Familial hemiplegic migraine type 1 (FHM-1), a rare hereditary form of migraine with aura and hemiparesis, serves as a good model for exploring migraine pathophysiology. The FHM-1 gene encodes the pore-forming CaV2.1 subunit of human P/Q-type voltage-gated Ca2+ channels (VGCCs). Some FHM-1 mutations result in a decrease of whole cell P/Q-type current density in transfected cells/neurons. Questions remain as to whether and how these mutations may increase the gain of the trigeminal nociceptive pathway underlying migraine headache. Here, we investigated the effects of T666M, the most frequently occurring FHM-1 mutation, on VGCC currents and neuronal excitability in trigeminal ganglion (TG) neurons. We expressed human wild-type and T666M CaV2.1 subunits in cultured TG neurons from CaV2.1 knockout mice and recorded whole cell VGCC currents in transfected neurons. Currents mediated by individual VGCC subtypes were dissected according to their pharmacological and biophysical properties. TG neurons were sorted into three subpopulations based on their soma size and their affinity to isolectin B4 (IB4). We found that the T666M mutation did not affect total or surface expression of CaV2.1 proteins but caused a profound reduction of P/Q-type current in all subtypes of TG neurons. Interestingly, a compensatory increase in CaV3.2-mediated low-voltage-activated T-type currents only occurred in small IB4-negative (IB4−) TG neurons expressing T666M subunits. Current-clamp recordings showed that the T666M mutation resulted in hyperexcitability of the small IB4− TG population. Taken together, our results suggest a possible scenario through which FHM-1 mutations might increase the gain of the trigeminal nociceptive pathway.
- low-voltage activated T-type current
- neuronal excitability
migraine is one of the most common neurovascular disorders that afflicts about 16% of the general population and 28 million people in the United States, with women more frequently affected than men (Lipton et al. 2001). Patients suffer from recurrent attacks of unilateral, often throbbing, headache accompanied by other neurological symptoms such as aura (Headache Classification Subcommittee 2004). Research on the pathophysiology of migraine was greatly facilitated by the discovery of gene mutations associated with familial hemiplegic migraine type 1 (FHM-1), a rare hereditary form of migraine with aura and hemiparesis (Headache Classification Subcommittee 2004; Ophoff et al. 1996). Many of the symptoms of headache and aura in FHM-1 are identical to those of the general forms of migraine, indicating common neurobiological pathways. Elucidating the functional consequences of FHM-1 mutations is not only crucial to the understanding of this rare disease but may offer a foothold for understanding general forms of migraine.
The FHM-1 gene encodes the pore-forming CaV2.1 subunit of P/Q-type Ca2+ channels. Based on their voltage dependence of activation, Ca2+ channels fall into two categories: high voltage-activated (HVA), including L-, N-, P/Q-, and R-type voltage-gated Ca2+ channels (VGCCs) that require strong depolarization for activation, and low-voltage-activated (LVA) T-type Ca2+ channels that can be triggered by much milder depolarization (Tsien and Wheeler 1999). Most excitable cells, including the primary afferent neurons in nociception, express multiple types of HVA and LVA Ca2+ channels (Borgland et al. 2001; Ikeda and Matsumoto 2003; Morikawa et al. 2006; Scroggs and Fox 1992). Ca2+ entry through P/Q-type channels is critical for excitation-secretion coupling in many presynaptic terminals as well as postsynaptic Ca2+ signaling in dendrites and cell bodies of many neurons (Cao 2006; Tsien and Wheeler 1999). Most FHM-1 mutations cause changes of amino acids in the conserved functional domains of the CaV2.1 subunit, such as within the pore or voltage sensor (de Vries et al. 2009). Consequently, all the FHM-1 mutations studied so far alter the biophysical properties of P/Q-type channels in one way or another (Pietrobon 2010; Pietrobon and Striessnig 2003). Single channel analysis of 8 FHM-1 mutations shows a negative shift of voltage-dependent activation, suggesting a gain-of-function (GOF) effect (Pietrobon and Striessnig 2003). However, when expressed in cultured hippocampal and cerebellar granule neurons as well as HEK cells, these same FHM-1 mutations by and large result in a decrease of peak whole cell P/Q-type current densities, accompanied by a negative shift of voltage-dependent activation in some but not all cases (Adams et al. 2009; Barrett et al. 2005; Cao et al. 2004; Cao and Tsien 2005; Hans et al. 1999; Kraus et al. 1998; Mullner et al. 2004; Tottene et al. 2002, 2005). It is not clear whether these discrepancies arise from other biophysical properties altered by the FHM-1 mutation, differences in expression systems, recording protocols, the animal species of CaV2.1, or the splice variant in which the mutations are introduced. Indeed, previous studies show that the effects of individual FHM-1 mutations on P/Q-type channel function depend on the CaV2.1 splice variants and auxiliary subunits (Adams et al. 2009; Mullner et al. 2004). Thus it is important to study the functional consequences of FHM-1 mutations in neurons in the trigeminal nociceptive pathway underlying migraine headache.
Recently, knockin (KI) mice carrying the R192Q and S218L mutations have been generated. P/Q-type currents show a negative shift of voltage dependence of activation and an increase in current density in cortical and cerebellar granule neurons. The KI mice exhibit decreased threshold for cortical spreading depression, the substrate for migraine aura and possibly a trigger for headache (Eikermann-Haerter et al. 2009; van Den Maagdenberg et al. 2004, 2010). In addition, the release of calcitonin gene-related peptide (CGRP) as well as the activity of ATP-gated P2X3 receptor is significantly enhanced in TG neurons from R192Q KI mice (Ceruti et al. 2011; Nair et al. 2010). However, a recent study suggests that P/Q-type current density, voltage dependence, as well as the release of CGRP in dural afferent neurons are not altered in R192Q KI mice (Fioretti et al. 2011). These data suggest that GOF mutations as represented by R192Q and S218L may result in hyperexcitability of brain areas that respond to migraine triggers.
