were coinjected in Xenopus oocytes (Fig. 3). Yeast two-hybrid experiments showing the binding of Kv␤1-C to the N terminus of Kv1.2 strongly suggest that the dramatic increase in functional surface expression of this channel is due to the direct interaction of these subunits. Coexpression of Kv1.2 with either Kv␤1.2-N⌬20 or Kv␤1-C increased current amplitudes to similar extents, by factors of 4.19 and 4.18, respectively (n ϭ 4 -5 oocytes in one injection series). This indicates that the Kv␣1 expression domain consists of the Kv␤ C terminus alone. Coinjection of saturating amounts of a normally noninactivating Kv␤ subunit, Kv␤2, produced similar results (Fig. 3). The high degree of homology between the C-terminal domains of Kv␤1 and Kv␤2 subunits is reflected in the similar amplitude increases elicited by the two subunits.
The increase in current amplitude occurred in both low K ϩ (5 mM) and high K ϩ (50 mM) solutions ruling out external K ϩ as the major cause of the increase in the maximal conductance. As a test for specificity, we coexpressed Kv␤1.2 or Kv␤1.2-N⌬20 with HERG, the K ϩ channel responsible for the rapid component of the delayed rectifier K ϩ current in human heart. In two injection series, we found no significant changes in kinetics or current amplitude when HERG was coexpressed with Kv␤1.2 or Kv␤1.2-N⌬20 (data not shown).
Single Channel Measurements Demonstrate That the Increase in Kv1.2 Current Mediated by the Kv␤1 C Terminus Is Due to an Increase in the Number of Functional Channels at the Cell Surface-To determine the mechanism by which the Kv␤1 C terminus increased functional expression, we compared the single channel properties of Kv1.2 alone and in the presence of Kv␤1.2 or the noninactivating Kv␤1.2-N⌬20. In a previous report, we demonstrated that the properties of single Kv1.2 channels are altered by Kv␤1.2 (9). These were confirmed here at a test potential of ϩ70 mV and compared with single channel measurements of Kv1.2 and Kv␤1.2-N⌬20 to see whether the mutant Kv␤ altered single channel properties. Kv1.2 cRNA injected alone produced single channels with openings in the outward direction, short closures, and a mean amplitude of 1.49 Ϯ 0.10 pA (n ϭ 7). The channels were open and showed no apparent inactivation (Fig. 4A). The open time distribution was fitted with a single exponential to give an open time of 4.1 ms (Fig. 4D). The open probability (P o ) distribution gave a mean of 0.81 and showed virtually no inactivation as well (Fig. 4G). From seven patches, the mean open time (t) was 3.91 Ϯ 0.23 ms, and the P o was 0.80 Ϯ 0.03.
Expression of Kv1.2 with Kv␤1.2 resulted in single channels with completely different kinetics as described previously. Channels opened frequently in the beginning of the trace followed by long closures before reopening (Fig. 4B). The amplitude of the openings was reduced to 1.13 Ϯ 0.14 pA (n ϭ 8), presumably because blocking events happened very fast, preventing the channels from reaching full amplitude under our recording conditions. Subsequently, the mean open time de- creased to 0.59 ms, and the mean open probability was reduced dramatically to 0.05 (Fig. 4, E and H). The mean of eight patches gave the following values: t ϭ 0.66 Ϯ 0.08 and P o ϭ 0.08 Ϯ 0.02. When Kv1.2 was coexpressed with Kv␤1.2-N⌬20, we observed single channels with similar properties to Kv1.2 alone (Fig. 4, compare C and A).  (11). Thus, removal of the ball from Kv␤1.2 would be expected to eliminate alterations in activation and deactivation as well as inactivation. To test this, activation curves of Kv1.2 were generated in the absence and the presence of inactivating and noninactivating Kv␤ subunits (Fig. 5A). In the presence of Kv␤1.2, Kv1.2 currents exhibited a leftward shift of almost 12 mV in the activation curve compared with Kv1.2 expressed alone (Fig. 5A). On the other hand, neither of the noninactivating Kv␤ fragments, the Kv␤1 C terminus nor Kv␤2, produced any significant changes in the activation curve of Kv1.2.
