A Targeting Motif Involved in Sodium Channel Clustering at the Axonal Initial Segment

See allHide authors and affiliations

Science  27 Jun 2003:
Vol. 300, Issue 5628, pp. 2091-2094
DOI: 10.1126/science.1085167


The sorting of sodium channels to axons and the formation of clusters are of primary importance for neuronal electrogenesis. Here, we showed that the cytoplasmic loop connecting domains II and III of the Nav1 subunit contains a determinant conferring compartmentalization in the axonal initial segment of rat hippocampal neurons. Expression of a soluble Nav1.2II-III linker protein led to the disorganization of endogenous sodium channels. The motif was sufficient to redirect a somatodendritic potassium channel to the axonal initial segment, a process involving association with ankyrin G. Thus, this motif may play a fundamental role in controlling electrical excitability during development and plasticity.

Neuronal action potentials are generated at the axonal initial segment (AIS), and their saltatory conduction in myelinated axons occurs via the nodes of Ranvier (1, 2). These processes require a precise distribution of voltage-gated sodium channels accumulated at high density in these two specific membrane microdomains defined by the segregation of the ankyrin G–βIV spectrin complex (36). At the molecular level, sodium channels from rat brain are composed of an α subunit, the pore-forming protein Nav1, and the auxiliary subunits, β2 and β1 or β3 (2). The fact that β subunits interact with proteins of the L1 family of adhesion molecules (7, 8), with ankyrin G (9), and with tenascin R and C (2) raises the possibility that β subunits contribute to sodium channel localization at specific sites in the neuronal membrane. However, tagged β subunits expressed in hippocampal neurons are uniformly distributed (10), indicating that information for sorting and clustering of sodium channels is probably carried by the pore-forming protein Nav1.

In the central nervous system, Nav1.2 sodium channels are distributed on unmyelinated axons, including the AIS (11, 12). To define the molecular determinant accounting for sodium channel compartmentalization at the AIS, we assessed whether any of the large intracellular regions of Nav1.2 contained sufficient information for sorting and specific membrane organization. We analyzed the surface distribution of chimeric proteins containing CD4 fused to one of the cytoplasmic regions of Nav1.2 (Fig. 1A) (13) after transfection in rat hippocampal neurons. During in vitro development, these cells have the intrinsic ability to form an AIS acting as a diffusion barrier (14, 15) and to accumulate sodium channels in the proximal region of the axon (16), as observed in cultured spinal motoneurons (17). A chimera containing the C terminus of Nav1.2 (CD4–Nav1.2 Ct) is targeted to axons (13). However, at the steady state, the surface distribution of CD4–Nav1.2 Ct is located in more distal regions of the axon than is ankyrin G (13). Thus, the C terminus of Nav1.2 does not contain sufficient information for sodium channel localization at the AIS. We further analyzed the other cytoplasmic regions. The addition of the II-III linker from the Nav1.2 subunit to the human CD4 receptor deleted of its cytoplasmic tail (CD4ΔCt; Fig. 1B) resulted in a surface distribution of the chimera markedly restricted to the AIS (Fig. 1C), which was identified by the absence of a somatodendritic marker, MAP2, and by staining for ankyrin G (3) (Fig. 1D). In contrast, CD4ΔCt was uniformly distributed at the cell surface (Fig. 1C). Endogenous sodium channels and CD4–Nav1.2 II-III were nonextractable by detergent before cell fixation and immunostaining (Fig. 1, E and F), like the axonal cell adhesion molecule L1 (14).

Fig. 1.

The cytoplasmic II-III region of Nav1.2 contains an AIS localization signal. (A) Schematic representation of the neuronal sodium channel Nav1.2. Four homologous domains (I to IV) are connected by large cytoplasmic linkers (I-II, II-III, and III-IV); color code as in (B). (B) Diagram of CD4ΔCt and CD4–Nav1.2 II-III constructs. The cytoplasmic II-III linker of Nav1.2 (amino acids 983 to 1203; green) was fused to the human CD4 receptor with a deleted C terminus (CD4ΔCt). (C and D) CD4ΔCt displays a nonpolarized surface distribution, whereas the surface localization of CD4–Nav1.2 II-III is restricted to the AIS of transfected rat hippocampal neurons. (E and F) Like endogenous sodium channels, CD4–Nav1.2 II-III is resistant to Triton X-100 extraction before cell fixation and immunostaining. Somatodendritic domains and AIS were identified by staining for MAP2 (C) and ankyrin G (D to F). Scale bar, 20 μm.

