|Home | About | Journals | Submit | Contact Us | Français|
The Intracellular Fibroblast Growth Factor (iFGF) subfamily includes four members (FGFs 11–14) of the structurally related FGF superfamily. Previous studies showed that the iFGFs interact directly with the pore-forming (α) subunits of voltage-gated sodium (Nav) channels and regulate the functional properties of sodium channel currents. Sequence heterogeneity among the iFGFs is thought to confer specificity to this regulation. Here, we demonstrate that the two N-terminal alternatively spliced FGF14 variants, FGF14-1a and FGF14-1b, differentially regulate currents produced by Nav1.2-and Nav1.6 channels. FGF14-1b, but not FGF14-1a, attenuates both Nav1.2 and Nav1.6 current densities. In contrast, co-expression of an FGF14 mutant, lacking the N-terminus, increased Nav1.6 current densities. In neurons, both FGF14-1a and FGF14-1b localized at the axonal initial segment, and deletion of the N-terminus abolished this localization. Thus, the FGF14 N-terminus is required for targeting and functional regulation of Nav channels, suggesting an important function for FGF14 alternative splicing in regulating neuronal excitability.
Four members (FGFs 11–14) of the intracellular Fibroblast Growth Factor (iFGF) subfamily, also referred to as fibroblast growth factor homologous factors (FHFs), share the core structure of FGFs, but form a distinct subfamily because they are not secreted from cells and do not interact with known FGF receptors (Itoh and Ornitz, 2008; Ornitz and Itoh, 2001). The iFGFs are highly expressed in embryonic and adult central and peripheral nervous systems, suggesting important roles in the normal development of the nervous system (Goldfarb, 2005; Smallwood et al., 1996). Importantly, Fgf14−/− mice show ataxia, dystonia, and cognitive impairment (Goldfarb et al., 2007; Wang et al., 2002; Wozniak et al., 2007). In humans, an autosomal dominant missense mutation, FGF14F145S, results in a progressive spinocerebellar ataxia, SCA27, and cognitive impairment (Brusse et al., 2006; Van Swieten et al., 2003).
The structure of the iFGFs has been resolved into a conserved FGF core domain and less conserved amino (N−) and carboxy (C−) termini (Smallwood et al., 1996). Diversity within the iFGF subfamily is further enhanced by N-terminal alternative splicing (Goldfarb, 2005; Munoz-Sanjuan et al., 2000; Smallwood et al., 1996; Wang et al., 2000). For example, FGF14-1a (FHF4) has a 64 amino acid N-terminus that is highly homologous to those of FGF12-1a and FGF13-1a (46% amino acid identity), whereas FGF14-1b, the most abundant form expressed in the central nervous system, has a unique 69 amino acid N-terminus (Munoz-Sanjuan et al., 2000; Wang et al., 2000). Thus, it is predicted that alternative N-termini may regulate the function of these factors.
Considerable evidence has shown that iFGFs interact directly with the pore-forming subunit of voltage-gated sodium channels and modulate gating properties and the density of these channels (Laezza et al., 2007; Liu et al., 2001; Liu et al., 2003; Lou et al., 2005; Rush et al., 2006b; Wittmack et al., 2004). Interestingly, the functional regulation of iFGF and Nav channel pairs can be isoform-specific. For example, co-expression with FGF14-1b markedly reduces Nav1.5 and Nav1.1 current densities and regulates voltage-dependence of inactivation of the two channels in opposite directions, whereas FGF14-1a suppresses Nav1.5, but not Nav1.1, currents and depolarizes activation of Nav1.1 but not Nav1.5 (Lou et al., 2005). Moreover, while both FGF13-1a (FHF2A) and FGF13-1b (FHF2B) increase Nav1.6 current density, FGF13-1a enhances the accumulation of inactivated channels, while FGF13-1b inhibits accumulation of inactivated channels (Rush et al., 2006b); similar effects on Nav1.2 have also been observed (Lampert, Dib-Hajj and Waxman, unpublished observations).
Sodium channels Nav1.2 and Nav1.6, are abundantly expressed in the cerebellum and the hippocampus (Schaller and Caldwell, 2000, 2003; Schaller et al., 1995), two of the brain regions most affected by loss of Fgf14 in mice or by the FGF14F145S mutation in SCA27 patients (Brusse et al., 2006; Dalski et al., 2005; Goldfarb et al., 2007; Laezza et al., 2007; Van Swieten et al., 2003), suggesting that the activity of one or both of these two sodium channels might be regulated by FGF14 in vivo. Additionally, FGF14-1a and -1b and FGF13 have been shown to localize at the axonal initial segment (AIS) and nodes of Ranvier (Goldfarb et al., 2007; Laezza et al., 2007; Rush et al., 2006b; Wittmack et al., 2004), two distinct sub-cellular domains that contain a high density of Nav1.2 and Nav1.6 channels and that control the generation and propagation of action potentials in neurons (Peles and Salzer, 2000; Poliak and Peles, 2003). Functional analysis suggests that the FGF14F145S mutant protein acts as a dominant negative and interferes with wild type FGF14 interactions with Nav channels. The expression of FGF14F145S in cultured hippocampal neurons impairs neuronal excitability, a phenotype that has been linked to reduced sodium current density (Laezza et al., 2007). The lack of FGF14 in cerebellar granule and Purkinje neurons from Fgf14−/− mice results in decreased neuronal excitability (Goldfarb et al., 2007; Shakkottai et al., 2009). These studies have revealed important roles for FGF14 in regulating neuronal activity via regulation of Nav1.2 and Nav1.6 channels.
