The sialic acid component of the beta 1 subunit modulates voltage-gated sodium channel function

Voltage-gated sodium channels (Nav) are responsible for initiation and propagation of nerve, skeletal muscle, and cardiac action potentials. Nav are composed of a pore-forming alpha subunit and often one to several modulating beta subunits. Previous work showed that terminal sialic acid residues attached to alpha subunits affect channel gating. Here we show that the fully sialylated beta1 subunit induces a uniform, hyperpolarizing shift in steady state and kinetic gating of the cardiac and two neuronal alpha subunit isoforms. Under conditions of reduced sialylation, the beta1-induced gating effect was eliminated. Consistent with this, mutation of beta1 N-glycosylation sites abolished all effects of beta1 on channel gating. Data also suggest an interaction between the cis effect of alpha sialic acids and the trans effect of beta1 sialic acids on channel gating. Thus, beta1 sialic acids had no effect gating on the of the heavily glycosylated skeletal muscle alpha subunit. However, when glycosylation of the skeletal muscle alpha subunit was reduced through chimeragenesis such that alpha sialic acids did not impact gating, beta1 sialic acids caused a significant hyperpolarizing shift in channel gating. Together, the data indicate that beta1 N-linked sialic acids can modulate Nav gating through an apparent saturating electrostatic mechanism. A model is proposed in which a spectrum of differentially sialylated Nav can directly modulate channel gating, thereby impacting cardiac, skeletal muscle, and neuronal excitability.


Introduction
The importance of voltage-gated sodium channels (Na v ) in action potential initiation and propagation is well established. The orchestrated activation and inactivation gating of sodium channels is vital to normal neuronal signaling, skeletal muscle contraction, and regular heart rhythms. Even small syncopations from this normal gating rhythm may alter cellular excitability and whole animal physiology significantly, leading to such disorders as long QT syndrome and epilepsy (1)(2)(3)(4)(5)(6)(7).
Na v are complex transmembrane glycoproteins that are comprised of a large α subunit that forms the pore through which ions can pass (8)(9)(10). Ten α subunit isoforms have been cloned from excitable tissues, with orthologues present in a wide range of species (11). While the α subunit is sufficient to form functional channels when expressed alone, it is often associated with a β 1 subunit that modulates sodium channel activity -the exact manner by which β 1 alters channel function is still under investigation (12)(13)(14)(15).
Isoforms of the α subunit undergo extensive glycosylation. Estimates indicate that 15-40% of the total Na v molecular weight is carbohydrate (16)(17)(18). Approximately 40-45% of the added carbohydrate residues are sialic acid moieties, resulting in the addition of an estimated 100 sialic acid residues per α subunit molecule (16;17).
attached to β 1 alter α subunit gating through a novel trans-regulatory mechanism?
While the β 1 gene product expressed in heart, skeletal muscle, and brain is identical, the level of posttranslational modification appears to vary among cell types (35). In addition, expression of β 1 is tightly regulated over the course of development. Depending on the tissue, expression is first observed in the first four days after birth and increases to maximal, maintained, levels after 2-4 weeks (35)(36)(37)(38)(39). Thus, expression of β 1 , an important modulator of Na v gating, is developmentally regulated. Also, β 1 is heavily and differentially glycosylated. If β 1 glycosylation alters α subunit gating, then channel gating might be modulated differently by various levels of β 1 sialic acid among excitable tissues and from one developmental stage to another.
Given that α subunit sialic acids impact channel gating, we wished to test the hypothesis that sialic acids attached to the β 1 subunit are involved in modulating channel function. To this end, we expressed four different Na v α subunit isoforms in the presence or absence of β 1 in a cell line that essentially fully sialylates proteins, and in a mutant daughter cell line that is unable to sialylate proteins. Our data indicate a novel mechanism by which β 1 can modulate Na v gating in a saturating, sialic acid-dependent manner, and that β 1 sialic acids account for all effects of β 1 on sodium channel gating.
In addition, the data indicate for the first time, that within a single membrane, transmembrane protein function is modulated by the sugars attached to a second membrane protein. 7 sodium channel vector and 8-15% pGreen Lantern Fluorescent Protein (GFP, Gibco) or the β 1 subunit vector) and then incubated for 24 hours. The transfection mix was replaced with the appropriate growth medium and cells were incubated for a further 48 hours before use.

