+ Functional reciprocity between Na channel Nav1.6 and β1 subunits in the coordinated regulation of excitability and neurite outgrowth

William J. Brackenburya, Jeffrey D. Calhouna, Chunling Chena, Haruko Miyazakib, Nobuyuki Nukinab, Fumitaka Oyamab, Barbara Ranschtc, and Lori L. Isoma,1

aDepartment of Pharmacology, University of Michigan Medical School, Ann Arbor, MI 48109-5632; bMolecular Neuropathology Group, RIKEN Brain Science Institute, Saitama, 351-0198 Japan; and cBurnham Institute for Medical Research, La Jolla, CA 92037

Edited* by William A. Catterall, University of Washington School of Medicine, Seattle, WA, and approved December 17, 2009 (received for review September 2, 2009) Voltage-gated Na+ channel (VGSC) β1 subunits regulate cell–cell adhesion, fyn kinase, and contactin (22, 23). Axonal path finding adhesion and channel activity in vitro. We previously showed that and fasciculation are disrupted in Scn1b null mice and zebrafish β1 promotes neurite outgrowth in cerebellar granule neurons scn1bb morphants (22, 24). Thus, β1 functions as a CAM to reg- (CGNs) via homophilic cell adhesion, fyn kinase, and contactin. Here ulate neuronal migration, axon guidance, and fasciculation during + we demonstrate that β1-mediated neurite outgrowth requires Na CNS development. Disruptions in these critical processes likely β current (INa) mediated by Nav1.6. In addition, 1 is required for high- contribute to disease in patients with SCN1B mutations. The next fi frequency action potential ring. Transient INa is unchanged in step in understanding the physiological role of β1 in vivo is to Scn1b β ( 1) null CGNs; however, the resurgent INa, thought to under- β β fi investigate whether 1-mediated modulation of INa and 1-medi- lie high-frequency ring in Nav1.6-expressing cerebellar neurons, is ated cell–cell adhesion are interrelated. reduced. The proportion of axon initial segments (AIS) expressing The aims of the present study were to assess the role of I in Na 1.6 is reduced in Scn1b null cerebellar neurons. In place of Na 1.6 Na v v β1-mediated neurite outgrowth and to investigate the effect(s) of at the AIS, we observed an increase in Na 1.1, whereas Na 1.2 was v v β1onI and neuronal excitability. We demonstrate that β1 and unchanged. This indicates that β1 is required for normal localization Na NEUROSCIENCE Nav1.6 have reciprocal functions in vivo, such that β1-mediated of Nav1.6 at the AIS during the postnatal developmental switch to Na 1.6-mediated high-frequency firing. In agreement with this, β1is neurite outgrowth requires Nav1.6-mediated INa, and Nav1.6 v fi β normally expressed with α subunits at the AIS of P14 CGNs. We localization and consequent high-frequency ring require 1. propose reciprocity of function between β1 and Na 1.6 such that Our results demonstrate a requirement for partnering of specific v α β β1-mediated neurite outgrowth requires Nav1.6-mediated INa, and VGSC and subunits during CNS development in vivo. Nav1.6 localization and consequent high-frequency firing require β1. We conclude that VGSC subunits function in macromolecular Results and Discussion signaling complexes regulating both neuronal excitability and Tetrodotoxin and the Scn8a Null Mutation Inhibit β1-Mediated migration during cerebellar development. Neurite Outgrowth. Electrical activity mediated by VGSCs regu- lates neuronal migration and development (25). The possible role cell adhesion | cerebellum | resurgent current | axon initial segment | of INa in β1-mediated neurite outgrowth has not yet been studied, action potential however. Similarly, the potential involvement of β1 in activity- dependent neuronal development is unknown. To investigate + β oltage-gated Na channels (VGSCs), composed of one pore- whether INa is involved in 1-mediated neurite outgrowth, we Vforming α subunit and two β subunits (1), are