To confirm a role for the II-III linker in organizing sodium channels at the AIS, we examined whether its overexpression perturbed endogenous sodium channels. Na+ current was measured by whole-cell patch-clamp recording on hippocampal neurons expressing either green fluorescent protein (GFP) or a chimera in which Nav1.2 II-III was fused to the C terminus of GFP (GFP–Nav1.2 II-III). The peak amplitudes of outward Na+ currents recorded in GFP–Nav1.2 II-III–positive neurons were reduced by 50% (n = 29) relative to neurons expressing GFP alone (n = 30) (Fig. 2, A and B). No modification of inactivation kinetics or of the voltage dependence of activation or inactivation was observed (fig. S1). The distribution of GFP–Nav1.2 II-III differed remarkably from that of GFP. The latter was uniformly distributed throughout dendrites, axons, and soma (Fig. 2C). In contrast, GFP–Nav1.2 II-III was highly concentrated at the AIS but absent from dendrites, although faint fluorescence was detected in soma (Fig. 2C). This distribution was displayed by 71.5% of transfected cells (n = 76). We next quantified the number of AISs positive for sodium channel clustering. Nearly all hippocampal neurons expressing GFP alone (96%; n = 69) displayed staining of endogenous sodium channels at the AIS (Fig. 2C, inset). In contrast, only 14% of cells (n = 76) expressing GFP–Nav1.2 II-III exhibited detectable sodium channel staining. These findings suggest that the II-III linker is critical for sodium channel organization at the AIS.

Fig. 2.

Perturbation of sodium channel organization by expression of GFP–Nav 1.2 II-III. (A and B) Effect of GFP–Nav1.2 II-III expression on the peak amplitude of Na+ current. Whole-cell outward Na+ currents were recorded from hippocampal neurons expressing either GFP or GFP–Nav1.2 II-III. Superimposed current traces evoked by depolarization from –90 mV to +40 mV are shown. (C) GFP was visualized in soma and throughout dendrites and axons (arrowhead), whereas GFP–Nav1.2 II-III was concentrated in the AIS. Inset: Clustering of endogenous channels was immunodetected in GFP-positive neurons but was absent in GFP–Nav1.2 II-III–positive neurons. Scale bar, 20 μm.

To identify the critical determinant at the sequence level, we generated mutants of CD4–Nav1.2 II-III containing either a truncated C terminus or internal deletions (Fig. 3A). Analysis of the steady-state distribution of mutants defined a 27-residue motif between amino acids 1102 and 1028 that governed chimera compartmentalization at the AIS (Fig. 3A). The AIS motif contained clusters of acidic residues similar to sorting motifs that regulate the trafficking of the endoprotease furin (18) and aquaporin 4 (19). The critical role of the AIS motif was further supported by an internal deletion (Δ1098-1111; Fig. 3, A and C). Impaired compartmentalization of mutated CD4–Nav1.2 II-III chimeras in the AIS was systematically associated with a loss of resistance to detergent extraction (Fig. 3A). This suggests that the sorting motif we have identified may also act as a retention motif, as in the case of furin (18) and the vesicular monoamine transporter VAMT2 (20). The AIS motif identified in Nav1.2 is highly conserved in Nav1.6, the nodal sodium channel type (21), and in the neuronal sodium channels Nav1.1 and Nav1.3 (22) (Fig. 3B). It is also relatively well conserved in Nav1.4 and Nav1.5, the skeletal and cardiac muscle sodium channels (22). Nav1.6 was consistently visualized at the AIS of cultured hippocampal neurons (Fig. 3D), as reported in vivo in Purkinje neurons (5). When the cytoplasmic II-III linker of Nav1.6 was fused to CD4, the resulting chimera was highly enriched at the AIS (Fig. 3E) (80% of cells displayed cell surface CD4 staining restricted to the AIS; n = 429, seven independent experiments) and was resistant to detergent extraction. During the biogenesis of nodes of Ranvier, Nav1.2 is replaced by Nav1.6 (23, 24). It is thus conceivable that the AIS motif is involved in sodium channel clustering at the nodes of Ranvier and possibly in the unexpected localization of neuronal sodium channels in the heart (25).

Fig. 3.

Identification of the AIS motif of neuronal sodium channels. (A) Schematic representation of mutations generated in Nav1.2 II-III to define the AIS motif. For each mutant, the percentage of transfected CD4-positive hippocampal neurons in which staining was restricted to the AIS is indicated, taking as 100% the total population of transfected neurons (i.e., CD4-positive neurons). Data are means ± SD from three to seven different experiments; n denotes the total number of neurons analyzed for each mutation. In parallel, the resistance of mutated proteins to detergent extraction (Triton X-100) was determined. (B) Alignment of the AIS motif sequences in rat neuronal voltage-dependent sodium channels. Gray shading indicates similarity between amino acids; nonconserved residues are underlined. Amino acid abbreviations: A, Ala; D, Asp; E, Glu; F, Phe; G, Gly; I, Ile; L, Leu; M, Met; N, Asn; P, Pro; R, Arg; S, Ser; T, Thr; V, Val. (C) Surface expression of CD4–Nav1.2 II-III mutants. (D) Nav1.6 clustering at the AIS of 4-week-old hippocampal neurons. (E) Addition of the intracellular II-III region of Nav1.6 restricts surface distribution of CD4 to the AIS. Scale bar, 20 μm.

