INTRODUCTION

Potassium (K⁺) channels are responsible for numerous important neuronal processes that include setting the resting membrane potential, controlling electrical excitability, and shaping action potential duration and waveform (1, 2, 3). Molecular cloning approaches have led to the identification of a large family of distinct voltage-gated K⁺ channels that can be divided into four subfamilies, Shaker (Kv1), Shab (Kv2), Shaw (Kv3), and Shal (Kv4). Mutagenesis studies have revealed structural determinants within the core K⁺ channel protein that underlie many of its functional properties (4, 5, 6, 7, 8). For example, the S4 transmembrane domain, which contains evenly spaced positively charged arginine and lysine residues, has been implicated as a voltage sensor region in both Na⁺ and K⁺ channels (9, 10, 11, 12, 13). Other regions of the core K⁺ channel protein such as the S4-S5 cytoplasmic linker are also involved in voltage-dependent activation because mutations in this region shift the voltage required for half-maximal activation dramatically (14, 15).

In contrast with the functional motifs uncovered within the core protein, much less is known about the role of posttranslational modifications in determining K⁺ channel function. Both voltage-gated Na⁺ and some K⁺ channels are glycoproteins that in comparison with other membrane glycoproteins contain unusually large amounts of posttranslationally attached sialic acids. Compositional analysis of purified Na⁺ channels (16, 17, 18, 19) and sialidase sensitivity of some Shaker-type K⁺ channels from brain (20, 21, 22) have suggested that they contain 120-220 negatively charged sialic acids per functional molecule. Glycosylation is thought to be important for membrane protein folding and cell surface targeting and in some cases function (23). For example, sialidase treatment of Na⁺ channels shifted its steady state activation curve to more positive voltages when they were analyzed in planar lipid bilayers (24). This observation suggested that the negatively charged channel sialic acids play a role in voltage gating by influencing the local electric field detected by the voltage sensor on the molecule. In support of such a mechanism, earlier studies suggested that fixed cell surface negative charges can be screened by raising extracellular divalent cations with the result that activation curves for ion channels are shifted to positive voltages (25, 26).

The aim of this report was to examine the role of glycosylation in affecting the activation properties of the rat brain Kv1.1 (RBK1 or RCK1) channel. Kv1.1 is a delayed rectifier K⁺ channel that was originally cloned from rodent brain cDNA libraries (27, 28, 29, 30) and is a heavily sialidated protein (20). We investigated the possible role of sialic acids by comparing the functional properties of Kv1.1 in control and glycosylation-deficient CHO¹ cell lines (Lec mutants). Evidence is presented that suggests reducing the sialic acid content of Kv1.1 channels has significant functional effects on its activation parameters but not on its cell surface expression.

EXPERIMENTAL PROCEDURES

CHO cell lines were obtained from American Type Culture Collection (Rockville, MD) and maintained in -modified minimum essential medium with 10% fetal bovine serum at 37 °C under 5% CO₂. The Kv1.1 cDNA was subcloned into the eucaryotic expression vector pZEM222 and transfected into CHO cells by the calcium phosphate method as extensively described (31). Twelve neomycin-resistant (800 µg/ml G418) colonies were selected from each cell line and maintained in medium containing 200 µg/ml G418 (31). Two selected clones from each cell line expressing 1-3nA of current at 60 mV were used in this study. For sialidase treatment, cells were grown on coverslips for 2-3 days and treated with sialidase at 2 units/ml or 5 units/ml at 37 °C for 2 h. The medium was then removed, and fresh medium was added. Affinity-purified exosialidase (Clostridium perfringens) was purchased from Sigma (type X).

CHO cell lines were plated on glass coverslips and examined for the presence of voltage-gated K⁺ currents 2-3 days later using standard whole cell recording methods and voltage pulse protocols (32). Cells were viewed using Nomarski optics (500× magnification) at room temperature (21 °C) in our standard recoding solution containing (in mM) NaCl, 145.0; KCl, 5.3; MgSO₄, 1.0; CaCl₂, 5.4; glucose, 5.6; HEPES, 5.0 (pH 7.3). For the high Ca²⁺ recording solution, an osmotically equivalent amount of NaCl was decreased, whereas in the low Ca²⁺ solution NaCl was increased. Patch pipettes were pulled from glass hematocrit capillary tubing (VWR) and had tip resistances of 1.5-2.0 M when filled with an intracellular solution containing (in mM) KCl, 145; CaCl₂, 1.2; MgCl₂, 2.0; EGTA, 2.0; ATP, 0.3; glucose, 15.4; Na-HEPES, 5 (pH 7.3).