responsible for isolated CGNs from P14 wild-type (WT) mice and cultured them initiation and conduction of action potentials (APs) (2). Of the nine on monolayers of control Chinese hamster lung (CHL) cells, which α subunits (3), Nav1.1, Nav1.2, and Nav1.6 are found in the postnatal do not express β subunits, or on CHL cells stably expressing β1. CNS (4, 5) where they display developmentally regulated expression β1–β1 trans-homophilic adhesion increased the average neurite patterns in specialized neuronal subcellular domains. For example, length by 1.7-fold, as reported previously (P < 0.001) (23). Inclu- Nav1.1 and Nav1.2 are replaced during postnatal development at the sion of tetrodotoxin (TTX) during the assay inhibited β1-mediated Scn8a axon initial segment (AIS) and nodes of Ranvier by Nav1.6. neurite outgrowth; this occurred across the entire neurite length null mice display motor dysfunction, ataxia, and lethality by post- distribution (Fig. 1 A and B). TTX had no effect on fibroblast natal day (P) 21, suggesting that this developmental switch to Nav1.6 growth factor (FGF; 20 ng/mL)-mediated neurite outgrowth, Scn8a expression in brain is critical (6, 7). null retinal ganglion which is β1-independent (P < 0.001; Fig. 1 C and D) (22). Fur- neurons display impaired excitability, demonstrating that Nav1.6 is thermore, FGF had no additive effect on the neurite outgrowth of vital for high-frequency firing (8–10). Thus, VGSC-driven neuronal CGNs plated on β1-expressing monolayers, suggesting that FGF- activity is important for proper CNS development, although the mediated and β1-mediated neurite outgrowth mechanisms ulti- underlying mechanism(s) are not well understood. P n VGSC β1 subunits are multifunctional molecules that modulate mately activate a common signaling pathway ( = 0.57; = 300). channel kinetics and gating, regulate channel cell surface expres- sion, and participate in cell–cell adhesion in vitro (11). Scn1b (β1) null mice are ataxic, experience spontaneous seizures, and exhibit Author contributions: W.J.B. and L.L.I. designed research; W.J.B., J.D.C., and C.C. per- β formed research; W.J.B. performed and analyzed all experiments except Western blot a prolonged cardiac QT interval, demonstrating that 1 modulates analysis, which was performed by J.D.C. and C.C.; H.M., N.N., and F.O. provided β1 and electrical excitability in vivo (12, 13). Consistent with this, human β4 antibodies; B.R. provided Cntn mice; W.J.B., J.D.C., and L.L.I. analyzed data; and W.J.B. mutations in SCN1B result in epilepsy and arrhythmia (14–21). As and L.L.I. wrote the paper. a member of the Ig superfamily of cell adhesion molecules The authors declare no conflict of interest. (CAMs), β1 mediates cellular aggregation, cytoskeletal recruit- *This Direct Submission article had a prearranged editor. ment, and extracellular matrix interactions in vitro (11). β1 also 1To whom correspondence should be addressed. E-mail: [email protected]. promotes neurite outgrowth in cerebellar granule neurons This article contains supporting information online at www.pnas.org/cgi/content/full/ (CGNs) via a lipid raft mechanism involving trans homophilic cell 0909434107/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.0909434107 PNAS Early Edition | 1of6 Downloaded by guest on September 26, 2021 CGNs (Fig. 1H), demonstrating that, similar to Scn1b neurons (22), Scn8a null neurons are not deficient in downstream mechanisms of neurite outgrowth that converge between lipid raft-mediated and non–raft-mediated pathways (26). β1 expression was unchanged in the cerebella of Scn8a null mice, suggesting that the lack of response of Scn8a null CGNs to the β1 monolayer was not due to a reduction in β1 in the CGNs (Fig. 1I). These findings suggest that β1-mediated, but not FGF-mediated, neurite out- growth requires INa and Nav1.6 expression.

