A New Mode of Regulation of N-type Inactivation in a Caenorhabditis elegans Voltage-gated Potassium Channel*

N-type inactivation in voltage-gated K+ (Kv) channels is a widespread means to modulate neuronal excitability and signaling. Here we have shown a novel mechanism of N-type inactivation in a Caenorhabditis elegans Kv channel. The N-terminal sequence of KVS-1 contains a domain of 22 amino acids that resembles the inactivation ball in A-type channels, which is preceded by a domain of eighteen amino acids. Wild type KVS-1 currents can be described as A-type; however, their kinetics are significantly (∼5-fold) slower. When the putative inactivation ball is deleted, the current becomes non-inactivating. Inactivation is restored in non-inactivating channels by diffusion of the missing inactivation domain in the cytoplasm. Deletion of the domain in front of the ball speeds inactivation kinetics ∼5-fold. We conclude that KVS-1 is the first example of a novel type of Kv channel simultaneously possessing an N-inactivating ball preceded by an N inactivation regulatory domain (NIRD) that acts to slow down inactivation through steric mechanisms.

N-type inactivation in voltage-gated K ؉ (Kv) channels is a widespread means to modulate neuronal excitability and signaling. Here we have shown a novel mechanism of N-type inactivation in a Caenorhabditis elegans Kv channel. The N-terminal sequence of KVS-1 contains a domain of 22 amino acids that resembles the inactivation ball in A-type channels, which is preceded by a domain of eighteen amino acids. Wild type KVS-1 currents can be described as A-type; however, their kinetics are significantly (ϳ5-fold) slower. When the putative inactivation ball is deleted, the current becomes non-inactivating. Inactivation is restored in non-inactivating channels by diffusion of the missing inactivation domain in the cytoplasm. Deletion of the domain in front of the ball speeds inactivation kinetics ϳ5-fold.

We conclude that KVS-1 is the first example of a novel type of Kv channel simultaneously possessing an N-inactivating ball preceded by an N inactivation regulatory domain (NIRD) that acts to slow down inactivation through steric mechanisms.
A-type channels constitute an important group of voltagegated K ϩ (Kv) 2 channels that play a prominent role in the control of neuronal and muscular excitability, synaptic input, and neurotransmitter release (1)(2)(3)(4). The name "A-type" stems from the typical profile of these currents, rapid activation at subthreshold voltages followed by fast inactivation (4). In many neurons, A-type channels are usually silent at resting membrane potentials. They, however, transiently activate during the decay of the after-hyperpolarization phase of the action potential, delaying depolarization (5). Thus, A-type inactivation is a key physiological feature that allows control of the neuronal firing frequency by regulating the interval between consecutive action potentials (4). In the late 70s, Armstrong and Bezanilla (6), in an effort to explain the mechanism of A-type inactivation, proposed the "ball-and-chain" model, which postulates that a positively charged inactivation particle (the ␣-ball) on a tether prevents the movement of ions by physically occluding the pore. Using the Drosophila channel Shaker, Hoshi, Zagotta, and Aldrich (7,8) identified the location and molecular composition of the ball. This is composed of the first Ϯ20 amino acids in the N terminus (thus the name "N-type" inactivation) followed by 40 or more residues constituting the chain. The inactivation ball possesses two essential chemical characteristics ( Fig. 1): the first half is composed of hydrophobic residues, and the rest has a net positive charge that pushes the ball toward its channel receptor site upon depolarization (9,10). Because of the dynamic nature of N-type inactivation, the details of the mechanisms have been unknown for a long time. Recently, using crystallography, Mackinnon and colleagues (11,12) have proposed that N inactivation occurs as a sequential reaction; the ball initially binds to the T1 domain surface by electrostatic interactions, and then it enters through the lateral portals that connect the cytoplasm to the inner pore and reaches the inner cavity to eventually impair ion movement.