Whole-cell membrane currents were recorded from CHO cells with an Axopatch-1 patch clamp (Axon Instruments, Burlingame, CA) and digitized using a Cheshire data interface (Indec Systems, Inc., Sunnyvale, CA) and a computer equipped with an LSI-11/73 processsor (Digital Equipment Corp., Maynard, MA). Original data acquisition and analysis programs were written by Dr. V. Dionne and modified when required. After establishing the whole-cell configuration, cells were held at 70 mV, and mean linear membrane leak current was assessed by averaging current responses to 10-mV hyperpolarizations. Membrane capacitance was measured from the ratio of the resulting capacitative charge displacement transient and the net applied voltage and compensated using the patch clamp amplifier controls. Series resistance (R_s) was simultaneously estimated from the ratio of the voltage step and the resulting peak capacitative current. R_s values were 3.0-4.0 M and were not compensated by patch clamp circuity. The voltage-induced currents were filtered at 5 kHz (3 db cut-off) and digitized at 100 µs. Linear leak currents were scaled and digitally subtracted from records. Each test voltage value was corrected for the V introduced by R_s (V = I R_s; I is the net voltage-gated current). It should be noted that all cell lines had similar peak K⁺ currents and R_s values. Thus any possible errors in voltage values due to R_s would have little effect on the changes in V1/2 among them.

To assay for voltage-gated currents, a family of test depolarizations were applied from 40 to 60 mV in 7-mV increments to record sustained K⁺ currents. Sustained conductance values (G) were obtained from the mean value of the leak-subtracted sustained current (I) at the end of each corrected test pulse potential (V_p) using Ohm's law (G = I/(V_p  EK)) and a predicted Nernst K⁺ equilibrium potential (EK) of 84 mV. The sustained conductance values were then plotted as a function of test potential, and the points fit to a single Boltzmann equation G = G_m/{1 + exp[(V_p  V1/2)/a]} using an automated nonlinear least squares curve fitting routine (Levenberg-Marquardt method) included in a Macintosh-based data analysis program (KaleidaGraph, Synergy Software, Reading, PA). In this equation, G_m is the maximum sustained conductance, V1/2 is the test potential where G/G_m = 0.5 and a represents the slope of the voltage dependence of activation (given by a = kT/ze; with T being the absolute temperature, k being Boltzmann's constant, e being the electron charge, and z being the valence of the gating charge). A Boltzmann raised to the 4th power (33) gave similar activation parameter values but a somewhat better fit to the points. Predicted estimates for standard errors of the activation parameters, V1/2 and a, obtained from the fits were <5%, and correlation coefficients were >0.95.

Analysis of variance was used to assess statistical differences of a control value compared with other test values in a group. A p < 0.05 was considered significant.

Antibodies to a synthetic peptide to the N terminus of Kv1.1 (amino acids 5-27) were raised in rabbits and characterized for their specificity as described (34, 35). CHO cell lines were grown until confluent on 100-mm plates. Total membranes were prepared by homogenizing the cells in 50 mM K⁺ phosphate buffer (pH 7.4) containing EDTA (2 mM), pepstatin A (1 mm), 1,10-phenanthroline (1 mM), phenylmethylsulfonyl fluoride (0.2 mM), and iodoacetamide (1 mM). The homogenate was centrifuged at 35,000 × g for 1 h to pellet membranes and stored at 85 °C. 25% of membranes collected from a 100-mm plate were used for each lane on a denaturing SDS gel. The proteins were electrotransferred to nitrocellulose (Bio-Rad), and the filter was blocked with 8% nonfat milk and then incubated in affinity-purified Kv1.1 antibody at 1:2000 overnight. After extensive washing, horseradish peroxidase-linked anti-rabbit secondary antibodies were added, and the bound antibodies were detected using chemiluminscence (Amersham Corp., ECL detection kit) and Kodak XAR 5 film.