Scn1b Null CGNs Have Impaired Excitability. Given that Scn1b mutations result in in vivo (12–21), and that β1- mediated neurite outgrowth in CGNs requires INa and Nav1.6, we postulated that β1 might regulate electrical excitability in the cerebellum. To test this, we recorded APs in P12-13 WT and Scn1b null CGNs in cerebellar slices by whole-cell patch clamping. WT and Scn1b null CGNs typically had small membrane capacitance (Table S1) and lacked spontaneous firing. Both WT and Scn1b null CGNs displayed normal inward and outward currents (Fig. 2A and Table S1). Resting potential, input resistance, and AP parameters were similar in the WT and Scn1b null CGNs (Table S1). Both WT and Scn1b null CGNs were capable of sustained repetitive firing, although some failed within the 500-ms pulse (Fig.

Fig. 1. β1-mediated neurite outgrowth is inhibited by TTX and the Scn8a null mutation. (A) Neurite length of WT CGNs grown on CHL or CHL-β1 monolayers and treated with/without TTX (10 μM) for 48 h (n = 20). (B) Neurite distribution (%) plotted against neurite length for CGNs in A.(C) Neurite length of WT CGNs grown on CHL monolayers and treated with/ without TTX and/or FGF (20 ng/mL) for 48 h (n = 300). (D) Neurite distribution (%) plotted against neurite length for CGNs in C.(E) Neurite lengths of WT and Scn8a null CGNs grown on CHL or CHL-β1 monolayers (n = 300). (F) Neurite distribution (%) plotted against neurite length for CGNs in E.(G) Neurite lengths of WT and Scn8a null CGNs grown on CHL monolayers and treated with/without FGF for 48 h (n = 300). (H) Neurite distribution (%) plotted against neurite length for CGNs in G. Data are mean ± SEM. ***P < 0.001, ANOVA with Tukey’s post hoc test. (I) Western blot of WT and Scn8a Fig. 2. Electrical excitability is impaired in Scn1b null CGNs in vivo. (A) Voltage null cerebellar , using anti-β1. Anti-α-tubulin was used as (V)-gated currents in a (i) WT and (ii) Scn1b null CGN following depolarization a loading control. to 0 mV. (B) Membrane V recordings from (i) WT and (ii) Scn1b null CGNs fol- lowing current (I) injection. Upper traces represent CGNs that sustained repetitive firing during 500 ms; lower traces represent CGNs that failed. (C)AP Thus, INa is specifically required for β1-mediated, but not FGF- firing rate plotted as a function of I, normalized to AP threshold for WT (filled mediated, neurite outgrowth. circles) and Scn1b null (open circles) CGNs. Data are mean ± SEM (n ≥ 15). *P < We next used neurons isolated from Scn8a null mice to test 0.05; **P < 0.01; ***P < 0.001, t test. (D) Distribution of APs across the 500-ms I injection for WT (dark bars) and Scn1b null (light bars) CGNs. For each CGN, the whether Nav1.6 is required for INa-dependent, β1-mediated neurite β number of APs was counted in each 100-ms interval for the sweep giving outgrowth. 1 in the monolayer increased the neurite length of WT fi ± ≥ < CGNs by 1.7-fold after 24 h (P < 0.001). In contrast, no increase in maximum ring frequency. Data are mean SEM (n 16). P 0.001 between WT and Scn1b null (two-way ANOVA). (E) Instantaneous firing frequency (Inst. neurite length was seen in Scn8a null CGNs grown on β1-expressing E F freq., in Hz), calculated as the reciprocal of the interval between two adjacent monolayers (Fig. 1 and ). FGF-dependent neurite outgrowth was APs, for WT (filled circles) and Scn1b null (open circles) CGNs, plotted for the not affected by the Scn8a null mutation (Fig. 1G). The neurite first 10 AP intervals. Data are expressed as mean ± SEM (n ≥ 11). P < 0.001 length distribution was increased by FGF in both WT and Scn8a null between WT and Scn1b null (two-way ANOVA).

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.0909434107 Brackenbury et al. Downloaded by guest on September 26, 2021 2B). The number of spikes escalated in response to increasing cur- that might underlie their impaired excitability. To address this fi Scn1b rent, although the ring rate was consistently reduced in null issue, we recorded INa in CGNs cultured for 14 DIV with axonal C CGNs at all input intensities (Fig. 2 ). Accordingly, the maximum processes intact (30, 31). As in dissociated CGNs, transient INa was fi Scn1b ring rate across the 500-ms current pulse was reduced in null unchanged in 14 DIV Scn1b null CGNs (Table S4). Persistent INa CGNs (P < 0.05), as was the mean number of spikes fired before also was unchanged in 14 DIV Scn1b null CGNs (Fig. 3 A and C). P < failure ( 0.01) (Table S1). In addition, the maximum number of In CGNs, resurgent INa flows after membrane repolarization, APs fired within each 100 ms interval was smaller in Scn1b null due to the release of an endogenous open-channel blocker, which P < D fi CGNs ( .001; Fig. 2 ), and the maximum instantaneous ring rapidly terminates the transient INa during strong depolarizations frequency (inverse of the interspike interval) was significantly reduced in Scn1b null CGNs (P < 0.001; Fig. 2E). These findings suggest that β1 regulates CGN excitability in situ. CGNs lacking β1 were less able to sustain prolonged high- frequency firing over a range of input intensities. Furthermore, CGNs lacking β1 were less efficient in generating APs at short interspike intervals. We postulated that the impaired excitability of Scn1b null CGNs might be due to altered INa, altered VGSC expression, or both. We then tested these possibilities.

Resurgent, but not Transient, INa Is Reduced in Scn1b Null CGNs. We recorded INa in P14 dissociated CGNs under whole-cell voltage clamp 24 h after plating. Both WT and Scn1b null dissociated CGNs expressed robust INa (Fig. S1A). The absence of β1 had no effect on peak INa density, voltage dependence or kinetics (Fig. S1 B–D and Table S2). Contactin, which is required for β1-mediated neurite outgrowth, interacts with the β1 Ig loop and increases VGSC cell surface expression (22, 27). Cntn null mice have significantly reduced brain levels of β1, are ataxic, and survive only to P21, sug- gesting that contactin is required for normal cerebellar excitability NEUROSCIENCE (22, 28). However, here the Cntn null mutation had no effect on transient INa in P12 dissociated CGNs (Fig. S2 and Table S3). One problem encountered when recording INa in acutely dis- sociated neurons is that processes present in vivo are lost during the extraction (29). This may hinder the ability to detect any alteration in axonal VGSC expression in Scn1b null CGNs in vivo

