Abstract
T-type Ca2+ channels (T-channels) are involved in the control of neuronal excitability and their gating can be modulated by a variety of redox agents. Ascorbate is an endogenous redox agent that can function as both an anti- and pro-oxidant. Here, we show that ascorbate selectively inhibits native Cav3.2 T-channels in peripheral and central neurons, as well as recombinant Cav3.2 channels heterologously expressed in human embryonic kidney 293 cells, by initiating the metal-catalyzed oxidation of a specific, metal-binding histidine residue in domain 1 of the channel. Our biophysical experiments indicate that ascorbate reduces the availability of Cav3.2 channels over a wide range of membrane potentials, and inhibits Cav3.2-dependent low-threshold-Ca2+ spikes as well as burst-firing in reticular thalamic neurons at physiologically relevant concentrations. This study represents the first mechanistic demonstration of ion channel modulation by ascorbate, and suggests that ascorbate may function as an endogenous modulator of neuronal excitability.
Introduction
Ascorbate (ascorbic acid, vitamin C) is a ubiquitous redox agent present at particularly high concentrations in neurons (Rice, 2000), but its interaction with the ion channels that control neuronal excitability is poorly understood. We have shown previously that T-type Ca2+ channels (T-channels) can be modulated by a variety of redox agents (Todorovic et al., 2001; Nelson et al., 2005; Joksovic et al., 2006); thus, we reasoned that ascorbate might also affect the gating of these channels. Ascorbate often functions as an antioxidant, readily scavenging reactive oxygen and nitrogen species. However, ascorbate can also function as a pro-oxidant by catalyzing the formation of several reactive oxygen species (ROS). Ascorbate generates ROS by reducing transition metals by a one-electron transfer mechanism, and oxygen by a two-electron transfer mechanism; the products of these reactions can subsequently interact to create several powerful ROS including hydroxyl radicals (·OH) (Stadtman, 1991). Collectively, this process is termed metal-catalyzed oxidation (MCO) (Stadtman, 1991; Stadtman, 1993). A unique feature of proteinaceous MCO is that only one or at most a select number of amino acids are modified. This site specificity has been attributed to the formation of ROS directly at metal binding sites, where they attack the functional groups of nearby residues. Histidine residues are particularly prone to MCO, in part because they are often structural components of metal binding sites (Stadtman, 1993). Here we present evidence that ascorbate inhibits Cav3.2, but not Cav3.1 or Cav3.3 T-channels by initiating the MCO of a unique, metal-binding histidine residue in domain 1 of the channel. Because Cav3.2 channels play an important role in tuning the excitability of several neuronal populations (White et al., 1989; Nelson et al., 2005; Vitko et al., 2005; Joksovic et al., 2006), ascorbate may function as an endogenous modulator of neuronal firing under physiological or pathological conditions.
Materials and Methods
DRG and thalamic neurons.
Acutely dissociated dorsal root ganglion (DRG) cells were prepared from adolescent rats as described previously (Nelson et al., 2005). For recording, cells were plated onto uncoated glass coverslips, placed in a culture dish, and perfused with external solution. Thalamic slices were prepared from adolescent rats as previously described (Joksovic et al., 2006). For recording, slices were placed in a recording chamber and perfused with external solution. All experiments were performed at room temperature.
HEK293 cells.
Human embryonic kidney 293 (HEK293) cells were grown in DMEM/F12 media (Invitrogen, Grand Island, NY) supplemented with fetal calf serum (10%), penicillin G (100 U/ml), and streptomycin (0.1 mg/ml). Cav3.1 (Perez-Reyes et al., 1998), human Cav3.2 (Vitko et al., 2005), Cav3.3 (Lee et al., 1999b), and GGHH (Welsby et al., 2003) plasmids were as described previously. HHGG, GHGG, HGGG, Cav3.2(H191Q), and Cav3.1(Q172H) plasmids were constructed as described previously (Kang et al., 2006) and subcloned into pcDNA3 (Invitrogen). To construct the Cav3.2(H191C) plasmid, a human Cav3.2 cDNA (GenBank accession number AF051946) contained in pcDNA3 was mutated using oligonucleotide primers and a QuickChange site-directed mutagenesis kit (Stratagene, LA Jolla, CA). The primers were obtained from Operon (Alameda, CA) and used without purification. The full-length cDNA was reassembled using a fragment of the mutated clone. The sequence corresponding to this fragment, as well as the flanking regions, was verified by automated sequencing. The full-length rat Cav3.2 cDNA was generated by full-length RT-PCR using total RNA isolated from adult rat thalamic tissue and oligonucleotides 5′-GATAAGCTTATGACCGAGGGCACG-3′ and 5′-CGCTCTAGACTACACAGGCTCATC-3′. The resulting ∼7 kb PCR products were subcloned into pGEM T-Easy vector and then into pCNA 3.1 zeo(+) (Invitrogen). The full-length thalamic cDNA was confirmed by DNA sequencing and compared with the reported PubMed sequences for Cav3.2 cDNAs from multiple organisms as well as the rat Cav3.2 genomic DNA sequence. The His 191 residue examined in the present study is conserved across the Cav3.2 T-type channels in all species in the database including chicken, dog, mouse, rat and human. Cells were transiently cotransfected using Lipofectamine 2000 (Invitrogen) at a 10:1 molar ratio with a plasmid encoding CD8 antigen and incubated with polystyrene microbeads coated with antiCD8 antibody (Invitrogen). After 48 h, cells with bound microbeads were selected for recording.