Here, we show that both FGF14-1a and FGF14-1b interact with the intracellular C-terminal domains of Nav1.2 and Nav1.6, FGF14-1b suppresses Nav1.2 and Nav1.6 current densities, whereas FGF14-1a has no effect on Nav1.2 or Nav1.6 currents. Additionally, expression of an FGF14 variant with an N-terminus derived from that of FGF12-1b no longer is targeted to the AIS. We conclude that the FGF14 N-terminal domain contributes to both the ability of FGF14 to functionally regulate Nav channels and to the subcellular localization of FGF14 in neurons.
Generation and characterization of the Fgf14-1b-Gfp, Fgf14-1a-Gfp and Fgf14-1bF145S-Gfp constructs were described previously (Laezza et al., 2007; Lou et al., 2005) Fgf14-ΔNT-Gfp, lacking the N-terminal domain, was amplified by PCR from Fgf14-1b and cloned into 5’ BamHI and 3’ Not I sites of the pQBI vector (Quantum Technologies). To provide FGF14-ΔNT with a transcription initiation codon, the amino acid sequence “MESK”, corresponding to the N-terminal sequence encoded by the first exon of Fgf12-1b) was added using PCR with the following primers: forward: TACCGGTACCATGGAGAGCAAGGATCCCCAGCTCAAG; reverse: GTTTAGCGGCCGCCTGTGGTCTTGCACTTG. Thus, the FGF14-ΔNT protein is a chimera of Fgf12-1b N-terminus and the core and C-terminus of Fgf14. The Fgf14 plasmids with C-terminal Myc tags were described previously (Laezza et al., 2007; Lou et al., 2005). The human Nav1.2 (Scn2a, aa 1777–2005) and mouse Nav1.6 (Scn8a, aa 1765–1978) were amplified by PCR from cDNA clones and ligated in frame into the pCruz-HA plasmid (Santa Cruz, Inc.) (see Fig. 5A).
The plasmid pcDNA3-Nav1.6R, which encodes full-length mouse Nav1.6, was described previously (Herzog et al., 2003) and contains a substitution of tyrosine 371 by serine (Y371S) which renders Nav1.6 channels resistant to micromolar concentrations of tetrodotoxin when the channel is produced in mammalian cell lines (HEK293 and ND7/23), or in DRG neurons (Cummins et al., 2005; Herzog et al., 2003; Rush et al., 2006a; Wittmack et al., 2004; Wittmack et al., 2005).
All reagents were purchased from Sigma Chemical Company (St. Louis, MO) unless otherwise noted. HEK293 cells stably expressing rat Nav1.2 (HEK-Nav1.2 cells) were maintained in Dulbecco’s Modified Eagle’s Medium (Invitrogen, Carlsbad, CA), supplemented with 10% fetal bovine serum, 100 U/ml penicillin and 100 µg/ml streptomycin, 500 µg/ml G418 (Invitrogen) and incubated at 37°C with 5% CO2, as previously described (Laezza et al., 2007). Cells were transfected at 90–100% confluence using Lipofectamine 2000 (Invitrogen), according to manufacturer’s instructions. Transfection of the ND7/23 cells was as described previously (Rush et al., 2006a). Briefly, ND7/23 cells were plated at low density on 35 mm dishes 24 hours before transfection. Cells were transfected with 2 µg of Nav1.6R and EGFP or Nav1.6R and Fgf14-GFP, and electrophysiology was performed 24 hours following transfections. Only cells with robust green fluorescence and currents that larger than 50 pA were included in the analysis.
Transfected cells were dissociated and re-plated at low-density approximately 12 hr post-transfection. Recordings were obtained at room temperature (20–22°C) ~12–18 hours post-transfection using an Axopatch 1D amplifier (Axon Instruments, Union City, CA). Borosilicate glass pipettes with resistance of 1–2 MΩ or 2.5–5 MΩ were made using a P-87 Micropipette Puller (Sutter Instruments, Novato, CA). Recording solutions: extracellular (mM), 140 NaCl, 3 KCl, 1 MgCl2, 1 CaCl2, 20 HEPES, pH 7.3; intracellular, 140 CsF, 1 EGTA, 10 NaCl, 10 HEPES, pH 7.3. After seal formation and cell break-in, membrane capacitance in each cell was calculated by integrating the area under the capacitative transients recorded in response to a 5 ms 10 mV hyperpolarizing test pulse from a holding potential of −70 mV. Capacitative transients and series resistances were compensated electronically by 80–90%. Data were acquired at 50 kHz and filtered at 5 kHz prior to digitization and storage. All experimental parameters were controlled with Clampex (version 9.2) (Axon Instruments) using a Digidata 1322A analog-digital interface (Axon Instruments).
Voltage-dependent inward currents were evoked by brief (50–100 ms) voltage steps to test potentials between −100 mV and +60 mV from various holding potentials (between −130 mV and −90 mV). Steady-state inactivation of Nav channels was determined using a paired-pulse protocol. From the holding potential, cells were stepped to varying test potentials between −130 mV and −10 mV (prepulse of 750 ms) prior to a test pulse to −10 mV or −20 mV. Current densities were determined by dividing the amplitudes of the sodium currents (INa) at each test potential by the membrane capacitance (in the same cell). Current-voltage relations were generated by plotting peak current density as a function of the test potential. The sodium channel conductance GNa was calculated using the following equation:
where INa is the amplitude of the current at voltage Vm, and Erev is the Na+ reversal potential. Steady-state activation curves were derived by plotting normalized GNa as a function of test potential and fitted using the Boltzmann equation:
where GNa,Max is the maximum conductance, Va is the membrane potential at half-maximal activation, Vm is the membrane voltage and k is the slope factor.