Accessory subunits of the Kv␤ family can convert non-inactivating Kv␣ subunits into A-type channels (13). These may also possess an inactivation ball, or ␤-ball, that functions identically to an ␣-ball. Thus, Kv␤1 and Kv␤3 interact with the Kv1 family of ␣-subunits and induce rapid inactivation in these channels (13)(14)(15)(16); however, there is one exception. Kv1.6 possesses an N-terminal domain termed the N inactivation prevention (NIP) domain, which neutralizes the ␤-ball of the Kv␤ subunit through electrostatic forces (17). A KChIP isoform (KChIP4a) also exhibits a K-channel suppressor (KIS) domain, which abolishes fast inactivation in the Kv4.3 channel (18). The KIS domain, however, profoundly differs from the NIP domain. Whereas NIP antagonizes the inactivation ball, KIS acts on residues around the inner mouth of the pore to alter the gating of Kv4.3. To date, no other cases of channels and/or accessory subunits possessing regulatory domains of N-type inactivation have been reported in the literature.
Here we have shown the first example of an ␣-subunit that simultaneously possesses an inactivation ball and a domain that regulates its function. KVS-1 is a voltage-gated K ϩ channel that operates in the nervous system of the nematode Caenorhabditis elegans to maintain and modulate cell sensitivity (19). Even though the KVS-1 current can be described as A-type, its inactivation is considerably slower than normal. We discovered that, in fact, KVS-1 possesses both a functional inactivation ball and a receptor. In addition, KVS-1 exhibits a domain in front of the ball (here named N inactivation regulatory domain (NIRD)) that modulates the kinetics of inactivation. NIRD is different from NIP and KIS in several respects. It co-exists with the ball in the same subunit and slows down inactivation by a mechanism that is best described as steric.
* This work was supported by National Institutes of Health Grant R01GM68581-01 (to F. S.). 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. 1 To whom correspondence should be addressed. Thus, KVS-1 provides the first example of a novel mechanism of regulation of N-type inactivation.

EXPERIMENTAL PROCEDURES
Molecular Biology-KVS-1 mutants were constructed by PCR. For biochemistry, KVS-1 was epitope-tagged by replacing the terminal stop codon with nucleotides encoding HA residues (YPYDVPDYA-STOP). The constructs were subcloned into pcI-neo vector (Promega) for expression in Chinese hamster ovary (CHO) cells. All sequences were confirmed by automated DNA sequencing. Transcripts were quantified with spectroscopy and compared with control samples separated by agarose gel electrophoresis stained with ethidium bromide.
Electrophysiology-CHO cells were transiently transfected with cDNA using the Superfect kit (Qiagen) and studied 24 -36 h post-transfection. Data were recorded with an Axopatch 200B amplifier (Axon), a Dell personal computer, and Clampex software (Axon) after cells were filtered at f c ϭ 1 kHz and sampled at 2.5 kHz. Bath solution was (in mM) 4 KCl, 100 NaCl, 10 Hepes (pH ϭ 7.5 with NaOH), 1.8 CaCl 2 and 1.0 MgCl 2 . The pipette solution included 100 KCl, 10 Hepes (pH ϭ 7.5 with KOH), 1.0 MgCl 2 , 1.0 CaCl 2 , and 10 EGTA (pH ϭ 7.5 with KOH). Whole-cell currents were evoked by 200-ms voltage sweeps from a holding potential of Ϫ80 to ϩ120 mV in 20-mV increments. The time course of activation and inactivation was fitted to a single exponential function as follows, where I 0 is a constant and is the value at which Equation 1 ϭ 0.37.
Macroscopic conductance (G) curves were calculated as, where I is the macroscopic current and V rev is the resting potential. G/G max curves were fitted to the Boltzmann function, where V is the membrane voltage, V 1/2 is the value of the voltage at which Equation 2 ϭ 0.5 and V S is the slope coefficient (in mV).
Data are presented as means Ϯ S.E. The number of determinations is indicated by n. Statistical significance of unpaired groups was estimated with the Student's t test.
Synthetic Peptide-Peptides encoded by the sequence MVIP-STRKGHAASRDNFGNSNH (⌬19 -40) and MSTERLMY-LDLNKGNQIAMVIPSTRKGHAASRDNFGNSNH (⌬1-40) were synthesized dry at the University of Medicine and Dentistry of New Jersey peptide synthesis core facility. Prior to the experiment, the peptides were freshly diluted in the pipette solution to a final concentration of 10 nM.