Electrophysiology.
Recording electrodes were pulled from borosilicate glass microcapillary tubes (Drummond Scientific, Broomall, PA), and had resistances from 1 to 4 MΩ when filled with internal solution. Recordings were made using an Axopatch 200B patch-clamp amplifier (Molecular Devices, Foster City, CA). Digitization of membrane voltages and currents were controlled using a Digidata 1322A interfaced with Clampex 8.2 (Molecular Devices). We analyzed data using Clampfit 8.2 (Molecular Devices) and Origin 7.0 (Microcal Software, Northampton, MA). Currents were low-pass filtered at 2–5 kHz. Series resistance and capacitance values were taken directly from readings of the amplifier after electronic subtraction of the capacitive transients. Series resistance was compensated to the maximum extent possible (usually 50–80%). Multiple independently controlled glass syringes served as reservoirs for a gravity-driven perfusion system. The external solution for DRG and HEK293 cell experiments contained (in mm) 152 tetraethylammonium (TEA)-Cl, 10 BaCl2, and 10 HEPES, adjusted to pH 7.4 with TEA-OH. The external solution for thalamic slice experiments contained (in mm) 130 NaCl, 26 NaHCO3, 10 glucose, 2.5 MgCl2, 2 CaCl2, 1.25 NaH2PO4, and 0.001 TTX, equilibrated with a mixture of 95% O2 and 5% CO2. In some experiments, TTX was omitted to study burst firing of nucleus reticularis thalami (nRT) neurons. The internal solution for voltage-clamp experiments with DRG neurons and thalamic slices contained (in mm) 135 tetramethylammonium-OH, 40 HEPES, 10 EGTA, and 2 MgCl2, adjusted to pH 7.2 with hydrofluoric acid. The internal solution for voltage-clamp experiments with HEK293 cells contained (in mm) 110 Cs-MeSO4, 14 Cr-PO4, 10 HEPES, 9 EGTA, 5 Mg-ATP, and 0.3 Tris-GTP, adjusted to pH 7.3 with CsOH. The internal solution for current-clamp experiments with thalamic slices contained (in mm) 130 KCl, 40 HEPES, 5 MgCl2, 2 MgATP, 1 EGTA, 0.1 Na3GTP, adjusted to pH 7.3 with KOH.
Analysis.
Statistical comparisons were made using paired or unpaired t tests or Mann–Whitney U tests where appropriate. All data are expressed as mean ± SEM and p values are reported only when statistically significant (<0.05). The percentage reductions in peak current at various ascorbate and Cu2+ concentrations were used to generate concentration–response curves. Mean values were fit to the following Hill–Langmuir function: where PBmax is the maximal percentage inhibition of peak current, IC50 is the concentration that produces 50% inhibition, and h is the apparent Hill–Langmuir coefficient for inhibition. The fitted values are reported with >95% linear confidence limits. The voltage dependencies of activation and steady-state inactivation were described with single Boltzmann distributions of the following forms: where Imax is the maximal activatable current, V50 is the voltage where half the current is activated or inactivated, and k is the voltage dependence (slope) of the distribution.
Results
We used whole-cell patch-clamp recordings to determine the effects of ascorbate on voltage-gated Ca2+ currents present in acutely dissociated DRG neurons. Ascorbate rapidly inhibited T-currents (Fig. 1A), but had no effect on high-voltage-activated Ca2+ currents even at 50-fold higher concentrations (data not shown). T-current inhibition was dose-dependent and partial with an effective concentration (IC50) of 6.5 μm and a maximal reduction in peak current of 70% (Fig. 1B). The inhibition induced by brief exposure to ascorbate was partially reversible after washout, with ∼75–90% recovery occurring after exposures of less than a minute; the effects of prolonged exposures were significantly less reversible (data not shown).
