Abstract
Expression cDNA clones of α1B-1 or α1E-3 subunits coding for human neuronal N-(Cav2.2) or R-subtype (Cav2.3) Ca2+ channels, respectively, was combined with α2-bδ and β3-a Ca2+ channel subunits, and transfected into human embryonic kidney cells for transient expression to determine whether specific types of neuronal voltage-sensitive Ca2+ channels are affected differentially by methylmercury (MeHg) and Hg2+. For both Ca2+ channel subtypes, MeHg (0.125-5.0 μM) or Hg2+ (0.1-5 μM) caused a time- and concentration-dependent reduction of current. MeHg caused an initial, rapid component and a subsequent more gradual component of inhibition. The rapid component of block was completed between 100 and 150 s after beginning treatment. At 0.125 to 1.25 μM, MeHg caused a more gradual decline in current. Apparent IC50 values were 1.3 and 1.1 μM for MeHg, and 2.2 and 0.7 μM for Hg2+ on N- and R-types, respectively. For N-type current, effects of Hg2+ were initially greater on the peak current than on the sustained current remaining at the end of a test pulse; subsequently, Hg2+ blocked both components of current. For R-type current, Hg2+ affected peak and sustained current approximately equally. Kinetics of inactivation also seemed to be affected by Hg2+ in cells expressing N-type but not R-type current. Washing with MeHg-free solution could not reverse effects of MeHg on either type of current. The effect of Hg2+ on N- but not R-type current was partially reversed by Hg2+-free wash solution. Therefore, different types of Ca2+ channels have differential susceptibility to neurotoxic mercurials even when expressed in the same cell type.
Voltage-sensitive Ca2+ channels play crucial roles in a number of cellular functions, including neurotransmitter release, gene expression, growth cone elongation, and dendritic action potential generation (Catterall, 1998, 2000). Various Ca2+ channelopathies resulting in neuronal or neuromuscular disorders are caused by mutations in genes coding for Ca2+ channel subunits (for review, see Meir and Dolphin, 2002). At least six distinct subtypes of Ca2+ channels (L, N, T, P, Q, and R) have already been identified based on their differential biophysical, molecular biological, and pharmacological properties (Tsien et al., 1995). Neuronal Ca2+ channels contain four subunits: α1, β, α2, and δ. The α1 subunit is the pore-forming, voltage-sensing, and ligand-binding component. cDNAs for at least seven distinct α subunits for high-voltage activated Ca2+ channels: α1A-1F, α1S, have been cloned. Four different β subunits and two different α2 subunits regulate assembly and modulate the kinetic parameters of the channel. The presence of several isoforms and splice variants further complicates the functional expression characteristics and classification of high-voltage activated Ca2+ channels (Brust et al., 1993; DeWaard and Campbell, 1995; McEnery et al., 1998; Pan and Lipscombe, 2000). Cells typically coexpress several types of Ca2+ channels, often with similar subcellular localization, providing a highly regulated degree of control over Ca2+-dependent cell functions, but confounding analyses of the properties of distinct Ca2+ channel subtypes in isolation. Because of their portal location within the plasma membrane, Ca2+ channels are potentially susceptible to the actions of a number of polyvalent heavy metal-type toxicants and serve as entry paths into the cell for heavy metals (Kiss and Osipenko, 1994; Atchison, 2003). Because of the crucial roles that Ca2+ channels play in key cellular functions, toxicant effects on Ca2+ channels could have significant deleterious consequences for neuronal function.
Methylmercury (MeHg) and inorganic mercury (Hg2+) are environmental neurotoxicants that differ chemically in ionic charge, ionic radii, and lipophilicity. Together, these factors can impact the manner in which these mercurials affect a given cellular function. Neurotoxic mercurials act on a number of cellular targets. In several neuronal systems, cellular effects of MeHg and Hg2+ are similar, yet distinct (Atchison et al., 1986; Hare and Atchison, 1992; Hewett and Atchison, 1992; Yuan and Atchison, 1994). The exact mechanisms by which these mercurials exert neurotoxicity are not known with certainty.