Electrophysiology and data analysis
I Na were recorded at room temperature (~22°C) using established whole cell patch clamp techniques, pulse protocols, data analyses and solutions as previously described (20;23). For the Ca 2+ perfusion studies, seals were formed in the bath solution containing 2.0 mM Ca 2+ . Cells were first perfused with the 2.0 mM Ca 2+ bath solution, and then with the 0.2 mM Ca 2+ bath solution to determine directly the shift in V a with a 10-fold decrease in external Ca 2+ concentration.
Although series resistance was compensated 95-98% for all data, the smaller current produced using the low sodium solutions further minimized any remaining series resistance error, resulting in < 1mV error. All data shown are recorded at least 5 minutes after attaining whole cell configuration to assure complete dialysis of the intracellular solution. All solutions were filtered using Gelman 0.2 µm filters immediately prior to use.
8 from the holding potential. At each test potential, steady-state whole cell conductance was determined by measuring the peak current at that potential and dividing by the driving force (i.e., difference between the membrane potential and the observed reversal potential). The maximum conductance generated by each cell was used to normalize the data for each cell to its maximum conductance by fitting the data to a single Boltzmann distribution (equation 1, solving for maximum conductance). The average V a +/-SEM values listed in Table 1 were determined from these single Boltzmann distributions. The normalized data were then averaged with those from other cells, and the resultant average conductance-voltage curve was fit via least squares using the following Boltzmann relation: where V is the membrane potential, V a is the voltage of half activation, and K a is the slope factor.

Measurement of inactivation time constants
Inactivation time constants were determined by fitting the current traces used to measure G-V relationships. Attenuating currents from 90 to 10% of the peak values were fit to a single exponential function to determine the time course of fast inactivation.

Steady state inactivation curves (h inf )
Voltage dependence of steady state inactivation was determined by first prepulsing the membrane for 500 ms from the holding potential, then stepping to +60 mV for 5 ms, and then returning to the holding potential. The prepulse voltages ranged from -130 mV to +10 mV in 10 mV increments. The currents from each cell were normalized to the maximum current measured by fitting each single cell data to a single Boltzmann distribution (equation 2, solving for maximum current), from which the mean V i +/-SEM values listed in Table 1 were determined. The normalized data for many cells were then averaged and fit to equation (2), from which the average h inf curves describing steady state inactivation for the channel population were calculated.

Recovery from inactivation
Cells were stepped to +60 mV for 10 ms from the holding potential and then returned to the recovery potential for varying duration ranging from 1 to 20 ms in 1 ms increments. Following this recovery pulse, the potential was again stepped to +60 mV for 10 ms. The peak current measured during the two +60 mV depolarizations were compared, and the fractional peak current remaining during the second depolarization was plotted as a function of the recovery pulse duration. This represents the fraction of channels that recovered from inactivation during the recovery interval. Time constants of recovery were determined by fitting the data to single exponential functions.

Results β 1 modulates gating of three of four Na v α subunits
The external domain of β 1 is critical for correct modulation of sodium current mediated by different Na v channel α subunit isoforms (26)(27)(28)(29). Reports indicate effects of β 1 ranging from increased fast inactivation rate to hyperpolarizing and even small depolarizing shifts in the voltage dependence of steady state channel gating(3;12;13;31;32;40). These varied effects apparently depend on the α subunit and the cellular expression system used to study β 1 function. In order to minimize the variation observed among cellular expression systems, we co-expressed β 1 with one of four different α subunits in Chinese hamster ovary (CHO) cells to compare directly α subunit isoform-specific effects of β 1 on sodium current. Figure 1 shows the average Na v conductance voltage relationships (G-V) recorded from cells expressing one of four Na v α subunits: the adult skeletal muscle isoform (Na v 1.4), the cardiac isoform (Na v 1.5), a peripheral nerve isoform (Na v 1.7), and a brain isoform (Na v 1.2), in the presence or absence of β 1 . Note that Na v 1.4 activation is unaffected by β 1 , whereas the G-V curves for Na v 1.2, Na v 1.5, and Na v 1.7 are shifted in the hyperpolarized direction by 8-9 mV when β 1 is present. Figure 2 shows that the voltage dependence of steady state channel availability (h inf ) is similarly affected by β 1 , with the voltages of half inactivation (V i ) for Na v 1.2, Na v 1.5, and Na v 1.7 shifted 6-9 mV in the hyperpolarized direction in the presence of β 1 , while V i for Na v 1.4 is unaffected by β 1 .
To describe more fully the effects of β 1 on α subunit gating, we examined the rates of fast inactivation and recovery from fast inactivation. As shown in Figures 3 and  4, the effects of β 1 on gating kinetics are similar to its effects on steady state parameters. Thus, β 1 had no effect on the gating of Na v 1.4, but caused a nearly uniform hyperpolarization of all measured gating parameters for Na v 1.2, Na v 1.5, and Na v 1.7.
Mean values ± SEM as well as statistical analyses for all parameters and conditions measured in this study are shown in Table 1.