+ Fig. 3. Resurgent Na current (INa) is reduced in Scn1b null CGNs cultured for 14 days in vitro (DIV). (A) Whole-cell INa from WT (black), and Scn1b null (gray) CGNs elicited by 60-ms depolarizing voltage (V) pulses to −40 mV normalized – to peak transient I (%). Inset, I persisting for 50–55 ms after onset of depola- Fig. 4. Nav1.6 expression is reduced at the AIS of Scn1b null CGNs. (A D) rization. (B) Resurgent I from WT (black) and Scn1b null (gray) CGNs elicited by 100× Z-series confocal projections of WT (i) and Scn1b null (ii) 14 DIV CGNs α repolarization to −30 mV following a depolarizing pulse to +30 mV for 20 ms, labeled with anti-ankyrinG (AnkG) (red) and subunit antibodies (green): α μ normalized to peak transient I at +30 mV (%). (C) Persistent I measured as the Nav1.6 (A), pan- subunit (B), Nav1.2 (C), and Nav1.1 (D). (Scale bar: 20 m.) mean I at 50–55 ms after onset of depolarization (Left) and resurgent I (Right). Arrows point to AIS expressing AnkG.(E) Proportion of AnkG-expressing CGN Data are presented as mean ± SEM (n ≥ 19). *P =0.01(t test). (D) Western blots AIS that also express α subunits for WT (dark bars) and Scn1b null (light bars) of WT and Scn1b null cerebellar membrane protein using anti-β4 and anti-β2. (n =60fields of view taken from three mice of each genotype). Data are Anti–α-tubulin was used as a loading control. mean ± SEM. ***P <.001, Mann-Whitney test.

Brackenbury et al. PNAS Early Edition | 3of6 Downloaded by guest on September 26, 2021 (32). This resurgent INa is proposed to promote repetitive firing We next investigated Nav1.6 expression in cerebellar cry- by rapidly restoring VGSC availability (32). We found that osections from P14 WT and Scn1b null mice. Although Nav1.6 was Scn1b resurgent INa was reduced in 14 DIV Scn1b null CGNs (P = 0.01; expressed in the inner granular layer of both WT and null Fig. 3 B and C). A peptide corresponding to the cytoplasmic sections, it was not possible to resolve the localization of Nav1.6 to domain of β4 mediates resurgent current in Purkinje neurons in specific compartments due to the tight compaction of CGNs within A vitro, implicating β4 in this mechanism (33). β4 protein levels this region (Fig. 5 ). It was possible to determine Nav1.6 expres- were unchanged in Scn1b null cerebella, suggesting that the sion in the soma and at the AIS of Purkinje neurons, however (Fig. A Scn1b reduced resurgent INa in Scn1b null CGNs was not due to altered 5 ) (36). Consistent with 14 DIV CGNs, the proportion of β4 expression (Fig. 3D), although the absence of β1 may change null Purkinje neurons expressing Nav1.6 at the AIS was sig- the conformation of α–β4 interactions (34). The level of β2 nificantly reduced (P < 0.001; Fig. 5 A and B). These defects in Scn1b protein was reduced by 39% in Scn1b null cerebella (P < 0.05; Nav1.6 expression were observed in all three null mice β Fig. 3D), although its subcellular distribution remained studied compared with WT littermates. 1 was robustly expressed unchanged (Fig. S3). In summary, reduced resurgent I in Scn1b at the AIS of WT calbindin-positive Purkinje neurons in cerebellar Na α null CGNs is consistent with reduced repetitive firing and sup- slices and colocalized at the AIS of 14 DIV CGNs with subunits C D Scn1b (Fig. 5 and ). Thus, Nav1.6 expression at the AIS may require ports the hypothesis that the impaired excitability of null β CGNs in situ is caused by altered I . clustering of 1 in this region, similar to the CAMs neurofascin-186 Na and NrCAM, which also are β1-binding partners (37–39). Although cerebellar neurons express TTX-sensitive VGSCs, Nav1.6 Is Reduced at the AIS of Scn1b Null CGNs. We next tested the idea that impaired excitability of Scn1b null CGNs may be caused TTX-resistant (TTX-R) Nav1.8 is up-regulated under pathophy- – by altered VGSC expression. Given that Na 1.6 is essential for siological conditions (40 42). TTX-R Nav1.5 channels also are up- v Scn1b high-frequency repetitive firing and carries resurgent I in Pur- regulated in null ventricular myocytes (12). Consequently, we Na investigated whether TTX-R channels are up-regulated in CGNs in kinje neurons (8, 10), we hypothesized that axonal localization of β Scn1b the absence of 1. TTX (500 nM) completely blocked the INa in both Nav1.6 might be altered in null CGNs. Consistent with Scn1b n previous work, ankyrin G (Ank ) and Na 1.6 were strongly coex- WT and null CGNs in cerebellar slices ( = 5 each), indi- G v cating no up-regulation of TTX-R channels in the absence of β1. pressed at the AIS in 14 DIV WT CGNs (Fig. 4A) (31, 35). In In summary, the proportion of AIS expressing Na 1.6 was contrast, the proportion of Ank -positive AIS expressing Na 1.6 v G v significantly reduced in the absence of β1. In addition, there was was reduced in Scn1b null CGNs (P < 0.001; Fig. 4 A and E). Our an increase in the proportion of AIS expressing Nav1.1 in Scn1b data cannot rule out the possibility that Nav1.6 may still be weakly Scn1b null CGNs with no changes in Nav1.2. Our electrophysiological expressed at the AIS of null CGNs, albeit at levels below the data, together with the observation that β1 is expressed at the limit of detection. Nonetheless, our results suggest that β1is fi AIS of WT CGNs and Purkinje neurons, support the hypothesis required for ef cient localization of Nav1.6 at the AIS. We propose Scn1b β that impaired excitability of null CGNs is caused by both that 1 may provide an additional link between AnkG and Nav1.6 altered I and altered VGSC expression. during AIS maturation. In Scn8a null mice, there is a neuron- Na fi speci c compensatory increase in Nav1.1 and Nav1.2 at the AIS A Model for Activity-Dependent Neurite Outgrowth Regulated by β1