RESULTS

Kv1.1 cDNA was stably transfected into two control cell lines (K1, parental of Pro5 and Pro5, parental of Lec mutants) and two mutant lines (Lec8 and Lec2) (36). The Lec mutants were selected for their ability to grow in lectin supplemented medium, and they have been shown to produce truncated cell surface glycoproteins (37): Lec8 (glycoproteins with reduced galactose content, due to decreased UDP-galactose transport into the Golgi (38)) and Lec2 (glycoproteins with reduced sialic acid content, due to decreased CMP-sialic acid transport into the Golgi (39)). Evidence suggests that galactose addition is required before terminal sialic acids are attached to N- or O-linked glycoproteins in the trans Golgi (23). The general glyco-structures on proteins are: N-linked, AsparagineGlcNAc₂-Mannose₃-GlcNAc₂-Galactose₃-Sialic acid₃; O-linked, Serine/Threonine-GalNAc_1-3-Galactose_1-3-Sialic acids_1-3. Therefore both Lec mutants would be predicted to produce truncated Kv1.1 glycoproteins having a reduced sialic acid content but not necessarily a complete reduction.

Whole-cell recording techniques were used to characterize the K⁺ currents expressed by the CHO cells (the standard recording solution contained 5.4 mM Ca²⁺). Kv1.1 transfected CHO cells exhibited large outward currents of 1-3 nA that showed a voltage dependence of activation and noninactivating characteristics of delayed rectifier K⁺ channels (Fig. 1). Also consistent with the expression of K⁺ channels, the outward currents in control transfected CHO cells were inhibited by classic K⁺ channel blockers such as tetraethylammonium (IC₅₀ = 0.5 mM, n = 5) and dendrotoxin (IC₅₀ = 5 nM, n = 5). In contrast with the large K⁺ currents recorded from transfected cells, identical depolarizations of nontransfected Pro5 and Lec mutants produced either no net currents (70% of cells) or small inward, and more rarely outward, currents that did not exceed 80 pA (30% of cells) (n = 10-15 for each cell line). Nontransfected K1 cells expressed similar peak currents but in 80% of cells.

In order to examine the voltage dependence for activating Kv1.1 channels expressed in CHO cells, fractional conductance was plotted as a function of the step depolarization and fit to the Boltzmann equation. The voltage at half-maximal activation (V1/2) and the slope for the voltage dependence of activation (a) were compared among the cell lines. Significant positive shifts in V1/2 values were recorded for Kv1.1 channels expressed in both Lec mutants (shifted 16 mV for Lec8 and 8 mV for Lec2) compared with Pro5 (Fig. 2 and Table I). The slope of the voltage dependence of activation was slightly decreased in Lec8 but not significantly different in Lec2 versus control (Fig. 2 and Table I). Kv1.1 activation kinetics were slower for both Lec mutants as expected from a positive shift in their V1/2 values (Fig. 1). Indeed, the activation rise time (10-90% value) (30) was significantly slower in both Lec mutants compared with control (Table I). Furthermore these activation rise time differences do not take into account the additional longer delay prior to onset of the current following the voltage pulses that is seen in both Lec mutants, particularly Lec8 (Fig. 1). All activation parameters obtained for Lec8 were significantly different compared with Lec2 (at least p < 0.05). In contrast to the changes in gating parameters, the maximal Kv1.1 whole-cell conductance relative to cell membrane capacitance did not differ significantly in the Lec mutants versus control (Table I). These findings indicate that the cell surface density of Kv1.1 channels in the Lec mutants was similar to that of control if the single channel conductance is similar but that a greater depolarization was required for activation.

Table I. Electrophysiological parameters of Kv1.1/CHO cells Whole-cell recording techniques were used to obtain sustained K+ currents from the Kv1.1/CHO cells. V1/2, a, and Gm were obtained as described under ``Experimental Procedures.'' Each value is a mean ± S.E. The asterisks indicate significant differences from Pro5 control using the analysis of variance test (**, p < 0.01; ***, p < 0.005). Cell line (n) V1/2a ab Rise timec Gm/Cmd mV ms pS/pF K1 (20)  10.9  ± 0.7  9.2  ± 0.4 6.7  ± 0.5 1465  ± 238 Pro5 (22)  10.5  ± 1.0  10.0  ± 0.3 7.3  ± 0.3 1266  ± 153 Lec8 (14) 5.4  ± 1.9***  12.4  ± 0.3*** 10.5  ± 0.3*** 1511  ± 521 Lec2 (20)  2.2  ± 1.3***  11.0  ± 0.3 9.0  ± 0.3** 901  ± 122 a  V1/2 is the voltage in mV at which the conductance is one-half of its maximal value. b  The a value is the slope of the normalized conductance versus voltage plot and corresponds to the change in voltage to cause an e-fold increase in conductance. c  The activation rise time is the time in milliseconds required for the whole-cell current to rise from 10 to 90% of its final value following a depolarization from 70 to 10 mV (29). d  Maximal conductance (Gm) is expressed relative to the membrane capacitance (Cm).