On the other hand, is it possible that FHM-1 may also arise from a deficiency in P/Q-type channel activity? Some data indicate that this might be the case. In rats, blocking P/Q-type channel activity in midbrain periaqueductal grey facilitates nociceptive transmission evoked by dural stimulation (Knight et al. 2002). A P/Q-type channel blocker also increases spontaneous firing of nociceptive neurons at the level of cervical/medullary dorsal horn (Ebersberger et al. 2004). The incidence of migraine in patients with episodic ataxia type 2, another channelopathy associated with loss-of-function (LOF) CaV2.1 mutations, is strikingly higher (>50%) than that in general population (Jen et al. 2004, 2007; Pietrobon 2005). Recently, a large-scale gene deletion causing the truncation of 350 amino acids at the C terminus of CaV2.1 has been identified in both episodic ataxia type 2 patients as well as patients with spontaneous hemiplegic migraine (Labrum et al. 2009). These findings suggest that LOF CaV2.1 mutations may also underlie the migraine phenotype. Several important questions, however, have not yet been addressed. For example, how do these mutations affect P/Q-type currents in neurons in the trigeminal nociceptive pathway? Does the change in P/Q-type currents alter the contributions of other VGCCs to whole cell currents? Is neuronal excitability affected by the expression of mutant channels? Answering these questions is a logical step in a bottom-up approach to understanding the pathophysiology of FHM-1.
Here, we investigated the effects of the FHM-1 T666M mutation, a threonine to methionine substitution at amino acid 666 of the human CaV2.1 subunit, on VGCC currents in trigeminal ganglion (TG) neurons, the primary afferent neurons in the trigeminal nociceptive pathway. We have found that the T666M mutation results in a decrease of P/Q-type current in all subtypes of TG neurons. Interestingly, the expression of the T666M mutant channels leads to a selective increase in LVA T-type current in small-diameter TG neurons that do not bind to isolectin B4 (IB4). Moreover, the excitability of this TG subpopulation is significantly increased by the expression of T666M mutant channels. This study provides the first detailed characterization of VGCC currents and neuronal excitability in TG neurons expressing T666M mutant channels.
MATERIALS AND METHODS
All procedures in this study were approved by the Animal Studies Committee at Washington University in St. Louis. CaV2.1 wild-type and knockout (KO) pups were obtained by crossing heterozygous breeders on the C57BL/6 background. The genotype was determined by PCR of tail DNA as described previously (Jun et al. 1999). The breeders were maintained on a 12-h light-dark cycle with constant temperature (23–24°C) and humidity (45–50%) as well as food and water ad libitum at the animal facility of Washington University in St. Louis.
Human CaV2.1 constructs.
All constructs for mammalian expression of hWT, R192Q, and T666M mutant CaV2.1 subunits have been used in previous studies (Barrett et al. 2005; Cao et al. 2004; Cao and Tsien 2005). Briefly, the cDNA of human wild-type (hWT) CaV2.1 BI-1 (V1) splice variant was cloned in plasmid pCDNA3.1(+) (Invitrogen), downstream of the CMV promoter. The splice variant content at the seven loci is the following: Δ10A (−V + G), 16+17+, Δ17A (−VEA), −31* (−NP), 37a (EFa), 43+44+, and Δ47 (Soong et al. 2002). This construct was used as backbone to generate human CaV2.1 cDNA with the T666M or R192Q mutation, using the Stratagene QuickChange site-directed mutagenesis kit. All PCR-generated cDNA fragments and linker regions were completely sequenced to verify the mutation and to make sure that no extra errors were introduced. The enhanced green fluorescent protein (EGFP)-tagged CaV2.1 constructs were generated by fusing hWT and T666M cDNAs in frame at the C terminal of EGFP cDNA (pEGFP-C1 plasmid; Clontech). To quantify surface expression of CaV2.1 subunits, a hemagglutinin (HA)-tag was inserted in-frame into the domain II S5-S6 loop of EGFP-tagged CaV2.1 constructs as previously described (Barrett et al. 2005).
Primary culture of neonatal mouse TG neurons and transfection.
TG tissues were collected from postnatal days 1–2 (P1–P2) mice and were treated with 5 mg/ml trypsin for 15 min. Neurons were dissociated by triturating with fire-polished glass pipettes and were seeded on Matrigel-coated coverslips. The culture medium contained 5% FBS, 10 ng/ml nerve growth factor, and 10 ng/ml glial cell line-derived neurotrophic factor, and was replaced every 3 days. Neurons were transfected at 1 day in vitro (DIV) using lipofectamine 2000 (Invitrogen). For electrophysiology experiments, neurons were transfected with constructs encoding untagged hWT or T666M CaV2.1 subunits along with pEGFPC1 plasmid (1:1 molar ratio). Transfected neurons were identified by EGFP fluorescence and were recorded between 3–6 DIV. To compare the total and surface level of hWT and T666M subunits, neurons were transfected with constructs encoding EGFP- and HA-tagged hWT or T666M CaV2.1 and were fixed 2 days posttransfection.
Size distribution and the affinity to IB4 of TG neurons.
The differential interference contrast (DIC) images of cultured TG neurons were captured using a Nikon TE2000S inverted microscope equipped with a CoolSnapEZ camera and were analyzed with SimplePCI software (Hamamatsu). Soma diameter was converted from cross-sectional area measured from the DIC image. At the end of the electrophysiological recording, neurons were incubated with Alexa Fluor 594-conjugated IB4 (3 μg/ml) for 10 min. The Alexa Fluor 594 fluorescence on soma membrane was detected after 10-min perfusion to wash off unbound IB4.
Image analysis of EGFP-tagged CaV2.1 proteins.
One-DIV TG neurons were transfected with constructs encoding EGFP-tagged hWT or T666M CaV2.1 subunits. Two days posttransfection, neurons were fixed with 4% paraformaldehyde at 4°C for 10 min. EGFP images were captured through a ×40 objective on a Nikon TE2000S inverted epifluorescent microscope equipped with a CoolSnapEZ camera (Photometrics). The intensity of EGFP signal was quantified with SimplePCI software (Hamamatsu).
Immunostaining of cultured TG neurons.
The immunostaining of IB4 and CGRP was performed in 1-DIV-cultured TG neurons. The coverslips were washed with PBS. Neurons were fixed by 4% formaldehyde for 5 min followed by PBS wash. The coverslips were incubated in blocking buffer (PBS with 10% normal goat serum and 0.3% Triton X-100) for 1 h and were then incubated with the CGRP antibody (1:1000; Millipore) and FITC-conjugated IB4 (2 μg/ml; Sigma) in blocking buffer at 4°C overnight. Following three rinses by PBS and three washes by the blocking buffer (20 min each), the coverslips were incubated with AlexaFluor 568-conjugated goat anti-rabbit secondary antibody (1:1000; Invitrogen) in blocking buffer for 1 h and washed again three times in PBS.