For steady-state inactivation, normalized current amplitude (INa/INa,Max) at each test potential was plotted as a function of prepulse potential (Vm) and fitted using the following (Boltzmann) equation:
where Vh is the membrane potential at half-maximal inactivation and k is the slope factor.
Data analyses were performed using Clampfit 9.2 (Axon Instruments), Prizm 4 software (GraphPad, San Diego, CA) and SigmaStat (Jendel Corporation). Results are presented as means ± SEM and the statistical significance of differences between groups was assessed using the Student’s t test and was set at P < 0.05.
Whole-cell voltage-clamp recordings (Hamill et al., 1981) from ND7/23 cells transiently transfected with Nav1.6R together with EGFP, or with one of the GFP-tagged Fgf14 (FGF14-1a, FGF14-1b, or FGF14-ΔNT) plasmids, were performed with an EPC-9 amplifier (HEKA electronics, Lambrecht/Pfalz, Germany) using fire polished 0.5–1.5 MΩ electrodes (World Precision Instruments, Inc, Sarasota, FL, USA). The pipette solution contained (in mM): 140 CsF, 10 NaCl, 1 EGTA, and 10 HEPES; 302 mosmol (pH 7.4, adjusted with CsOH) and the extracellular bath contained (in mM): 140 NaCl, 3 KCl, 10 glucose, 10 HEPES, 1 MgCl2, 1 CaCl2, 0.0003 TTX; 310 mosmol (pH 7.4, adjusted with NaOH). Tetrodotoxin (TTX) was added to the bath solution to block endogenous TTX-S sodium currents, which are present in ND7/23 cells (Rush et al., 2006b; Wittmack et al., 2004) All recordings were conducted at room temperature (~21°C). The pipette potential was adjusted to zero before seal formation, and the voltages were not corrected for liquid junction potential. Capacitive transients were cancelled, and series resistances were compensated at 10 µs by 65–95%. Leak currents were subtracted digitally online using hyperpolarizing voltage steps applied after the test pulse (P/4 procedure). Currents were acquired using Pulse software (HEKA electronics, Lambrecht/Pfalz, Germany), filtered at 10 kHz and sampled at 20–62.5 kHz; the longer inactivation protocol was sampled at 4 kHz.
Voltage-clamp recordings were obtained 4 min after establishing the whole-cell configuration. Standard current-voltage (I–V) relations were obtained using 40 ms pulses from a holding potential of −120 mV to a range of test potentials (−100 to +60 mV) in 5 mV steps with 5 seconds between pulses. Activation curves were obtained by calculating the conductance, G, at each voltage, V, as described above. To examine the voltage-dependences of steady-state inactivation of the currents, a series of 500 ms pre-pulses (−120 to −10 mV) from a holding potential of −120 mV, each followed by a 40 msec depolarization to −10 mV, were presented; voltage steps were presented at 10 sec intervals. Normalized curves were fitted using the Boltzmann equation (above).
For recovery from inactivation, two pulse protocols of 40 and 10 ms voltage steps to −10 mV and 0 mV from the holding potential of −120 mV were implemented, with variable recovery times (0.1 to 6533 ms) at −80 mV between the two pulses. Following normalization of data obtained in different cells, curves were fitted with a single rising exponential function for each cell
and the recovery time constant τ was calculated as the mean from all cells.
Cumulative (frequency-dependent) inactivation was examined by applying a 2 ms test pulse to −10 mV 20 times at frequencies between 0.5 Hz and 10 Hz from a holding potential of −80 mV. Current responses were normalized to the first recorded pulse and the currents at the 20th pulses were compared.
All data are presented as mean ± SEM. The statistical significance (p<0.05) of observed differences among groups was determined by either Student's t-test or ANOVA using GraphPad Prism software (GraphPad Software, San Diego, CA, USA). Post hoc analysis for multiple comparisons was done using Dunnett's test.
Hippocampal neuron cultures were prepared from embryonic day 18 (E18) rat embryos using previously described methods (Goslin et al., 1998), except that hippocampi were dissected and dissociated using papain. Briefly, following trituration through a Pasteur pipette, neurons were plated at low density (1–5 × 105 cells/dish) on poly-L-lysine-coated coverslips in 60 mm culture dishes in MEM supplemented with 10% horse serum. After 2–4 hr, coverslips (containing neurons) were inverted and placed over a glial feeder layer in serum-free MEM with 0.1% ovalbumin, and 1 mM pyruvate (N2.1 media; Invitrogen) separated by ~1 mm wax dot spacers. The presence of the spacer prevented contact between the neurons on the coverslips and the glial feeder layer. To prevent the overgrowth of the glia, cultures were treated with cytosine arabinoside (5 µM; Calbiochem, La Jolla, CA) at day 3 in vitro (DIV). Transfections were performed using Lipofectamine 2000 (Invitrogen) at DIV9.