Disruption of function of voltage-sensitive Ca2+ channels is a prominent effect of acute exposure to low concentrations of both MeHg (Shafer and Atchison, 1991; Leonhardt et al., 1996; Sirois and Atchison, 1996, 2000; Shafer, 1998) and Hg2+ (Büsselberg et al., 1991; Weisenberg et al., 1995). MeHg blocks Ba2+ currents (IBa) carried through multiple subtypes of Ca2+ channels in primary cultures of cerebellar granule cells and in rat pheochromocytoma (PC12) cells (Shafer and Atchison, 1991; Sirois and Atchison, 2000). Hg2+ also alters function of several types of Ca2+ channels at low micromolar concentrations (Büsselberg et al., 1994; Leonhardt et al., 1996; Szucs et al., 1997). However, the actions of mercurials on Ca2+ channels may be more complex than mere block of function. In PC12 cells, very low concentrations of Hg2+ increase amplitude of current carried through voltage-sensitive Ca2+ channels (Rossi et al., 1993), whereas in cerebellar granule cells and NG108-15 cells, MeHg causes an increase in fura-2 fluorescence, which is dependent, at least in part, on extracellular Ca2+, and which is delayed by nifedipine, ω-conotoxin GVIA, and Ni2+ (Hare and Atchison, 1995; Marty and Atchison, 1997). Moreover, treatment of rodents with Ca2+ channel blockers prevents the toxic effects of MeHg (Sakamoto et al., 1996), and Ca2+ channel blockers delay the onset of cerebellar granule cell death with MeHg (Marty and Atchison, 1997; Gasso et al., 2001). Finally, in cells lacking Ca2+ channels, the onset of intracellular action of MeHg is delayed, suggesting that Ca2+ channels provide a path of entry for MeHg into the cell (Edwards et al., 2002). Therefore, mercurials seem to interact with voltage-sensitive Ca2+ channels in a complex manner.
Because of the numerous and important roles that voltage-sensitive Ca2+ channels play in neuronal function, disruption of function of voltage-sensitive Ca2+ channels may be a significant contributory factor in mercurial-induced neurotoxicity. There are few published reports comparing the effects of different mercurials on function of voltage-sensitive Ca2+ channels (Hewett and Atchison, 1992; Szucs et al., 1997; Schirrmacher et al., 1998), and no comparative study on the effects of these two forms of mercury on defined types of voltage-sensitive Ca2+ channels exists.
The goal of the present study was to determine whether specific types of voltage-sensitive Ca2+ channels were affected differentially by MeHg or Hg2+. We compared the effect of MeHg and Hg2+ on N-(Cav2.2) and R-(Cav2.3) types of voltage-sensitive Ca2+ channels expressed using cDNA copies of their genes transferred into human embryonic kidney cells (HEK293). These cells are nonexcitable and commonly used for heterologous expression of membrane proteins, including voltage-dependent Ca2+ channels (Williams et al., 1994; Perez-Garcia et al., 1995; Quefurth et al., 1998). Expression cDNA clones of α1B-1 or α1E-3 subunit were combined with α2-bδ and β3-a Ca2+ channel subunits of human neuronal origin for transient expression of N- and R-subtypes, respectively, of high-voltage activated Ca2+ channels. Jellyfish green fluorescent protein (GFP) was used as a cotransfection reporter.
Materials and Methods
Materials. HEK293 cells (ATCC CRL-1573) were purchased from American Type Culture Collection (Manassas, VA). All reagents were pure or ultrapure laboratory grade unless specifically noted otherwise. cAMP, EGTA, HEPES, ATP, ω-conotoxin GVIA, and tetrodotoxin were all obtained from Sigma-Aldrich (St. Louis, MO). Stock solutions (5 mM) of methylmercuric chloride (ICN Pharmaceuticals, Costa Mesa, CA) and (10 mM) HgCl2 were prepared weekly in double-distilled water, from which test solutions were prepared daily in extracellular solution (see below). Plasmids containing expression cDNA clones of human neuronal Ca2+ channel subunits were generously provided by Dr. Kenneth A. Stauderman of SIBIA Neurosciences (San Diego, CA), now Merck Research Laboratories. α1E-3 (Williams et al., 1994) and β3-a subunit clones (Mark Williams, Merck Research Laboratories, personal communication) were isolated from hippocampus; α1B-1 was isolated from the IMR32 cell line (Williams et al., 1992b), and α2Bδ was isolated from brainstem and basal ganglia (Williams et al., 1992a).
