Abstract
Extracellular Ca2+ robustly potentiates the acetylcholine response of α4β2 nicotinic receptors. Rat orthologs of five mutations linked to autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE)—α4(S252F), α4(S256L), α4(+L264), β2(V262L), and β2(V262M)—reduced 2 mM Ca2+ potentiation of the α4β2 1 mM acetylcholine response by 55 to 74%. To determine whether altered allosteric Ca2+ activation or enhanced Ca2+ block caused this reduction, we coexpressed the rat ADNFLE mutations with an α4 N-terminal mutation, α4(E180Q), that abolished α4β2 allosteric Ca2+ activation. In each case, Ca2+ inhibition of the double mutants was less than that expected from a Ca2+ blocking mechanism. In fact, the effects of Ca2+ on the ADNFLE mutations near the intracellular end of the M2 region—α4(S252F) and α4(S256L)—were consistent with a straightforward allosteric mechanism. In contrast, the effects of Ca2+ on the ADNFLE mutations near the extracellular end of the M2 region—α4(+L264)β2, β2(V262L), and β2(V262M)—were consistent with a mixed mechanism involving both altered allosteric activation and enhanced block. However, the effects of 2 mM Ca2+ on the α4β2, α4(+L264)β2, and α4β2(V262L) single-channel conductances, the effects of membrane potential on the β2(V262L)-mediated reduction in Ca2+ potentiation, and the effects of eliminating the negative charges in the extracellular ring on this reduction failed to provide any direct evidence of mutant-enhanced Ca2+ block. Moreover, analyses of the α4β2, α4(S256L), and α4(+L264) Ca2+ concentration-potentiation relations suggested that the ADNFLE mutations reduce Ca2+ potentiation of the α4β2 acetylcholine response by altering allosteric activation rather than by enhancing block.
ADNFLE is a monogenic partial epilepsy linked to four α4 and two β2 nicotinic subunit mutations (Steinlein et al., 1995, 1997; Hirose et al., 1999; De Fusco et al., 2000; Phillips et al., 2001; Leniger et al., 2003). ADNFLE seizures occur primarily during slow-wave sleep and seem to originate in the frontal lobe (Scheffer et al., 1995). However, the physiological mechanism that generates these seizures has not been established.
At physiological concentrations (1.5-2 mM), Ca2+ potentiates the neuronal nicotinic agonist response by binding to an N-terminal site in the receptor protein (Galzi et al., 1996; Le Novere et al., 2002; Rodrigues-Pinguet et al., 2003). Allosteric Ca2+ activation of the neuronal nicotinic receptors is conserved across species (rat, human, chicken) and across nicotinic receptor subtypes (Mulle et al., 1992; Vernino et al., 1992; Galzi et al., 1996; Steinlein et al., 1997; Liu and Berg, 1999). Adding 2 to 2.5 mM Ca2+ to the extracellular saline increases the wild-type α4β2 acetylcholine response by 300 to 500% (Steinlein et al., 1997; Figl et al., 1998; Rodrigues-Pinguet et al., 2003). A common feature of the ADNFLE mutations is that they reduce 2 to 2.5 mM Ca2+ potentiation of the α4β2 acetylcholine response by 50 to 72%, at acetylcholine concentrations ≥30 μM (Steinlein et al., 1997; Figl et al., 1998; Rodrigues-Pinguet et al., 2003). Because the extracellular Ca2+ concentration in the mammalian brain is normally 1.5 to 2 mM (Egelman and Montague, 1999), a decrease in Ca2+ potentiation of the α4β2 acetylcholine response could contribute to ADNFLE seizures either by (1) reducing α4β2-mediated inhibitory transmitter release in the cortex (McNamara, 1999) or (2) shifting the balance between α4β2-mediated excitatory and inhibitory transmitter release during bouts of high-frequency cortical synaptic activity, in favor of excitatory transmitter release (Rodrigues-Pinguet et al., 2003).
At concentrations outside the physiological range (≥20 mM), Ca2+ also reduces the single-channel conductance of the α4β2 receptor (Buisson et al., 1996). The ability of Ca2+ to both block and potentiate α4β2 nicotinic receptors means that at least two distinct molecular mechanisms could account for the effects of the ADNFLE mutations on Ca2+ potentiation. The mutations could 1) interfere with allosteric Ca2+ activation or 2) enhance the potency of uncompetitive or noncompetitive Ca2+ block. There are several classes of possible Ca2+ blocking mechanisms. We refer to them simply as “noncompetitive.”
The α4 N-terminal mutation α4(E180Q) eliminates allosteric Ca2+ activation of the α4β2 receptor (Rodrigues-Pinguet et al., 2003). To decide whether the ADNFLE mutations reduce Ca2+ potentiation by interfering with allosteric Ca2+ activation or by enhancing Ca2+ block, we constructed double-mutant receptors containing this mutation and one of five rat orthologs—α4(S252F), α4(S256L), α4(+L264), β2(V262L), and β2(V262M)—of the human ADNFLE mutations—α4(S248F), α4(S252L), α4(776ins3), β2(V287L), and β2(V287M). We refer to these receptors as α4(E180Q):ADNFLE double mutants. The rat α4(+L264) mutation is a 3-base pair insertion that adds a leucine at position 264 to the α4 amino acid sequence.
If the Ca2+ blocking mechanism is correct, then the percentages by which Ca2+ blocks the acetylcholine response of α4(E180Q):ADNFLE double-mutant receptors and by which the corresponding single ADNFLE mutations reduce Ca2+ potentiation of the wild-type acetylcholine response should be the same. On the other hand, if the mutations reduce Ca2+ potentiation of the acetylcholine response by altering allosteric Ca2+ activation, then the α4(E180Q) mutation should nullify the effects of the ADNFLE mutations, and Ca2+ should not block the α4(E180Q):ADNFLE double-mutant acetylcholine response any more than it blocks α4(E180Q)β2 acetylcholine response. We also measured the effects of the α4(S256L) and α4(+L264) mutations on the Ca2+ concentration-potentiation relation for the acetylcholine response, the effects of 2 mM Ca2+ on the wild-type, α4(+L264)β2, and α4β2(V262L) single-channel conductance, and the voltage dependence of the β2(V262L)-mediated reduction in Ca2+ potentiation. Finally, we determined whether eliminating the fixed negative charge in the extracellular ring at the extracellular pore entrance could reverse the effects of the ADNFLE mutations on Ca2+ potentiation. The results show that the ADNFLE mutations reduce Ca2+ potentiation of the α4β2 acetylcholine response by disrupting allosteric Ca2+ activation of the receptor rather than by enhancing Ca2+ block.