(8). We investigated whether a similar up-regulation occurs in and Nav1.6. The present results, combined with previous work, show Scn1b null CGNs. Both WT and Scn1b null CGNs had strong pan- that β1-mediated neurite outgrowth requires fyn kinase, contactin, α B E subunit immunoreactivity at the AIS (Fig. 4 and ). In addi- Nav1.6, and INa (22). We hypothesized that a complex containing tion, Nav1.2 was robustly expressed at the AIS of both WT and these may be present at the growth cone. We found that Scn1b null CGNs (Fig. 4 C and E). In contrast, there was an up- whereas VGSC α subunits and β1 were most highly expressed at the regulation of Nav1.1-positive AIS in Scn1b null CGNs (P < 0.001; AIS of 14 DIV WT CGNs, they also were expressed in the soma, Fig. 4 D and E). We concluded that Nav1.1, but not Nav1.2, com- along the neurite, and at the growth cone, defined by phalloidin pensated for the loss of Nav1.6 at the AIS of Scn1b null CGNs. labeling of F-actin (Fig. 6A) (22). Similar to β1, contactin was

Fig. 5. Nav1.6 expression is reduced at the AIS of Scn1b null Purkinje neurons. (A)100× Z-series confocal projections of WT (i) and Scn1b null (ii) P14 cerebellum. AIS are labeled with

anti-AnkG (red) and anti-Nav1.6 (green). Insets,3× zoom of white boxes, showing colocalization of AnkG and Nav1.6 at Purkinje neuron AIS in WT, but not in Scn1b null. Arrows

indicate Scn1b null Purkinje neuron AIS expressing both AnkG and Nav1.6. (B) Proportion of AnkG-expressing Purkinje neu- ron AIS that also express Nav1.6 for WT (dark bars) and Scn1b null (light bars) littermates (n =60fields of view from three mice of each genotype). Data are mean ± SEM. ***P < 0.001, t test. (C)100× Z-series confocal projection of WT Purkinje neuron in situ labeled with anti-calbindin (red) and anti-β1 (green). The arrow shows β1atAIS.(D)100× Z-series confocal projection of WT 14 DIV CGN labeled with anti-pan VGSC α subunit (red) and anti-β1 (green). The arrow indicates β1at AIS. (Scale bars: 20 μm.)