Cell Culture and Transfection. HEK293 cells were grown at 37°C in Eagle's minimal essential medium fortified with 1 mM sodium pyruvate, 0.1 mM nonessential amino acids, 2 mM l-glutamine, 1.5 g/l sodium bicarbonate, 10% (v/v) fetal bovine serum, and penicillin/streptomycin/amphotericin B (at a final concentration of 100 U/ml penicillin G, 100 μg/ml streptomycin sulfate, and 0.25 μg/ml amphotericin B; Invitrogen, Carlsbad, CA) in a 5% CO2 environment. One day before gene transfer, cells were plated at a density of 5 × 105 on 35-mm culture dishes. Cells were transfected with a mixture of plasmids containing human neuronal α1B-1, or α1E-3 plus α2-bδ and β3-a Ca2+ channel subunits and a jellyfish GFP cDNA sequence using FuGENE 6 (Roche Diagnostics, Indianapolis, IN) following the manufacturer's instruction. Reactions contained a total of 3 μl of FuGENE 6 and 1 μg of plasmid DNA containing the three subunits in 1:1:1 M ratio, with GFP plasmid at 20% of the total DNA. Two days were allowed for transient expression of proteins at which time the cells were examined for GFP expression. Cells from dishes with a reasonable number of green fluorescent cells (usually 20%), were replated at a low density of sufficiently isolated individual cells to facilitate recording.
Ca2+Channel Current Recording. Before recording, culture medium was removed, cells were rinsed twice with extracellular solution, and then replenished with 1 ml of extracellular recording solution. The extracellular solution contained 117 mM tetraethylammonium chloride, 20 mM BaCl2, 1 mM MgCl2, 25 mM d-glucose, 10 mM HEPES, and 0.001 mM tetrodotoxin; pH was adjusted to 7.2 at room temperature (∼23-25°C) with tetraethylammonium hydroxide. The osmolarity of the solution was 310 mOsM. Patch-clamp pipettes with resistances between 6 and 8 MΩ were prepared from 1.5-mm i.d. glass capillaries (WPI, Sarasota, FL) using a two-stage micro-electrode puller (PP-830; Narishige, Tokyo, Japan) and fire-polished using a Narishige MF-830 microforge. Intracellular (pipette) solution contained 140 mM CsCl, 10 mM EGTA, 10 mM HEPES, 2 mM ATP-Mg, and 1 mM cAMP; pH was adjusted to 7.2 with CsOH.
The tight seal whole cell configuration of the patch-clamp technique (Hamill et al., 1981) was used on fluorescent green cells to record IBa through transiently expressed Ca2+ channels. Whole cell currents were recorded using an Axopatch-1D amplifier (Axon Instruments, Inc., Foster City, CA), sampled at 10 kHz and filtered at 2 kHz (-3 dB, four-pole Bessel filter; Axon Instruments, Inc.) and acquired on-line using the pClamp6 program (Axon Instruments, Inc.). Pipette and cell capacitance was compensated online in all experiments. Series resistance was also compensated in the range of 60 to 80%. Extracellular media were exchanged using a gravity-fed bath perfusion system (BPS-4; ALA Scientific Instruments, Westbury, NY); the flow rate was 5 × 10-3 ml/s. The distance of the flow pipette from the cell was maintained at approximately 150 μm. All experiments were carried out at room temperature (23-25°C).