To quantify the surface expression of CaV2.1 subunits, 1-DIV TG neurons were transfected with constructs encoding EGFP- and HA-tagged hWT or T666M CaV2.1 subunits. Two days posttransfection, neurons were fixed with 4% paraformaldehyde at 4°C for 10 min. The coverslips were incubated with a monoclonal anti-HA antibody (Covance; 1:500) at 4°C overnight, followed by staining with AlexaFluor 594-conjugated goat anti-mouse secondary antibody (1:1000; Invitrogen). Triton X-100 was omitted in all solutions.
After immunostaining, the coverslips were mounted with the Crystal/Mount medium (Sigma) and stored at 4°C. The images were examined and captured through a ×40 objective on a Nikon TE2000S inverted epifluorescence microscope equipped with a CoolSnapEZ camera (Photometrics). The intensity of fluorescent signals was quantified with SimplePCI software (Hamamatsu).
One-DIV neurons were transfected with constructs encoding untagged hWT or T666M CaV2.1 subunits along with pEGFPC1 plasmid (1:1 molar ratio). Transfected neurons were identified by EGFP fluorescence and were recorded between 3–6 DIV with 5 mM Ba2+ as charge carrier. Most (>95%) EGFP-positive cells expressed CaV2.1 subunits (Cao et al. 2004; Cao and Tsien 2005). Whole cell patch-clamp recordings were performed at room temperature with a MultiClamp 700B amplifier (Molecular Devices). pClamp 10 (Molecular Devices) was used to acquire and analyze data. Cell capacitance and series resistance were constantly monitored throughout the recording.
Ba2+ current recording.
The recording chamber was perfused with extracellular solution (0.5 ml/min) containing the following (in mM): 20 CsCl, 140 TEA-Cl, 5 BaCl2, 10 HEPES, 25 glucose, pH 7.3 with TEA-OH, and 310 mosmol/kgH2O. The pipette solution contained the following (in mM): 110 CsCl, 10 EGTA, 4 ATP-Mg, 0.3 GTP-Na, 25 HEPES, 10 Tris-phosphocreatine, 20 U/ml creatine phosphokinase, pH 7.3 with CsOH, and 290 mosmol/kgH2O. Recording pipettes had <3.5 MΩ resistance (average 2.7 MΩ). Series resistance (<15 MΩ, average 7.5 MΩ) was compensated by 80%. Current traces were corrected with online P/6 trace subtraction. Signals were filtered at 1 kHz and digitized at 10 kHz except for the tail current measurement experiments (filtered at 10 kHz and digitized at 100 kHz). The DIC images of neurons were captured before patch-clamp to calculate soma diameter. At the end of the recording, neurons were incubated with Alexa Fluor 594-conjugated IB4 (3 μg/ml) to test for IB4 affinity.
To measure the current-voltage relationships (I-V curves) of HVA Ca2+ channels, neurons were held at −80 mV and were depolarized from −80 to +100 mV (10 mV increments) for 40 ms and then repolarized back to −80 mV. Amplitudes of tail currents were normalized to the largest tail and data were fitted by a single Boltzmann function. To dissect whole cell Ba2+ current through individual subtypes of HVA Ca2+ channels, we bath-applied blockers for each VGCC while evoking Ba2+ currents with ramp depolarization pulses from −80 to +100 mV at 1.8 mV/ms every 10 s. L-, N-, and P/Q-type currents were identified as currents sensitive to their specific blockers: nifedipine (10 μM), ω-conotoxin-GVIA (ω-CTx-GVIA; 2 μM), and ω-agatoxin-IVA (ω-Aga-IVA; 0.5 μM), respectively. The remaining Cd2+-sensitive (100 μM) current was taken as R-type current. For each neuron, the peak current was normalized by the membrane capacitance (a measure of cell surface area) to reflect current density. This paradigm has been used to dissect HVA VGCC components in TG and dorsal root ganglion (DRG) neurons in previous studies (Borgland et al. 2001; Ikeda and Matsumoto 2003; Mintz et al. 1992; Morikawa et al. 2006; Scroggs and Fox 1992).
We used the following protocol to dissect LVA T-type currents, building on the knowledge that the majority of T-type VGCCs are inactivated at more depolarizing holding potentials (Perez-Reyes 2003). First, we blocked most, if not all, HVA VGCC currents by bath-applying a cocktail of channel blockers, including 10 μM nifedipine, 2 μM ω-CTx-GVIA, 0.5 μM ω-Aga-IVA, and 0.2 μM SNX482 [for a subset of R-type current (Fang et al. 2007; Iftinca et al. 2007)]. Next, we recorded Ba2+ currents at −40 mV for 40 ms by depolarization from either −60- or −110-mV holding potential (Vh). Current through T-type VGCCs was then calculated by digitally subtracting currents measured from −60 mV Vh from those measured from −110 mV Vh. This eliminated the residual HVA VGCC current that was not blocked by the cocktail of blockers (Zhang et al. 2002).
Voltage-dependent activation of T-type Ca2+ channels in TG was studied with tail currents. HVA VGCC currents were minimized with the cocktail of channel blockers. Neurons were held at either −60 or −110 mV and were depolarized from −80 to +30 mV (10-mV increments) for 40 ms and then repolarized back to Vh. After digital subtraction (see above), the amplitude of tail currents was normalized to the largest tail and data were fitted by a single Boltzmann function.
DIC images of transfected neurons were captured to calculate soma diameter. Neurons were incubated with Alexa Fluor 594-conjugated IB4 (3 μg/ml) to test for IB4 affinity. The transfected, small IB4-negative (IB4−) TG neurons were used in current-clamp experiments. Prestaining of IB4 does not alter properties of primary sensory neurons (Stucky and Lewin 1999). The recording chamber was perfused with Tyrode solution (0.5 ml/min) containing the following (in mM): 130 NaCl, 2 KCl, 2 CaCl2, 2 MgCl2, 25 HEPES, 30 glucose, pH 7.3 with NaOH, and 310 mosmol/kgH2O. The pipette solution contained the following (in mM): 130 K-gluconate, 7 KCl, 2 NaCl, 1 MgCl2, 0.4 EGTA, 4 ATP-Mg, 0.3 GTP-Na, 10 HEPES, 10 Tris-phosphocreatine, 20 U/ml creatine phosphokinase, pH 7.3 with KOH, and 290 mosmol/kgH2O. Recording pipettes had <4.5 MΩ resistance. Series resistance (<20 MΩ) was not compensated. Signals were filtered at 10 kHz and digitized at 50 kHz. After whole cell access was established, the amplifier was switched to current-clamp mode to measure resting membrane potential (Vrest). The input resistance (Rin) was calculated by measuring the change of membrane potential in response to a 20-pA hyperpolarizing current injection from Vrest. Neurons were excluded from analysis if the Vrest was higher than −40 mV or Rin was <200 MΩ.