Treatment of control cells with exosialidase, which cleaves sialic acids sequentially, would be predicted to shift V1/2 to positive voltages if sialic acids are responsible for the shift in the Lec mutants. Indeed, sialidase treatment (5 units/ml) had a significant effect on Pro5 and shifted V1/2 about 7 mV compared with control (Fig. 3), whereas there was no significant change in slope factor (10.5 ± 0.7 mV). Sialidase-treated Pro5 also exhibited a slower activation rise time (8.8 ± 0.4 ms) versus control (7.1 ± 0.5 ms, p < 0.01). In contrast, sialidase treatment of Lec8 had no significant effect (Fig. 3). These results suggest that sialic acids on control cells, presumably on Kv1.1 channels, are sensitive to sialidase treatment and that Lec8 appears to be devoid of these carbohydrates.

It has been long known that changing extracellular ionized Ca²⁺ from its physiological value of 1.5 mM in mammals results in altered electrical excitability of nerve and muscle tissue (3, 25). In high Ca²⁺ the V1/2 for Na⁺ and K⁺ channels is shifted to positive voltages, whereas in low Ca²⁺ it is shifted to negative values. This effect of Ca²⁺ has been proposed to be due to its interaction with fixed negative charges on ion channels and/or glycophospholipids on the cell surface by direct binding and/or screening (25). To examine whether sialic acids are the Ca²⁺-sensitive negative charges on the cell surface, we determined the V1/2 of CHO/Kv1.1 cells in high (25 mM) and low (1 mM) Ca²⁺ to compare with the values recorded in the control solution (5.4 mM). If sialic acids are the negative charges that Ca²⁺ is affecting, then Pro5 would be more sensitive to Ca²⁺ changes than the Lec mutants. Indeed, the V1/2 for Pro5 was shifted about 15 mV in the positive direction in high Ca²⁺ and 7 mV in the negative direction in low Ca²⁺ compared with the control solution (Fig. 4A). In contrast, this same treatment had no effect on the V1/2 of Lec8. However high Ca²⁺ did shift the V1/2 about 7 mV in Lec2 versus control, whereas low Ca²⁺ had no significant effect. Furthermore, in low Ca²⁺ (which is physiological) there were more dramatic positive shifts in V1/2 for both Lec mutants (23 mV for Lec8 and 14 mV for Lec2) versus control Pro5, whereas in high Ca²⁺ this parameter was similar in all cell lines (Fig. 4A). In high Ca²⁺ the slope factor (Fig. 4B) and activation rise time (Fig. 4C) parameters were similar for all cell lines. These results suggest that the titratable Ca²⁺-sensitive negative charges on Pro5 are sialic acids, presumably on the Kv1.1 channel, and these charges are absent from Lec8 and reduced in Lec2.

Total CHO membrane proteins were run on a SDS gel for immunoblotting with Kv1.1 antibodies (34, 35) to determine the pattern of glycosylation of Kv1.1 membrane proteins. The major immunoreactive membrane protein from all cell lines had a molecular weight similar to the Kv1.1 core polypeptide (calculated 56,000) (Fig. 5). This finding suggests that in the steady state the bulk of Kv1.1 proteins are in internal RER membranes and only slightly processed. Bands of higher M_r were also detected indicating that the protein was glycosylated. Pro5 exhibited a diffuse Kv1.1 band that had a M_r of about 80,000 (Fig. 5, lane 2), which was similar to that of the rat brain Kv1.1 protein (Fig. 5, lane 1). A fraction of this Pro5 80,000 band presumably represents plasma membrane proteins that were processed to a similar extent as in brain. The diffuse Kv1.1 banding pattern is due to different degrees of glycosylation of the core protein. Both Lec mutants expressed truncated Kv1.1 glycoproteins with a M_r that was less than the diffuse Pro5 80,000 band (Fig. 5, compare lane 2 with lanes 3, 4, and 5), which is consistent with a glycosylation deficiency. A fraction of the band that had the highest M_r presumably represents plasma membrane proteins. Furthermore the most processed Lec2 Kv1.1 protein (70,000-75,000) exhibited a diffuse banding pattern and appears to be glycosylated to a greater extent than Lec8 Kv1.1 proteins (59,000-62,000) (Fig. 5, compare lane 4 with lanes 3 and 5). This apparent additional glycosylation is probably due to sialic acid addition, because they are the terminal residues added to both N- and O-linked glycoproteins. Thus the Lec2 strain that we have, although a sialic acid deficient mutant, appears to have an incomplete block of sialic acid addition. The Pro5 80,000 protein has a M_r that is 5,000-10,000 and 18,000 larger than the most processed Lec2 and Lec8 proteins, respectively. Nontransfected cell lines did not show Kv1.1 antibody reactivity (Fig. 5, lanes 6-9). Preincubating the Kv1.1 antibody with its specific peptide blocked the immunoreactivity, demonstrating the specificity of the reaction (data not shown).

