Modes of regulation of shab K+ channel activity by the Kv8.1 subunit.

The Kv8.1 subunit is unable to generate K+ channel activity in Xenopus oocytes or in COSm6 cells. The Kv8.1 subunit expressed at high levels acts as a specific suppressor of the activity of Kv2 and Kv3 channels in Xenopus oocytes (Hugnot, J. P., Salinas, M., Lesage, F., Guillemare, E., Weille, J., Heurteaux, C., Mattéi, M. G., and Lazdunski, M. (1996) EMBO J. 15, 3322-3331). At lower levels, Kv8.1 associates with Kv2.1 and Kv2.2 to form hybrid Kv8.1/Kv2 channels, which have new biophysical properties and more particularly modified properties of the inactivation process as compared with homopolymers of Kv2.1 or Kv2.2 channels. The same effects have been seen by coexpressing the Kv8.1 subunit and the Kv2.2 subunit in COSm6 cells. In these cells, Kv8.1 expressed alone remains in intracellular compartments, but it can reach the plasma membrane when it associates with Kv2.2, and it then also forms new types of Kv8.1/Kv2.2 channels. Present results indicate that Kv8.1 when expressed at low concentrations acts as a modifier of Kv2.1 and Kv2.2 activity, while when expressed at high concentrations in oocytes it completely abolishes Kv2.1, Kv2.2, or Kv3.4 K+ channel activity. The S6 segment of Kv8.1 is atypical and contains the structural elements that modify inactivation of Kv2 channels.

Voltage-gated K ϩ channels (Kv) serve a wide range of functions including the regulation of cardiac pacemaking, action potentials, and neurotransmitter release in excitable tissues, as well as hormone secretion, cell proliferation, cell volume regulation, and lymphocyte differentiation in non-excitable tissues (2). The diversity of K ϩ channel functions is reflected by the diversity of K ϩ channel structures. Delayed rectifier K ϩ channels are constituted by 2 types of subunits called ␣ and ␤ (3). A large number of genes encode the different ␣ pore-forming subunits (18 genes cloned in mammals (4)), and an increasing number of ␤-auxiliary subunits are being discovered (5)(6)(7)(8). Different gene expressions in different cells or at different times of development associated with the possible formation of heteromultimeric channels containing different types of ␣and ␤-subunits probably allow individual cells to acquire their own characteristics of K ϩ current properties.
All subunits belonging to the Kv1 to Kv4 subfamilies have been functionally expressed in Xenopus oocytes. Elicited K ϩ currents display a large variety of electrophysiological characteristics reminiscent of the variety of K ϩ currents recorded in vivo. Particularly, large variations in inactivation characteristics have been observed ranging from inact ϳ30 ms for a very fast inactivating channel, such as (Kv1.4) (15), to inact ϳ15 s for a slow inactivating one, such as Kv2.1 (16). K ϩ channel inactivation occurs by at least two distinct mechanisms. The N-type inactivation is usually quite rapid with a time constant in the milliseconds range. It occurs by a "ball and chain" mechanism involving the NH 2 -terminal cytoplasmic domain, which acts as a tethered blocker that occludes the pore channel in its open state and causes inactivation (17)(18)(19). C-type inactivation is generally slower and appears to involve the COOH-terminal sequence of the ␣-subunits (20 -22).
Although the Kv5.1 and Kv6.1 proteins apparently have the structural hallmarks of functional K ϩ channel ␣-subunits, their expression in oocytes fails to induce K ϩ currents (13). The Kv8.1 subunit also displays the structural characteristics of a K ϩ channel subunit and is highly expressed in the brain. As Kv5.1 and Kv6.1, it does not lead to expression of K ϩ currents when produced in Xenopus oocytes. However, upon coexpression it is able to specifically abolish K ϩ currents generated by channels formed by Kv2.1, Kv2.2, and Kv3.4 subunits (1), and this inhibition is associated with the formation of multimers with these other subunits.