Immunoprecipitations-CHO cells were washed with 10 ml of ice-cold phosphate-buffered saline and lysed with ϳ2 ml of ice-cold radioimmune precipitation assay buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% IGEPAL CA-630, 0.5% (w/v) deoxycholate, and 0.1% (w/v) SDS) and freshly added 10 mM iodoacetamide, phosphatase, and protease inhibitors for 30 min at 4°C. Cell lysates were centrifuged for 60 min at 4°C, and the supernatant was mixed with HA-conjugated beads (Roche Applied Science) and rocked at 4°C for 3 h. The beads were washed three times with ice-cold TBST (8.8 g of NaCl, 0.2 g of KCl, 3 g of Tris base, 500 l of Tween 20, pH to 7.4, H 2 O to 1 liter) and incubated in SDS sample buffer at ϳ90 -95°C for 15 min.
Membrane Biotinylation-30 h after transfection, the CHO cells were washed three times with phosphate-buffered saline at room temperature (22-25°C), and cell surface proteins were biotinylated by 1.0 mg/ml impermeant biotin analog EZ-link sulfo-NHS-Lcbiotin (Pierce) in phosphate-buffered saline. After incubation at 4°C for 1 h, the cells were washed five times with ice-cold phosphate-buffered saline to remove any remaining biotinylation reagent. Cells were then harvested in radioimmune precipitation assay buffer. Lysate proteins were precipitated by streptoavidin-agarose-linked beads. The precipitated KVS-1 was detected by monoclonal anti-HA antibodies (Roche Applied Science).

RESULTS
A 22-amino-acid domain (⌬19 -40) ( Fig. 1) in the N terminus of KVS-1 (dubbed K1 in this "Result" section) exhibited homology to the N-inactivating domains in several A-type ␣and ␤-subunits (7,9,(35)(36)(37)(38). This suggests that inactivation in K1 might proceed through a N-type mechanism. To determine whether ⌬19 -40 is an inactivation ball, we constructed a panel of K1 truncation mutants and used the whole-cell configuration of the patch clamp to characterize currents.
KVS-1 Possesses an N Inactivation Ball-In CHO cells, wild type (wt) K1 channels express robust voltage-dependent K ϩ currents exhibiting fast activation and inactivation ( Fig. 2A). The voltage dependence of the extent of inactivation, expressed as the ratio between the residual and the peak current, is sharp (ϳ12.7 Ϯ 0.7 mV) with a midpoint at 45.5 Ϯ 7 mV (Fig. 2B). The time course of inactivation, which is best fit with a single exponential function (Equation 1), weakly depends on the membrane voltage (Fig. 2C). Thus, the characteristic of the K1 current is A-type overall. However, compared with normal fast inactivating channels, the kinetics of K1 inactivation are considerably slower (ϳ5-fold). Therefore, the time constant of inactivation represents the best parameter to describe changes in inactivation in wt or mutant K1 channels. Complete or par-FIGURE 1. The secondary structure of KVS-1 exhibits an N inactivation ball. Sequence alignment of the N terminus of KVS-1 with that of ␣-subunit Drosophila Shaker (red, identity 14%; green, strong similarity 27%; blue, weak similarity 18%), C. elegans Shaker-related SHL-1 (identity 9%; strong similarity 22%; weak similarity 17%), human Kv3.4 (identity 14%, strong similarity 14%; weak similarity 32%), human Kv1.4 (identity 13%; strong similarity 21%; weak similarity 21%), and with ␤-subunit Kv␤1.1 (identity 18%; strong similarity 14%; weak similarity 23%) was computed with ClustalW software. tial deletion of the putative inactivation domain ( Fig. 2A, K1⌬40) (K1⌬30 not shown) suppressed most of the inactivation. The current exhibited a residual time-dependent decline, suggesting that multiple inactivation mechanisms may co-exist in K1. The magnitude of the current in K1⌬40 and K1⌬30 (not shown) was roughly 2-fold larger than in wt channels. The level of K1⌬40 protein as well as its expression in the plasma membrane were not significantly different (Fig. 2F). This ruled out effects due to altered channel biosynthesis, trafficking, or abundance in the membrane. The midpoint for voltage activation (V 1/2 ) was shifted ϳ8 mV to the left in the K1⌬40 mutant compared with wt channels (Fig. 2D). In contrast, the kinetics of activation did not significantly differ in the two channel types (Fig. 2E). These observations are consistent with the notion that the current increase in K1⌬40 and K1⌬30 was because of the removal of N-type inactivation. Under normal conditions, N inactivation does not allow the current to reach saturation. Thus, an increase in the current upon removal of the ball was expected, and it was, in fact, observed in several A-type channels (e.g. see Ref. 20). Alternative mechanisms such as increased unitary current are also possible, but they seem unlikely consid-ering that the N terminus is not expected to contribute to the ion conduction pathway.