Materials and Methods
Oocyte Expression. Stage V to VI Xenopus laevis oocytes were surgically isolated as described previously (Quick and Lester, 1994). Female X. laevis frogs were anesthetized by immersion in 0.2% tricaine methanesulfonate (Sigma, St. Louis, MO), pH 7.4, for 45 to 60 min, and the ovarian lobes were extracted through a small abdominal incision. The oocyte follicular layer was removed using Type A collagenase (1-2 h in a 2 mg/ml collagenase solution; Sigma). To increase nicotinic receptor expression, the rat α4-1 (Goldman et al., 1987) and β2 inserts (Deneris et al., 1988) were subcloned into a vector containing a 5′-untranslated region from the alfalfa mosaic virus that enhanced protein translation and a long 3′ poly A tail (Figl et al., 1998). We used the QuikChange single and multiple site-directed mutagenesis kit (Stratagene, La Jolla, CA) to construct the rat α4 and β2 mutations and verified them by DNA sequencing. Capped cRNA was synthesized in vitro using the mMessage mMachine RNA transcription kit (Ambion, Austin, TX). After a 24-h incubation in a modified Barth's solution containing 96 mM NaCl, 5 mM HEPES, 2.5 mM sodium pyruvate, 2 mM KCl, 1.8 mM CaCl2, 1 mM MgCl2, and 0.6 mM theophylline (Sigma) with 2.5 μg/ml Gentamicin (Sigma) and 5% horse serum, pH 7.4 (Irvine Scientific, Santa Ana, CA), the isolated oocytes were injected with rat α4 and β2 cRNA.
Whole-Oocyte Electrophysiology. Injected oocytes were incubated for ≥24 h in the modified Barth's solution at 15-19°C before electrophysiological recordings were attempted. We voltage-clamped the oocytes with two 3-MΩ, KCl-filled microelectodes (1.5- to 4-MΩ resistance) using a GeneClamp 500 voltage clamp (Axon Instruments, Union City, CA). During the voltage-clamp recordings, the oocytes were continually superfused with a nominally Ca2+-free saline (ND98) containing 98 mM NaCl, 1 mM MgCl2, and 5 mM HEPES, pH 7.4, at 20-23°C, unless otherwise stated. We added 0.1 to 2 mM CaCl2 to the ND98 solution to measure Ca2+-induced changes in the acetylcholine response. Acetylcholine was applied to the oocytes using a U-tube microperfusion system (Cohen et al., 1995). The time constant for solution exchange was ∼0.5 s. The voltage-clamp currents were digitized using a personal computer equipped with a DigiData 1322A analog-to-digital interface and pCLAMP version 8 software (Axon Instruments). To avoid aliasing, the voltage-clamp currents were filtered at one fourth to one fifth of the sampling frequency (typically 20 Hz) with an 8-pole, low-pass Bessel filter. We measured the current-voltage relation of the acetylcholine-induced current by ramping the membrane potential between -120 and +50 mV over a 0.5-s interval, in the presence and absence of 1 μM acetylcholine, and digitally subtracting the ramp current in the presence of acetylcholine from that in the absence of acetylcholine. We used Dunnett's test (SigmaStat ver. 1; SPSS Inc., Chicago, IL) to determine whether the ADNFLE mutations significantly affected Ca2+ potentiation of the peak acetylcholine response. The errors reported in the text are S.E.M. or S.E. (for fitted parameters). We used a previous approximation for the variance of the ratio of two random variables (Mood et al., 1974) to calculate the S.E.M. for the mutant-induced fractional reductions in Ca2+ potentiation of the acetylcholine response.
Desensitization Analysis. To compare desensitization of the wild-type and mutant responses, we fit their desensitizing phases to the sum of an exponential component and a constant term using the curvefitting routine in pCLAMP version 8.2. From these fits, we obtained an apparent time constant of desensitization (τD) and the percentage of steady-state desensitization of the response (SSD), defined as where Iexp and IC are the fitted amplitudes of the exponential component and constant, respectively. We analyzed the effects of the mutations and Ca2+ on these two parameters using a two-way analysis of variance with receptor type and Ca2+ concentration as factors. Post hoc comparisons were carried out on the ranked data using the Student-Newman-Keuls test.
BAPTA Injections. To prevent Ca2+ influx through the wild-type and mutant nicotinic receptors from activating endogenous Ca2+-activated Cl- currents, we injected the oocytes with 50 nl of a 100 mM K4BAPTA solution buffered to pH 7.4 with 10 mM HEPES, 5 to 10 min before recording from them (Haghighi and Cooper, 2000; Rodrigues-Pinguet et al., 2003). Assuming an oocyte volume of 500 nl, these injections produced a final intracellular BAPTA concentration of ∼10 mM.
Single-Channel Recordings. We recorded single wild-type and mutant channels in cell-attached patches at 20-23°C using the patchclamp option of the GeneClamp 500 voltage clamp. Patch electrodes were pulled from borosilicate capillary tubing (1.6-mm o.d., 0.80 mm i.d.; Garner Glass Company, Claremont, CA) using a modified Kopf electrode puller (Hamill et al., 1981). To patch onto the oocytes, we removed the vitelline membrane with forceps manually after incubating the oocytes in a hypertonic solution (ND98 plus 100 mM NaCl) for 20 to 30 min on a rotating shaker. The ionic composition of the pipette-filling and bathing solution used for the recordings was 98 mM KCl, 1 mM MgCl2, and 5 mM HEPES, pH 7.4. The patch electrode resistance was 10 to 30 MΩ. Because the extracellular K+ concentration was 98 mM, the oocyte resting potentials in the patch experiments were near 0 mV. We added 10 nM acetylcholine to the pipette-filling solution to activate the wild-type and mutant channels with minimal desensitization. To measure the effects of Ca2+ on the single-channel current, we added 2 mM CaCl2 to the pipette-filling and bathing solution. Single-channel currents were recorded digitally using the Digidata 1322A acquisition system and pCLAMP version 8 software. The single-channel data were sampled at 10 kHz and low-pass filtered at 2 kHz before digitization. We used pCLAMP version 9 to construct and fit all-points histograms of the single-channel current to the sum of two or more Gaussian components. The single-channel current amplitudes were obtained from the differences between the means of these Gaussian components. We used a two-way analysis of variance with conductance state and Ca2+ concentration as factors to determine 1) whether the amplitudes of the small and large single-channel currents were significantly different and 2) whether Ca2+ affected the single-channel currents. Post hoc comparisons were carried out using the Student-Newman-Keuls test.