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.0909434107 Brackenbury et al. Downloaded by guest on September 26, 2021 localized to the soma, along the neurite, and at the growth cone (Fig. (Fig. 7A) (11, 22, 25). At this stage, the AIS is not yet formed, + 6B). Interestingly, whereas α subunit immunoreactivity was clearly and Nav1.6 is proposed to conduct Na into microdomains highest at the AIS, contactin was fairly constant along the axon (Fig. surrounding channel complexes in response to localized mem- 6B). The arrangement of α subunits, β1, and contactin at the AIS, brane depolarizations during neurite extension. In our model, this + fl fi β along the axon, and at the growth cone is consistent with their localized Na in ux is speci cally required for 1-mediated neu- A functions within complexes in these regions that regulate electrical rite outgrowth (Fig. 7 ). Furthermore, in this model VGSC complexes are present along the entire neurite, permitting adhe- activity on the one hand and neurite outgrowth on the other hand. sion and β1-dependent fasciculation in vivo (22). β1-independent, We propose that in immature CGNs, including those used in FGF-mediated neurite outgrowth also would occur. our neurite outgrowth experiments, Nav1.6, β1, contactin, fyn β1 is required for normal localization of Nav1.6 at the AIS, for kinase and AnkG are present along the neurite and at the growth resurgent INa, and for repetitive firing. We postulate that reci- cone to promote β1-mediated neurite extension and migration procity of function occurs between β1 and Nav1.6, in which β1is required for Nav1.6-dependent high-frequency firing on the one hand, and Nav1.6 is required for β1-mediated neurite outgrowth on the other hand. Given that both Scn1b and Scn8a null mice die shortly after the transition to high-frequency firing should it occur, this unique reciprocal relationship between β1 and Nav1.6 likely is critical to postnatal development (8, 13). Thus, we propose a second, more complex scenario in CGNs in vivo in which the AIS has formed, represented by the 14 DIV CGNs. In this model, complexes of α subunits, β1, contactin, fyn kinase, and AnkG are expressed along the entire axon, but are con- NEUROSCIENCE

Fig. 7. A model for INa involvement in β1-mediated neurite outgrowth. (A) In immature CGNs lacking AIS, complexes containing Nav1.6, β1, and con- tactin are present throughout the neuronal membrane in the soma, neurite, and growth cone. Localized Na+ influx is necessary for β1-mediated neurite extension and migration (11, 22, 25). VGSC complexes along the neurite are proposed to participate in cell–cell adhesion and fasciculation (22). (B) In 14

DIV CGNs, β1 is also required for Nav1.6 expression at the AIS, and sub- sequent high-frequency AP firing through modulation of resurgent INa (8, 10, 45). Electrical activity may further promote β1-mediated neurite out- growth at or near the growth cone in vivo (8, 10, 45). Thus, the devel- + opmental functions of β1 and Nav1.6 are complementary, such that Na influx carried by Nav1.6 is required for β1-mediated neurite outgrowth and Fig. 6. α subunits, β1, and contactin are present at the growth cone. 100× Z- β1 is required for normal expression/activity of Nav1.6 at the AIS. Fyn kinase series confocal projections of WT 14 DIV CGNs labeled with pan-α subunit and AnkG also are likely present in all complexes (11), but they are only (green), anti-β1(A; red) or anti-contactin (B; red), and Alexa 594–conjugated shown once in each panel for clarity. The FGF-mediated, β1-independent phalloidin (magenta). (Scale bar: 10 μm.) he panels on the right show 4× neurite outgrowth pathway is shown as well. Other CAMs that regulate digital zoom views highlighting growth cone (arrows). These distributions neurite outgrowth and also may interact with β1 in this system (11, 26), have were observed in all three mice studied. been omitted for clarity.

Brackenbury et al. PNAS Early Edition | 5of6 Downloaded by guest on September 26, 2021 centrated at the AIS and the growth cone to promote β1-mediated Neurite Outgrowth. Neurite outgrowth assays were conducted as described neurite extension and migration (Fig. 7B). As in the immature previously (22). Additional details are provided in SI Materials and Methods. + CGNs, here Nav1.6 permits localized Na influx required for β1- mediated neurite outgrowth. In addition, electrical activity gen- Cerebellar Slice Recording. Cerebellar slices were prepared and electro- physiological recordings performed and analyzed as described previously (12, erated at the AIS may provide a depolarizing signal to open Nav1.6 channels at the growth cone, further promoting β1-mediated 30, 44). Additional details are provided in SI Materials and Methods.

neurite outgrowth (22). Interestingly, CGNs do not display spon- + taneous APs (43), and thus synaptic input likely would be required Whole-Cell Recording of Na Currents in CGNs in Vitro. Cerebellar tissue from P12–14 mice was dissociated as described previously (22). Voltage clamp for activity-dependent β1-mediated neurite outgrowth in vivo. fi β recordings were performed using standard methods. Additional details are In conclusion, our ndings indicate that 1 is required for given in SI Materials and Methods. high-frequency firing in CGNs, by regulating localization of Nav1.6 to the AIS and contributing to resurgent INa. We propose Immunocytochemistry and Immunohistochemistry. Procedures for antibody + fl β that localized Na in ux carried by Nav1.6 is required for 1- labeling and confocal microscropy have been described previously (22). β mediated neurite outgrowth. In conclusion, 1 and Nav1.6 per- Additional details are given in SI Materials and Methods and Figs. S4 and S5. form complementary roles, interacting from within distinct regions in the neuron to regulate excitability and neurite exten- Western Blot Analysis. SDS/PAGE and transfer to nitrocellulose were per- sion in a coordinated fashion. As a result, VGSCs function in formed as described previously (22). Additional details are provided in macromolecular complexes that participate in signaling on SI Materials and Methods. multiple time scales to regulate excitability, adhesion, neurite outgrowth, and migration in the developing CNS. ACKNOWLEDGMENTS. We thank Dr. Miriam Meisler for providing the Medtg mice, T. J. O’Shea for technical assistance, and Drs. Luis Lopez-Santiago Materials and Methods and Yukun Yuan for helpful discussions. This work was supported by National Institutes of Health (NIH) Grant R01 MH059980 (to L.L.I.), Animals. Mice were maintained in accordance with the guidelines of the National Multiple Sclerosis Society Grant RG2882 (to L.L.I.), NIH Grant University of Michigan Committee on the Use and Care of Animals. Details of R01 NS38297 (to B.R.), and a University of Michigan Organogenesis Post- mouse strains are given in SI Materials and Methods. doctoral Fellowship (to W.J.B.).