Rat hippocampal neurons (DIV 10) were fixed in fresh 4% paraformaldehyde and 4% sucrose in phosphate-buffered saline (PBS) for 15 minutes and permeabilized with 0.25% Triton-X-100. After blocking with 10% BSA for ~30 minutes at 37°C, neurons were incubated at room temperature for 12–16 hr with a mouse monoclonal Pan-Nav channel antibody (diluted 1:100, Sigma, St Louis, MO) and polyclonal chicken anti-βIV spectrin (diluted 1:1000, gift from Dr. M. Komada (Tokyo Institute of Technology, Tokyo, Japan). All antibodies were diluted (to the concentrations noted) in PBS containing 3% BSA. Following incubations with the primary antibody combinations, neurons were washed three times in PBS and incubated for 2 hr at 37°C with appropriate secondary antibodies: Alexa 647-conjugated goat-anti-mouse IgG1 (1:500) together with aminomethylcoumarin (AMCA)-conjugated-goat-anti-rabbit IgG (1:100, Vector Laboratories, Burlingame, CA) for the cells stained with the mouse monoclonal pan Nav channel antibody and the rabbit polyclonal anti-MAP2 antibodies. Coverslips were then washed (three times) with PBS and mounted in elvanol (Tris-HCl, glycerol, and polyvinyl alcohol with 2% 1,4-diazabicyclo[2,2,2] octane). Images were acquired using an Axio Imager epifluorescence microscope (Zeiss, Oberkochen, Germany) with a 63× objective. Images were acquired with an Axio Cam MRm using the Axio Vision software (Zeiss).
Transfected HEK293 cells were washed with PBS and lysed in lysis buffer containing (in mM): 20 Tris-HCl, 150 NaCl, 1% NP-40. Protease inhibitor cocktail (set #3, Calbiochem) was added immediately before cell lysis. Cell extracts were collected and sonicated for 20 sec and centrifuged at 4°C, 15,000g for 15 min. Rabbit anti-Myc-agarose beads (Sigma, St. Louis, MO) were incubated with the supernatants 2 hours at 4°C with agitation and then washed 3 times in PBS, 0.1 % NP-40 buffer before adding 2x sample buffer containing 50 mM tris(2-carboxyethyl)phosphine (TCEP) and 4M urea. Mixtures were heated for 10 min at 65°C and resolved on 4–15% polyacrylamide gels (BioRad, Hercules, CA). Resolved proteins were transferred to PVDF membranes (Millipore, Bedford, MA) for 1.5 hours at 4°C and blocked in tris-buffered saline (TBS) with 3% nonfat dry milk and 0.1% Tween-20. Membranes were then incubated in blocking buffer containing the monoclonal anti-Myc (1:1000) or anti-HA (1:1000, Sigma, H9856) antibody overnight. Washed membranes were incubated with goat anti-mouse-HRP (1:5000) and visualized with the Super Signal Fempto chemiluminescence substrate (Pierce, Rockford, IL).
For the quantitative analysis of the AIS expression of GFP, FGF14-1a-GFP, FGF14-1b-GFP, or FGF14-ΔNT-GFP, a mask of the axonal initial segment was first generated by thresholding the βIV spectrin images. The mask of the corresponding region was then applied to the unthresholded GFP images. Fluorescence intensity in dendrites was directly measured on the GFP images. The AIS enrichment index (AEI) was defined as the ratio of total fluorescence intensity/area in the AIS divided by the total fluorescence intensity/area in the dendrites (Laezza et al., 2007; Lou et al., 2005). Analysis was done using ImageJ software (National Institutes of Health, Bethesda, Maryland, USA, http://rsb.info.nih.gov/ij/, 1997–2008).
To determine whether FGF14 modulates the functional properties of Nav1.2-encoded currents, Gfp, Fgf14-1a-Gfp or Fgf14-1b-Gfp was transiently transfected into an HEK293 cell line stably expressing Nav1.2 (Laezza et al., 2007). In control cells expressing GFP, robust fast activating and inactivating inward voltage-dependent sodium currents, INa, were recorded in response to depolarizing voltage steps (Fig. 1A). In Nav1.2 cells co-expressing FGF14-1b-GFP, however, mean ± SEM (n = 28) peak INa density was significantly (p<0.001) lower (−91 ± 11 pA/pF) than in GFP-expressing control cells (−242 ± 20 pA/pF, n=37). Interestingly, co-expression of FGF14-1a-GFP, or of a construct lacking the FGF14 N-terminus, FGF14-ΔNT-GFP, which resembles the 1b splice forms of other iFGFs (Lou et al., 2005), did not result in any measurable differences in Nav1.2 current densities (Fig. 1A,D, Table 1).
The voltage-dependences of INa activation and steady-state inactivation were determined by plotting normalized current amplitudes as a function of the test potential (activation) or the pre-pulse potential (inactivation), as detailed in Methods. These analyses revealed that expression of FGF14-1b-GFP resulted in a small (+4 mV), but statistically significant (p<0.05) depolarizing shift in the voltage dependence of Nav1.2 activation, without measurably affecting steady-state inactivation of the current, whereas expression of FGF14-1a-GFP produced a statistically significant (p<0.01) depolarizing shift (+9 mV) in the voltage-dependence of inactivation. Expression of FGF14-ΔNT-GFP, in contrast, resulted in a statistically significant (p<0.001) hyperpolarizing shift (−5 mV) in the voltage-dependence of steady-state inactivation without changing the properties of activation. The N-terminal domain of FGF14, therefore, influences the modulatory effects of FGF14 on Nav1.2-encoded currents.