Modulation of Na v gating by β 1 requires β 1 sialic acids
Sodium channel α subunits undergo extensive glycosylation, often capped with sialic acid residues (SA) (16)(17)(18). Previous work showed that Na v α subunit sialylation is an important process by which sodium channel gating is modulated in an isoform-and developmental-dependent manner (21)(22)(23). The β 1 subunit is also heavily glycosylated, with four potential N-glycosylation motifs within its extracellular N-terminal domain (24).
To determine if the observed isoform-specific shifts in Na v gating produced by β 1 are caused by β 1 sialic acids, we co-expressed each of the α subunits with β 1 in two well characterized CHO cell lines that produce proteins with differing amounts of attached sialic acids (41)(42)(43). The Pro5 cell line allows normal CHO cell protein sialylation while Lec2 cells, deficient in the CMP-sialic acid transporter, produce proteins that are essentially non-sialylated. Figures 1 and 2 show the G-V and h inf curves for each α subunit ± β 1 ± SA. When expressed in Lec2 cells, the less sialylated Na v 1.4 gates at more depolarized potentials than the fully sialylated Na v 1.4 expressed in Pro5 cells irrespective of the presence or absence of β 1 . These data indicate that gating of Na v 1.4 is dependent on α  subunit sialic acids but not on β 1 sialic acids. Gating of Na v 1.2, Na v 1.5 and Na v 1.7 are not significantly affected by sialylation in the absence of β 1 , suggesting that α subunit sialic acids do not alter gating of these isoforms significantly.
There is a small, consistent, but generally insignificant, SA-dependent shift in Na v 1.2 gating, while Na v 1.5 and Na v 1.7 gating show no dependence on sialic acids. In contrast, β 1 expression induces a 6-10 mV hyperpolarizing shift in G-V and h inf curves for Na v 1.2, Na v 1.5 and Na v 1.7 when expressed in the fully sialylating Pro5 cell line. This β 1 induced shift in gating is not present when β 1 sialylation is reduced. When β 1 is coexpressed with α in Lec2 cells, the observed G-V and h inf relationships for each isoform are nearly identical to those measured for the α subunit alone. These data indicate that β 1 sialic acids can alter directly Na v 1.2, Na v 1.5, and Na v 1.7 gating. Figures 3 and 4 confirm that the β 1 induced shifts in all measured gating parameters for these three α subunits are caused by β 1 sialic acids. Here we observed that Na v 1.4 gating is not affected by the presence of β 1 .
Na v 1.4 is predicted to have more N-glycosylation than Na v 1.2 or Na v 1.7, and we showed previously that Na v 1.4 is more heavily glycosylated than Na v 1.5 in this system(23), consistent with work performed in other laboratories (44). The majority of Na v 1.4 N-glycosylation sites are present in the DIS5-S6 extracellular linker, and previous work demonstrated that the sialic acid effects on Na v 1.4 gating are localized to this domain (23). Given that β 1 can modulate gating of channels with putatively less glycosylation than Na v 1.4 (e.g., Na v 1.2, Na v 1.5, and Na v 1.7), we examined the effects of β 1 on the Na v 1.4 DIS5-S6 loop chimera, hSkM1P1, that is less glycosylated than wild-type Na v 1.4 (23;27). As shown in Figure 5, gating of hSkM1P1 alone is not dependent on sialic acid. However, co-expression of β 1 imposes a sialic acid dependent hyperpolarizing shift in hSkM1P1 gating similar in magnitude and direction to the effects of β 1 sialic acids on gating of Na v 1.2, Na v 1.5, and Na v 1.7. This, together with the lack of effect of β 1 on Na v 1.4 gating observed here would suggest there is a potential saturating limit to the amount (location) of β 1 sialic acids that can affect α subunit channel gating.
More precisely, apparently combinations of the sialic acids associated with α subunit IS5-S6 and with β 1 determine the overall impact of these sugars on channel gating, and this effect is saturating. sialic acids contribute to a negative surface potential, then α subunits that are sensitive to β 1 sialic acids would be predicted to be more sensitive to external Ca 2+ concentration when co-expressed with β 1 than when expressed alone. As shown in Figure 6, the external Ca 2+ dependent shift in sodium channel V a is largest when hSkM1P1 is coexpressed with the fully sialylated β 1 subunit. When hSkM1P1 is expressed alone (± SA) or when co-expressed with β 1 under conditions of reduced sialylation, the Ca 2+ dependent shift in V a is significantly smaller and nearly identical for these three conditions. indistinguishable (data not shown). Thus, all β 1  sialic acid dependent effects on α subunit gating can be assigned to β 1  N-linked sialic acids.