To elicit IBa, the membrane was depolarized by a positive voltage step from the holding potential (-90 mV). Conventional voltage clamp step protocols were used to examine the amplitude of IBa over a range of membrane potentials and to examine the voltage dependence of activation of IBa and of steady-state inactivation of IBa. In experiments designed to examine the effects of Ni2+, Cd2+, MeHg, Hg2+,or ω-conotoxin GVIA on IBa, the membrane was repeatedly (0.1 Hz) depolarized (from -90 mV) by a voltage step to 0 mV (R-type) or to 20 mV (N-type). Each depolarization was preceded by a hyperpolarizing voltage step (one-quarter the amplitude of the depolarizing step) to permit measurement of leakage current. After establishing that the current elicited by each depolarizing step was eventually constant, the test compound (modulator) was applied to the cell as mentioned above. Unless noted otherwise, in all experiments IBa amplitude was measured after subtraction of leak current from the whole current elicited by the activating depolarization.
Statistical Analysis. For each experiment recordings were made from a minimum of five cells, and these cells were derived from at least three independent transfections. Values are expressed as the mean ± S.E.M. The number of replicates is indicated for each experiment in the figure legend. Curve fitting of data points on graphs was performed using Origin software (OriginLab Corp., Northampton, MA).
Results
Characteristics of Recombinant Transiently Expressed α1Band α1ESubunit-Mediated Ca2+Channel Currents in HEK293 Cells. Basic biophysical and pharmacological qualities of Ca2+ channels containing α1B (Cav 2.2) or α1E (Cav 2.3) subunit in combination with α2-bδ and β3-a subunits and expressed transiently in HEK293 cells were similar to those of N-or R-type Ca2+ channels, respectively, as expressed in their native environment. A constant set of α2-bδ and β3-a subunits was used in all experiments so that we could focus on the comparative actions of the two mercurials on the principal phenotype-defining subunit of voltage-sensitive Ca2+ channels, the α1 subunit. Representative tracings of IBa elicited from these cells along with their current-voltage relationships and voltage dependence of steady-state inactivation and activation are shown in Fig. 1, A to C (for α1B) and Fig. 2, A to C (for α1E), respectively. For α1B-containing channels, current seemed to activate at approximately -20 mV, reached peak amplitude at about +20 mV and reversed at +60 mV. For channels containing α1E subunit, current activated at -40 mV, reached a peak amplitude at 0 mV and reversed at approximately +60 mV. The voltage for half-maximal activation (V1/2) was 0 mV in α1B-containing channels (Fig. 1C) and -14.8 mV in α1E-containing channels (Fig. 2C), respectively. The isochronal inactivation (h) was determined using an 8-s inactivating prepulse. The inactivation curves are shown (Figs. 1C and 2C) with slope parameter (k) of 12.0 mV and V1/2 of -64.6 mV for α1B-containing channels, and k of 9.2 mV and V1/2 of -69.6 mV for α1E-containing channels, respectively.
Figure 3 demonstrates that IBa expressed in cells transfected with α1B subunit was sensitive to inhibition by ω-conotoxin GVIA. At 1 or 10 μM, ω-conotoxin GVIA caused a rapid reduction of current amplitude; the rate with which block occurred was concentration-dependent. However, ultimately, both concentrations blocked virtually all current. The effect of ω-conotoxin block was not reversed by washing the cell with toxin-free solution. Thus, the IBa recorded from HEK293 cells expressing human neuronal α1B subunits combined with a fixed complement of α2δβ3 subunits has biophysical and pharmacological properties consistent with those mediated by N-type Ca2+ channels. We tested α1E subunit-expressing HEK293 cells with known specific antagonists (nimodipine and ω-agatoxin IVA) for the L- or P/Q-type channels and found the IBa to be resistant to block by both of these agents (data not shown), consistent with the characteristics of R-type current. Therefore, we used varying concentrations of Cd2+ and Ni2+ (both of which are known to block current carried through all known subtypes of voltage-sensitive Ca2+ channels) on the α1E subunit-containing cells (Fig. 4, A and B). The block of IBa from these cells by both divalent cations was concentration-dependent. A stepwise reduction of current was seen after addition of successively higher concentrations of Cd2+ or Ni2+; this effect seemed to plateau until the next higher concentration of metal was added. The reduction of amplitude of IBa was complete at 100 μM Cd2+ or 1 mM Ni2+. These two observations together indicate that the pharmacological characteristics of α1E subunit-mediated IBa from these cells is consistent with that of the R-type. An interesting observation at the lowest concentrations of Cd2+ (0.1 μM; Fig. 4A) and Ni2+ (1 μM; Fig. 4B) was an apparent slight increase rather than reduction in amplitude of IBa. This effect was more prominent with Cd2+.