Similar experiments were completed to examine the effects of FGF14 on Nav1.6R encoded currents. The effects of the FGF14 N-terminal variants on the properties of Nav1.6R current were examined in the neuronal cell line, ND7/23, transiently transfected with Nav1.6R, a tetrodotoxin-resistant variant of Nav1.6 (Herzog et al., 2003), together with GFP or one of the GFP-tagged FGF14 variants indicated above. Recordings were obtained from GFP-positive cells in the presence of 300 nM TTX to block the endogenous TTX-S currents present in ND7/23 cells (Wittmack et al., 2004). Rapidly activating and inactivating sodium channel currents were recorded in all cells (Fig. 2A). In ND7/23 cells co-expressing Nav1.6R and GFP, Nav1.6R currents (−74 ± 8 pA/pF; n=30) with properties similar to those recorded from DRG neurons transfected with Nav1.6R (Herzog et al., 2003), were obtained. Similar peak Nav1.6R current density (−81± 24 pA/pF; n=12) was obtained in cells co-expressing GFP or FGF14-1a-GFP and Nav1.6R (Fig. 2A,D). In contrast, co-expression of FGF14-1b with Nav1.6R resulted in significantly (p<0.01) reduced current densities (−14 ± 3 pA/pF; n=15). Indeed, the residual Nav1.6R current in cells co-expressing FGF14-1b-GFP should be considered an overestimate, as Nav1.6R currents were not detectable in 40 of 55 GFP-positive ND7/23 cells co-expressing FGF14-1b-GFP and Nav1.6R. In contrast, Nav channel currents <50 pA were only observed in 2 of 26 GFP-positive ND7/23 cells expressing Nav1.6R and FGF14-1a-GFP and in 2 of 14 GFP-positive ND7/23 cells expressing Nav1.6R and FGF14-ΔNT-GFP. Interestingly, co-expression of FGF14-ΔNT-GFP with Nav1.6R significantly (p<0.01) increased current density by almost 3 fold, to 168 ± 20 pA/pF (n=12; Fig. 2A,D).
Experiments were also completed to investigate the effects of FGF14-1a-GFP, FGF14-1b-GFP and FGF14-ΔNT on activation and steady-state fast-inactivation of Nav1.6R currents. As shown in Fig. 2C, E and Table 2, Nav1.6R activation was not affected by co-expression of FGF14-1a-GFP or FGF14-1b-GFP, although there was a depolarizing shift in the voltage-dependence of activation when Nav1.6R and FGF14-ΔNT were co-expressed. In contrast, all FGF14 constructs induced a significant depolarizing shift in the voltage-dependence of steady-state inactivation (Fig. 2B, F).
Nav1.6 channels play an important role in maintaining high frequency firing in cerebellar Purkinje cells (Levin et al., 2006), and loss of Fgf14 severely impairs the spontaneous firing properties of Purkinje neurons (Shakkottai et al., 2009). Additional experiments were completed, therefore, to explore further the potential functional effects of FGF14 on Nav1.6R-encoded Na currents. The expression of Nav1.6R in ND7/23 cells produced a current that recovers from fast-inactivation with properties similar to those of DRG neurons (Herzog et al., 2003). Previous studies have shown that FGF13-1a-GFP slowed Nav1.6R recovery from inactivation (Rush et al., 2006b), whereas FGF13-1b-GFP did not (Wittmack et al., 2004). We investigated the effects of FGF14 on the kinetics of recovery from inactivation of Nav1.6R currents using a two pulse protocol, as shown in Fig. 3A. Cells were held at −120 mV and two depolarizing steps to −10 mV were applied. These two pulses were separated by a hyperpolarizing voltage step to −80 mV of variable duration, during which channels recovered from fast-inactivation. Cells co-expressing Nav1.6R and GFP recovered quickly, with the majority of the current being fully available after 100 ms. Co-expression of Nav1.6R with FGF14-1b-GFP or FGF14-ΔNT resulted in recovery from inactivation with similar kinetics. In contrast, co-expression of FGF14-1a-GFP with Nav1.6R resulted in a much slower recovery from inactivation, with full recovery requiring more than one second (Fig. 3A). Analyses of the recovery data (Fig. 3B) revealed that the mean time constants of Nav1.6R recovery were similar in cells co-expressing Nav1.6R with GFP (17 ± 1.1 ms; n = 23); FGF14-1b-GFP (50 ± 22 ms; n = 7) and FGF14-ΔNT-GFP (13 ± 1 ms; n = 10), as illustrated in Fig. 3B. The large variability in the case of FGF14-1b is related to the very small Nav1.6R currents that are recorded under this condition. In contrast, co-expression with FGF14-1a-GFP caused a significant (p<0.05) increase in the time constant of Nav1.6R current recovery: 346 ±19 ms (n=9) (Fig. 3B).