Results
The α4(E180Q) Mutation Abolishes Ca2+ Potentiation of the α4β2 Acetylcholine Response. To ensure that the α4(E180Q) mutation eliminated allosteric Ca2+ activation of the α4β2 receptor, we measured its effects on 2 mM Ca2+ potentiation of the α4β2 1 mM acetylcholine response. We used 1 mM acetylcholine for these experiments instead of the 30 μM concentration used previously (Rodrigues-Pinguet et al., 2003) to maximize any potential open-channel Ca2+ block of the ADNFLE mutant receptors and to minimize any effects that mutant-induced shifts in acetylcholine potency might have on Ca2+ potentiation of the acetylcholine response. Adding 2 mM Ca2+ to a nominally Ca2+-free saline (ND98; see Materials and Methods) increased the peak 1 mM wild-type acetylcholine response by 310 ± 30% (mean ± S.E.M.) at -50 mV (Fig. 1B; Table 1). In contrast, adding 2 mM Ca2+ reduced the α4(E180Q)β2 1 mM acetylcholine response by 10 ± 1% (Fig. 1, C and E; Table 1). Thus, consistent with previous experiments using 30 μM acetylcholine (Rodrigues-Pinguet et al., 2003), the α4(E180Q) mutation completely abolished 2 mM Ca2+ potentiation of the α4β2 1 mM acetylcholine response (Fig. 1; Table 1). It is interesting that a β2 mutation—β2(E177Q)—homologous to the α4(E180Q) mutation did not reduce α4β2 Ca2+ potentiation (Fig. 1, D and E; Table 1).
The α4(S252F) and α4(S256L) Mutations Obviously Alter Allosteric Ca2+ Activation. To determine whether the ADNFLE mutations reduced Ca2+ potentiation by enhancing Ca2+ block of the receptor or by altering allosteric Ca2+ activation, we compared the effects of 2 mM Ca2+ on the ADNFLE single- and α4(E180Q):ADNFLE double-mutant acetylcholine responses. Ca2+ affected the double mutants with ADNFLE mutations closer to the intracellular end of the M2 region—α4(E180Q:S252F)β2 and α4(E180Q: S256L)β2—differently from those with ADNFLE mutations near the extracellular end of M2—α4(E180Q:+L264)β2, α4(E180Q)β2(V262M), and α4(E180Q)β2(V262L). We shall discuss the results for the ADNFLE mutations closer to the intracellular end first. Consistent with their previously reported effects on Ca2+ potentiation of the 30 μM acetylcholine response (Rodrigues-Pinguet et al., 2003), the single α4(S252F) and α4(S256L) mutations significantly reduced 2 mM Ca2+ potentiation of the α4β2 1 mM acetylcholine response (Fig. 2, A and B; Table 1). Adding 2 mM Ca2+ increased the α4(S256L)β2 and α4(S252F)β2 1 mM acetylcholine responses by only ∼40% (Fig. 2, A, B, and E) as opposed to a 310% increase for the wild-type (Fig. 1, B and E). Thus, the Ca2+-induced mutant increases were ∼60% less than that of the wild-type. If enhanced Ca2+ block accounted entirely for the effects of the α4(S252F) and α4(S256L) mutations on Ca2+ potentiation, then 2 mM Ca2+ should have also reduced the α4(E180Q:S252F)β2 and α4(E180Q:S256L)β2 1 mM acetylcholine responses by ∼60%. However, 2 mM Ca2+ had little effect on the α4(E180Q:S252F)β2 and α4(E180Q: S256L)β2 responses (Fig. 2, C, D, and F; Table 1). In fact, it reduced the α4(E180Q:S252F)β2 1 mM acetylcholine response by only 10 ± 1%, and it increased the α4(E180Q: S256L)β2 1 mM acetylcholine response by 10 ± 4%. Therefore, the α4(S252F) and α4(S256L) mutations seem to reduce α4β2 Ca2+ potentiation by diminishing allosteric Ca2+ activation of the α4β2 receptor rather than by enhancing Ca2+ block.
The α4(S256L) and α4(S252F) mutations did not significantly affect the amplitude of the 1 mM acetylcholine response in the absence of Ca2+ (P > 0.05). In 0 mM Ca2+, the peak amplitudes of the wild-type, α4(S256L)β2, and α4(S252F)β2 1 mM acetylcholine responses were 1400 ± 550 (n = 10), 830 ± 280 (n = 4), and 1000 ± 460 nA (n = 4), respectively.
