|Home | About | Journals | Submit | Contact Us | Français|
Voltage-gated Na+ channel (VGSC) β1 and β2 subunits are multifunctional, serving as both channel modulators and cell adhesion molecules (CAMs). The purpose of this study was to determine whether VGSC β3 subunits function as CAMs. The β3 extracellular domain is highly homologous to β1, suggesting that β3 may also be a functional CAM. We investigated the trans homophilic cell adhesive properties of β3, its association with the β1-interacting CAM contactin, as well as its ability to interact with the cytoskeletal protein ankyrin. Our results demonstrate that, unlike β1, β3 does not participate in trans homophilic cell-cell adhesion or associate with contactin. Further, β3 does not associate with ankyrinG in a heterologous system. Previous studies have shown that β3 interacts with the CAM neurofascin-186 but not with VGSC β1. Taken together, these findings suggest that, although β1 and β3 exhibit similar channel modulatory properties in heterologous systems, these subunits differ with regard to their homophilic and heterophilic CAM binding profiles.
Voltage-gated ion channels are multi-functional . In addition to regulating electrical excitability through ion conduction, some voltage-gated ion channels contribute to processes as diverse as intracellular signaling, transcriptional regulation, scaffolding, and cell adhesion without requiring changes in ion flux [5, 6, 24]. For example, VGSC β subunits regulate channel transcription, cell surface expression, and subcellular localization, modulate channel currents, and participate in cell-cell adhesion [4, 8, 14, 17, 22, 25, 31, 33].
VGSCs isolated from mammalian neurons are heterotrimers, composed of a single α subunit, one non-covalently linked β subunit (β1 or β3), and one disulfide-linked β subunit (β2 or β4) [7, 31]. VGSC α subunit cDNAs express functional channels in heterologous systems. However, for tetrodotoxin-sensitive α subunits, the currents characteristic of these channels expressed in isolation are different from native currents. Co-expression of β with α subunits in heterologous systems shifts the voltage-dependence of activation and inactivation, changes the rates of inactivation and recovery from inactivation, and increases channel cell surface expression [5, 6, 31]. VGSC β subunits and the β subunits of Ca2+ and K+ channels are functionally homologous in terms of channel modulation [2, 19]. However, Ca2+ and K+ channel β subunits are not structurally homologous to VGSC β subunits. Of this group, only the VGSC β subunits contain extracellular Ig domains [18, 19] and only the VGSC β1 and β2 subunits have been shown to function as cell adhesion molecules (CAMs) in addition to their roles in channel modulation [5, 6]. An important question is whether all of the four known VGSC β subunits function as CAMs.
β1 and β2 function as CAMs: β1 and β2 interact with the extracellular matrix protein tenascin to influence cell migration [37, 41]; β1 and β2 participate in trans homophilic cell-cell adhesion resulting in cellular aggregation and ankyrin recruitment [26, 27]; the β1 extracellular Ig domain interacts with the neuronal and glial CAMs contactin, VGSC β2, neurofascin-155, neurofascin-186, and NrCAM, but interestingly not with VGSC β3 [21, 29, 30]; β1 interactions with contactin and neurofascin-186 result in increased channel cell surface expression [21, 29, 30]; and β1 promotes neurite extension as a result of homophilic adhesion [4, 12].
The purpose of the present study was to determine whether VGSC β3 subunits function as CAMs. The β3 extracellular domain is highly homologous to β1, suggesting that β3 may also be a functional CAM. Consistent with this hypothesis, β3 associates with the CAM neurofascin-186 [29, 34]. We investigated the trans homophilic cell adhesive properties of β3, its association with the β1-interacting CAM contactin, as well as its ability to interact with the cytoskeletal protein ankyrin. Our results demonstrate that, unlike β1, β3 does not participate in trans homophilic cell-cell adhesion or associate with contactin. Further, β3 does not associate with ankyrinG in a heterologous system. These findings are relevant to recent findings showing that the Scn3b null phenotype is subtle and that the null mice are viable . This is in contrast to the severe neurological phenotype of Scn1b null mice , supporting the hypothesis that physiological role of β1 in vivo may include both channel modulation and cell-cell adhesion, while the mechanism of β3 function may be less dependent on cell adhesion and more limited to channel modulation.