To test neuronal excitability, neurons were held at Vrest and were injected with 1-s depolarizing currents in 25-pA incremental steps until at least one action potential (AP) was elicited. The rheobase was defined as the minimum amount of current to elicit ≥1 AP. The first AP elicited using this paradigm was used to measure AP threshold (the membrane potential at which dV/dt exceeds 10 V/s), amplitude and half width. The amplitude of afterhyperpolarization (AHP) was measured from the single AP elicited by injecting 0.5- to 2-nA depolarizing current for 0.5 ms. To measure spike frequency in response to suprathreshold stimuli, we injected neurons with 1-s depolarizing currents at one-, two-, and threefold rheobase.
Nifedipine (Sigma-Aldrich) was dissolved in 100% ethanol to generate 10 mM stock solution. All peptide Ca2+ channel blockers were purchased from Peptides International (Louisville, KY), reconstituted in 1 mg/ml cytochrome c (Sigma-Aldrich) at ×100 concentrations and stored at −80°C in aliquots. Before the application of Ca2+ channel blockers, the perfusion of the recording chamber was stopped and cytochrome c was added to the bath to 0.1 mg/ml final concentration. Subsequently, the blockers were applied sequentially and cumulatively to the recording chamber.
All data are reported as means ± SE. Data were analyzed with Clampfit (Molecular Devices) and Origin (OriginLab) softwares. Statistical significance between experimental groups was assessed by a two-tail t-test or ANOVA with post hoc Bonferroni test. Statistical significance between I-V curves and between spike frequency curves was assessed by repeated-measures ANOVA (RM ANOVA) with post hoc Bonferroni test.
Whole cell VGCC currents in cultured TG neurons from wild-type and CaV2.1 KO mice.
Our first objective was to test whether CaV2.1 KO TG neurons can be used as the platform to study the effect of the FHM-1 T666M mutation on VGCC currents, avoiding interference from endogenous P/Q-type Ca2+ channels. To this end, we cultured TG neurons from neonatal (P1–P2) wild-type and CaV2.1 KO mice and measured whole cell VGCC currents with 5 mM Ba2+ as the charge carrier to avoid complications from Ca2+-dependent modulation of VGCCs. When subjected to a ramp depolarization from −80 to +100 mV, both wild-type and CaV2.1 KO TG neurons exhibited typical VGCC I–V curves (Fig. 1A). The inward currents started around −40 mV, peaked around 0 mV, and reversed around +50 mV. However, the total Ba2+ current density of CaV2.1 KO TG neurons was reduced by ∼30% relative to that of wild-type neurons between −10 mV to +20 mV (RM ANOVA with post hoc Bonferroni test, P < 0.05; Fig. 1A).
TG neurons express multiple types of HVA Ca2+ channels (L-, N-, P/Q-, and R-type). We asked whether the loss of P/Q-type Ca2+ channels solely accounts for the decrease of total VGCC current in CaV2.1 KO TG neurons. Are the current densities from other HVA Ca2+ channels altered in CaV2.1 KO neurons as well? To address these questions, we measured whole cell Ba2+ currents through P/Q-type and other types of HVA VGCCs. Neurons were depolarized with the ramp protocol every 10 s. Subtype-specific VGCC blockers nifedipine (10 μM), ω-CTx-GVIA (2 μM), and ω-Aga-IVA (0.5 μM) were sequentially and cumulatively introduced into the recording chamber to block L-, N-, and P/Q-type Ca2+ channels, respectively (Fig. 1C). The remaining Cd2+-sensitive (100 μM) current was taken as R-type current. Little or no Ba2+ current could be detected after Cd2+ application. The addition of each blocker produced a stepwise reduction in current size that was readily quantified (Fig. 1C). This paradigm has been used to dissect HVA VGCC components in TG and DRG neurons previously (Borgland et al. 2001; Ikeda and Matsumoto 2003; Mintz et al. 1992; Morikawa et al. 2006; Scroggs and Fox 1992).
As expected, CaV2.1 KO TG neurons did not express P/Q-type currents (Fig. 1, B and D), whereas ∼30% of whole cell VGCC currents in wild-type TG neurons (142.6 ± 6.8 pA/pF at 0 mV) was mediated by P/Q-type channels (41.0 ± 4.1 pA/pF at 0 mV; Fig. 1D). On the other hand, L-, N-, and R-type current densities were not significantly different between wild-type and KO neurons (Fig. 1D). Thus other HVA Ca2+ channels did not compensate for the absence of P/Q-type channels in CaV2.1 KO neurons. These results indicate that CaV2.1 KO TG neurons in culture are a valid system for exploring the functional significance of FHM-1 mutations.
Similar expression level of hWT and T666M CaV2.1 subunits in TG neurons.
Next, we tested whether human CaV2.1 subunits can be expressed in cultured TG neurons at an efficiency adequate for electrophysiological study. Numerous studies have demonstrated that cell classifications are very important for the interpretation of data from primary sensory neurons in in vitro experiments. The profile of endogenous VGCCs and auxiliary subunits varies between each subgroup of primary sensory neurons (Borgland et al. 2001; Ikeda and Matsumoto 2003; Morikawa et al. 2006; Scroggs and Fox 1992). To account for the heterogeneity in cultured TG neurons, we first sorted neurons by soma diameter into small- (<25 μm) and medium-diameter (25–35 μm) groups. We used 25 μm as a cut-off value based on the previous studies that the average diameter of small neurons in mouse TG is around 20 μm (Borgland et al. 2001; Catacuzzeno et al. 2008). Most TG neurons in our cultures were small-diameter neurons (Fig. 2A open bars and Table 1). Very few TG neurons have soma diameter larger than 35 μm (Fig. 2A), as the cell bodies of the large-diameter proprioceptors are localized in the mesencephalic trigeminal tract nucleus in the brain stem. Secondly, we divided the small-diameter neurons into IB4-positive (IB4+) and IB4− groups, based on their ability to bind to fluorescently-labeled IB4 (Stucky and Lewin 1999). The binding of IB4 is commonly used to define the nonpeptidergic populations of primary afferents: sensory neurons that express little or very low levels of neuropeptides (Julius and Basbaum 2001; Snider and McMahon 1998). On the other hand, most small IB4− neurons express CGRP, a neuropeptide that plays important role in migraine pathophysiology (Ho et al. 2010; Julius and Basbaum 2001; Snider and McMahon 1998).