The Effects of the α4(+L264), β2(V262M), and β2(V262L) Mutations Are More Complex. The single α4(+L264), β2(V262M), and β2(V262L) ADNFLE mutations reduced Ca2+ potentiation of the acetylcholine response by an amount similar to that of the α4(S252F) and α4(S256L) mutations (Fig. 3, A, C, E, and G; Table 1) but Ca2+ inhibited the α4(E180Q:+L264)β2, α4(E180Q)β2(V262M), and α4(E180Q)β2(V262L) double-mutant receptors more than the α4(E180Q)β2 receptor. Adding 2 mM Ca2+ to the extracellular solution increased the α4(+L264)β2 1 mM acetylcholine response by only 10 ± 2%, it increased the α4β2(V262M) response by only 40 ± 4%, and it actually reduced the α4β2(V262L) response by 20 ± 4% (Fig. 3G and Table 1). Thus, similar to α4(S252F) and α4(S256L) mutations, the α4(+L264), β2(V262M), and β2(V262L) mutations reduced 2 mM Ca2+ potentiation of the acetylcholine response by 55 to 75% compared with the wild-type. However, in contrast to the α4(S252F) and α4(S256L) mutations, adding 2 mM Ca2+ reduced the α4(E180Q:+L264)β2, α4(E180Q)β2(V262M), and α4(E180Q)β2(V262L) double-mutant 1 mM acetylcholine responses by 30 to 50% (Fig. 3, B, D, F, and H; Table 1). Thus, Ca2+ inhibited the α4(E180Q:+L264)β2, α4(E180Q)-β2(V262M), and α4(E180Q)β2(V262L) responses significantly (P < 0.05) more than the α4(E180Q)β2 response (Fig. 3H; Table 1). Nevertheless, Ca2+ inhibition of the double-mutant responses was less than that predicted from the corresponding single-mutant reductions in Ca2+ potentiation by a pure blocking mechanism (Fig. 4).
The α4(+L264) and β2(V262L) mutations did not significantly affect the peak amplitude of 1 mM acetylcholine response in 0 mM Ca2+ (P > 0.05). However, the β2(V262M) mutation significantly (P < 0.05) increased it. The peak amplitudes of the α4β2(V262M), α4(+L264)β2, and α4β2(V262L) 1 mM acetylcholine responses in 0 mM Ca2+ were 4000 ± 2000 (n = 5), 3000 ± 2000 (n = 4), and 500 ± 240 nA (n = 4), respectively. Thus, the α4β2, α4(+L264), β2(V262L), and (V262M) mutations did not reduce the amplitude of the 1 mM acetylcholine response in 0 mM Ca2+.
Altered Allosteric Ca2+ Activation Explains the Effects of the ADNFLE Mutations on the Ca2+ Concentration-Potentiation Relation Better than Enhanced Ca2+ Block. There are two possible interpretations of the effects of Ca2+ on the α4(E180Q:+L264)β2, α4(E180Q)β2(V262M), and α4(E180Q)β2(V262L) acetylcholine responses. First, the α4(+L264), β2(V262M), and β2(V262L) mutations could reduce α4β2 Ca2+ potentiation by both enhancing Ca2+ block and inhibiting allosteric Ca2+ activation of the receptor. Second, the α4(E180Q) mutation could interact with these particular ADNFLE mutations to convert the allosteric Ca2+ binding site from a positive to a negative site. To decide between these two alternatives, we compared the effects of two of the ADNFLE mutations on the Ca2+ concentration-potentiation relation and asked whether they were better explained by a mutant-induced change in allosteric Ca2+ activation of the receptor or enhanced Ca2+ block. We chose the α4(S256L) mutation as representative of the ADNFLE mutations closer to the intracellular end of M2 and, the α4(+L264) mutation as representative of those closer to the extracellular end. We used 30 μM acetylcholine for these experiments rather than 1 mM to avoid cumulative receptor desensitization by repeated acetylcholine applications.
Consistent with a small or negligible Ca2+ block of the wild-type receptor between 0.1 and 2 mM Ca2+, a simple hyperbolic binding function (eq. 1) fit the wild-type data well (R2 = 0.98; Fig. 5, A and B): where the ICa/I0 is the ratio of the peak 30 μM acetylcholine response with added extracellular Ca2+ to that in 0 mM added Ca2+, the [Ca2+]o is the added extracellular Ca2+ concentration, the EC50 is the half-maximal [Ca2+]o for potentiation of the acetylcholine response, and the Pmax is the maximum value of the ICa/I0. The fitted wild-type Pmax and EC50 were 3.6 ± 0.1 and 0.4 ± 0.1 mM (df = 6), respectively. To determine whether enhanced Ca2+ block could account for the effects of the ADNFLE mutations on the Ca2+ concentration-potentiation relation, we multiplied the wild-type hyperbolic binding function (eq. 1) by a simple inhibitory binding isotherm and fit this function (eq. 2) to the α4(S256L)β2 and α4(+L264)β2Ca2+ concentration-potentiation relations (Fig. 5, A and B): where the IC50 is the half-maximal inhibitory concentration for Ca2+ block of the mutant receptor and the expression, describes allosteric Ca2+ activation of the receptor (assuming that it is unaffected by the mutations). This function fit the α4(S256L)β2 and α4(+L264)β2 Ca2+ concentration-potentiation data poorly (R2 = 0; Fig. 5, A and B) because it predicted a relief from Ca2+ block; thus, a relief from the mutant-mediated reductions in Ca2+ potentiation as the [Ca2+]o approached zero. The fitted IC50 values for Ca2+ block of the α4(S256L)β2 and α4(+L264)β2 receptors were 0.8 ± 0.3 mM (df = 4) and 1.27 ± 0.03 mM (df = 4), respectively. Thus, enhanced Ca2+ block cannot account for the effects of the α4(S256L) and α4(+L264) mutations on the α4β2 Ca2+ concentration-potentiation relation.
A model (eq. 3, below) assuming that the mutations altered allosteric Ca2+ activation of receptor fit the data better (Fig. 5, C and D), where the KCa is the equilibrium constant for Ca2+ binding to the open receptor state (see the online Supplement for the derivation of eq. 3). Because the parameters KCa and Pmax were highly covariant, it was appropriate to fix the Pmax beforehand and to let only the KCa vary during the fit. To fit eq. 3 to the wild-type data, we fixed the Pmax at 3.6 (estimated from fitting eq. 1 to the wild-type data). With this constraint, eq. 3 fit the wild-type data as well as eq. 1 (R2 = 0.98; Fig. 5, C and D) and yielded a KCa of 108 ± 3 μM (df = 7). Visual inspection of the α4(S256L)β2 and α4(+L264)β2 data indicated that both mutations reduced the Pmax for Ca2+ potentation (Fig. 5, C and D). Fixing the Pmax at 1.6 gave reasonable fits to the α4(S256L)β2 and α4(+L264)β2 data (R2 = 0.75, 0.56, respectively; Fig. 5, C and D) and yielded KCa values of 1.2 ± 0.4 mM and 200 ± 100 μM (df = 4), respectively. Thus, a mutant-induced change in allosteric Ca2+ activation accounts for the effects of ADNFLE mutations on the Ca2+ concentration-potentiation relation better than enhanced Ca2+ block. Therefore, an interaction between the ADNFLE mutations closer to the extracellular end of M2 and the α4(E180Q) mutation that creates a negative Ca2+ allosteric site may account for their more complex effects.