1. Catterall WA (2000) From ionic currents to molecular mechanisms: The structure and 25. Brackenbury WJ, Djamgoz MB, Isom LL (2008) An emerging role for voltage-gated function of voltage-gated sodium channels. Neuron 26:13–25. Na+ channels in cellular migration: Regulation of central nervous system development 2. Hille B (1992) Ionic Channels of Excitable Membranes (Sinauer Associates, Sunderland, and potentiation of invasive cancers. Neuroscientist 14:571–583. MA), 2nd Ed. 26. Maness PF, Schachner M (2007) Neural recognition molecules of the immunoglobulin 3. Goldin AL, et al. (2000) Nomenclature of voltage-gated sodium channels. Neuron 28:365–368. superfamily: Signaling transducers of axon guidance and neuronal migration. Nat 4. Trimmer JS, Rhodes KJ (2004) Localization of voltage-gated ion channels in mammalian Neurosci 10:19–26. brain. Annu Rev Physiol 66:477–519. 27. McEwen DP, Meadows LS, Chen C, Thyagarajan V, Isom LL (2004) 5. Vacher H, Mohapatra DP, Trimmer JS (2008) Localization and targeting of voltage- beta1 subunit–mediated modulation of Nav1.2 currents and cell surface density is dependent ion channels in mammalian central neurons. Physiol Rev 88:1407–1447. dependent on interactions with contactin and ankyrin. J Biol Chem 279:16044–16049. 6. Burgess DL, et al. (1995) Mutation of a new sodium channel , Scn8a, in the mouse 28. Berglund EO, et al. (1999) Ataxia and abnormal cerebellar microorganization in mice mutant “motor endplate disease.” Nat Genet 10:461–465. with ablated contactin gene expression. Neuron 24:739–750. 7. Kohrman DC, et al. (1995) Insertional mutation of the motor endplate disease (MED) 29. Raman IM, Bean BP (1999) Ionic currents underlying spontaneous action potentials in locus on mouse 15. Genomics 26:171–177. isolated cerebellar Purkinje neurons. J Neurosci 19:1663–1674. 8. Van Wart A, Matthews G (2006) Impaired firing and cell-specific compensation in 30. Goldfarb M, et al. (2007) Fibroblast growth factor homologous factors control neurons lacking nav1.6 sodium channels. J Neurosci 26:7172–7180. neuronal excitability through modulation of voltage-gated sodium channels. Neuron 9. Rush AM, Dib-Hajj SD, Waxman SG (2005) Electrophysiological properties of two 55:449–463. axonal sodium channels, Nav1.2 and Nav1.6, expressed in mouse spinal sensory 31. Osorio N, et al. (2005) Differential targeting and functional specialization of sodium neurones. J Physiol 564:803–815. channels in cultured cerebellar granule cells. J Physiol 569:801–816. 10. Raman IM, Sprunger LK, Meisler MH, Bean BP (1997) Altered subthreshold sodium 32. Afshari FS, et al. (2004) Resurgent Na currents in four classes of neurons of the cerebellum. currents and disrupted firing patterns in Purkinje neurons of Scn8a mutant mice. JNeurophysiol92:2831–2843. Neuron 19:881–891. 33. Grieco TM, Malhotra JD, Chen C, Isom LL, Raman IM (2005) Open-channel block by the 11. Brackenbury WJ, Isom LL (2008) Voltage-gated Na+ channels: Potential for beta cytoplasmic tail of sodium channel β4 as a mechanism for resurgent sodium current. subunits as therapeutic targets. Expert Opin Ther Targets 12:1191–1203. Neuron 45:233–244. 12. Lopez-Santiago LF, et al. (2007) Sodium channel Scn1b null mice exhibit prolonged QT 34. Aman TK, et al. (2009) Regulation of persistent Na current by interactions between β and RR intervals. J Mol Cell Cardiol 43:636–647. subunits of voltage-gated Na channels. J Neurosci 29:2027–2042. 13. Chen C, et al. (2004) Mice lacking sodium channel β1 subunits display defects in neuronal 35. Zhou D, et al. (1998) AnkyrinG is required for clustering of voltage-gated Na channels excitability, sodium channel expression, and nodal architecture. J Neurosci 24:4030–4042. at axon initial segments and for normal action potential firing. J Cell Biol 143: 14. Wallace RH, et al. (2002) Generalized epilepsy with febrile seizures plus: Mutation of 1295–1304. the sodium channel subunit SCN1B. Neurology 58:1426–1429. 