Similarly, deactivation of Kv1.2 was significantly slowed in the presence of Kv␤1.2 but neither the Kv␤1 C terminus nor Kv␤2 produced large changes in deactivation compared with Kv1.2 alone (Fig. 5B). Furthermore, neither the Kv␤1 C terminus nor Kv␤2 produced any significant changes in the activation time constants of Kv1.2 (data not shown). Therefore, the kinetic changes in Kv1.2 currents that have been observed in the presence of Kv␤1.2 are due to the presence of a functional ball peptide on the Kv␤ N terminus plugging the Kv1.2 pore. When the Kv␤1.2 N terminus is removed or when the normally noninactivating Kv␤2 subunit is tested, the kinetics of Kv1.2 are unaltered even though the current amplitudes are dramatically increased (see Fig. 3). Thus, there is a clear separation in functional effects between the structurally distinct Kv␤1.2 Nand C-terminal domains.
The Maximal Fraction of Kv1.2 Current Blocked by the Kv␤1.2 N Terminus Is Much Larger than Previously Suspected-To quantify the block by the Kv␤1.2 N terminus, we initially considered fractional block or inactivation as the difference between peak current and steady state current divided by the peak current as done previously (2,9,11). The fractional block is plotted as a function of Kv␤1.2 concentration in Fig.  6A. A maximal peak to steady state inactivation of 0.66 was determined similar to results reported previously (9). A problem with this method is that it implies that 34% of the total current remained unblocked. This is clearly not consistent with single channel measurements, which show that the open probability was reduced to 0.08 in the presence of Kv␤1.2. We assumed that this inconsistency was due to a reduction in the peak current, masked by the dramatic concomitant increase The line of fit was generated using a maximum likelihood method as described previously (10). Note that the x axis is on a log scale and most of the data is at the longer time points. The line of fit follows the data very closely at these longer times. G-I, corresponding open probabilities. Oocytes were held at Ϫ80 mV, and 1000 traces were recorded in the cellattached configuration with 5 mM K ϩ in the pipette during pulses to ϩ70 mV of 100-ms duration at a frequency of 2 Hz. produced by Kv␤1-C, thereby overestimating the amount of sustained outward current. A more accurate measure of the blocking effect of Kv␤1.2 was determined from the steady state currents to eliminate bias of current block due to normalization with the peak current. First, we measured the steady state current levels as a function of Kv␤1.2-N⌬20 and Kv␤1.2 cRNA concentrations (Fig. 6B). We then assumed that the steady state current in the presence of Kv␤1.2-N⌬20 reflected the current amplitude that would be measured in the presence of Kv␤1.2 if there were no block. The amount of steady state current block at each concentration of RNA was then calculated in the following way, I ss current block ϭ (mean I ss Kv␤1.2N⌬20 Ϫ mean I ss Kv␤1.2)/mean I ss Kv␤1.2N⌬20, where I ss is the steady state current amplitude. The relation of steady state current block versus concentration is shown in Fig. 6C. The maximum fractional current block determined in this way was 0.94. This is much greater than the maximum block of 0.66 estimated from the ratio of peak to steady state in Fig. 6A. The maximal fractional current block determined by both methods saturates at 100 ng/l. The saturation point is not affected by normalization with the peak.