To test what percentage of TG neurons express CGRP, we immunostained TG cultures with a CGRP antibody and FITC-labeled IB4. Table 1 shows that 42% of total TG neurons belonged to the small IB4− subgroup and exhibited CGRP immunoreactivity (CGRP-ir). Only 9% of total TG neurons were small IB4− but devoid of CGRP-ir. Thus the majority (82%) of the small IB4− neurons in our culture expressed CGRP, consistent with a previous study (Durham and Russo 1999). On the other hand, 72% of neurons containing CGRP-ir belonged to the small IB4− TG subpopulation (Table 1). In addition, 9% of total TG neurons bound IB4 and contained CGRP-ir, in line with a previous report (Price and Flores 2007).
To test for transfection efficiency, we expressed EGFP in TG neurons from wild-type mice. All three subpopulations of TG neurons could be transfected (Fig. 2, B–D). More than 80% of transfected neurons were small-diameter neurons (Fig. 2A, filled bars). The size distribution of transfected and untransfected TG neurons was comparable (Kolmogorov-Smirnov test, P = 0.5)
We proceeded to test whether the FHM-1 T666M mutation, a threonine to methionine substitution at amino acid 666 of the human CaV2.1 subunit (Fig. 3A), alters the level of CaV2.1 protein in TG neurons. The T666M mutation was introduced into human CaV2.1 as described previously (Cao et al. 2004). We expressed EGFP-tagged hWT and T666M CaV2.1 subunits in wild-type TG cells and monitored EGFP fluorescence in the soma. Most of the transfected neurons were of small diameter (18.5 ± 1.1 μm for hWT and 17.3 ± 0.8 μm for T666M group, respectively). Neurons in both groups exhibited diffuse EGFP fluorescence throughout the soma (Fig. 3B), similar to the immunostaining pattern of endogenous P/Q-type channels (Nair et al. 2010). Quantitative analysis showed a comparable level of somatic EGFP fluorescence intensity in TG neurons expressing hWT or T666M CaV2.1 subunits (Fig. 3C), suggesting that the T666M mutation does not affect the expression level of CaV2.1 proteins in TG neurons.
Next, we examined whether the T666M mutation affects the trafficking of CaV2.1 subunits to the plasma membrane by expressing EGFP- and HA-tagged hWT or T666M CaV2.1 subunits in wild-type TG neurons. The EGFP tag was at the N terminus, and the HA epitope was inserted within the extracellular S5-S6 loop of domain II. We measured HA immunoreactivity in transfected, nonpermeabilized neurons as an index of the amount of CaV2.1 subunits on the plasma membrane. EGFP signals were observed in both the plasma membrane and intracellular compartments (Fig. 3D, left). In contrast, HA immunoreactivity was restricted to the cell surface (Fig. 3D, middle). The hWT- and T666M-expressing neurons displayed similar levels of surface HA immunoreactivity (Fig. 3E), indicating that the T666M mutation does not affect the ability of the channel to traffic to the plasma membrane.
T666M mutation results in decrease of P/Q-type current density in all subtypes of TG neurons.
We went on to investigate the effects of the T666M mutation on P/Q-type currents in TG neurons. We transfected neonatal CaV2.1 KO TG culture to express EGFP alone, EGFP with hWT, or EGFP with T666M CaV2.1 subunits. Whole cell VGCC currents in transfected TG neurons were measured 2–5 days posttransfection. Neurons were depolarized with a ramp protocol and P/Q-type current was identified as the Ba2+ current sensitive to 0.5 μM ω-Aga-IVA blockade (Fig. 4A). CaV2.1 KO neurons expressing EGFP alone did not exhibit P/Q-current (Fig. 4, A and B, EGFP group), whereas neurons expressing hWT subunits displayed I–V curves typical of P/Q-type channels, with peak inward current between −10 and 0 mV (Fig. 4, A and B, hWT group). The mean current density for hWT channels was 119.6 ± 7.7 pA/pF at −5 mV, threefold the size of the endogenous P/Q-type current in wild-type TG (39.3 ± 4.4 pA/pF; Fig. 5B; P < 0.05). Thus CaV2.1 KO TG neurons can support the functional expression of human CaV2.1 subunits at threefold their normal level without introduction of extra auxiliary subunits.
In neurons expressing T666M CaV2.1 subunits, P/Q-type current densities were significantly smaller across a broad voltage range (RM ANOVA with post hoc Bonferroni test, P < 0.05; −20 to +30 mV) relative to those expressing hWT channels (Fig. 4B), consistent with previous studies of the T666M mutation in other types of neurons and cell lines (Barrett et al. 2005; Cao et al. 2004; Hans et al. 1999; Tottene et al. 2002). When depolarized to 0 mV, the current density of T666M channels (17.3 ± 2.2 pA/pF) was only 15% of that generated by hWT channels (111.2 ± 7.3 pA/pF; P < 0.05; Fig. 4B), indicating that T666M has a LOF effect on P/Q-type current in TG neurons.
In previous studies, the FHM-1 R192Q mutation exhibits a GOF effect at the single channel level as well as on whole cell current density in neurons from R192Q KI mice, yet it causes a reduction of whole cell P/Q-type current density when expressed in cultured hippocampal and cerebellar granule neurons from CaV2.1 KO mice (Cao et al. 2004; Hans et al. 1999; Tottene et al. 2002; van Den Maagdenberg et al. 2004). This has led to the proposal that transfected neurons are not appropriate to study the effects of ion channel mutations on the whole cell current density, as any alteration of current density might be an artifact due to the overexpression of the exogenous channels (van Den Maagdenberg et al. 2004). We addressed this question by recording whole cell P/Q-type current densities in CaV2.1 KO TG neurons overexpressing hWT or R192Q subunits. We used 2 mM Ba2+ as the charge carrier in anticipation of a GOF effect of the R192Q mutation. Indeed, TG neurons expressing R192Q subunits exhibited significantly larger P/Q-type current densities compared with neurons expressing hWT channels (Fig. 4C; RM ANOVA with post hoc Bonferroni test, P < 0.05; 0 to +20 mV). When depolarized to 0 mV, the current density of R192Q channels (81.7 ± 11.7 pA/pF) was 60% greater than that generated by hWT channels (50.9 ± 6.3 pA/pF; Fig. 4C). We conclude that expressing exogenous P/Q-type channels in CaV2.1 KO TG culture is a valid method to investigate the effects of FHM-1 mutations on whole cell P/Q-type current densities. It is likely that the T666M mutation exhibits a similar LOF effect on endogenous P/Q-type channels in TG neurons.