Perhaps this effect is due to these cations actually entering the Ca2+ channels based on their relatively similar ionic radii and electrical charges. These same cations become inhibitory at higher concentrations.
Time- and Concentration Dependence of MeHg Effect on IBa. As shown in Figs. 5A and 6A for representative traces, and in composite data from a number of preparations in Figs. 5B and 6B, MeHg caused a rapid, concentration- and time-dependent reduction in current from both α1B and α1E subunit-containing channels. As shown in Fig. 5A for cells expressing α1B-type current, the effect of 5 μM MeHg was initially greater on the peak current than on the sustained current remaining at the end of a test pulse. However, with additional time, MeHg blocked both components of current apparently equally. For cells expressing α1E subunit, the degree of block of current by 5 μM MeHg seemed to be approximately the same for both peak and sustained current (Fig. 6A). The effect of different concentrations of MeHg on whole cell IBa with either subunit is shown as a time course in Figs. 5B and 6B and was determined after at least 5-min stabilization of the control current. Addition of MeHg concentrations between 0.125 (or 0.25 for α1E subunit) and 5.0 μM to the extracellular solution resulted in a concentration-dependent block in the rate and magnitude of IBa after leak subtraction. For both kinds of channels, the block of peak current by MeHg was again concentration- and time-dependent. MeHg concentrations in excess of 1 μM caused complete block of current within 2 to 6 min. Submicromolar concentrations also caused measurable inhibition (Figs. 5B and 6B). The rate at which inhibition occurred at micromolar concentrations seemed to consist of two distinct components: one which was initially more rapid, and the other, which was more gradual (Figs. 5B and 6B). Depending upon the concentration of MeHg used, the rapid component was completed between 100 and 150 s after beginning treatment. Lower concentrations of MeHg (0.125-1.25 μM) caused a more gradual decline in current amplitude. Inhibition of current at 1.25 μM MeHg seemed ultimately to reach the same plateau as at higher concentrations. At 0.125 μM in α1B- and 0.25 μM in α1E-containing channels, inhibition never reached the same level as at higher concentrations over the 10-min recording period. Because lower concentrations of MeHg caused a slow, progressive block of current that didn't reach a plateau, we could only estimate an apparent IC50 value for MeHg. This was done by comparing the percentage of block at each concentration at a set time point. At 200 s, MeHg blocked ∼50% of both types of current at 1.3 (N-type) and 1.1 μM (R-type).
Concentration Dependence of Hg2+Effect on IBa. Figures 7 and 8 show the effects of Hg2+ on IBa from HEK293 cells expressing α1B and α1E subunit-containing Ca2+ channels, respectively. Figures 7A and 8A depict representative current traces showing effects of different concentrations of Hg2+ on current after 2-min exposure. Hg2+ caused a rapid, concentration-dependent reduction in current from both α1B and α1E subunit-expressing channels. As shown in Fig. 7A for cells expressing α1B-type current, the effect of 0.5, 1, and 5.0 μM Hg2+ was initially greater on the peak current than on the sustained current remaining at the end of a test pulse.
However, with additional time, Hg2+ blocked both components of current. For cells expressing α1E subunit, the degree of block by 0.1, 0.5, and 1 μM Hg2+ seemed to be approximately the same for both peak and sustained current (Fig. 8A). Furthermore, the inactivation kinetics seems to be affected by Hg2+ in cells expressing α1B-type current. Figures 7B and 8B depict the degree of reduction of normalized IBa as a function of concentration of Hg2+. The IC50 values for Hg2+-induced block of α1B and α1E-type current were 2.2 and 0.7 μM, respectively.