The marked slowing of Nav1.6R current recovery observed in cells expressing FGF14-1a-GFP, compared to the other FGF14 proteins in these experiments (Fig. 3), suggested that the effects of the FGF14 variants on Nav1.6R currents would be particularly dramatic during repetitive stimulation. To test this hypothesis directly, the peak Nav1.6R currents in the presence of the different FGF14 proteins were measured during short test pulses to −10 mV at a range of stimulation frequencies (0.5–10 Hz) from a holding potential of −80 mV (Fig. 4). These experiments revealed that co-expression of FGF14-1a-GFP results in marked reductions in peak Nav1.6R current amplitudes at all stimulation frequencies tested. At a stimulation frequency of 5 Hz for example, the peak current amplitude was reduced markedly (by ~50%) in the second stimulus of the train in cells co-expressing FGF14-1a-GFP, whereas current amplitudes were unchanged in cells co-expressing GFP, FGF14-1b-GFP or FGF14-ΔNT-GFP (Fig. 4A). Co-expression with FGF14-1b-GFP prevented the slight decrease in peak Nav1.6R currents observed at 5 Hz in cells expressing Nav1.6R and GFP. In the presence of FGF14-1a-GFP, the decline in Nav1.6R current amplitudes was even more dramatic at higher stimulation frequencies, reaching 80% at a frequency of 10 Hz (Fig. 4B, Table 3). Thus, the N-terminal domain of FGF14 is essential not only for modulating Nav1.6 current density, but also for regulating recovery from inactivation of this channel and, therefore, its availability during repetitive stimulation.
Different members of the iFGF family have been shown to interact directly with the intracellular C-terminal tails of Nav1.5, Nav1.6 and Nav1.9 (Goetz et al., 2009; Liu et al., 2001; Wittmack et al., 2004). The C-termini of Nav1.2 and Nav1.6 are highly conserved, especially in the membrane proximal region (Fig. 5A). Previous studies on FGF12-1b-GFP demonstrated that this interaction is mediated by amino acid residues 1–41 in the N-terminal end of the conserved FGF core domain (Liu et al., 2001; Liu et al., 2003). These residues show 71% identity with the FGF14 core domain and 78% amino acid identity with the FGF13-1b core domain, suggesting that a common binding region might mediate interaction with Nav channels.
We investigated whether N-terminal FGF14 splice variants interact with the C-terminal tails of the sodium channels Nav1.2 and Nav1.6. For these experiments, Fgf14-1a-Myc, Fgf14-1b-Myc, Fgf14-ΔNT-Myc or Fgf14-1bF145S-Myc were co-transfected together with a plasmid expressing the C-terminal tail of Nav1.2 or Nav1.6 tagged with the HA epitope (HA-Nav1.2-C-tail and HA-Nav1.6-C-tail) into HEK293 cells. As shown in Figure 5B, the HA-Nav1.2 C-terminal tail immunoprecipitated with FGF14-1b-Myc, FGF14-1a-Myc, and FGF14-ΔNT-Myc but not with the negative control, FGF14F145S-Myc, which has been previously shown not to interact with full length Nav1.2 (Laezza et al., 2007). Similarly, the HA-Nav1.6 C-terminal tail immunoprecipitated with FGF14-1b-Myc and FGF14-1a-Myc, but not with FGF14F145S-Myc. Interestingly, although FGF14-ΔNT-Myc immunoprecipitated with the HA-Nav1.2 C-tail, this protein was not as efficient in immunoprecipitating the HA-Nav1.6-C-tail. This was unexpected because FGF14-ΔNT-Myc co-expression significantly (and by nearly threefold) increased Nav1.6R current densities (Fig. 2A–D). One possible explanation for this apparent discrepancy is that the interaction of FGF14-ΔNT-Myc with the C-terminal tail of Nav1.6R is of lower affinity (than for the other FGF14 constructs) and was, therefore, not detected in the immunoprecipitation assay. Alternatively, FGF14-ΔNT-Myc could interact with another region (i.e., other than the C terminal tail) of the Nav1.6 protein. Collectively, these experiments demonstrate that FGF14-1a-Myc and FGF14-1b-Myc interact with the C-terminal tails of both Nav1.2 and Nav1.6 and suggest that, despite the differences in primary sequence, both FGF14 splice forms retain the ability to interact with a common intracellular C-terminus of Nav channels.
In neurons, Nav1.2 and Nav1.6 channels are concentrated at the axon initial segment (AIS) and at nodes of Ranvier (Boiko et al., 2001; Boiko et al., 2003; Kaplan et al., 2001). Interestingly, recent evidence indicates that native FGF14 is concentrated at the AIS in hippocampal neurons (Laezza et al., 2007) and co-localizes with native Nav channels and with β-IV spectrin, an AIS marker (Komada and Soriano, 2002; Nishimura et al., 2007). Similar to FGF14, endogenous FGF13 has also been shown to co-localize with Nav channels at the AIS and at nodes of Ranvier (Goldfarb et al., 2007; Wittmack et al., 2004), raising the possibility that iFGFs might share common molecular determinants important for their sub-cellular localization.
The results presented above demonstrate that the alternatively spliced N-terminal domains of FGF14 confer different regulatory effects on Nav current densities and gating properties. Intriguingly, however, previous findings indicate that both FGF14-1a-GFP and FGF14-1b-GFP are enriched at the AIS when heterologously expressed in hippocampal neurons (Lou et al., 2005), suggesting that the sub-cellular targeting and the functional effects (on Nav channels) of FGF14 might be mediated by different protein domains.