All experimental procedures are included in the on-line supplemental materials.
To characterize the full scope of β3 function, we generated a GFP-tagged β3 plasmid, allowing for β3 detection using either β3 antiserum or a commercial anti-GFP antibody. We co-expressed β3-GFP with Nav1.2 to determine if the addition of the epitope tag interfered with β3-mediated channel modulation. β3-GFP cRNA was coinjected with Nav1.2 cRNA in Xenopus oocytes. As shown in Supplemental Fig. 1 and Supplemental Table 1, β3-GFP modulated Na+ currents by shifting the half-voltage of inactivation in the hyperpolarizing direction, increasing the rate of inactivation, and increasing the peak current amplitude compared with currents expressed by Nav1.2 alone (Nav1.2: −1035 nA, Nav1.2 + β3-GFP: −2160 nA), similar to previous reports with untagged β3 [32, 35]. Thus, we used β3 and β3-GFP interchangeably in all subsequent experiments.
We used Drosophila S2 cells previously to demonstrate that VGSC β1 and β2 subunits function in trans homophilic cell-cell adhesion resulting in ankyrin recruitment [26, 27]. S2 cells are ideal for this type of experiment, as they are free-floating in suspension culture, and express no endogenous CAMs. cDNAs encoding putative CAMs to be tested are transfected into S2 cells under an inducible promoter and clonal lines established. Induction of protein expression followed by mechanical shaking allows a detailed analysis of the time course of cellular aggregation. To investigate the trans homophilic cell adhesive properties of β3, we expressed β3 or β3-GFP in Drosophila S2 cells and established stable, clonal cell lines using soft agar cloning techniques as in . We verified expression of β3 or β3-GFP in these clonal lines by Western blot analysis with anti-β3 or anti-GFP, respectively (Fig. 1A, anti-β3; anti-GFP data not shown). Immunocytochemical analysis of S2-β3 and S2-β3-GFP cell lines using anti-β3 or anti-GFP antibodies showed robust β3 expression at the cell surface (Fig. 1B, anti-β3; anti-GFP data not shown). In spite of this expression, however, we were not able to detect β3-mediated cellular aggregation under any condition, including increasing the cell density in the assay, indicating that unlike β1, β3 does not mediate trans homophilic cell adhesion in S2 cells. In parallel experiments and under the same conditions, S2-β1 cells exhibited efficient aggregation as described previously [26, 27] (data not shown), ensuring that the cell density in the experiment was sufficient to allow aggregation to occur.
We next asked whether β3 could interact with the neuronal and glial CAM contactin, as shown for β1 [21, 29, 30]. CHL cells stably expressing contactin were stably cotransfected with β1, β2, or β3-GFP. Triton X-100 solubilized cell lysates were immunoprecipitated with either non-immune IgG or anti-β-subunit antibodies. Contactin could be co-immunoprecipitated only from cells expressing β1 but not β2, in agreement with previous results , or β3-GFP (Fig. 2A). To determine whether the presence of a VGSC α subunit altered the ability of β3 to associate with contactin, we immunoprecipitated solubilized cells expressing Nav1.2, β3, and contactin (Fig. 2B). While β3-GFP could be immunoprecipitated with anti-β3, anti-contactin was not able to precipitate β3, demonstrating that β3 and contactin do not associate either in the presence or absence of the ion-conducting pore. While these data demonstrated that β3 and contactin do not associate in a heterologous system, they did not indicate whether β3 could associate with a contactin-containing complex in brain, where many other channel-associating proteins are present [5, 6]. To address this question, adult rat brain membrane preparations were solubilized in Triton X-100 and immunoprecipitated with either non-immune IgG, anti-contactin, or anti-β3, as indicated. The anti-β3 immunoprecipitate was positive for contactin (Fig. 2C), suggesting that, while β3 and contactin likely do not associate directly, a β3-contactin containing complex exists in brain. We postulate that the most likely β3 binding partner in brain is neurofascin-186, a CAM localized with the VGSC complex at nodes of Ranvier and axon initial segments that associates with both β1 and β3 subunits as well as ankyrin [10, 11, 16, 34].