To assess the effects of ascorbate on the biophysical properties of DRG T-currents, we measured current–voltage (I–V) relationships in the presence and absence of ascorbate (Fig. 1C,D). Ascorbate significantly reduced T-currents at all potentials between −70 and 20 mV, but had no significant effect on the kinetics of current activation or inactivation (Fig. 1E,F). Ascorbate did induce a 5 mV rightward shift in half-maximal activation (V50) with a decrease in voltage dependence, as well as a 5 mV leftward shift in the V50 of inactivation, but without an effect on voltage dependence (Fig. 1G,H).
Cav3.2 is overwhelmingly the predominant T-channel isoform expressed in DRG neurons (Talley et al., 1999; Chen et al., 2003). Therefore we also examined the effects of ascorbate on intact neurons in brain slices of thalamic nuclei known to express diverse T-channels (Talley et al., 1999; Joksovic et al., 2006). Ascorbate significantly inhibited T-currents from reticular (nRT, Cav3.2 and Cav3.3), but not ventrobasal (VB, Cav3.1) or laterodorsal (LD, Cav3.1 and Cav3.3) thalamic neurons, suggesting selective inhibition of Cav3.2 (Fig. 2A–C,G). We confirmed this using both recombinant human and rat T-channels heterologously expressed in HEK293 cells. Ascorbate significantly inhibited Cav3.2 currents, but had no effect on either Cav3.1 or Cav3.3 currents even at a 1000-fold higher concentration (Fig. 2D–F,H,I). Additionally, ascorbate induced gating shifts in recombinant human Cav3.2 currents (Fig. 3) that were nearly identical to those observed in native rat DRG neurons (Fig. 1).
To determine the molecular substrate for ascorbate inhibition of Cav3.2, we constructed chimeras between Cav3.1 (α1G, insensitive) and Cav3.2 (α1H, sensitive) and screened for sensitivity to ascorbate. The chimeras were named using letters to represent the α-subunit donor for each of the four channel domains (Fig. 4). GGHH currents were insensitive to ascorbate, whereas HHGG currents were sensitive, indicating that critical residues are located within domains 1 or 2 of Cav3.2 (Fig. 4A,B,E). We then constructed two single-domain chimeras, exchanging either domain 1 or 2. GHGG currents were insensitive to ascorbate whereas HGGG currents were sensitive. Thus, domain 1 of Cav3.2 is required for ascorbate inhibition (Fig. 4C–E).
Notably, ascorbate inhibition of the chimeras mirrors their reported inhibition by Ni2+ (Kang et al., 2006). This is of interest because similar to ascorbate, Ni2+ is one of the few agents capable of discriminating among T-channel isoforms, as Cav3.2 is ∼20-fold more sensitive than Cav3.1 or Cav3.3 (Lee et al., 1999a; Jeong et al., 2003; Kang et al., 2006). Furthermore, mutation of a single extracellular histidine (H) residue to glutamine (Q) at position 191 (H191Q) greatly reduces the Ni2+ sensitivity of Cav3.2, indicating that H191 is part of a high-affinity Ni2+ binding site (Kang et al., 2006). Based on these observations and studies demonstrating that ascorbate-induced MCO can selectively modify histidine residues at metal-binding sites (Uchida and Kawakishi, 1990; Stadtman, 1993; Zhao et al., 1997; Hovorka et al., 2002), we hypothesized that H191 is also important for ascorbate inhibition of Cav3.2. Consistent with this, the effects of ascorbate were completely abolished in Cav3.2(H191Q) (Fig. 4F). The biophysical parameters of Cav3.2(H191Q) currents were similar to Cav3.2, but were unaffected by ascorbate, indicating the mutation disrupted ascorbate sensitivity, but not basic channel gating (data not shown). We next attempted to confer ascorbate sensitivity to Cav3.1 by performing the analogous reverse mutation (Q172H). Figure 4F shows that Cav3.1(Q172H) currents were significantly inhibited by ascorbate, confirming the requirement of H191 for sensitivity.
Similar to Ni2+, Cav3.2 is significantly more sensitive to inhibition by other divalent transition metals such as Cu2+ and Zn2+ than Cav3.1 or Cav3.3 (Jeong et al., 2003; Traboulsie et al., 2007). Hence, H191 may not be a Ni2+-exclusive binding site, and may also be critical for high-affinity Cu2+ and Zn2+ binding. Because highly redox reactive metals such as Cu2+ and Fe3+ are much more likely to play a role in MCO than relatively redox inactive metals such as Ni2+ or Zn2+ (Stadtman, 1991, 1993), we examined whether the Cu2+ sensitivity of Cav3.2 is also disrupted by the H191Q mutation. As shown in Figure 4G, Cav3.2(H191Q) was >40-fold less sensitive to Cu2+ than Cav3.2, supporting the hypothesis that H191 is part of a general high-affinity metal binding site located on the external surface of Cav3.2.