To assess whether the N-terminal domain of FGF14 regulates the sub-cellular targeting of this factor, GFP, FGF14-ΔNT-GFP, FGF14-1a-GFP or FGF14-1b-GFP was transfected into cultured hippocampal neurons. Neurons were fixed and immunolabeled with an anti β-IV spectrin antibody to identify the AIS and imaged for GFP fluorescence (Fig. 6). As previously demonstrated, these experiments revealed that FGF14-1a-GFP and FGF14-1b-GFP were enriched at the AIS (Laezza et al., 2007; Lou et al., 2005) and co-localized with endogenous β-IV spectrin (Fig. 6D–F, G–I). In contrast, FGF14-ΔNT-GFP (Fig. 6J–L) was uniformly distributed throughout the cell soma, dendrites and axon, showing no enrichment at the AIS (Fig. 6M), similar to GFP expressing neurons (Fig. 6A–C). The simplest interpretation of these findings is that the presence of the N-terminal domain is necessary for proper targeting of FGF14 to the AIS.
Accumulating evidence has established a link between the iFGFs and the regulation of Nav channel activity and localization (Laezza et al., 2007; Liu et al., 2003; Lou et al., 2005; Rush et al., 2006b; Wittmack et al., 2004). However, the functional consequence of these interactions on sodium currents are both iFGF and Nav channel isoform-dependent. In vitro, FGF12-1b, FGF13-1a and FGF13-1b have all been shown to either have no effect or to up-regulate Nav1.5 or Nav1.6 current amplitudes in heterologous expression systems (Liu et al., 2003; Rush et al., 2006b; Wittmack et al., 2004). In contrast, FGF14-1b, the major neuronal form of Fgf14, was shown previously to inhibit heterologously expressed Nav1.1 and Nav1.5 currents (Lou et al., 2005; Wang et al., 2000). These observations suggest that the physiological activity of FGF14-1b may be fundamentally different from FGF14-1a and other iFGFs and further that the unique FGF14-1b N-terminus is responsible for the profound difference in functional properties.
The results of the experiments presented here demonstrate that Nav1.2 and Nav1.6 current densities are reduced in cells co-expressing FGF14-1b, whereas FGF14-1a co-expression did not measurably affect peak Nav1.2 or Nav1.6 current densities. We also show here that both FGF14-1a and FGF14-1b interact with the intracellular C-terminal domains of Nav1.2 and Nav1.6 channels and that both forms of FGF14 are targeted to the AIS of hippocampal neurons. Taken together with the results of previous studies showing that FGF14-1b interacts with Nav1.1 and attenuates Nav1.1 currents (Lou et al., 2005), the results here demonstrate that FGF14-1b can interact with and similarly modulate the properties of all Nav channels thus far tested. The function of FGF14-1b in attenuating the amplitudes of heterologously expressed neuronal Nav channels appears to be unique compared to other iFGFs and to FGF14-1a, in the HEK 293 and ND7/23 cell expression systems. The N-terminal 1b domain of FGF14 appears essential, as substitution with the similarly sized 1a domain, or deletion of the N-terminus, abolishes the ability of FGF14 to reduce sodium current densities. These results also suggest that FGF14-1b may also play an important role in the regulation of neuronal Nav channels in vivo.
In addition to the distinct effects on Nav channel current densities, FGF14-1a and FGF14-1b also differentially modify the voltage-dependences of activation and inactivation of neuronal Nav channel currents. Co-expression with FGF14-1a (but not FGF14-1b), for example, shifts the voltage-dependence of inactivation of Nav1.2 currents and FGF14-1b (but not FGF14-1a) modifies the voltage-dependence of Nav1.2 current activation. Both FGF14-1a and FGF14-1b, however, produce substantial (~+10 mV) shifts in Nav1.6R current inactivation; shifts that could alter Nav1.6R channel availability near the resting potential and impact neuronal excitability directly. Interestingly, the observed effect is similar to that reported previously for FGF13-1a (Rush et al., 2006b). These data suggest that there may be some conserved functions among the 1a N-terminal iFGF isoforms.
Other studies showed that FGF14-1b, but not FGF14-1a, suppressed Nav1.1 currents expressed in heterologous cells (Lou et al., 2005; Rusconi et al., 2009). However, both FGF14-1a and FGF14-1b effectively inhibited Nav currents carried by Nav1.1-R191G, a C-terminal mutant found in General Epilepsy Febrile Syndrome (Rusconi et al., 2009). This suggests that interactions between different iFGF isoforms and the C-terminal domain of Nav channels may contribute to phenotypes associated with human sodium channelopathies. Future studies will be needed to address possible interactions between iFGFs and mutations in other Nav channel α subunits.
It has previously been demonstrated that FGF14-1a and FGF14-1b directly interact with full-length Nav1.1, Nav1. 2 and Nav1.5 in heterologous expression systems and with native Nav channels expressed in the Neuro2a neuroblastoma cell line (Laezza et al., 2007; Lou et al., 2005). FGF12 and FGF13 also reportedly interact directly with Nav channels (Liu et al., 2001; Liu et al., 2003; Wittmack et al., 2004). The interaction between FGF13-1b and Nav1.6 is mediated via the intracellular C-terminal domain of Nav1.6 (Goetz et al., 2009; Wittmack et al., 2004). To begin to understand the structure/function relationship between FGF14 and neuronal Nav channels, experiments were designed to test whether different forms of FGF14 could interact with the C-terminal domains of Nav1.2 and Nav1.6. The results of these experiments revealed that FGF14-1a and FGF14-1b interact with the C-terminal domains of both Nav1.2 and Nav1.6 α subunits.