β1 and β2 recruit ankyrin to points of cell-cell contact in response to trans homophilic cell-cell adhesion . The interaction of β1 with ankyrin is dependent on the intracellular tyrosine residue Y181 . Yeast-two-hybrid constructs generating the intracellular domain of β1 do not interact with constructs generating ankyrin [3, 23], suggesting that extracellular homophilic β1-β1 adhesion is required for intracellular signaling. The intracellular domain of β3 contains a tyrosine residue at the position corresponding to β1Y181 (β3Y174), suggesting that β3 may also interact with ankyrin. However, if homophilic cell-cell adhesion is required to signal intracellular ankyrin recruitment by VGSC β subunits, then this interaction with β3 may not take place. To test this, CHL cells were transiently transfected with either β1-GFP or β3-GFP and ankyrinG-GFP (Fig. 3). Cell homogenates from confluent dishes in which all cells were in contact were solubilized in Triton X-100 and immunoprecipitated either with nonimmune IgG or anti-ankyrinG. Blots were probed with anti-GFP to detect either β1-GFP or β3-GFP (Fig. 3, arrow). In agreement with previous results, ankyrinG associated with β1 . These results also demonstrated that the presence of the GFP epitope tag on the carboxyl terminus of β1 did not disrupt its ability to interact with ankyrinG. Under the same conditions, β3 did not associate with ankyrinG despite robust cellular expression levels. Taken together, these results suggest that trans homophilic cell-cell adhesion is required for VGSC β subunit-mediated ankyrin recruitment. Further, that β1 and β3 differ significantly in their signaling properties when expressed in heterologous systems.
In spite of its similarity to β1, β3 does not exhibit trans homophilic cell adhesive interactions. The crystal structure of the extracellular domain of MPo [28, 36] has been used previously to predict structure-function relationships for VGSC β1 subunits [28, 36]. Comparison of the β3 Ig domain with the Ig domains of β1, β2, and MPo reveals a number of important amino acid differences that may explain the absence of β3 trans homophilic adhesive activity (Fig. 4). Four proline residues are present in the β3 Ig loop domain that are not present in the corresponding positions of the other three molecules: β3P16, located in the A’-B connecting loop of the Ig domain; β3P63, located in the C”-D connecting loop of the Ig domain; β3P109, located in F-G connecting loop of the Ig domain; and β3P118, located within β sheet G. P16, P63, and P109 are positioned in previously described flexible loop segments connecting the β sheets [38, 40]. The presence of prolines in these regions may add rigidity to the flexible loops of the β3 Ig domain, limiting its ability to interact with other CAMs. A MPo mutation found in Charcot-Marie-Tooth patients, S49L, is located in the putative homophilic adhesive interface of MPo [36, 39]. Alignment of this region of MPo with the corresponding regions of β1, β2, and β3 shows a lysine residue at this position in β1, a methionine residue in β2, and the absence of a corresponding residue in β3 (Fig. 4, “-”). If MPoS49 is critical for adhesive interactions, then the lack of a corresponding residue at this position in β3 may contribute to its inability to participate in trans homophilic adhesive interactions.
In summary, these data demonstrate that, in spite of its high degree of similarity to β1, β3 does not mediate trans homophilic cell adhesion resulting in ankyrin recruitment. While β1 and β3 both bind to the CAM neurofascin-186 [30, 34], they do not bind to each other , and of the two, only β1 binds to the CAM contactin [21, 29, 30]. Taken together, these results support the idea that, in spite of similar channel modulatory properties, the structures of the β1 and β3 Ig loop domains are significantly different and their ability to transduce intracellular signals differ as well. This hypothesis is supported by results showing that Scn1b, encoding β1, but not Scn3b, encoding β3, is expressed in cancer cells where it modulates cell adhesion and migration . In contrast, Scn3b, unlike Scn1b, is up-regulated in response to DNA damage and mediates a p53-dependent apoptotic pathway . Thus, these two VGSC β subunits may play very different physiological roles in vivo. Finally, the present results in heterologous systems may shed light on recent data showing that that phenotype of Scn3b null mice is mild compared to Scn1b null mice . It is possible that Scn1b compensates for Scn3b deletion but not vice-versa.
This work was supported by NIH R01MH59980 to LLI. DPM was supported by University of Michigan NIH training grant GM07767 and by an individual NRSA predoctoral fellowship NS43067.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.