Next, we tested whether the T666M mutation shifts the voltage dependence of activation of P/Q-type current in TG neurons by an analysis of ω-Aga-IVA-sensitive tail currents (Fig. 4D). Tails of P/Q-type current were evoked by repolarizing steps to −80 mV after brief depolarizing steps to various voltage levels. The activation curves of hWT and T666M channels overlapped with each other (Fig. 4D). Neither the midpoint voltage (V50,tail) nor the slope of the activation curve differed between the two groups (Table 2). Thus displacements in voltage-dependent activation were not responsible for the overall decrease of current density in T666M channels.
We further examined whether the T666M mutation reduces P/Q-type current density in all subtypes of TG neurons. We measured soma diameter as well as IB4 affinity of each neuron after recording. The scatterplot of P/Q-type current density at −5 mV showed a clear segregation between neurons expressing hWT and T666M CaV2.1 subunits regardless of the TG subtype to which they belong (Fig. 4E). These results indicate that T666M is a LOF mutation in mediating P/Q-type current in all TG neurons.
Total Ba2+ current densities were significantly reduced in all subtypes of TG neurons expressing T666M mutant channels.
Here we tested whether the total Ba2+ current density as well as currents through individual VGCC subtypes is affected in TG neurons expressing T666M channels. Following the same strategy as described above, we used type-specific blockers to pharmacologically dissect individual HVA Ca2+ channel components in CaV2.1 KO TG neurons expressing hWT or T666M subunits (Fig. 5A). Consistent with the profound reduction of P/Q-type current density, TG neurons expressing T666M mutants displayed a ∼50% reduction of total Ba2+ current density (102.0 ± 4.9 pA/pF) relative to those expressing hWT subunits (217.0 ± 13.0 pA/pF; P < 0.001; Fig. 5B) when depolarized to −5 mV. On the other hand, L-, N-, and R-type current densities were not significantly different between wild-type TG neurons and CaV2.1 KO neurons expressing EGFP alone, EGFP with hWT, or EGFP with T666M subunits (Fig. 5B), suggesting that the magnitude of P/Q-type currents does not affect current densities through other HVA Ca2+ channels.
Does the T666M mutation cause similar changes of whole cell Ba2+ current in all subtypes of TG neurons? We found that in all three subtypes of TG neurons, the expression of T666M subunits leads to a reduction of total Ba2+ current density across a broad voltage range (RM ANOVA with post hoc Bonferroni test, P < 0.05; −10 mV to +20 mV; Fig. 5, C–E), compared with the corresponding hWT groups. Interestingly, only the small IB4− T666M-expressing neurons showed a leftward shift of I–V curve (Fig. 5C). Tail current analysis indicated that the V50,tail was significantly shifted to more negative voltage in small IB4− TG neurons expressing T666M subunits (−25.3 ± 1.7 mV) compared with that in small IB4− neurons expressing hWT subunits (−17.0 ± 1.1 mV, p < 0.01, Fig. 5C, inset, and Table 3). The T666M mutation did not affect voltage-dependent activation of whole cell Ba2+ currents in small IB4+ or medium-sized TG neurons (Table 3).
Deficiency of P/Q-type current results in a selective increase in LVA T-type current in small IB4− subpopulation of TG neurons.
The leftward shift of V50,tail in small IB4− T666M-expressing neurons may result from an increase in currents mediated by LVA T-type Ca2+ channels. A change in T-type currents has been observed in several types of neurons from mice carrying spontaneous CaV2.1 mutations (Nahm et al. 2005; Xie et al. 2007; Zhang et al. 2002). We therefore compared the T-type current densities in CaV2.1 KO TG neurons expressing hWT or T666M channels.
The protocol we used to dissect T-type currents is based on the knowledge that the majority of T-type channels are inactivated at more depolarized potentials (Perez-Reyes 2003). First, we minimized HVA VGCC currents with a cocktail of type-specific blockers, including 10 μM nifedipine, 2 μM ω-CTx-GVIA, 0.5 μM ω-Aga-IVA, and 0.2 μM SNX482 for a subset of R-type current (Fang et al. 2007; Iftinca et al. 2007). Next, we recorded Ba2+ currents by depolarizing neurons to −40 mV from either −60 or −110 mV Vh (Fig. 6A). T-type currents were then calculated by digitally subtracting currents measured from −60 mV Vh from those measured from −110 mV Vh (Fig. 6A). This eliminated the residual HVA VGCC component (Zhang et al. 2002). We cannot exclude the possibility that the T-type currents recorded by this protocol may still contain a fraction of SNX482-insensitive R-type current that has similar biophysical properties as the T-type currents (Tottene et al. 2000). On the other hand, previous studies show that TG neurons express a low level of R-type channels (Borgland et al. 2001; Fioretti et al. 2011). Moreover, the SNX482-insensitive R-type currents with T-type-like properties have not been reported in neonatal TG neurons.
We first examined T-type current densities in small IB4− neurons, the TG subpopulation that is highly relevant to migraine pathophysiology (Ho et al. 2010). CaV2.1 KO neurons expressing EGFP alone exhibited a substantial amount of T-type current (13.5 ± 2.5 pA/pF; Fig. 6B, EGFP group). Expression of hWT subunits significantly reduced T-type current densities (2.8 ± 0.9 pA/pF; P < 0.05; Fig. 6B, hWT group) to the level comparable to that of wild-type small IB4− TG neurons (4.8 ± 1.7 pA/pF; Fig. 6B). On the other hand, T-type current densities in neurons expressing T666M subunits (11.9 ± 2.2 pA/pF; Fig. 6B, T666M group) were comparable to that of KO neurons but more than threefold larger than that of hWT-expressing neurons (P < 0.05). Our results suggest that deficiency in P/Q-type channel activity leads to an increase in T-type currents in the small IB4− TG neuron population. Furthermore, a scatterplot of T- vs. P/Q-type current densities recorded from the same neurons showed an inverse correlation between the size of P/Q- and T-type currents in small IB4− TG neurons (r = −0.85; P < 0.01; Fig. 6C).
Next, we tested whether the magnitude of T-type current in other subtypes of TG neurons was also changed as the result of deficient P/Q-type channels. Interestingly, T-type current densities in small IB4+ and medium-sized TG neurons were comparable among the hWT, T666M, and EGFP groups (Fig. 6B). Thus expression of the T666M mutant channels results in deficient P/Q-type activity in all subtypes of TG neurons, yet the increase in T-type current density only occurs in the small IB4− population.