The ADNFLE Mutations Enhance Steady-State De-sensitization in 2 mM Ca2+. In the absence of any added extracellular Ca2+, previous results show that the five ADNFLE mutations we studied do not consistently affect acetylcholine-induced desensitization (Figl et al., 1998; Rodrigues-Pinguet et al., 2003). However, we only tested the effects of three of the mutations—α4(S256L), β2(V262L), and β2(V262L)—in 2 mM Ca2+ previously. In 2 mM Ca2+, these three mutations increase steady-state desensitization of the 30 μM acetylcholine response (Rodrigues-Pinguet et al., 2003). The effects of the α4(S252F) and α4(+L264) mutations on desensitization in 2 mM Ca2+ have not previously been reported. To determine whether all five mutations affect desensitization in 2 mM Ca2+, we analyzed desensitization of the 1 mM acetylcholine responses obtained in the Ca2+ potentiation experiments above. The desensitizing phase of these responses primarily reflects the fast component of de-sensitization because we only applied acetylcholine for 5 to 16 s to minimize cumulative receptor desensitization by repeated agonist applications. The time course of desensitization of these responses was adequately fit by the sum of a single exponential component and a constant term (Fig. 6, A and B). As in previous studies (Figl et al., 1998; Rodrigues-Pinguet et al., 2003), the ADNFLE mutations did not consistently affect the SSD or τD of desensitization (see Materials and Methods) in 0 mM Ca2+. However, all five mutations significantly (P < 0.05) increased the SSD in 2 mM Ca2+ (Fig. 6C; Table 2). In 2 mM Ca2+, the SSD of the mutant receptors was 38 to 77% larger than that of the wild-type. Thus, the ADNFLE mutations increased steady-state desensitization in 2 mM Ca2+. The mutations did not consistently affect the τD in 0 or 2 mM Ca2+ (Fig. 6D; Table 2).
Consistent with previous results (Rodrigues-Pinguet et al., 2003), Ca2+ did not affect the SSD of the wild-type 1 mM acetylcholine response. However, it significantly (P < 0.05) reduced the τD by 51% (Fig. 6, C and D; Table 2). Thus, Ca2+ increased the speed, but not the depth, of fast wild-type desensitization. It is interesting, the α4(E180Q) mutation eliminated this effect (Table 2).
Combining the α4(E180Q) and ADNFLE mutations only significantly affected desensitization of the α4β2(V262M) and α4(+L264)β2 responses (Table 2). Combination with the α4(E180Q) mutation significantly (P < 0.05) reduced the SSD of the α4β2(V262M) response in 0 and 2 mM Ca2+ and brought the mutant SSDs closer to the wild-type values (Table 2). It also significantly (P < 0.05) increased the τD in 2 mM Ca2+ but it did not eliminate the effects of Ca2+ on the mutant τD (Table 2). Thus, the α4(E180Q) mutation reduced steady-state desensitization of the α4β2(V262M) response in both 0 and 2 mM Ca2+ and, reduced the effects of Ca2+ on its time course of desensitization. Combining the α4(E180Q) with the α4(+L264) mutation significantly (P < 0.05) reduced the α4(+L264)β2 SSD in 0 mM Ca2+ but, did not significantly affect the SSD in 2 mM Ca2+ or, the τD in 0 or 2 mM Ca2+ (Table 2). Thus, the α4(E180Q) mutation did not consistently affect desensitization of the ADNFLE mutant responses.
The β2(V262L)-Induced Reduction in Ca2+ Potentiation Is Voltage-Independent. Another way to test the hypothesis that ADNFLE mutations near the extracellular end of M2 enhance Ca2+ block of the α4β2 channel is to examine the voltage dependence of their effects on Ca2+ potentiation. If the mutations enhance Ca2+ block at a site within the membrane electric field, then their effects on Ca2+ potentiation should be voltage-dependent. To measure the effects of the β2(V262L) mutation on Ca2+ potentiation at positive (+30 mV) and negative (-50 mV) membrane potentials, we used a second mutation—α4(E245Q)—expected to relieve inward rectification of the α4β2 acetylcholine response. Severe inward rectification of the wild-type acetylcholine response precludes accurate measurements of Ca2+ potentiation at positive membrane potentials. To overcome this limitation, we used the α4(E245Q) mutation. Similar to the previously reported α4(E245A) mutation (Haghighi and Cooper, 2000), the α4(E245Q) mutation relieved inward rectification of the α4β2 acetylcholine-induced current-voltage relation (n = 5; Fig. 7A), allowing us to easily measure responses at positive potentials. It is interesting that the α4(E245Q) mutation also significantly reduced 2 mM Ca2+ potentiation of the α4β2 1 mM acetylcholine response at -50 mV (P < 0.05; Fig. 7B), but its effects were less severe than the ADNFLE mutations (Table 1). Ca2+ increased the α4(E245Q)β2 1 mM acetylcholine response by 90 ± 30% (n = 5) at -50 mV (Fig. 7B) and by 80 ± 20% (n = 10) at +30 mV (Fig. 7, D and F). Thus, consistent with previous results for native neuronal nicotinic receptors (Amador and Dani, 1995), Ca2+ potentiation of the α4(E245Q)β2 acetylcholine response was voltage-independent. Coexpression of the α4(E245Q) and β2(V262L) mutations yielded a double-mutant receptor that responded robustly to 1 mM acetylcholine at both negative and positive membrane potentials (Fig. 7, C, E, and F). If the β2(V262L) mutation reduces Ca2+ potentiation by enhancing Ca2+ block at a site within the membrane electric field, then positive membrane potentials should relieve the block and thus relieve the β2(V262L)-mediated reduction in α4(E245Q)β2 Ca2+ potentiation. However, the Ca2+-induced change in the α4(E245Q)β2(V262L) acetylcholine response at -50 mV (-50 ± 10%, n = 5) was not significantly (P > 0.05) different from that at +30 mV (-30 ± 10%, n = 13) (Fig. 7, C, E, and F), consistent with previous results showing that the α4(S252F) and α4(+L264) ADNFLE mutations similarly affect Ca2+ potentiation at membrane potentials of -100 and -50 mV (Figl et al., 1998). Thus, enhanced Ca2+ block of the α4(E245Q)β2(V262L) receptor at a site inside the membrane electric field cannot account for the reduced Ca2+ potentiation of this receptor.