Synthetic ⌬19 -40 Peptide Restores N-type Inactivation in K1⌬40 Mutants-The ball-and-chain model makes predictions that can be tested: if deletion of the inactivating ball eliminates inactivation, then diffusion of the missing piece into the cytoplasm should restore inactivation. Therefore, to corroborate the notion that ⌬19 -40 is an inactivation ball and to ascertain whether K1 possesses a functional ball receptor, we synthesized a synthetic ⌬19 -40 peptide and assessed its ability to inactivate K1⌬40 mutant channels. Addition of the inactivation peptide to the pipette solution restored inactivation in K1⌬40 channels in a time-dependent fashion, consistent with progressive diffusion of the peptide into the cytoplasm (Fig. 3). As expected, we did not observe any change in the current in the absence of the peptide (Fig. 3B). Moreover, diffusion of a synthetic peptide formed by the first 40 amino acids of KVS-1 (⌬1-40) restored wild type inactivation in K1⌬40 currents ( 120 ϭ 51 Ϯ 7 ms) (Fig. 3B). We conclude that K1 possesses an ␣-inactivation ball and a functional receptor. Therefore, N-type inactivation is a primary (although not the only) mechanism contributing to the transient behavior of the K1 current.
NIRD Hinders N-type Inactivation-The kinetics of inactivation induced by the ⌬19 -40 peptide were considerably faster than in wt channels (compare them in Fig. 4B). This suggests that the domain in front of the ball might slow down N inactivation in wt channels. As expected, deletion of this domain (Fig.  4, K1⌬18) conferred full A-type attributes to the current. The inactivation in K1⌬18 currents did not significantly differ from that in K1⌬40 currents modified by the inactivating peptide ( 120 ϭ 11.2 Ϯ 3 versus 10.0 Ϯ 4 ms) (Fig. 4B). We conclude that ⌬18 specifically acts to modulate N-type inactivation without apparently affecting other gating parameters. The time course of activation was fitted to a single exponential function. Symbols are the same as described for A. F, total protein expression and surface expression of wt and K1⌬40 epitope tagged to the HA tag in the C terminus. Western blots were stained with an anti-HA primary antibody or with anti-tubulin for control and then visualized with a horseradish peroxidase chemiluminescence-coupled secondary antibody. Steric Mechanisms Underlie NIRD-mediated Inhibition of Inactivation-To gain insight into the mechanisms underlying NIRD function, we probed the role of electrostatic interactions, which have been shown to play a key role for NIP function (13). Surprisingly, mutation of charged residues to neutral residues or to residues of opposite charge did not induce any apparent modification in the current (Table 1). Post-translational modifications such as phosphorylation and oxidation of methionine have also been shown to modulate N-type inactivation (21)(22)(23)(24). Notably, neutralization of a protein kinase C consensus site in NIRD (T3V) or mutation of its methionines (M7L, M19L) gave rise to mutant channels that did not significantly differ from wt (Table 1). Conversely, progressive deletion of NIRD gradually accelerated inactivation in a fashion that did not depend on the overall electrical charge of NIRD (Fig. 5). We conclude that the NIRD acts primarily to slow down N-type inactivation through steric mechanisms.

DISCUSSION
Electrophysiological and biochemical analyses support the following conclusions. First, KVS-1 is an A-type channel. It has both a functional N-inactivating ball and a ball receptor. Second, KVS-1 possesses a domain in front of the ball (NIRD) that modulates the kinetics of N inactivation. The primary sequence of NIRD as well as its function lead us to conclude that NIRD represents a new mode of A-type current modulation.