We next investigated whether the AIS motif of Nav1.2 was sufficient to relocalize a potassium channel in hippocampal neurons. We constructed a hemagglutinin (HA)–tagged Kv2.1-Nav1.2 chimera (Fig. 4A) in which the region encompassing the proximal restriction clustering (PRC) (26) motif was replaced by amino acids 1082 to 1203 of Nav1.2 II-III. HA-Kv2.1 displayed a somatodendritic distribution with “donut”-shaped clusters (Fig. 4A) (26). In contrast, HA–Kv2.1-Nav1.2 was restricted to the AIS (Fig. 4A) (97% of transfected cells; n = 194). We confirmed that HA–Kv2.1-Nav1.2 was inserted at the cell surface and was resistant to Triton X-100 extraction (figs. S2 and S3). Genetic studies have shown that the suppression of ankyrin G gene expression impaired sodium channel localization at the AIS (3, 5). Thus, we next addressed whether the AIS motif associated with ankyrin G by coexpressing 270-kD ankyrin G–GFP (27) either with HA–Kv2.1-Nav1.2 or HA-Kv2.1 in HEK cells. Ankyrin G–GFP coimmunoprecipitated with HA–Kv2.1-Nav1.2, but not with HAKv2.1 (Fig. 4B). The two partners colocalized in the plasma membrane and also intracellularly (Fig. 4C). These data provide evidence for an association between 270-kD ankyrin G–GFP and the AIS motif, a finding that was not observed with an in vitro binding assay (28). In hippocampal neurons, overexpressed ankyrin G–GFP was localized at the plasma membrane of soma, in addition to its normal distribution at the AIS (Fig. 4D) (27). In addition to colocalization in the AIS, coexpressed HA–Kv2.1-Nav1.2 was associated with mislocalized somatic ankyrin G and somatic clusters were clearly visible (Fig. 4D). This was observed in 80% of the cells coexpressing ankyrin G–GFP and HA–Kv2.1-Nav1.2 (n = 391, two independent experiments). In contrast, HA–Kv2.1-Nav1.2 remained confined in the AIS in 76% of cells coexpressing GFP alone (n = 406). In cells coexpressing HA-Kv2.1 and ankyrin G–GFP, somatic clusters of ankyrin G did not mirror HA-Kv2.1 aggregates (Fig. 4D). Thus, overexpression of ankyrin G resulted in a mislocalization of HA–Kv2.1-Nav1.2 in soma.

Fig. 4.

Involvement of ankyrin G in targeting membrane proteins bearing the AIS motif. (A) The AIS motif of Nav1.2 is sufficient to target the voltage-dependent channel Kv2.1 to the AIS. Top: Schematic representations of HA-Kv2.1 and HA–Kv2.1-Nav1.2. A segment of the C terminus of Kv2.1 encompassing the PRC motif was substituted by a segment of Nav1.2 containing the AIS motif. Bottom: In permeabilized neurons, HA-Kv2.1 was distributed both in soma and proximal dendrites, whereas HA–Kv2.1-Nav1.2 was restricted to the AIS (identified by ankyrin G staining). (B) Coimmunoprecipitation of 270-kD ankyrin G–GFP and HA–Kv2.1-Nav1.2. HEK cells were cotransfected either with ankyrin G–GFP and HA–Kv2.1-Nav1.2 or with ankyrin G–GFP and HA-Kv2.1, and then extracted with detergent. Extracts were subjected to immunoprecipitation with antibody to HA. Immunoprecipitation inputs and pellets were analyzed by Western blotting with antibody to GFP. (C) Colocalization of ankyrin G–GFP and HA–Kv2.1-Nav1.2 in HEK cells. (D) Overexpression of ankyrin G–GFP induces a mislocalization of HA–Kv2.1-Nav1.2. Hippocampal neurons were cotransfected with the indicated constructs and subjected to HA immunostaining 1 day later. Colocalization of ankyrin G–GFP and HA–Kv2.1-Nav1.2 is evident not only at the AIS (arrows) but also in the soma. GFP expression did not affect segregation of HA–Kv2.1-Nav1.2 at the AIS, whereas somatodendritic HA-Kv2.1 clusters did not colocalize with ankyrin G–GFP. Scale bar, 20 μm.

Our results offer insights into the intrinsic signals for sodium channel targeting and clustering at discrete sites in the neuronal plasma membrane. We propose that segregation is specified by a 27–amino acid motif within linker II-III of the pore-forming Nav1 proteins. This signal is sufficient to relocalize the somatodendritic cation channel Kv2.1, as well as the cytosolic protein GFP, to the AIS. However, additional motifs probably play a role in establishing the differential distribution of certain types of sodium channel in vivo. We also obtained evidence for the association of ankyrin G with the motif we have identified. From a mechanistic point of view, ankyrin G could be primarily involved in anchoring sodium channels at the plasma membrane. However, we cannot exclude the possibility that the two partners could be preassembled early in biogenesis and cotransported during sorting, as observed in the case of presynaptic proteins (29, 30).

Supporting Online Material

Figs. S1 to S3

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article