The single channel studies support the large reduction in current produced by Kv␤1.2 determined above. The maximal block of Kv1.2 by the N-terminal domain of Kv␤1.2 can be calculated from the change in open probability from 0.8, in the presence of the noninactivating Kv␤1.2-N⌬20, to 0.08, in the presence of Kv␤1.2 as follows, maximum block ϭ fractional change in open probability ϭ (0.8 Ϫ 0.08)/0.8 ϭ 0.90. This value is much higher than was predicted by previous macroscopic current data, which suggested a maximal block of only 60 -70% based on peak to steady state current (2,9). On the other hand, the value of 0.90 is very close to the value of 0.94 determined from fractional steady state current block (Fig. 6C, shown above). The single channel and macroscopic data confirm that the N-terminal domain of Kv␤1.2 is able to block much more available outward current than previously suspected and significantly reduce the peak current as suggested previously (2,9). This also implies that even small amounts of an inactivating Kv␤ subunit like Kv␤1.2 in channel complexes could result in significant reductions in open probability and outward current.  (4,6,7,11,12). We asked whether Kv␤1-C would also enhance Kv1.5 expression as observed with Kv1.2. Surprisingly, Kv1.5 currents were significantly decreased upon coexpression with Kv␤1-C as shown in Fig. 7A, where comparisons were made directly with Kv1.2. In four injection series, average Kv1.5 current levels decreased from 9.31 Ϯ 0.94 A (n ϭ 31) to 3.24 Ϯ 0.32 A (n ϭ 31) when expressed with the Kv␤1-C terminus. A decrease in Kv1.5 current levels was also noted upon coexpression with Kv␤2 (Fig. 7B). These data show that the Kv␣1 expression domain of Kv␤1 can produce different and even opposite effects on different Kv␣1 subunits. This differs from the Kv␣1-blocking domain, which effectively blocks FIG. 5. Kinetic alterations in Kv1.2 currents require the Kv␤1.2 N terminus. A, steady state activation curves. Curves were generated from peak tail currents, at Ϫ80 mV, normalized to the maximum, and plotted against test potential. Data were fitted to the following Boltzmann equation, I/I max ϭ 1 ϩ exp(V Ϫ V 0.5 )/k, where I is current, V is the test potential, and k is the Boltzmann constant. The following values were obtained for half-maximal activation (V 0.5 ) and k, respectively; Ϫ1.66 Ϯ 0.30 mV, 7.25 Ϯ 0.25 mV (Kv1.2, n ϭ 8, circles); Ϫ3.21 Ϯ 0.19 mV, 6.01 Ϯ 0.16 (Kv1.2 ϩ Kv␤1 C terminus, n ϭ 14, triangles); Ϫ2.69 Ϯ 0.30, 6.54 Ϯ 0.26 (Kv1.2 ϩ Kv␤2, n ϭ 11, inverted triangles); and Ϫ13.21 Ϯ 0.16, 6.42 Ϯ 0.14 (Kv1.2 ϩ Kv␤1.2, n ϭ 11, squares). B, deactivation time constants as a function of test potential. Oocytes were pre-pulsed to the following potentials for 150 ms; 58 Ϯ 0.6 mV (Kv1.2, n ϭ 6), 53 Ϯ 1.2 mV (Kv1.2 ϩ Kv␤1 C terminus), 51 Ϯ 0.75 mV (Kv1.2 ϩ Kv␤2, n ϭ 7), and 58 Ϯ 0.8 mV (Kv1.2 ϩ Kv␤1.2, n ϭ 6) and pulsed to different test potentials for 100 ms. Deactivation of the tail currents was fit with a single exponential function. All measurements in A and B were carried out in oocytes injected with 2 ng/l of Kv1.2 alone or with 250 ng/l of Kv␤1.2, Kv␤1 C terminus, or Kv␤2 using a high K ϩ NMDG ϩ solution (see "Experimental Procedures"). Symbols are as in panel A.