DISCUSSION

Kv1.1 channels expressed in Lec mutants and sialidase-treated control cells exhibited significant differences in voltage-dependent gating parameters versus control. Lec8 had the most dramatic Kv1.1 gating changes, whereas Lec2 and sialidase-treated control had an intermediate effect. An explanation for these findings may be that there was an incomplete block of sialic acid addition to the Lec2 Kv1.1 proteins, as suggested by immunoblot analysis and Ca²⁺ sensitivity of its V1/2 parameter and incomplete desialidation of Kv1.1 channels in control cells.

The effects of desialidation and high Ca²⁺ on Kv1.1 gating parameters will be discussed in the context of two theories advanced to explain the effects of Ca²⁺ on voltage-gated channels. However, we are not suggesting that desialidation and high Ca²⁺ have identical effects on channel gating. Frankenhaeuser and Hodgkin (25) proposed the surface potential (SP) theory, and Armstrong (40) proposed a hypothesis that may be called the gating stabilization theory (3). The SP theory proposes that Ca²⁺ ions bind to and/or screen negative charges on ion channels and/or glycophospholipids and thereby alter the local electric field detected by the channel's voltage sensor. In contrast, the gating stabilization theory proposes that Ca²⁺ ions bind to negatively charged amino acids on the voltage sensor, thereby stabilizing the channel's closed state and are then removed with depolarization resulting in the open state. Ca²⁺ has also been proposed to be necessary for the normal function of K⁺ channels (41, 42). It is generally considered that ion channel voltage-sensing regions experience a uniform transmembrane electric field. Thus the SP theory predicts that all voltage-dependent gating processes should be changed equally by increasing Ca²⁺. Studies have presented data in support of (43) and against (40, 44, 45) this view (for reviews see Refs. 3 and 46). However, if different channel gating determinants do not reside in a uniform electric field or if fixed negative surface charges are not homogeneously distributed in relation to voltage-sensing regions, then differential changes in gating parameters could result. Overall, the data may be more consistent with Ca²⁺ having two distinct actions: altering surface charge and affecting channel gating directly.

Positive V1/2 shifts of up to 23 mV and slower activation kinetics recorded in the Lec mutants and sialidase-treated control suggest that sialic acids either stabilize the channel's open state or destabilize its closed state. The most likely explanation for these results is that decreasing Kv1.1 sialic acid content reduces the negative surface potential, thereby reducing the negative potential, which is felt by its voltage sensor. A greater depolarization is now required to activate the channel. Although this is consistent with the SP theory, a purely electrostatic ``screening'' mechanism is probably an oversimplification. The sites and positions of sialic acids on Kv1.1 are unknown, but the SP theory would predict that functional ones are closest to the membrane surface. The SP theory requires that titrating surface charge be equivalent to changing the holding potential of the voltage clamp. In physiological Ca²⁺ Lec2 and sialidase-treated control are in accord with this requirement, as evidenced by their G/V curves that have similar slope factors as control, whereas Lec8 clearly is not, as evidenced by its decreased slope. One interpretation of this is that partial reduction of sialic acids on the channel is consistent with the SP theory, whereas complete reduction results in further changes in gating parameters that are not. A combination of the SP and gating stabilization theory may explain the effect of complete reduction of sialic acids on gating parameters. In order to assess any effects of the decrease in glycosylation on the voltage sensor, direct gating current measurements or lower limit estimates of gating valence via limiting slopes from ionic currents should be obtained.