To date, only two examples of N-terminal domains that can act to regulate N inactivation (NIP and KIS) have been reported. KIS affects multiple gating parameters in Kv4.3, including activation (18). In contrast, NIP specifically antagonizes N-type inactivation in heteromeric Kv1.6-Kv␤ channels (17). NIRD differs from NIP and KIS in several respects. First, the ball and NIRD co-exist in the same subunit. Second, NIRD does not abolish N inactivation nor does it affect other gating parameters. Its mode of action is also different from those of the other domains. Although NIP exploits electrostatic forces and KIS binds to the inner mouth of the pore, the mechanism by which NIRD hinders inactivation is best described as steric. The details of the mechanism are not completely elucidated. NIRD might either decrease the affinity of the ball for its receptor, or alternatively, it might hamper its progression into the pore of the channel. Nonetheless, in the light of the known three-dimensional structures of Kv channels, these findings unveil unexpected insights into the molecular mechanisms of N-type inactivation.
Suppression of N inactivation in KVS-1 revealed the existence of secondary inactivation mechanisms. This is not unexpected, as many Kv channels exhibit multiple forms of inactivation that involve slow and fast processes. For example, two forms of inactivation (N-type and C-type) co-exist in Shaker (25). C-type inactivation is produced by rearrangements of the external mouth of the pore that involve movements of S6 and the S5-S6 linker (26,27). The C terminus can also be a source of inactivation in Kv channels (28). The attributes of non N-type inactivation in KVS-1 are more consistent with a C-type mechanism than a C terminus mechanism, because this domain is remarkably small. Furthermore, accessory subunits might also contribute to secondary mechanisms of inactivation. In native FIGURE 4. NIRD slows down N-type inactivation. A, representative macroscopic K1⌬18 currents. Inset, wt and K1⌬18 single traces at ϩ120 mV. B, voltage dependence of the time constants of inactivation in K1⌬18 (triangles, n ϭ 14), in K1⌬40 in the presence of the synthetic ⌬19 -40 peptide (circles, n ϭ 5), and for comparison in wt channels (dotted line). C, recovery from inactivation in K1⌬18 channels was evaluated using 1.2-and 0.2-s depolarizing pulses at ϩ60 mV, spaced out by progressively longer periods at Ϫ80 mV (inset). The time course was best fit to a single exponential function with ϭ 9.1 Ϯ 1.2 ms for wt channels (n ϭ 7) and ϭ 9.7 Ϯ 2.6 ms (n ϭ 6) with K1⌬18 channels. D, macroscopic current densities calculated by normalizing the peak current at ϩ120 mV to the cell capacitance in cells transfected with K1 or K1⌬18. Averages were computed from groups of 10 -15 cells.

TABLE 1 Mutation of charged residues
Current density, at ϩ120 mV ( 120 ) and midpoint of the extent of inactivation for wild type and mutant KVS-1 channels were calculated as described under "Steric Mechanisms Underlie NIRD-mediated Inhibition of Inactivation" and fitted to Equation 1 ( 120   tissue, KVS-1 forms a complex with MPS-1, a bifunctional ␤-subunit that possesses kinase activity (29). When MPS-1 is co-expressed with KVS-1 in CHO cells, it accelerates the rate of inactivation through mechanisms that are independent of its enzymatic activity. This implies that MPS-1 might either interfere with N-type inactivation, with secondary mechanisms of inactivation, or induce de novo inactivation. Even though a complete elucidation of these issues is beyond the scope of this study, we can rule out that MPS-1 interferes with N-type inactivation. When MPS-1 was co-expressed with K1⌬40, the channel inactivated (data not shown). The most likely explanation for these effects is that they result from structural rearrangements in the channel protein induced by the incorporation of MPS-1. A-type inactivation is a key attribute of neuronal K ϩ channels. This explains why this gating characteristic can be controlled by multiple, often redundant mechanisms. It can be regulated through the formation of heteromeric complexes of distinct A-type ␣-subunits and/or ␣ and ␤-subunits by signaling molecules, including protein kinases and post-translational modifications (13-16, 21-24, 30 -33). An example has been reported in the literature of a Kv channel in which the ball can bind to phospholipids, thereby converting the channel from non-inactivating to inactivating and vice versa (34). We are currently testing whether in vivo deletion of NIRD is a physiological mechanism of regulation of neuronal excitability. Hence, exciting questions lay ahead that we expect will provide new insights into the molecular mechanisms determining K ϩ channel function.