Kv␣1 subunit currents, although the potency with which different N-terminal domains introduce block may vary (2,9). DISCUSSION Kv␤ subunits have multiple effects on voltage-gated channels of the Kv1 subfamily. Our results show that Kv␤1.2 can be divided into two structural and functional domains, one of which, the N terminus, acts as the Kv␣1-blocking domain by producing inactivation. Here we have shown that the Kv␤1 C terminus alone is able to bind to the Kv␣1 subunit and promote the alteration of current expression. Thus, the Kv␤1-C terminus acts as the Kv␣1 expression domain. Coexpression of Nterminal Kv␤1.2 deletion mutants with Kv1.2 resulted in noninactivating currents and increased current amplitude (up to 14-fold). Single channel amplitude and open probability were not affected by coexpression of Kv1.2 with the deletion mutant Kv␤1.2-N⌬20. Therefore, the observed increase in whole cell current amplitude was due to an increased number of functional channels in the membrane. An increase in the number of functional Kv1.2 channels must occur because the single channel open probability of Kv1.2 determined here and previously (9) is approximately 0.8, yet the increase in Kv1.2 currents by noninactivating Kv␤1.2 deletions was at least 4-fold. This increase is in agreement with experiments in mammalian cells showing enhanced transport of Kv1.2 channels to the membrane in the presence of Kv␤ subunits (13). Yeast two-hybrid interaction data presented here and previously (8,9) suggest that the increase in current results from a direct interaction between the subunits and have shown that the regions critical for interaction are located in the N terminus of Kv␣1 subunits and the Kv␤ subunit C terminus. This is in agreement with coimmunoprecipitation studies indicating that the interaction of Kv␤ subunits with Kv1␣ subunits does not require the Kv␤ N-terminal domain (19).
Macroscopic and single channel data have shown that the first 20 amino acids of the N terminus of Kv␤1.2 contain the ␤-ball peptide. This domain introduces current block (inactiva- FIG. 6. Dose dependence of inactivation, steady state current block, and increase in current levels of Kv1.2 produced by Kv␤1.2. A, inactivation due to Kv␤1.2 measured as the ratio of the difference between peak current and steady state current divided by the peak current. Note that maximum inactivation is about 65% and is close to saturation at 100 ng/l. B, whole cell steady state currents measured in oocytes injected with a constant amount of cRNA for Kv1.2 and increasing amounts of cRNA for Kv␤1.2 or Kv␤1.2-N⌬20. Note that saturation has not occurred at 100 ng/l. C, steady state current block versus increasing amounts of cRNA for Kv␤1.2. Note that the maximum current block is greater than 0.9 and is close to saturation at 100 ng/l. Steady state current block was obtained from the mean currents in B using the following calculation, I ss current block ϭ (mean I ss Kv␤1.2-N⌬20 Ϫ mean I ss Kv␤1.2)/mean I ss Kv␤1.2-N⌬20, where I ss is the steady state current. Oocytes were held at Ϫ80 mV and pulsed to ϩ70 mV, and steady state currents were measured at the end of a 100 ms (total number of eggs ϭ 88 in the one injection series). tion) as originally proposed based on homology with Kv␤1.1 (4). Similarly, the deletion of 34 amino acids from the N terminus of Kv␤1.1, a subunit that normally confers inactivation upon noninactivating K ϩ channels, resulted in a loss of inactivation (2). Previous experiments suggested that Kv␤1.2 not only introduces inactivation but shifts activation kinetics to more negative voltages and increases deactivation time constants of Kv1.1 and Kv1.5 (7,11,12). Modeling studies have suggested that the changes in activation and deactivation of Kv1.5 currents, like inactivation, are a result of the open channel block of Kv␤1.2 (11). Our data with Kv1.2 are consistent with this interpretation because the removal of 20 or more amino acids from the N terminus of Kv␤1.2 not only removed inactivation completely but also eliminated changes in activation and deactivation kinetics.
We found the true maximal fractional current block of Kv1.2 by Kv␤1.2 to be 0.94, much greater than 0.66, which was determined from the ratio of steady state to peak current (also called "inactivation"; see 2,9). The discrepancy between these values can be explained by a rapid open channel block (9, 11), which reduces the peak current leading to an underestimation of current block produced by Kv␤1.2 when calculated as a fraction of the peak current. Significant fractional block, as estimated from the ratio of steady state currents, can occur in the absence of any apparent inactivation as estimated from the ratio of steady state to peak current. Hence, the ratio of steady state to peak current is not a reliable indication of current block.