Previous data suggest that the membrane-proximal segment of the C-terminus of Nav1.5 and Nav1.9, which has ~70% sequence identity among Nav channels, is important for the interaction with FGF12 (Liu et al., 2001; Liu et al., 2003). In addition to direct effects on Nav channel activation and inactivation, the interaction with the iFGFs might be important in regulating the stability or the trafficking to the cell surface of assembled Nav channels, effects that could account for the observed changes in sodium current densities observed in cells co-expressing FGF14-1b. For example, iFGF binding could modulate the function of a potential di-leucine motif in the proximal region of the Nav1.2 C-terminal domain (Fig. 5A) that may contribute to trafficking of the channel to the axon (Garrido et al., 2001). Alternatively, iFGF binding may affect a PY motif (-PPSY) (Fig. 5A), which forms a binding site for Nedd4 and Nedd4-2, ubiquitin-protein ligases implicated in the endocytic removal of the channel from the plasma membrane (Fotia et al., 2004; Rougier et al., 2005). Recent evidence indicates that co-expression of the FGF14 mutant protein FGF14F145S, identified in a family afflicted with an autosomal dominant form of spinocerebellar ataxia (SCA27) (Brusse et al., 2006; Dalski et al., 2005), decreases expression of Nav channels at the AIS through a dominant negative effect on wild type FGF14, preventing endogenous wild type FGF14 from interacting with Nav channels (Laezza et al., 2007). This suggests a potential active role for endogenous FGF14 in regulating the stability and/or the trafficking of Nav channels to specific subcellular compartments. Further experiments will be needed to determine directly whether FGF14-Nav α subunit interactions mediate trafficking of Nav channels.
Endogenous FGF13 and FGF14 are localized to the AIS and co-localize with Nav channels in cerebellar granule and hippocampal neurons, respectively, suggesting that iFGFs might play a role in stabilizing or targeting native Nav channels to specific subcellular compartments (Goldfarb et al., 2007; Laezza et al., 2007). For example, iFGFs might affect the interaction of Nav channels with other proteins, such as ankyrin-G (Garrido et al., 2001; Garrido et al., 2003a; Garrido et al., 2003b). Alternatively, iFGFs, which interact with JIP2, which in turn interacts with kinesin light chain (Hirokawa and Takemura, 2005; Schoorlemmer and Goldfarb, 2002; Verhey et al., 2001), may play a role in the regulation of axonal transport. Whether the interaction of iFGFs with the C-terminal domain of Nav channels is sufficient to mediate localization of the channels remains to be determined.
An important finding of the present study is that the presence of the 1a or 1b N-terminal domain of FGF14 is necessary to control the sub-cellular targeting of FGF14 in hippocampal neurons. Indeed, deletion of the N-terminal domain eliminates the specific targeting of FGF14 to the AIS. Thus, the N-terminal domain of FGF14 not only modulates the regulation of Nav channel function by FGF14, but also has a distinct role in the sub-cellular localization of FGF14. One unexpected result is that both FGF14-1a and FGF14-1b are targeted to the AIS, even though the 1a and 1b protein domains share few conserved sequences (Lou et al., 2005; Wang et al., 2000). This suggests that specific sequences in the N-terminal domains are not necessary for AIS localization. Future experiments, focused on testing this hypothesis directly are warranted.
Taken together with previous work, the results presented here clearly demonstrate a direct role for FGF14 in regulating both the densities and the biophysical properties of Nav1.2 and Nav1.6 currents in heterologous cells. As previously reported, however, there is a dramatic difference between the effects of FGF14 on Nav current densities in heterologous cells and in neurons. In heterologous cells, FGF14-1b reduces sodium currents, and similar results have now been obtained for all channels tested. In contrast, in cultured hippocampal neurons, overexpression of FGF14-1a or FGF14-1b results in markedly increased Nav current densities (Lou et al., 2005), whereas genetic ablation of FGF14 in Fgf14−/− mice or expression of the SCA27-linked FGF mutation, FGF14F145S, results in impairment of excitability in the cerebellum and in the hippocampus, as well as in isolated hippocampal neurons in vitro (Goldfarb et al., 2007; Laezza et al., 2007; Shakkottai et al., 2009). These results suggest that endogenous FGF14 functions to promote activity of neuronal Nav channels rather than suppress activity as observed in heterogonous cells. This paradox likely results from the expression of different sets of proteins in heterologous cells and neurons that could differentially affect the end result of iFGF activity. Alternatively, the iFGFs or the sodium channels might be modified post-translationally and these modifications may be different in neurons and in heterologous cells.
Taken together with findings detailed in several recent studies, the results presented here emphasize the potential complexities of iFGF-mediated effects on Nav channels and on the regulation of neuronal excitability. In addition, cell type specific differences in iFGF expression, redundant and/or competing actions of the iFGFs, and/or heteromeric assembly among the various iFGFs might add further complexity to the mechanisms of iFGF action in neurons.
This work was supported by the Hope Center for Neurological Disorders, the National Ataxia Foundation, the McDonnell Center for Cellular and Molecular Neurobiology and NIH grant R01NS065761 (DMO and JMN). Work in SGW laboratory was supported in part by grants from the National Multiple Sclerosis Society, and the Rehabilitation Research and Development Service and Medical Research Service, Department of Veterans Affairs. We thank L. Li and Lynda Tyrrell for technical assistance and M. Komada (Tokyo Institute of Technology) for providing the affinity purified chicken anti-βIV-spectrin antibody.
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.