Ascorbate generates ROS through reactions that depend on transition metals. A point that has been repeatedly demonstrated by the use of metal chelators; in particular, ascorbate inhibition of T-currents in pancreatic β cells is completely abolished by previous application of the chelator diethylenetriaminepentaacetic acid (DTPA) (Parsey and Matteson, 1993). Of note, DTPA results in a large increase in baseline T-currents, indicating that channels are subject to tonic inhibition by trace metals (Parsey and Matteson, 1993). Similarly, we found that DTPA significantly enhanced T-currents in DRG neurons, and completely occluded the effects of subsequently applied ascorbate (Fig. 5A,E). In addition to chelators, previous studies in other systems have implicated MCO as a mechanism for the effects of ascorbate by showing that catalase, which decomposes hydrogen peroxide (H2O2) and thus prevents formation of ROS via MCO, protects against ascorbate (Uchida and Kawakishi, 1990; Zhao et al., 1997; Hovorka et al., 2002). We found that catalase completely prevented the inhibition of DRG T-currents by ascorbate (Fig. 5B,E), as did high concentrations of the ROS scavenger 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (c-PTIO) (Fig. 5C,E). These results, in combination with our mutational studies, suggest that ascorbate likely modulates Cav3.2 via MCO.
In general, electrophysiological solutions contain several trace metal contaminants, with Zn2+ predominating (Li et al., 1996; Paoletti et al., 1997; Thio and Zhang, 2006). This implies that the observed increase in Cav3.2 currents after the addition of DTPA largely results from relief of tonic Zn2+ inhibition. This is interesting because Zn2+ is relatively redox inactive and does not readily participate in MCO; in fact, total replacement of Cu2+ with Zn2+ can protect against MCO in some systems (Chevion, 1988). However, histidine residues have significantly higher affinity for Cu2+ than Zn2+ (Sundberg and Martin, 1974), thus many Zn2+–binding proteins are nonetheless susceptible to modification by MCO in the presence of even trace amounts of Cu2+. This was specifically demonstrated by Hovorka et al. (2002), who showed that Zn2+-insulin was less susceptible to MCO as the Cu2+/Zn2+ ratio was decreased from 1:1 to 1:10, but that even a 1:1000 ratio was unable to protect against ascorbate completely. Along these lines, we found that increasing the Cu2+/Zn2+ ratio in our extracellular solution by adding 300 nm Cu2+, resulted in a significant increase (12%) in the inhibition produced by ascorbate (Fig. 5D,E).
Collectively, our data indicate that H191 is a critical component of a high-affinity metal binding site located on the external surface of Cav3.2, and that ascorbate inhibits Cav3.2 via interaction with trace metal contaminants bound at this site, likely resulting in MCO. It is less clear which specific amino acid(s) are modified. H191 may be the residue modified by ROS, but because the affinity of metals for any single amino acid is relatively low, H191 is more likely only one part of a multiresidue binding site. Thus, it is possible that the ascorbate insensitivity of Cav3.2(H191Q) results solely from the disruption of metal binding, which prevents subsequent MCO of a neighboring residue in the binding site. To investigate this, we made another point mutation, Cav3.2(H191C), because cysteine residues have metal binding abilities comparable to histidines, but are less sensitive to MCO. Figure 6 shows that Cav3.2(H191C) channels were ascorbate-insensitive, but highly Cu2+-sensitive, strongly suggesting that the effects of ascorbate on Cav3.2 are the result of the MCO of H191 and not a neighboring residue.
Last, we examined the ability of ascorbate to modulate Cav3.2-dependent neuronal excitability in nRT neurons in intact brain slices. The ability of thalamic neurons to fire low-threshold Ca2+ spikes (LTSs) is dependent on T-currents and contributes to the ability of these neurons to fire bursts of action potentials (APs) in response to small membrane depolarizations such as those caused by excitatory postsynaptic potentials (Perez-Reyes, 2003). As shown in Figure 7, application of ascorbate at physiologically relevant concentrations (Rice, 2000) reversibly inhibited both the isolated LTSs as well as burst-firing in current-clamp recordings from nRT neurons, suggesting that ascorbate may function as an endogenous modulator of nRT excitability.