36. Jenkins SM, Bennett V (2001) Ankyrin-G coordinates assembly of the spectrin-based 15. Wallace RH, et al. (1998) Febrile seizures and generalized epilepsy associated with a membrane skeleton, voltage-gated sodium channels, and L1 CAMs at Purkinje neuron mutation in the Na+-channel beta1 subunit gene SCN1B. Nat Genet 19:366–370. initial segments. J Cell Biol 155:739–746. 16. Audenaert D, et al. (2003) A deletion in SCN1B is associated with febrile seizures and 37. McEwen DP, Isom LL (2004) Heterophilic interactions of sodium channel beta1 early-onset absence epilepsy. Neurology 61:854–856. subunits with axonal and glial cell adhesion molecules. J Biol Chem 279:52744–52752. 17. Scheffer IE, et al. (2007) Temporal lobe epilepsy and GEFS+ phenotypes associated 38. Dzhashiashvili Y, et al. (2007) Nodes of Ranvier and axon initial segments are ankyrin with SCN1B mutations. Brain 130:100–109. G–dependent domains that assemble by distinct mechanisms. J Cell Biol 177:857–870. 18. Meadows LS, et al. (2002) Functional and biochemical analysis of a sodium channel β1 39. Hedstrom KL, et al. (2007) Neurofascin assembles a specialized extracellular matrix at subunit mutation responsible for generalized epilepsy with febrile seizures plus type the axon initial segment. J Cell Biol 178:875–886. 1. J Neurosci 22:10699–10709. 40. Black JA, et al. (1999) Abnormal expression of SNS/PN3 sodium channel in cerebellar 19. Watanabe H, et al. (2008) Sodium channel beta1 subunit mutations associated with Purkinje cells following loss of myelin in the taiep rat. Neuroreport 10:913–918. and cardiac conductiondisease in humans. JClinInvest118:2260–2268. 41. Black JA, et al. (2000) Sensory neuron–specific sodium channel SNS is abnormally 20. Watanabe H, et al. (2009) Mutations in sodium channel β1- and β2-subunits associated expressed in the brains of mice with experimental allergic encephalomyelitis and with atrial fibrillation. Circ Arrhythmia Electrophysiol 2:268–275. humans with multiple sclerosis. Proc Natl Acad Sci USA 97:11598–11602. 21. Patino GA, et al. (2009) A functional null mutation of SCN1B in a patient with Dravet 42. Schaller KL, Caldwell JH (2003) Expression and distribution of voltage-gated sodium syndrome. J Neurosci 29:10764–10778. channels in the cerebellum. Cerebellum 2:2–9. 22. Brackenbury WJ, et al. (2008) Voltage-gated Na+ channel β1 subunit–mediated 43. D’Angelo E, De Filippi G, Rossi P, Taglietti V (1998) Ionic mechanism of electroresponsiveness neurite outgrowth requires Fyn kinase and contributes to central nervous system in cerebellar granule cells implicates the action of a persistent sodium current. J development in vivo. J Neurosci 28:3246–3256. Neurophysiol 80:493–503. 23. Davis TH, Chen C, Isom LL (2004) Sodium channel β1 subunits promote neurite 44. Yuan Y, Atchison WD (2003) Methylmercury differentially affects GABA(A) receptor– outgrowth in cerebellar granule neurons. J Biol Chem 279:51424–51432. mediated spontaneous IPSCs in Purkinje and granule cells of rat cerebellar slices. J 24. Fein AJ, Wright MA, Slat EA, Ribera AB, Isom LL (2008) scn1bb, a zebrafish ortholog of Physiol 550:191–204. SCN1B expressed in excitable and nonexcitable cells, affects motor neuron axon 45. Raman IM, Bean BP (1997) Resurgent sodium current and action potential formation morphology and touch sensitivity. J Neurosci 28:12510–12522. in dissociated cerebellar Purkinje neurons. J Neurosci 17:4517–4526.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.0909434107 Brackenbury et al. Downloaded by guest on September 26, 2021