This paper extends previous work which was limited to Kv8.1 expression in Xenopus oocytes. It reports an analysis of Kv8.1 expression in the COS mammalian cell line. In this system, the Kv8.1 subunit is normally retained in cytoplasmic compartments. It requires coexpression with Kv2.2 to bring the subunit to the plasma membrane. Kv8.1 then induces no inhibition of the Kv2.2 current in this system but instead produces a drastic modification of the kinetic properties of Kv2.2 and particularly of its inactivation. The same effect can in fact be seen on Kv2. 1  mented with 10% fetal calf serum and antibiotics (60 g/ml penicillin, 50 g/ml streptomycin). One day before transfection, 10 5 cells were plated onto 35-mm Petri dishes for an electrophysiology experiment and 15 ϫ 10 3 cells onto 15-mm plates for indirect immunofluorescence microscopy. The cells were transfected by a modification of the DEAEdextran/chloroquine method (23) using 0.6 g of supercoiled DNA per cm 2 of cell culture. For electrophysiology studies, we co-transfected a plasmid encoding the CD8 receptor, which allows direct visualization of transfected cells by antibody-coated beads (24).
Electrophysiology on COS Cells-Voltage-clamp experiments were carried out using the whole-cell suction-pipette technique. The intracellular (pipette) solution contained 150 mM KCl, 1 mM MgCl 2 , 2 mM EGTA, 10 mM HEPES-KOH, pH 7.2. The extracellular solution was 140 mM NaCl, 5 mM KCl, 2 mM MgCl 2 , 2 mM CaCl 2 , 10 mM HEPES-NaOH, pH 7.3. Pipettes were coated with Sylgard resin to reduce their capacity. Electrical signals were digitized and stored on a hard disk by a personal computer for further analysis. Experiments were carried out at room temperature (24 Ϯ 2°C).
Indirect Immunofluorescence Microscopy-The cells that were grown on glass coverslips were fixed for 15 min with 4% (v/v) paraformaldehyde in phosphate-buffered saline (PBS). 1 After rinsing twice with PBS, cells were permeabilized by incubation for 10 min with 0.1% Triton X-100. The sites of nonspecific binding were blocked by 2 h of incubation with 5% goat serum, 2% bovine serum albumin in PBS at room temperature. The cells were then incubated for 2 h with a mixture of M2 monoclonal antibody (1:50 dilution, Eastman Kodak Co.) or polyclonal antibodies against the Kv1.5 subunit (1:100) 2 with a BS solution (2% bovine serum albumin in PBS), followed by washing with PBS and incubation for 1 h with fluorescein isothiocyanate-conjugated goat antimouse Ig (1:100, Sigma) or fluorescein (5-[4,6-dichlorotriazin-2-yl)amino]fluorescein)-conjugated F(abЈ) 2 fragment of a goat anti-rabbit Ig (1:100, Immunotech) in BS. After washing in PBS then in 1 mM Tris-HCl, pH 7.5, cells were mounted in Vectashield Medium (Vector Laboratories, Inc.) and observed with a Leitz Aristoplan microscope (Wild Leitz) using an interference blue (fluorescein isothiocyanate) filter and a 40ϫ or oil-immersion 100ϫ lens.
Plasmid Constructions and Mutagenesis-The Kv8.1 hamster cDNA was cloned into the pRc/CMV vector (Invitrogen) for expression in COS cells. To introduce mutations in this plasmid, a two-step polymerase chain reaction method using sense and antisense mutant primers was used (25). The dominant-negative mutant Kv8.1⌬, obtained by this technique, bears a deletion from residue 426 to the COOH-terminal end of the protein. After each mutation the Kv8.1 open reading frame was entirely resequenced (dye terminator kit, Applied Biosystems). The construction of NH 2 -terminal tagged Kv8.1 subunit was described previously (1). To create a COOH-terminal tagged Kv8.1 protein the same scheme was used, but the polymerase chain reaction fragment was cleaved with ClaI restriction enzyme and inserted in a HindIII-(filled with Klenow DNA polymerase I large fragment) and ClaI-digested Flag-pRc/CMV vector.