Three genes (CaV3.1, CaV3.2, and CaV3.3) encode the pore-forming subunit of T-type Ca2+ channels. Which CaV3 subunit accounts for the upregulation of T-type current in small IB4− TG neurons? Previous studies indicate that CaV3.2 displays the highest sensitivity to Ni2+ blockade and is the predominant subtype expressed in DRG neurons (Bourinet et al. 2005; Chen et al. 2003; Perez-Reyes 2003; Talley et al. 1999). In CaV2.1 KO neurons, 10 μM Ni2+ blocked the majority of T-type currents (65.2 ± 3.8%; n = 3). Since Ni2+ at this concentration preferentially inhibits CaV3.2 channels (Perez-Reyes 2003), we conclude that deficiency of P/Q-type channel activity leads to a selective increase in CaV3.2-mediated T-type current in small IB4− TG neurons.
We went on to compare the I–V curves of T-type channels in transfected small IB4− TG neurons. After minimizing HVA VGCC currents with the cocktail of blockers, we recorded the I–V curves of each neuron from −110 mV Vh and −60 mV Vh. The T-type component was then derived by digitally subtracting currents measured from −60 mV Vh from those measured from −110 mV Vh. As expected, T-type currents initiated at a much more negative voltage (approximately −50 mV; Fig. 6D) than HVA Ba2+ currents (Fig. 3B). Between −50 and −10 mV, the T-type current densities of CaV2.1 KO neurons as well as T666M-expressing neurons were significantly larger than those of neurons expressing hWT channels (RM ANOVA with post hoc Bonferroni test, P < 0.05; Fig. 6D).
Moreover, we studied the voltage dependence of T-type channel activation in greater detail by tail current analysis. We observed a significant negative shift of V50,tail in CaV2.1 KO neurons and T666M-expressing neurons, compared with that in hWT-expressing neurons (P < 0.01; Fig. 6E and Table 2). Of note, the voltage dependence of activation of T-type channels did not differ between T666M and EGFP groups. We conclude that a negative shift of the voltage dependence of activation may account for the increase in T-type current density across a broad voltage range in small IB4− TG neurons with deficiency of P/Q-type channel activities.
Increased neuronal excitability in small IB4− TG neurons expressing T666M mutant channels.
It is well established that VGCCs, in particular the T-type Ca2+ channels, regulate neuronal excitability directly by providing a depolarizing ionic current as well as indirectly by modulating Ca2+-activated K+ currents that underlie AP repolarization, spike-frequency adaptation, and AHP in various cells. Here we investigated the effect of the T666M mutation on the excitability of small IB4− TG neurons. We expressed hWT or T666M channels in CaV2.1 KO TG neurons and measured the electrophysiological properties of the transfected, small IB4− TG neurons using current-clamp recording between 2–4 days posttransfection. The values of input resistance (Rin) and Vrest were all comparable between the hWT and T666M groups (Table 4).
To assess neuronal excitability, we held neurons at Vrest and injected 1-s depolarizing currents at 25-pA incremental steps to elicit APs. All neurons expressing hWT channels generated a single spike at the rheobase (the minimum amount of current to elicit an AP, 83 ± 7 pA). The rheobase of T666M-expressing neurons was significantly lower (38 ± 3 pA; P < 0.05; Table 4 and Fig. 7B). Furthermore, 75% of T666M-expressing neurons (6 out of 8) produced multiple spikes in response to rheobase current injection (P < 0.001, Fisher's exact test).
Next, we injected 1-s depolarizing current at one- to threefold rheobase of each neuron. In both hWT and T666M groups, the number of APs increased almost linearly in response to incremental current injection (Fig. 7C; P < 0.05 at 2∼3 × rheobase, one-way RM ANOVA with post hoc Dunnett test comparing to 1× rheobase in each group). However, compared with neurons expressing hWT channels, neurons expressing T666M mutant channels generated significantly more number of spikes in response to threshold as well as suprathreshold current injection (Fig. 7C; P < 0.05 at 1∼3 × rheobase, two-way RM ANOVA and post hoc t-test with Bonferroni correction). Taken together, we conclude that the FHM-1 T666M mutation results in increased excitability of small IB4− TG neurons.
On the other hand, the T666M mutation did not alter AP threshold (P = 0.32), amplitude, half width, or AHP amplitude to any significant degree in small IB4− TG neurons (Table 4), suggesting that P/Q- or T-type currents do not play a major role in the polarization or the repolarization process in small IB4− TG neurons.
We have described functional consequences of the FHM-1 T666M mutation on both VGCC currents and neuronal excitability in TG neurons. Our experiments demonstrate that the T666M mutation causes a dramatic reduction in P/Q-type current in all subtypes of TG neurons but did not affect the expression level of CaV2.1 proteins or their trafficking to the plasma membrane. A previous study of the T666M mutation in HEK cells indicates that the decreased ionic current seen with T666M channels can be attributed to a reduced number of channels available to undergo the voltage-dependent conformational changes needed for channel opening, not to fewer channel proteins expressed on the cell surface (Barrett et al. 2005). A similar mechanism may underlie the decrease of P/Q-type current densities in T666M-expressing TG neurons.
Our results also show that the T666M mutation does not affect the voltage dependence of activation of this CaV2.1 isoform in TG neurons, consistent with a previous study of the T666M mutation in the same CaV2.1 splice variant in HEK cells (Barrett et al. 2005). It is possible, though, that the space-clamp issue resulting from the extensive processes of transfected neurons prevents the detection of a small shift of activation (Hans et al. 1999). It is also possible that the T666M mutation may produce a hyperpolarizing shift of CaV2.1 channel activation in other types of neurons (Adams et al. 2009; Fioretti et al. 2011; Mullner et al. 2004).
At TG neuron terminals, P/Q-type channels mediate the release of both neurotransmitters and neuropeptides (Bao et al. 1998; Heinke et al. 2004; Hong et al. 1999). It is possible that the T666M mutant channels are impaired in contributing to neurotransmission in precise accord with their deficiency in supporting Ca2+ influx in TG neurons (Cao et al. 2004). Consequently, neurotransmission at T666M-expressing TG synapses would shift from P/Q- to N-dependence, compared with those expressing wild-type CaV2.1 channels. Nevertheless, the overall synaptic strength likely will be maintained, as suggested by previous studies on the effects of LOF CaV2.1 mutations on synaptic transmission (Cao et al. 2004; Cao and Tsien 2005; Zhou et al. 2003). It is pertinent to directly test the effects of T666M mutant channels on synaptic transmission at TG terminals in the future.