Discussion β 1 N-linked sialic acids fully account for effects of β 1 on Na v gating
Here we sought to determine more directly the role of the β 1 subunit in voltagegated sodium channel gating. We compared, in a single cellular system, the effects of Previous work from our laboratory demonstrated that the gating of Na v channel α subunits expressed in CHO cells is altered by changes in sialylation levels (20;23). We recently described how Na v 1.5 gating parameters were unaffected by α subunit sialic acids, whereas Na v 1.4 gated at more depolarized potentials when expressed in essentially non-sialylating Lec2 cells versus fully sialylating Pro5 cells (23). Here, we extended this study to question the role of sialic acids in the gating of two neuronal α subunits, Na v 1.2 and Na v 1.7. As shown in Figures 1-4, much like that observed for Na v 1.5 previously, reduced sialylation of Na v 1.7 had no effect on gating. Na v 1.2 gating was, in general, not significantly altered by α subunit sialic acids, although a consistent, small, depolarizing shift in gating was observed for Na v 1.2 expressed in Lec2 cells compared to gating in the fully sialylating Pro5 cells.
The β 1 subunit is predicted to have four N-linked glycosylation sites, at least three of which are glycosylated (24;34). We tested the hypothesis that β 1 sialic acids modulate Na v gating by co-expressing β 1 with the four α subunits in the fully sialylating Pro5 cells and compared channel function with those observed in the non-sialylating  We showed previously that α subunit sialic acids impose effects on gating through an apparent electrostatic mechanism. It is well established that Na v gating is sensitive to changes in external Ca 2+ concentrations, requiring a larger depolarization as external Ca 2+ levels increase. A surface potential theory is often assigned to this phenomenon. Thus, external negative surface charges produce a surface potential that alters the voltage sensed by the channel gating mechanism(49). As Ca 2+ levels increase, these charges will be screened effectively, minimizing their impact on gating.
The voltage sensed by the gating mechanism of the channel becomes more negative, moving away from the voltage of activation, and thus a larger depolarization is required to activate the channel. If sialic acids contribute to such a negative surface potential, then sensitivity to screening by external Ca 2+ should increase with the level of functional sialic acids. If β 1 sialic acids contribute to a negative surface potential, then gating of channels co-expressed with fully sialylated β 1 should be the most Ca 2+ sensitive as shown in Figure 6. The shift in V a with changing external Ca 2+ levels for hSkM1P1 alone ± SA, and for hSkM1P1 + β 1 expressed in Lec2 cells, were similar and significantly smaller than the shift observed for the fully sialylated hSkM1P1 + β 1 . Thus, β 1 sialic acids alter α subunit gating through an electrostatic mechanism, apparently contributing to an external negative surface potential. dependent on α sialic acids, but was dependent on β 1 sialic acids (trans effect). These data suggest that in this system there is a saturating limit to the contribution of sialic acids to channel gating, with Na v 1.4 sialic acids likely achieving saturation. Figure 5 provides further supporting evidence, showing that gating of the less glycosylated Na v 1.4 chimera, hSkM1P1, is no longer sensitive to sialic acid, but is sensitive to β 1 sialic acids. These data suggest that with hSkM1P1, Na v 1.4 functional sialylation is reduced to levels below saturation such that β 1 sialic acids might contribute to channel gating. Thus, it appears the combined effects of cis α subunit DIS5-S6 and trans β 1 subunit functional sialic acids on channel gating are saturating (see Figure 8).