Expression of the Kv2.2 subunit in COS cells was realized with a pCDNAI vector. To create NH 2 -terminal tagged Kv2.2 protein, a polymerase chain reaction fragment (PWO Taq polymerase, Boehringer Mannheim) containing Kv2.2 open reading frame and a 3Ј-ApaI restriction site was cloned in a HpaI-and ApaI-digested Flag-pRc/CMV vector.
Kv2.1 subunit mutations were generated by oligonucleotide-directed mutagenesis on a single-stranded template (26) derived from Bluescript II vector containing Kv2.1 cDNA (27). All mutational changes were checked by sequence determination.
cRNA Synthesis, Injection, and Electrophysiological Measurement in Xenopus Oocytes-Preparation of oocytes, cRNA synthesis, injection (50 nl), and electrophysiological measurements have been previously described (1,28). We have previously reported that in oocytes as well as in cell lines high magnitude currents could show pronounced modifications when compared with currents of lower intensity (28,29). Therefore, in all this work, attention was paid to always compare currents of similar and relatively low intensity.

Modification of Kv8.1 Subcellular Localization upon Kv2.2
Coexpression in COS Cells-Despite numerous attempts to express the Kv8.1 subunit in various expression systems including Xenopus oocytes and Chinese hamster ovary and COS cell lines, K ϩ currents could never be recorded (1). One possible reason for this lack of success is an inadequate cellular localization of the expressed Kv8.1 protein. To check that point, an 8-amino acid epitope, which can be recognized by a monoclonal antibody, was added to the Kv8.1 protein, and protein localization was examined by indirect immunofluorescence on Tritonpermeabilized COS cells (Fig. 1). Transfected cells expressing the Kv8.1 protein tagged on the NH 2 -terminal or on the COOHterminal end show a very strong fluorescent staining localized at the perinuclear region as well as in a fine reticular network extending through the cytoplasm (Fig. 1, A, B, and C). In the many experiments that have been done, a surface labeling of Kv8.1 was never observed. Confocal microscopy experiments also failed to detect the Kv8.1 protein at the plasma membrane (data not shown).
To verify that the COS cells were able to correctly express well known functional K ϩ channel ␣-subunits and integrate them at the plasma membrane, cells were also transfected with a Kv1.5-expressing vector (30). Protein detection in this case was made with polyclonal antibodies specific for Kv1.5. As shown in Fig. 1D, Kv1.5-expressing cells, which produce K ϩ currents with the expected biophysical properties (30), show a typical surface labeling unlike Kv8.1-expressing cells. 1 The abbreviation used is: PBS, phosphate-buffered saline. 2 F. Lesage, unpublished data. An NH 2 -terminal tagged Kv2.2 ␣-subunit was also prepared. Electrophysiological recording of cells transfected with this subunit show typical delayed outward rectifier K ϩ currents (not shown), and as expected, immunodetection indicates a clear surface membrane localization (Fig. 1E) in addition to a perinuclear accumulation.
As the Kv8.1 subunit is able to interact physically with the Kv2.2 protein (1), the subcellular localization of Kv8.1 was examined in cells transfected with a mixture (3:1 molecular ratio) of Kv2.2 and tagged Kv8.1-expressing plasmids. Immunofluorescent detection then revealed a different Kv8.1 localization pattern. A minority of cells showed a fluorescence distri-bution resembling that obtained with Kv8.1 alone, i.e. an apparent lack of surface membrane localization. In contrast, the majority of cells displayed a fluorescent staining resembling that observed for the Kv2.2 localization, i.e. a perimembranous staining combined with a cytoplasmic retention ( Fig.  1, F and G). The conclusion is that the Kv8.1 subunit is transported to the plasma membrane when it is coexpressed with the functional Kv2.2 subunit.