Effect of Wash with MeHg- or Hg2+-Free Solution on IBa. As shown in Fig. 9, the block by MeHg of either type of current could not be reversed by washing with MeHg-free solution. However, the Hg2+ induced reduction of IBa through α1B-mediated N-type channels was partially reversible (Fig. 10A), but the IBa reduction from α1E-mediated R-type channels was not reversed by washing with Hg2+-free solution (Fig. 10B).
Discussion
Previous studies have shown that MeHg (Shafer and Atchison, 1991; Marty and Atchison, 1997; Shafer, 1998; Sirois and Atchison, 2000) and Hg2+ (Büsselberg et al., 1994; Leonhardt et al., 1996; Szucs et al., 1997) affect the function of native Ca2+ channels in multiple types of cells. However, in most of these studies, the cells examined express more than one subtype of Ca2+ channels, and attempts to compare the actions of these metals on known subtypes of Ca2+ channel were incomplete and/or indirect. Until recently (Peng et al., 2002), there have been no studies of effects of mercurials on distinct subtypes of Ca2+ channel in isolation. The present study using transient expression of human neuronal Ca2+ channels is the first to characterize and compare the effects of the organomercurial MeHg and inorganic Hg2+ on current mediated by single, identified phenotypes (N- and R-type) of Ca2+ channel in isolation. Because the two types of expressed Ca2+ channel contained the same α2δ and β subunits, differential effects of the mercurials on expressed N- and R-type channel current must be due largely to the mercurial interactions with, or action on, the α1 pore-forming subunit of the channel. Our results support and extend several aspects of previous studies of mercurials on native N-type and R-type Ca2+ channels. Our results demonstrate that, first, MeHg is an irreversible and essentially equipotent inhibitor of current recorded from HEK293 cells expressing human neuronal α1B (N-type) or α1E (R-type) subunit containing Ca2+ channels. Second, although Hg2+ inhibits current mediated by the same channels, it has a more potent effect on recombinant R-type channels than N-type channels. Third, the effect of Hg2+ is partially reversible for recombinant N-type channels but is evidently irreversible for R-type channels.
Our recordings show that the voltage- and time-dependent characteristics of the current mediated by channels containing the α1B subunit and those containing the α1E subunit expressed in transfected cells resemble those of native N- and R-type Ca2+ channels, respectively, in neurons (Randall and Tsien, 1995). The different biophysical characteristics of native N-type channel current and current mediated by recombinant channels containing the α1B subunit and of native R-type channel current and current mediated by recombinant channels containing the α1E subunit may simply reflect differences of the α2δ and β subunits expressed in recombinant and native channels. Alternatively, these differences may be due to differences of the human α1 subunits expressed in cells studied here and the α1 subunits of the native channels of nonhuman derived neurons such as those studied by Randall and Tsien (1995). Furthermore, we have demonstrated that recombinant channels containing the α1B subunit are sensitive to ω-conotoxin GVIA as are native N-type Ca2+ channels, whereas those channels expressing the α1E subunit were highly sensitive to block by Cd2+ and to a lesser extent Ni2+ as is the case for native R-type channels.
The data presented here show that recombinant expressed N-type channels were more sensitive to the effects of MeHg than were presumptive N-type native channels in PC12 cells (Shafer and Atchison, 1991; Shafer, 1998) and dorsal root ganglion cells (Leonhardt et al., 1996) but were equally sensitive compared with presumptive N-type channels of rat cerebellar granule cells (Sirois and Atchison, 2000). Such differential sensitivity of native and recombinant expressed channels has also been reported for other toxins, including calcicludine effects on L-type channels (Stotz et al., 2000), ω-agatoxin IVA effects on P/Q-type channels (Bourinet et al., 1999), and peptide spider toxin DW13.3 effects on N-type channels (Sutton et al., 1998). These differential effects of mercurials and other toxins on recombinant and native channels may reflect differences between human and nonhuman α1 subunits (present in studies of recombinant channels and native channels, respectively), or between the accessory (α2δ and β) subunits of various native channels and recombinant channels.