Kv␤2, a noninactivating Kv␤ subunit, resembles the Kv␤1 C terminus with respect to functional modifications of Kv␣1 currents. As for Kv␤1-C, we observed large increases in current expression and little change in the kinetic parameters of Kv1.2 channels with Kv␤2. Kv␤2 also promoted dramatic decreases in Kv1.5 current expression. In yeast two-hybrid assays, Kv␤2 was also shown to interact with the N termini of Kv1.2 and Kv1.5. These results were not unexpected because there is extensive homology in the C-terminal domains of all Kv␤ subunits. A recent report showing that Kv␤2 and Kv␤1 bind to the same region on the Shaker N terminus emphasizes the similarity in the two subunits (20). However, in contrast to the Kv␤1 subfamily, a function for the Kv␤2 N terminus has not been described.
Surprisingly, the C terminus of Kv␤1 produced a significant decrease in Kv1.5 current levels in contrast to the dramatic increase observed with Kv1.2. Thus, the action of the Kv␣1 expression domain apparently differs depending upon the particular Kv␣1 subunit it is complexed with. The origin of these differences are presently unknown. Surface expression of Kv1.1, Kv1.3, and Kv1.6, in addition to Kv1.2, were all shown to increase upon coexpression with Kv␤2 but to a lesser extent than Kv1.2 (13). In contrast, the rate of maturation of Shaker channels actually decreased when coexpressed with Kv␤2 (15). In heterologous expression systems, Shaker subunits are fully glycosylated and transferred to the Golgi apparatus in the absence of Kv␤ subunits (21) unlike Kv1.2, which is poorly glycosylated and found predominantly in the endoplasmic reticulum (13). Taken together, this would suggest that Kv1.2, Kv1.1, Kv1.3, and Kv1.6 subunits exist predominantly in an immature form in the absence of Kv␤ subunits compared with Kv1.5 and Shaker. Kv␤ subunits would then promote transit of the former group to the cell surface resulting in increased expression. This hypothesis would not, however, explain the decrease in Kv1.5 expression we observed or the decrease in maturation of Shaker channels in the presence of Kv␤2 (15). The decrease in functional expression of Kv1.5 may be due to a nonspecific competition for factors required for expression between Kv1.5 and Kv␤ as suggested for Shaker (15) or may be due to other mechanisms specific to Kv1.5-Kv␤ interactions. Quantitative differences in Kv␣-Kv␤ interactions may also exist that contribute to the functional differences described. For example, Nakahira et al. (19) showed that significantly less Kv1.5 protein was coimmunoprecipitated with anti Kv␤ subunit antibodies compared with Kv1.2, possibly because of a lower affinity between Kv␤ and Kv1.5 subunits.
In summary, cytoplasmic Kv␤ subunits affect Kv␣1 channels in two ways through distinct protein domains. The Kv␤ C terminus, which is highly conserved among all subfamilies of Kv␤ subunits thus far described, acts as a Kv␣1 expression domain, and in the case of Kv1.2, increases the number of functional channels at the cell surface. Further experiments will be required to determine the origin of the decrease in Kv1.5 expression mediated by this domain. For Kv1.2, all of the apparent kinetic alterations can be attributed to the variable Kv␤1 N-terminal domains. Because all Kv␤ subunits share similar C-terminal regions, it is tempting to speculate that their original association with ion channels served to modulate K ϩ currents through alteration of expression. Kv␤2, for example, promotes alterations in expression, but the N terminus Oocytes were held at Ϫ80 mV and pulsed to ϩ70 mV for 150 ms. Current amplitudes were measured at the end of the step. Values represent the means Ϯ S.E. * and ** indicate a significant difference from the value of current for Kv1.2 and Kv1.5, respectively (p Ͻ 0.05, one-way analysis of variance). The numbers above each bar represent the total number of oocytes in that group. The data presented in A and B are from three and one injection series, respectively. Kv1.2, with and without Kv␤ subunits, was injected for comparison in each series. apparently does not alter channel kinetics. Distinct Kv␤ isoforms with unique N-terminal domains might then have evolved to fine tune channel function. Together, these considerations suggest that the relative levels of distinct Kv␣1 and Kv␤ subunits within a cell will have profound consequences on the composition and properties of native K ϩ channel complexes.