The Kv8.1 Subunit Modifies Kv2.2 Current Characteristics-The effect of Kv8.1/Kv2.2 coexpression was then explored electrophysiologically. Expression of the Kv2.2 subunit alone resulted in a typical delayed outward rectifier K ϩ current with a slow inactivation process (Fig. 2B). Expression of Kv8.1 alone did not give rise to a detectable K ϩ current (not shown). Cells were then transfected with a mixture of Kv2.2 and Kv8.1expressing vectors using 1:5 to 1:50 molecular ratios. Increasing amounts of Kv8.1 decreased the intensity of the Kv2.2 current ( Fig. 2A). However, this decrease was not specific because co-transfection of Kv2.2-expressing cells with increasing amounts of ␤-galactosidase produced the same effect ( Fig. 2A). This lack of specificity was confirmed by a coexpression of Kv2.2 with a non-functional deleted form of the Kv1.5 subunit (Kv1.5⌬) (30) which, in addition to being non-functional, cannot associate with Kv2.2 since Kv1 and Kv2 subunits cannot form Voltage-dependent currents, evoked by a step from Ϫ60 mV to ϩ30 mV, were recorded. A, intensities of the Kv2.2 current in the presence of increasing amounts of Kv8.1-containing plasmids, while maintaining the Kv2.1 plasmid amount constant, reduces the average peak current. Co-transfection of Kv2.2 with other plasmids containing ␤-gal or Kv1.5⌬ also results in a decrease in peak current. Kv8.1⌬ abolishes the Kv2.2 current. B, examples of K ϩ currents recorded at ϩ30 mV. Note the slowly activating component and reduced current inactivation when Kv2.2 is coexpressed with Kv8.1 (5:1 molecular ratio of Kv8.1/Kv2.2 ). In this representation currents were scaled so that their peak currents are at the same level. C, voltage dependence of the steady-state activation. Steady-state inactivation was measured by recording the peak outward current (ordinate) in response to voltage steps from a holding potential of Ϫ60 mV to varying potentials (abscissa). D, the activation phases of K ϩ currents were fitted with Ae * exp(t/ e ) ϩ Ah * (1 Ϫ exp(t/ h ) 4 . Co-transfection of Kv2.2 with Kv8.1 introduces a slow exponential component Ae. The ratio of Ae over Ah is shown as a function of membrane potential. E, co-transfection of Kv2.2 with Kv8.1 shifts the steady-state inactivation curve by 11 mV to the hyperpolarizing direction. Steady-state inactivation was measured by recording the peak current (ordinate) in response to a step from a 30-s conditioning prepulse and varying potential (abscissa) to ϩ30 mV. F, an estimation of inactivation kinetics was made by taking the ratio of the current inactivated during a 10-s pulse to ϩ30 mV over the peak current. Co-transfection of Kv2.2 with Kv8.1 reduces the inactivation ratio from 42 Ϯ 5% to 16 Ϯ 2% (p Ͻ 0.01%). heteropolymers (9,31,32). Again the same decrease of Kv2.2 current was observed with a Kv2.2/Kv1.5⌬ mixture in a 1:50 ratio ( Fig. 2A). To verify that Kv2.2 and Kv8.1 subunits interact in COS cells we created a truncated form of the Kv8.1 subunit, Kv8.1⌬, in which half of the S6 domain was deleted. This modified subunit, which conserves the NH 2 -terminal part that is required for interaction with the Kv2.2 subunit (1) but lacks a part of the S6 domain that is known to be essential for K ϩ channel expression (30), acts as a potent dominant-negative subunit since cells cotransfected with a mixture of the Kv2.2 and Kv8.1⌬ plasmids (1:5 ratio) showed no detectable K ϩ current ( Fig. 2A).