Eliminating Negative Charges in the Extracellular Ring Does Not Prevent the β2(V262L) Mutation from Reducing Ca2+ Potentiation. A ring of negatively charged residues located at the extracellular entrance to the channel pore (the extracellular ring) mediates external block of the muscle nicotinic receptor by the divalent cation Mg2+ (Imoto et al., 1988). The aligning residues in the rat α4β2 nicotinic receptor are a negatively charged Glu in the α4 subunit— α4(E266)—(Goldman et al., 1987) and a positively charged Lys in the β2 subunit—β2(K260)—(Deneris et al., 1988). Because α4(E266) lies near the extracellular entrance to the channel pore, enhanced Ca2+ binding to this negatively charged residue could potentially mediate a voltage-independent Ca2+ block of the ADNFLE mutant receptors. To test this hypothesis, we mutated the Glu at position 266 in the α4 subunit to an uncharged Gln [α4(E266Q)], coexpressed this mutation with either the wild-type β2 or β2(V262L) subunit, and compared the effects of 2 mM Ca2+ on the α4(E266Q)β2 and α4(E266Q)β2(V262L) 1 mM acetylcholine responses. Similar to its effects on the wild-type receptor, 2 mM Ca2+ increased the α4(E266Q)β2 1 mM acetylcholine response by 300 ± 10% (n = 4) (Fig. 8, A and C; Table 1). Hence, not all the α4β2 mutations in and near M2 significantly reduce Ca2+ potentiation. If enhanced Ca2+ block at the extracellular ring mediates the effects of the ADNFLE mutations on Ca2+ potentiation, then we would expect the α4(E266Q) mutation to relieve the β2(V262L)-mediated reduction in Ca2+ potentiation. However, coexpression of the β2(V262L) and α4(E266Q) mutations did not relieve the β2(V262L)-mediated reduction in Ca2+ potentiation (Fig. 8, B and C; Table 1). Similar to the α4β2(V262L) single-mutant receptor, 2 mM Ca2+ reduced the α4(E266Q)β2(V262L) double-mutant 1 mM acetylcholine response by 40 ± 2% (n = 4) (Table 1). Thus, the ADNFLE mutations do not seem to reduce Ca2+ potentiation by enhancing Ca2+ block at the α4β2 extracellular ring.
Ca2+ Does Not Reduce the Mutant Single-Channel Currents More Than the Wild-Type Currents. The ADNFLE mutations reduce 2 mM Ca2+ potentiation of the 1 mM acetylcholine response by 55 to 74% (Table 1). If uncompetitive or noncompetitive Ca2+ inhibition of the mutant receptors mediates this effect, then the expected Ki for Ca2+ binding to its inhibitory site on the mutant receptors is ≥700 μM. If diffusion limits Ca2+ binding to this inhibitory site (i.e., a forward rate constant for Ca2+ binding of 10-8 ms-1), then Ca2+ remains bound to this site for ≤14 μs (Hille, 2001). This brief residency time implies that if the ADNFLE mutations reduce Ca2+ potentiation by enhancing Ca2+ block of the receptor, then Ca2+ is expected to reduce the apparent single-channel conductance (gs) of the mutant receptors by 55 to 74%. To test this hypothesis, we measured the amplitudes of wild-type and mutant single-channel currents (is) in cell-attached patches at a pipette potential of +100 mV with and without 2 mM added Ca2+ in the pipette-filling solution (Fig. 9, A-J; Table 3). Consistent with previous reports (Charnet et al., 1992; Kuryatov et al., 1997; Figl et al., 1998), the rat α4β2 wild-type channels displayed multiple conductance states in 0 mM added Ca2+ (Fig. 9, A-D; Table 3). At a driving potential of -100 mV (pipette potential of + 100 mV), the is of the smaller conductance state (2.6 ± 0.1 pA, n = 5 patches) was significantly (P < 0.01) less than that of the larger state (4.4 ± 0.1 pA, n = 5) (Table 3). The internal and external K+ concentrations for the oocytes in these experiments were nearly equal (assuming an internal K+ concentration of ∼100 mM). Thus, the oocyte resting potential and α4β2 reversal potential was near 0 mV, and the calculated gs values for the two wild-type conductance states were 26 ± 1 and 44 ± 1 pS. These conductances closely matched those previously reported (34 ± 2 and 49 ± 1 pS) for rat α4β2 channels in outside-out patches in symmetrical 100 mM K+ (Figl et al., 1998). In contrast, the α4(+L264)β2 channels displayed a single conductance state with an is of 1.8 ± 0.1 pA (n = 6) in 0 mM Ca2+ (Fig. 9, I and J; Table 3). The corresponding gs (18 ± 1 pS) was similar to the value previously reported (14.7 ± 1 pS) for the small-conductance rat α4(+L264)β2 channels in outside-out patches with symmetrical 100 mM K+ (Figl et al., 1998). Similar to the wild-type channels, the α4β2(V262L) channels also displayed two conductance states in 0 mM added Ca2+. The is values associated with these two states at -100 mV (2.2 ± 0.1 pA, n = 4 and 3.3 ± 2 pA, n = 5) were significantly different (P < 0.01; Fig. 9, E-H; Table 3).