Discussion
Our results demonstrate the remarkably specific inhibition of Cav3.2 channels by ascorbate. Indeed, ascorbate is the most selective agent capable of discriminating among T-channel isoforms yet described. Thus, ascorbate may be a useful pharmacological tool for identifying the contribution of Cav3.2 current to total Ca2+ current, as well as to cellular excitability.
Additionally, our findings suggest that MCO is the likely molecular mechanism for inhibition of Cav3.2 channels by ascorbate. Several reaction products can result from the MCO of histidine, but the most common is 2-oxo-histidine, which is created by oxidation at the C-2 position of histidine's imidazole ring (Uchida and Kawakishi, 1990; Stadtman, 1993; Schoneich, 2006). The formation of 2-oxo-histidine has been demonstrated in small peptides in vitro and in vivo using high-resolution mass spectrometry (Schoneich, 2006). However, current technology is limited to small peptides and is unable to detect such small covalent modifications in larger proteins such as Ca2+ channels. Despite being unable to demonstrate the precise reaction product, our data collectively implicate MCO as the mechanism underlying the effects of ascorbate on Cav3.2.
One particularly interesting observation of this study was the significantly increased inhibitory potency of ascorbate against both recombinant human and rat Cav3.2 channels compared with native rat Cav3.2 channels. This discrepancy may be attributable to the interaction of native Cav3.2 channels with accessory subunits or other regulatory proteins, or the presence of endogenous redox buffers that are absent or altered in heterologous systems.
Another intriguing aspect of our findings concerns the observation that the effects of brief applications of ascorbate are at least partially reversible, which is interesting given the fact that MCO is a covalent modification. However, there are numerous biological examples of reversible covalent modification serving as a mechanism for the regulation of protein function (Rakitzis, 1990; Veal et al., 2007), including modulation of ion channel gating, as exemplified by the activation of TRPA1 channels by natural isothiocyanate-containing compounds (Hinman et al., 2006). Additionally, there are several reactions involving reversible covalent modifications of histidine residues in other systems (including several mediated by ascorbate), despite the fact that the reversibility is predicted to be energetically unfavorable (Farver et al., 1998; Njus et al., 2001). It is possible that endogenous enzymes with the ability to reverse the effects of MCO exist within neurons; however, additional experiments will be necessary to evaluate this possibility.
Ascorbate can function as a neuromodulator, is concentrated in both extra (100–500 μm) and intracellular (up to 10 mm) brain spaces, and can undergo dynamic changes in a variety of physiological and pathophysiological conditions (Rice, 2000). Thus, it is interesting to speculate on the functional significance of Cav3.2 modulation by ascorbate, especially considering Cav3.2-dependent LTSs are crucial to the synchronization of low-amplitude oscillations in the loop of mutually interconnected nRT, thalamic relay, and cortical neurons (Perez-Reyes, 2003). Additionally, modulation of Cav3.2 by genetic or pharmacological means has been shown to modulate LTSs and burst firing in nRT neurons as evidenced by altered excitability in both current-clamp recordings (Joksovic et al., 2006) and in silico simulations (Vitko et al., 2005). Thus, modulation of these channels by ascorbate under conditions that favor MCO may have important consequences for a variety of physiological processes and pathological conditions. Such conditions may include those associated with neuronal injury, epilepsy, Alzheimer's disease, Menke's disease, and Wilson's disease, where there are documented increases in extracellular (free) redox-reactive transition metals such as Cu2+ and Fe3+ (Strausak et al., 2001; Bush, 2003; Mathie et al., 2006). High levels of Cu2+ (200–400 μm) can also be released from synaptic terminals onto postsynaptic cells known to contain Cav3.2 channels in such brain areas as the hippocampus and olfactory bulb (Mathie et al., 2006). Future studies are necessary to investigate these possibilities, as well as to determine whether other ion channels such as NR1/NR2A NMDA receptors that are known to contain high-affinity metal binding sites (Herin and Aizenman, 2004), show inhibition by ascorbate (Majewska and Bell, 1990; Majewska et al., 1990), and undergo covalent modification by free radicals (Aizenman et al., 1990) are subject to similar regulation.
Footnotes
-
This work was supported by National Institutes of Health Grants NS054521 (M.T.N.), HL36977 (P.Q.B.), MH077791 and NS057105 (C.F.Z.), NS38691 (E.P.-R.), and GM075299 (S.M.T.), the Canadian Institutes of Health Research (T.P.S.), and the Heart and Stroke Foundation of Canada (L.S.D.). We thank Le Banh of the University of Virginia for helpful comments on this manuscript.
- Correspondence should be addressed to Slobodan M. Todorovic, Department of Anesthesiology, University of Virginia Health System, Box 800710, Charlottesville, VA 22908. st9d{at}virginia.edu