When compared with control cells expressing Kv2.2 alone, COS cells in which Kv2.2 and Kv8.1 are coexpressed display new electrophysiological properties. A typical Kv2.2/Kv8.1 current trace is shown in Fig. 2B. The steady-state activation curve of the Kv8.1/Kv2.2 current has a midpoint potential of 13.1 mV and a slope parameter, k, of 17.1 mV. These values are similar to those of the Kv2.2 current (midpoint, 7.1 mV; slope, 17.9 mV) (Fig. 2C). However, activation kinetics are clearly modified. The time course of the delayed outward Kv2.2 current can be fitted with the Hodgkin-Huxley model (33) (Fig.  2D). The Kv2.2 currents modified by the coexpression with Kv8.1 cannot be adequately fitted with the same equation, and an additional exponential term is required. The ratio of the amplitude of this exponential component compared with the Hodgkin-Huxley component is shown as a function of the depolarization potential in Fig. 2D. Coexpression of Kv1.5⌬ instead of Kv8.1 with Kv2.2 had no effect on the kinetics of the Kv2.2 current (not shown).
The inactivation of the K ϩ current expressed by Kv2.2 is drastically changed by the coexpression of Kv8.1 (Fig. 2, E and  F). The midpoint potential for Kv2.2 inactivation (E 0.5(inact) ) is Ϫ37.1 mV, the slope parameter, k, is 13.0 mV, and 43% of the current is inactivated during a 10-s pulse to ϩ30 mV. The inactivation of the Kv8.1/Kv2.2 current is less complete (16%) after the same pulse and displays a different voltage sensitivity (E 0.5(inact) at Ϫ48.1 mV) together with a slightly less steep voltage dependence (k ϭ 9.3 mV). Fig. 2F shows that these changes of inactivation are specific to the expression of Kv8.1 and are not observed upon coexpression with ␤-galactosidase or Kv1.5⌬.
Kv8  Fig. 3, B, C, and D. Their biophysical characteristics are given in Table I (Table I).
The Modification of the Inactivation Process of Kv2 Currents Is Mediated by the S6 Segment of Kv8.1-The slow inactivation process of other types of K ϩ channels appears to involve a constriction of the extracellular mouth of the channel (20 -22, 34), and critical amino acids controlling this type of inactivation have been localized in the pore domain and in the S6 domain. A first experiment to test the possible implication of the S6 segment of Kv8.1 in the modification of the inactivation process of the Kv2 current was made by coexpressing the Kv2.1 subunit in oocytes with a chimeric protein (designated as (NtH5Kv8)-Kv1), which contains the Kv8.1 sequence from the NH 2 terminus to the end of the pore domain followed by the S6 and COOH-terminal regions of the Kv1.3 subunit. This chimera which is inactive by itself is known to completely inhibit Kv2.1 current (1). A 50% inhibition of Kv2.1 currents occurred

subunit in Xenopus oocyte
Molecular ratios of injected plasmids are those indicated in Fig. 3. For Kv2.1/(NtH5Kv8)-Kv1.3 coinjection with an 0.1 ratio was used, corresponding to a 50% inhibition of Kv2.1 current. Activation parameters were elicited by 500-ms depolarizing pulses from Ϫ80 mV to ϩ60 mV in 10-mV increments with a holding potential at Ϫ80 mV. The current (I) was recorded at the end of the pulse and plotted against membrane potential (V m ). The membrane conductance (G) was calculated for a given command voltage (V m ) and peak current responses (I peak ) from the expression G ϭ I peak /(V m Ϫ E K ), where E K is the K ϩ reversal potential. G was normalized versus a maximum (G max ) and plotted against V m . These plots were fitted to a Boltzmann distribution of the form, G/G max ϭ 1/1 ϩ exp {(E 0.5(act) Ϫ V m )/k act }. E 0.5(act) is the midpoint of activation, and k act is the slope factor. act corresponds to the time constant of activation for a monoexponential fitting of the ϩ40-mV activation curve. Inactivation parameters were determined by 500-ms step depolarizations at ϩ30 mV preceded by 10-s prepulses from Ϫ120 mV to ϩ30 mV. Current (I) was recorded at the end of 500-ms pulses, normalized by a maximum (I max ), and plotted against prepulse potential (V m ). This representation is the inactivation curve that can be fitted with a form of the Boltzmann equation, I/I max ϭ 1/1 ϩ exp {(E 0.5(inact) Ϫ V m )/k inact }. E 0.5(inact) is the midpoint of inactivation, and k inact is the slope factor. The percent of inactivated current was calculated from the ratio of measured currents at 10 s and at 0.1 s from pulses at ϩ30 mV. Values are the mean Ϯ S.E. for five oocytes measured.  Fig. 4A. Kv8.1 is atypical in 4 positions in this segment: (i) an alanine is localized at position 412 instead of a glycine in all other functional Kv subunits (M1, Fig. 4A); (ii) an isoleucine at 420 replaces a conserved valine residue (M2); (iii) an alanine is found at 428 instead of a proline (M3) in other functional Kv subunits; (iv) at position 433, an arginine is found instead of a very conserved asparagine (M4).