Previous results show that 10 to 20 mM added extracellular Ca2+ reduces the conductance of neuronal nicotinic channels (Vernino et al., 1992; Amador and Dani, 1995; Buisson et al., 1996). Adding 2 mM Ca2+ to the pipette-filling solution reduced the is values for the two wild-type states by a small, but significant (P < 0.01), amount (Fig. 9, A-D; Table 3). It reduced the small is by 19 ± 5% and the large wild-type is by 9 ± 2%. The gs for the large wild-type conductance state in 2 mM Ca2+ (40 ± 4 pS, n = 6) (Table 3) was similar to that reported previously (∼46 pS) for the predominant conductance state of human α4β2 nicotinic receptors expressed in human embryonic kidney cells and measured in outside-out patches using 120 mM external Na+, 2 mM external Ca2+, and 120 mM internal K+ (Buisson et al., 1996). The addition of 2 mM Ca2+ also slightly reduced the is values for the mutant conductance states (Fig. 9, E-J; Table 3), but the reductions were not significant (P > 0.05). Thus, the α4(+L264) and β2(V262L) mutations do not enhance Ca2+-induced reductions in the α4β2 single-channel current.
Discussion
Less effective Ca2+ potentiation of the acetylcholine response is a common feature of all the ADNFLE mutations tested so far. Consistent with previous results for the 30 μM acetylcholine response (Steinlein et al., 1997; Figl et al., 1998; Rodrigues-Pinguet et al., 2003), rat orthologs of five of the six reported ADNFLE mutations reduce 2 mM Ca2+ potentiation of the 1 mM acetylcholine response by 55 to 74%. The effect of the sixth ADNFLE mutation—α4(T265I)—on Ca2+ potentiation has not been reported (Leniger et al., 2003). Contrary to previous suggestions (Rodrigues-Pinguet et al., 2003), a change in allosteric Ca2+ activation, rather than enhanced Ca2+ block, seems to be responsible for the ADNFLE mutant-induced reductions in Ca2+ potentiation. Several lines of evidence support this conclusion. Ca2+ at a concentration of 2 mM does not inhibit the α4(E180Q:S252F)β2 and α4(E180Q:S256L)β2 responses any more than it inhibits the α4(E180Q)β2 response. The effects of the α4(S256L) and α4(+L264) mutations on the Ca2+ concentration-potentiation relation are more consistent with altered allosteric Ca2+ activation than enhanced Ca2+ block. The effects of the β2(V262L) mutation on α4(E245Q)β2 Ca2+ potentiation are voltage-independent. The α4(E266Q) mutation also fails to relieve the β2(V262L)-mediated reduction in Ca2+ potentiation. Finally, 2 mM Ca2+ does not significantly reduce the α4(+L264)β2 and α4β2(V262L) single-channel conductances.
It is interesting that the five ADNFLE mutations we tested also increased the apparent steady-state desensitization of the 1 mM acetylcholine response in 2 mM Ca2+. Because our acetylcholine applications lasted only 5 to 16 s, this increase could reflect either a real increase in steady-state desensitization or a decrease in the relative amplitude of the slow component of desensitization. Regardless of its origin, this effect does not seem to be a common feature of the mutations because the sixth reported ADNFLE mutation—α4(T265I)— does not seem to affect desensitization of the acetylcholine response in Ca2+ (Leniger et al., 2003).
The effects of Ca2+ on the α4(E180Q:+L264)β2, α4(E180Q)β2(V262M), and α4(E180Q)β2(V262L) double-mutant responses allow two possible interpretations. The α4(+L264), β2(V262M), and β2(V262L) mutations could reduce Ca2+ potentiation of the acetylcholine response by a mixed mechanism involving both altered allosteric Ca2+ activation and enhanced Ca2+ block. On the other hand, these mutations could interact with the α4(E180Q) mutation to change the allosteric Ca2+ site from a positive to a negative one. The effects of the α4(+L264) mutation on the Ca2+ concentration-potentiation relation, and the failure of 2 mM Ca2+ to reduce the α4(+L264)β2 and α4β2(V262L) single-channel conductance more than the wild-type, support the latter interpretation. The voltage independence of the β2(V262L)-mediated reduction in α4(E245Q)β2 Ca2+ potentiation also shows that the β2(V262L) mutation does not enhance Ca2+ block at a site within the membrane electric field, and the failure of the α4(E266Q) mutation to relieve the β2(V262L)-mediated reduction in Ca2+ potentiation shows that the β2(V262L) mutation does not enhance Ca2+ block at negatively charged residues in the extracellular ring.
Consistent with previous results (Galzi et al., 1996; Rodrigues-Pinguet et al., 2003), the α4(E180Q) mutation abolishes Ca2+ potentiation of the α4β2 1 mM acetylcholine response. The homologous β2 mutation—β2(E177Q)— does not. The results for both mutations agree with a previous model of the α4β2 allosteric Ca2+ binding site (Le Novere et al., 2002) in which the Glu at α4(E180), but not that at β2(E177), directly contributes to Ca2+ binding. However, if combining the α4(E180Q) mutation with the ADNFLE mutations near the extracellular end of M2 produces a negative allosteric Ca2+ binding site, then the α4(E180Q) mutation cannot prevent Ca2+ binding to the α4β2 receptor.
Similar to α7 nicotinic receptors (Galzi et al., 1996), 2 mM Ca2+ increases the maximum α4β2 acetylcholine response ∼3-fold. If we assume that 1) Ca2+ does not amplify the neuronal nicotinic acetylcholine response by removing receptor inactivation (Amador and Dani, 1995) and 2) the ADNFLE mutations do not enhance Ca2+ block of the receptor, then two mechanisms could account for the effects of the ADNFLE mutations on the Ca2+-induced increase in acetylcholine efficacy. First, the ADNFLE mutations could increase acetylcholine efficacy in the absence of Ca2+ (decrease K2 in Scheme 1; see the online Supplement), thereby reducing the number of available receptors that could be opened by adding extracellular Ca2+. Second, the mutations could reduce the Ca2+ affinities of the closed, agonist-bound (K 3-1 in Scheme 1; see the online Supplement) and/or open α4β2 receptors (mK3-1 in Scheme 1; see the online Supplement), thereby reducing the ability of Ca2+ to increase acetylcholine efficacy. Because the rat α4(S252F) and α4(+L264) ADNFLE mutations do not affect surface antibody binding or increase the maximum (500 μM) acetylcholine response in 0 mM added Ca2+ (Figl et al., 1998), these mutations at least do not seem to reduce Ca2+ potentiation by increasing acetylcholine efficacy in the absence of Ca2+. Comparable data are not available for the α4(S256L), β2(V262M), and β2(V262L) mutations. However, our results show that adding 2 mM Ca2+ reduces the α4β2(V262L) 1 mM acetylcholine response by 20 ± 4%. At most, increasing the acetylcholine efficacy in the absence of Ca2+ could eliminate Ca2+ potentiation. It could not produce a Ca2+-induced decrease in the maximum acetylcholine response, as observed for the β2(V262L) mutation. Thus, this mutation also does not seem to reduce Ca2+ potentiation by increasing acetylcholine efficacy in 0 mM added Ca2+. However, we cannot rule out this mechanism for the α4(S256L) and β2(V262M) mutations.