The Ca²⁺-sensitive negative charges on Kv1.1 channels appear to be sialic acids because high Ca²⁺ had no effect on the V1/2 of Lec8. Although these data may be consistent with the SP theory, the effect of varying Ca²⁺ and Zn²⁺ on ion channel function appears to be more complex than the SP theory would imply (44, 45). In the present study, high Ca²⁺ did not change V1/2 or activation kinetics in Lec8, but it did increase the slope of the G/V curve to a value similar to that of control. This finding is more consistent with Ca²⁺ interacting directly with K⁺ channel amino acids to modify gating (44, 45). However, if Ca²⁺ blocks (or permeates (47)) K⁺ channels more at negative than positive voltages, it would contribute to increasing the G/V slope as it is raised.

An alternative explanation for changes in Kv1.1 channel gating may involve conformational changes that shift negatively charged amino acids to positions that no longer influence the voltage sensor. The cell surface negative charge on the Lec mutants and sialidase-treated control would also be predicted to be decreased. Any of these charges close to the Kv1.1 voltage sensor could conceivably affect activation parameters. In this regard, it has been suggested that the effect of Ca²⁺ on the V1/2 of sodium channels in planar lipid bilayers is due to both its interaction with negative charges on the channel molecule (70% of effect) and net negatively charged phospholipids (30% of effect) (48). Further studies are required to differentiate between effects of sialic acids on the channel versus on associated glycolipids.

A recent report has suggested that some cell lines, including CHO cells, express 2.1 K⁺ channel subunits (49). If this subunit is expressed in the CHO cells used in the present study, our conclusions would not be altered. Kv1.1 in CHO cells exhibited delayed rectifier currents, and this channel is not modified by 2.1 (50).

The functional expression levels of Kv1.1 in control and Lec mutants were similar, which suggests that galactose and sialic acid addition in the Golgi are not required for expression on the cell surface. The Kv1.1 channel has only one extracellular N-glycosylation site yet the glycoprotein has about 24 kDa of carbohydrate/monomer, of which 12-18 kDa is sialic acid (20). It is unlikely that 24 kDa of carbohydrate is attached to this one site, and the channel may have O-linked carbohydrates.

N-Glycosylation did not modify Shaker H4 K channel gating parameters when expressed in oocytes (51), although these channels synthesized in a cell-free rough microsomal system and analyzed in planar lipid bilayers required larger than expected depolarizations for activation (52). Functional expression of Kv1.1 channels was inhibited by tunicamycin, a N-glycosylation inhibitor (34, 53). This inhibition was probably due to an indirect effect because the Kv1.1 N-glycosylation mutant was expressed (53). This study concluded that N-glycosylation did not modify channel function, such as altering V1/2 or activation kinetics, when the control and mutant Kv1.1 gating properties were compared.

Additional studies have suggested that sialidase treatment of cells in culture modifies ion channel function or toxin sensitivity in the case of T- and L-type calcium channels (54, 55) and sodium channels (56), whereas other studies recorded no significant effects on sodium channels (57), Ca²⁺-activated K⁺ channels (58), inward rectifier K⁺ channels (59), or delayed rectifier K⁺ channels (60). In the last study it was estimated that 12% of surface sialic acid was resistant to cleavage. If sialic acids play a role in the function of most voltage-gated ion channels, which is unknown, then the negative results above may imply that sialidase-resistant sialic acids are functionally important.

Neurons express their unique complement of ion channels either as homo- and/or heteromultimers giving them their particular properties of electrical excitability. Functional diversity may be further increased by modulation of channel gating by phosphorylation/dephosphorylation (61, 62). Additional functional diversity may result from neurons expressing ion channels with varying degrees of sialidation, which may fine tune their excitability. This task could be accomplished through differential expression of a variety of glycotransferases as reported in tissues (63) and cell lines (64). An advantage to modifying gating via sialidation would be a finer tuning control of voltage dependence, shifting V1/2 to different voltages but retaining other channel core properties. This notion has relevance because mutagenesis studies suggest that amino acid changes in S4 can result in drastic and unpredictable gating alterations. Shifts in channel V1/2 of 5-10 mV would be expected to have physiological consequences in a neuron's response to incoming signals and its output response (3). Although heavy sialidation appears to be a common posttranslational modification to Na⁺ channels (16, 17, 18, 19), which also modifies its function (24, 65), it is unknown whether the numerous K⁺ channel subtypes or other ion channels that have been functionally characterized are also sialidated proteins. The present study suggests that the sialidation state of the Kv1.1 channel plays an important role in determining its particular properties of voltage-dependent gating. Thus the amino acid sequence as well as the degree of sialidation should be considered in any explanation of the functional differences among K⁺ channels.