Current traces recorded after expression of mutated Kv2.1 proteins in Xenopus oocytes are shown in Fig. 4B. Corresponding electrophysiological parameters are reported in Table II. All the mutations significantly alter current properties. Channel opening is most dramatically affected by mutation Kv2.1M3 (Pro replaced by Ala). This mutation produces a 30-fold decrease of the rate of activation ( act 501.5 ms versus 17.7 ms for wild type Kv2.1) together with a ϩ17.6 mV shift of the activation midpoint (ϩ5.4 mV versus Ϫ10.5 mV for wild type). Other mutations induce smaller modifications of the activation process ranging from a 1.6-fold acceleration for Kv2.1M2 ( act 10.9 ms) to a 1.5-fold decrease for Kv2.1M1 ( act 26.3 ms). Changes in activation midpoints ranged from a Ϫ13.7 mV shift for Kv2.1M7 (E 0.5(act) ϭ Ϫ4.7 mV) to a ϩ17.6 mV change for Kv2.1M3 (E 0.5(act) ϭ ϩ26.8 mV).
Except for the Kv2.1M5 mutation, all mutations produced a reduction of the inactivation process of the Kv2.1 channel. The most spectacular modifications are observed for K ϩ currents produced by Kv2.1M3 (P406A) and Kv2.1M7 (C394I) mutants that inactivate only by 2-10% during a 10-s depolarizing pulse to ϩ30 mV, while the control Kv2.1 current inactivates by 67% in the same conditions (Fig. 4B). Both mutants also show a considerable shift of the inactivation midpoint toward positive (Kv2.1M3, ϩ56 mV shift) or negative potentials (Kv2.1M7, Ϫ42 mV shift).
Kv2.2) and Kv3 (Kv3.4) subunits, and immunoprecipitation studies have shown that this is due to the formation of Kv8.1/ Kv2 heteropolymers. This paper first analyzes the properties of expression of Kv8.1 in the mammalian COS cell line. Again, no K ϩ current could be recorded from these cells when they expressed only Kv8.1. One important reason for this lack of expression is an inadequate subcellular localization of Kv8.1. Identification of the Kv8.1 protein with antibodies indicates that it remains in the endoplasmic reticulum and is apparently unable to reach the surface membrane (Fig. 1, A, B, and C). This is unlike other Kv subunits that generate K ϩ currents in these COS cells such as Kv1.5 and Kv2.2 which, as expected, are found to be localized at the plasma membrane (Fig. 1, D and E). Intracellular accumulation of Kv8.1 corresponds to an intrinsic property of this particular Kv channel subunit. However, Kv8.1 is transferred to the plasma membrane, if it is coexpressed with Kv2.2 ( Fig. 1, F and G), a subunit which interacts with Kv8.1 (1), and which, when expressed alone, forms functional K ϩ channels at the surface membrane. Similar situations where a channel protein is not able by itself to reach the plasma membrane and requires the association with other subunits for proper movement to the surface are not unusual. One similar example concerns subunits of the amiloride-sensitive Na ϩ channel (35). Since no known endoplasmic reticulum localization sequence is found in the Kv8.1 subunit, it is probable that endoplasmic reticulum retention arises from protein misfolding and/or inefficient assembly. Association with the Kv2.2 subunit, which has access by itself to the plasma membrane, corrects the defect and facilitates Kv8.1 transport. The Kv2.2/Kv8.1 complex, once it has reached the surface membrane, forms a functional K ϩ channel with activation and inactivation properties that are different from those of the Kv2.2 channel.