Previous results show that three rat ADNFLE mutations— α4(S256L), β2(V262M), and β2(V262L)—reduce α4β2 Ca2+ potentiation in an acetylcholine concentration-dependent manner (Rodrigues-Pinguet et al., 2003). These mutations significantly reduce Ca2+ potentiation of the 30 μM and 1 mM acetylcholine responses but not that at lower acetylcholine concentrations (10 or 50 nM). In contrast, the α4(S252F) and α4(+L264) mutations reduce Ca2+ potentiation of the 10 and 50 nM and 30 μM acetylcholine responses equally well. Although uncompetitive Ca2+ inhibition, at a site independent of the one mediating allosteric Ca2+ activation, could explain the acetylcholine concentration dependence of the α4(S256L), β2(V262M), and β2(V262L) effects on Ca2+ potentiation, the evidence above argues against this explanation. An alternative explanation is receptor heterogeneity. There are least two distinct α4β2 subtypes with different pharmacological properties and subunit stoichiometries (Zwart and Vijverberg, 1998; Buisson and Bertrand, 2001; Nelson et al., 2003; Zhou et al., 2003). The α4(S256L), β2(V262M), and β2(V262L) mutations may selectively reduce Ca2+ potentiation of the low-affinity, but not the high-affinity, α4β2 subtypes. A third possible explanation, at least for the effects of the α4(S256L) and β2(V262M) mutations, is that these mutations reduce Ca2+ potentiation by increasing acetylcholine efficacy in the absence of Ca2+. Such an increase would not affect Ca2+ potentiation at the foot of the acetylcholine concentration-potentiation relation (where a large fraction of additional receptors are available for opening), but it would reduce Ca2+ potentiation at higher acetylcholine concentrations, where a smaller fraction of additional receptors is available for opening.
Mechanism of Seizure Generation. The effects of the ADNFLE mutations on Ca2+ potentiation could facilitate seizure generation in two ways. If the mutations do not affect surface receptor expression or acetylcholine efficacy in 0 mM Ca2+, then their effects on Ca2+ potentiation should reduce α4β2-mediated inhibitory release in vivo. This loss could facilitate seizure generation by reducing α4β2-mediated lateral inhibition in the cortex (McNamara, 1999). The rat α4(S252F) and α4(+L264) mutations do not affect surface anti-α4 or -β2 antibody binding and do not increase the 500 μM acetylcholine response in the absence of added Ca2+ (Figl et al., 1998). Thus, these two mutations could reduce the peak acetylcholine response under physiological Ca2+ concentrations. Consistent with this prediction, the human α4(S248F), α4(S252L), and β2(V287M) mutations—orthologous to the rat α4(S252F), α4(S256L), and β2(V262M) mutations—reduce the maximum acetylcholine response in 2.5 mM added extracellular Ca2+ by ∼50% (Bertrand et al., 1998; Bertrand et al., 2002). However, these three mutations do not reduce the maximum acetylcholine response of mock heterozygous receptors (created by injecting X. laevis oocytes with a mixture of wild-type and mutant cRNA) (Bertrand et al., 2002), and ADNFLE patients are heterozygous for the mutant alleles. Cellular mRNA and protein regulation could also affect ADNFLE receptor expression in vivo. Hence, determining whether the ADNFLE mutations reduce the α4β2 acetylcholine response in vivo will ultimately require measurements in knock-in mice.
The ADNFLE mutant-mediated reductions in Ca2+ potentiation could also alter the balance between α4β2-mediated excitatory and inhibitory transmitter release during bouts of synchronized, high-frequency firing in the cortex (Rodrigues-Pinguet et al., 2003). Presynaptic nicotinic receptors facilitate both excitatory and inhibitory neurotransmitter release in the central nervous system (Wonnacott, 1997; Alkondon et al., 2000). Because of the limited extracellular space in the brain, synchronous repetitive firing may selectively deplete Ca2+ from the extracellular space around excitatory synapses (Vassilev et al., 1997; Egelman and Montague, 1999; Rusakov and Fine, 2003). Ca2+ regulation of the neuronal nicotinic agonist response may act as a negative feedback mechanism to down-regulate nicotinic receptor-mediated excitatory transmitter release during high-frequency synaptic firing (Amador and Dani, 1995). In general, Ca2+ depletion at excitatory synapses may protect against α4β2-initiated seizures by ensuring that α4β2-mediated inhibitory transmitter dominates α4β2-mediated excitatory transmitter release during bouts of synchronized, high-frequency cortical firing, such as sleep spindles. Reducing α4β2 Ca2+ potentiation could impair the ability of α4β2-mediated lateral inhibition to contain the spread of α4β2-mediated excitation in the cortex and thus allow α4β2-mediated excitation to initiate a seizure.
Footnotes
-
This work was supported by grants from the National Institute of Health (NS0438000 and NS011756, to B.N.C. and H.A.L., respectively and a minority supplement (NS0438000-04S1 to N.O.R.).
-
Portions of this work were published previously by UMI publishers in a doctoral dissertation by N.O.R.
-
Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org.
-
doi:10.1124/mol.105.011155.
-
ABBREVIATIONS: ADNFLE, autosomal dominant nocturnal frontal lobe epilepsy; [Ca2+]o, added extracellular Ca2+ concentration; is, single-channel current; gs, single-channel conductance; SSD, percentage of steady-state desensitization; τD, time constant of desensitization.
-
↵ The online version of this article (available at http://molpharm.aspetjournals.org) contains supplemental material.
- Received January 18, 2005.
- Accepted May 17, 2005.
- The American Society for Pharmacology and Experimental Therapeutics