A notable finding in this study is the inverse correlation between the size of P/Q- and T-type currents in small IB4− TG neurons but not other TG subpopulations. Deficiency of P/Q-type channel activity, either due to the complete lack of CaV2.1 or the expression of T666M subunits, results in a selective increase in T-type current density in small IB4− TG neurons. Our result is similar to a previous study demonstrating that thalamic neurons from mice carrying the LOF CaV2.1 mutation tottering exhibit a substantial increase in T-type current density that results from the shift of voltage-dependent inactivation of T-type current in a depolarized direction (Zhang et al. 2002). In our case, the increase in T-type current density likely results from a displacement of the voltage dependence of activation of CaV3.2 channels. An alteration of CaV3 mRNA expression has also been observed in cerebellar neurons from mice carrying LOF CaV2.1 mutations (Nahm et al. 2005; Xie et al. 2007). Further work is needed to determine whether the number of T-type Ca2+ channels is altered in small IB4− TG neurons expressing T666M subunits. Our results stress that cell classifications are very important for the interpretation of data from primary afferent neurons in in vitro experiments. Future studies are necessary to clarify the mechanisms responsible for the cell-specific alteration of T-type current in TG neurons.
Of note, we recorded VGCC currents in 3- to 6-DIV TG neurons maintained in media with trophic factors, raising the concern that prolonged exposure to trophic factors may alter the expression level and the properties of Ca2+ channels. Previous studies have identified T-type currents in acutely dissociated TG neurons from adult mice (Borgland et al. 2001; Ikeda and Matsumoto 2003). Fioretti et al. (2011) reported that 27% of small TG neurons express T-type currents. Borgland et al. (2001) reported that 35% of TG neurons with prominent T-type currents bound IB4. The rest (65%) of T-type current-containing neurons presumably belonged to the small IB4− or medium-sized population. We found T-type currents in all subpopulations of TG neurons from neonatal wild-type mice with similar mean T-type current density (Fig. 6B), suggesting that the expression of T-type currents in cultured neurons from neonatal mice is comparable to that in acutely dissociated TG neurons from adult mice.
In addition to characterizing the effects of T666M mutation on VGCC currents in TG neurons, our study provided direct evidence that the expression of T666M mutant channels results in hyperexcitability of small IB4− TG neurons. This has important implications for understanding the contribution of VGCCs to migraine pathophysiology. The activation and sensitization of TG neurons innervating the dura and cerebral blood vessels are the first steps in the onset of migraine headache (Burstein 2001; Strassman et al. 1996). Subsequently, the afferent activity reaches the central terminals of TG neurons and triggers the release of neurotransmitters and neuropeptides at the cervical/medullary dorsal horn. From there, the signals are conveyed to the second order neurons and eventually to the cortex, where the perception of headache is formed (Burstein 2001). A substantial percentage of IB4− TG neurons projecting to dura and intracranial vessels express CGRP, a neuropeptide that plays an important role in migraine pathophysiology (Ho et al. 2010). Taken together, our results suggest a possible scenario through which LOF FHM-1 mutations increase the gain of the trigeminal nociceptive pathway underlying migraine headache.
Both the reduction of HVA Ca2+ currents and the increase in T-type currents in DRG neurons have been reported in animal models of persistent pain (Abdulla and Smith 2001a; Jagodic et al. 2008, 2007; McCallum et al. 2006). The reduction of HVA Ca2+ currents may attenuate Ca2+-sensitive K+ currents and, in turn, increase neuronal excitability (Abdulla and Smith 2001b; Hogan 2007). The contribution of T-type VGCCs to controlling the excitability and firing patterns of DRG neurons is well established (Nelson et al. 2006; Perez-Reyes 2003). DRG neurons from CaV3.2 KO mice completely lacked T-type Ca2+ current (Chen et al. 2003). Compared with wild-type mice, CaV3.2 KO mice displayed significantly attenuated nociceptive responses in both visceral and cutaneous inflammatory pain models (Choi et al. 2007). On the other hand, chronic peripheral nerve injury in rats induced an upregulation of T-type currents and consequently an increased cellular excitability in DRG neurons (Jagodic et al. 2008; Jagodic et al. 2007). Pharmacologically blocking peripheral T-type channels as well as knocking-down of CaV3.2 expression in DRG neurons attenuated nociceptive response after nerve injury (Bourinet et al. 2005; Nelson et al. 2006). Future experiments are needed to determine which VGCC subtype(s) directly account for the hyperexcitation of small IB4− TG neurons expressing T666M mutant subunits. Whether the T666M mutation increases the excitability of all TG neurons or preferentially the small IB4− TG population also merits further investigation.
In summary, our results show that the FHM-1 T666M mutation results in deficient P/Q-type channel activity in all subtypes of TG neurons but a selective increase in LVA T-type currents in the small IB4− TG population. Furthermore, expression of T666M mutant channels leads to hyperexcitability of small IB4− TG neurons.
This work was supported by National Institute of Neurological Disorders and Stroke Grant 1R21-NS-066202–01, American Heart Association Scientist Development Grant 0735081N, and a departmental start-up fund (to Y.-Q. Cao).
No conflicts of interest, financial or otherwise, are declared by the author(s).
Author contributions: J.T. and Y.-Q.C. conception and design of research; J.T., P.L., Z.X., H.Z., and B.R.G. performed experiments; J.T., P.L., Z.X., H.Z., and B.R.G. analyzed data; J.T., P.L., Z.X., H.Z., and Y.-Q.C. interpreted results of experiments; J.T., P.L., Z.X., H.Z., and Y.-Q.C. prepared figures; J.T., P.L., Z.X., H.Z., B.R.G., and Y.-Q.C. approved final version of manuscript; Y.-Q.C. drafted manuscript; Y.-Q.C. edited and revised manuscript.
We thank Drs. Robert W. Gereau IV and Joe Henry Steinbach for helpful discussions during the preparation of this manuscript. We thank Drs. Gina M. Story and Robert W. Gereau IV for helping us establish TG culture and transfection protocols. We also thank Emily Guhl for technical assistance.
Current address for J. Tao: Dept. of Neurobiology, Medical College of Soochow University, Suzhou, China.
- Copyright © 2012 the American Physiological Society