The ability of the cell to produce differentially sialylated proteins leads to a spectrum of Na v functional sialic acid levels that directly modulate channel gating
There have been numerous α and β subunits identified to date. Each α subunit has a distinct glycosylation pattern that potentially can modulate current in an isoform specific manner (21;23). We now show that sialylation of the β 1 subunit also modulates Na v function. Expression of β 1 is regulated over the course of development, and is apparently differently processed among excitable tissues (35)(36)(37)(38). As suggested by the model shown in Figure 8, it is possible to envisage a scenario in which different Na v αβ 1 subunit combinations are differentially functionally sialylated in various tissues, over the course of development, perhaps pathologically, and/or even chronically within a single cell, potentially leading to modulation of Na v gating.
Thus, the cell would have two complementary methods to modulate sodium current through differential sialylation. Changing the level of sialylation through regulation of sialyltransferase activity might alter the function of some channels chronically. For example, sodium currents in adult rat cortical and dorsal root ganglion neurons are less sensitive to sialic acid than sodium currents from neonatal neurons (22;50). A developmental decrease in Na v glycosylation was observed in one study (22). cause shifts in the window current voltage range that may be responsible for such maladies as epilepsy and arrhythmias (LQTS)(1;4;5). In addition, β 1 sialic acids slow the rate to recovery from fast inactivation which would directly affect action potential relative refractory periods. That is, if β 1 sialic acids induce slower recovery rates, then 20 at a short time following initialization of a Na v kinetic cycle, the percentage of inactivated channels will increase. This will directly affect the rate at which subsequent depolarizations might lead to activation of Na v sufficient to cause an action potential to fire. Thus, we report novel findings that are relevant to the modulation of voltage-gated sodium channel activity that will directly impact how one's heart, muscle, and brain function.

Trans-regulation of membrane proteins by carbohydrate structures attached to a second protein: a unique process?
In addition, a more global phenomenon can be described for the first time. Our data indicate that β 1 sialic acids directly alter α subunit function, consistent with a mechanism by which transmembrane protein function can be modulated solely and  Conductance-voltage (G-V) relationships for four voltage-gated sodium channel α subunits ± β 1 as expressed in two CHO cell lines with varying abilities to sialylate proteins. Expression in Pro5 cells allows normal CHO cell sialylation, while expression in Lec2 cells prevents sialylation. Data are the mean normalized peak conductance ± SEM at a membrane potential, and curves are fits of the data to single Boltzmann relationships. n = 9-13 for each condition tested (see Table 1). Significant differences among conditions tested throughout are listed in Table 1  Steady state channel availability (h inf ) curves for the four α subunits ± β 1 ± sialic acid.
Data are the mean normalized peak current (Ī)+/-SEM measured during a maximally depolarizing test pulse following a 500 ms prepulse to the plotted potentials. Symbols, lines, sample numbers, and panels are as described in Figure 1.  sample numbers, and panels are as described in Figure 1. Inset to Panel C: Typical plot of the fractional recovery measured using a standard twin pulse protocol to determine the rate of recovery from fast inactivation at a -120 mV recovery potential for Na v1.7 ± β 1 ± SA. Data are the mean ± SEM fractional current measured during a second depolarizing test pulse following the plotted interval at -120 mV from the original depolarizing test pulse. Lines are exponential fits of the data from which the τ rec were determined. for each condition (see Table 1    are pictured along the maximum of the curve. In a second example, Na v 1.7 alone can not contain many DIS5-S6 functional sialic acids, and therefore Na v 1.7 expressed alone must lie somewhere along the minimum portions of the curve. When β 1 is co-expressed with Na v 1.7, the levels of functional sialic acids increase, causing a hyperpolarization in V a , and hence, Na v 1.7 + β 1 is shown somewhere along the rising slope of the curve.
While the data presented here show uniform effects of β 1 sialic acids on measured voltage dependent Na v gating parameters, there is no reason, a priori, for this to occur.
The impact of β 1 sialic acids on specific Na v gating parameters will depend on several factors including the extent of coupling among channel kinetic states and/or any inherent voltage dependence of the transition from one state to another. Thus, for example, the interaction of β 1 sialic acids with a specific α subunit may place the channel at a different position along the concentration response curves for shifts in V a versus shifts in V i . This would likely be observed as differences in the impact of β 1 sialic acids on V a and V i . Na v 1.4, Na v 1.5, and hSkM1P1, and at -30 mV for Na v 1.2 and Na v 1.7. To determine the effects of β 1 sialic acids, significance was tested using a two-tailed student's t test, with each condition compared to the parameter measured for the fully sialylated α subunit alone. Significance is demarcated as follows: * = significant (p<0.1); ** = highly significant (p<0.005).