Coexpression of Kv8.1 and Kv2.2 proteins in COS cells does not lead to a specific K ϩ current reduction as compared with expression of Kv2.2 alone. Only large Kv8.1/Kv2.2 plasmid ratios (50:1) lead to smaller K ϩ currents. However, the latter effects are not specific since they are also observed in Kv2.2expressing cells transfected with a high ␤-galactosidase/Kv2.2. The reduction of the K ϩ current magnitude is then due to a decrease of Kv2.2 expression due to competition between the two plasmids at the transfection, transcription, and/or translation levels. Conversely, in Xenopus oocytes, dramatic K ϩ current reductions eventually leading to complete inhibition of Kv2.   TABLE I. Activation parameters were elicited by 500-ms pulses (3 s for M3) from Ϫ80 mV to ϩ60 mV in 10-mV increments with a holding potential at Ϫ80 mV (Ϫ120 mV for M1 and M7). E 0.5(act) is the midpoint of activation, and k act is the slope factor. The time constants for activation ( act ) are determined at ϩ40 mV. Inactivation parameters were determined by 500-ms step depolarizations from Ϫ80 mV to ϩ30 mV preceded by 20-s prepulses from Ϫ80 mV (Ϫ120 mV for M1 and M7) to ϩ30 mV (ϩ70 mV for M3). E 0.5(inact) ) is the midpoint of inactivation, and k inact is the slope factor. The percent of inactivated current has been calculated for a pulse at ϩ30 mV at 10 s after the peak current. Values are the mean Ϯ S.D. for 6 to 11 oocytes measured. The inactivation process of voltage-gated K ϩ channels is important for defining the shape and for the integration of electrical signals and can occur over a wide range of time scales. It is probable that in the regions of the mammalian brain that express Kv8.1 and Kv2 subunits at high levels (40), an interaction of Kv8.1 with Kv2 subunits will result in the generation of K ϩ currents similar to those recorded in both the Xenopus oocyte and COS cells. In fact, a current (I K(slow) ) with properties similar to those of the Kv8.1/Kv2 assembly in COS cells has been described recently in CA3 pyramidal cells of rat hippocampus, where its role would be to control discharge onset after a period of membrane hyperpolarization (41).
The interesting question is whether one can observe in the functioning nervous system both the inactivation effect of Kv8.1, which occurs in COS cells and in Xenopus oocytes, and the total suppression of the activity of both Kv2 (Kv2.1 and Kv2.2) and Kv3 channels observed in Xenopus oocytes expressing large amounts of Kv8.1. If it were the case, adequate modulations of Kv8.1 expressions in the nervous system could produce drastic changes in the electrophysiological identity of neurons, i.e. slowing K ϩ channel inactivation in some and totally inhibiting K ϩ channels in others. One can even suspect that the two types of modulations could take place in a single neuron in different localizations (soma, dendrites, synapses). The effects described in this paper might be associated with drastic long term modifications of K ϩ channel expression, which would lead to drastic changes in the function of the neurons carrying the Kv8.1 channels together with its partners of the Kv2 and Kv3 families. Since the Kv8.1 subunit is abundantly expressed in the hippocampus, its capacity to produce long term changes in electrical signals might have a role in long term potentiation and memory processes.