The present study of mercurials on recombinant Ca2+ channels also confirms previous observations concerning several aspects of the effect of MeHg on native channels. First, as has been observed for native channels in cerebellar granule cells (Sirois and Atchison, 2000) and PC12 cells (Shafer and Atchison, 1991), the effect of MeHg on recombinant α1B-(N-type) and α1E (R-type)-containing channels was not voltage-dependent. Second, the reduction of current during exposure to MeHg was irreversible regardless of whether the current was mediated by channels containing the α1B (N-type) or α1E (R-type) subunit or by various types of native Ca2+ channel. Third, it is also noteworthy that the effect of MeHg proceeds to complete reduction of the current mediated by channels containing the α1B (N-type) or α1E (R-type) subunit and by at least some native channels (given a sufficient duration of exposure to MeHg), although this effect was not observed for MeHg effects on recombinant α1C (L-type) subunit-containing channels (Peng et al., 2002). Fourth, the effect of Hg2+ on native channels, which has variously been reported to be reversible or irreversible (Leonhardt et al., 1996), was found in the present study to be partially reversible for channels containing the α1B (N-type) subunit, but evidently irreversible, for channels containing the α1E (R-type) subunit. Although mechanisms associated with reversibility were not examined, this observation may help to explain the variability of reversal seen in studies with Hg2+ of native channels, inasmuch as, depending upon the complement of channel phenotype in the preparation being studied, the actions of Hg2+ may be more or less reversible. In the case of the irreversible effect of MeHg or Hg2+, it is possible that the mercurial binds covalently to the channel (so that the mercurial cannot be removed) or interacts in some way with the channel that is irreversible (even after removal of the mercurial). Again, differences concerning the reversibility of a mercurial effect on native channels and on recombinant channels may reflect differences between the α1 subunit or the accessory units of the native and recombinant Ca2+ channels. Regardless of the reversibility of the interaction of Hg2+ or MeHg with the channel, the specific site of interaction between mercurial and channel is unclear. Hg2+ or MeHg could act directly in the channel pore, have nonspecific or specific interactions with negatively charged groups on the cell surface, or act at a regulatory or allosteric binding site on the intra- or extracellular surface of the channel. However, it should be noted that Shafer (1998) reported that intracellular application of MeHg did not affect Ca2+ channel current in PC12 cells. Another difference between the effects of Hg2+ on recombinant and native channels is that Hg2+ seems to increase Ca2+ current through native channels at low concentration (Manalis and Cooper, 1975; Rossi et al., 1993); this phenomenon was not observed in our study of recombinant channels.
In summary, both MeHg and Hg2+ perturb the function of heterologously expressed, recombinant human neuronal N-type and R-type Ca2+ channels at low micromolar concentrations, well within the range of concentrations known to cause disruption of function of corresponding native channels as well as toxicity from these agents in vivo. MeHg was an equipotent inhibitor of human neuronal N-type and R-type Ca2+ channels expressed in HEK293 cells. However, there seem to be subtle differences on the effects of Hg2+, which vary somewhat depending on the type of α1 subunit.
Acknowledgments
We acknowledge the generous contribution of the Ca2+ channel cDNA clones by SIBIA Neurosciences (now Merck Research Laboratories, San Diego, CA). Thanks are also due to Dr. Peter J. R. Cobbett (Department of Pharmacology and Toxicology, Michigan State University) for critical reading of the manuscript and valuable discussion. The excellent secretarial assistance of Mallory Koglin and Kelly O'Brien is especially appreciated.
Footnotes
-
↵1 R.K.H. and S.-Q.P. contributed equally to this study.
-
This study was supported by National Institutes of Health Grants R01ES03299 and R01ES05822 (to W.D.A.). A preliminary report of these findings was presented at the 2001 Annual Meeting of the Society of Toxicology in San Francisco, CA and published in Toxicologist60:185.
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
DOI: 10.1124/jpet.103.049429.
-
ABBREVIATIONS: MeHg, methylmercury; IBa, Ba2+ current; HEK, human embryonic kidney; GFP, green fluorescent protein; V1/2, voltage for half-maximal activation.
- Received January 22, 2003.
- Accepted May 23, 2003.
- The American Society for Pharmacology and Experimental Therapeutics