HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 538/1/25    most recent 2001.012897v1 Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Jatzke, C. Articles by Wollmuth, L. P. Search for Related Content PubMed PubMed Citation Articles by Jatzke, C. Articles by Wollmuth, L. P.

Journal of Physiology (2002), 538.1, pp. 25-39
© Copyright 2002 The Physiological Society
DOI: 10.1113/jphysiol.2001.012897

Voltage and concentration dependence of Ca²⁺ permeability in recombinant glutamate receptor subtypes

Claudia Jatzke, Junryo Watanabe * and Lonnie P. Wollmuth

	ABSTRACT

Top Abstract Introduction Methods Results Discussion References

The channels associated with glutamate receptor (GluR) subtypes, namely N-methyl-D-aspartate receptors (NMDARs), and Ca²⁺-permeable -amino-3-hydroxy-5-methyl-4-isoxazolepropionate receptors (AMPARs) and kainate receptors (KARs), are to varying degrees permeable to Ca²⁺. To compare the mechanism of Ca²⁺ influx, we measured Ca²⁺ permeability relative to that of Na⁺ (P_Ca/P_Na) using fractional Ca²⁺ currents (P_f) and reversal potential measurements over a wide voltage and Ca²⁺ concentration range in recombinant NMDAR NR1-NR2A, AMPAR GluR-A(Q) and KAR GluR-6(Q) channels. For NR1-NR2A channels, P_Ca/P_Na derived from P_f measurements was voltage independent but showed a weak concentration dependence. A stronger concentration dependence was found when P_Ca/P_Na was derived from changes in reversal potentials on going from a Na⁺ reference solution to a solution with Ca²⁺ as the only permeant ion ('biionic' condition). In contrast, P_Ca/P_Na was concentration independent when derived from changes in reversal potentials on going from a Na⁺ reference solution to the same solution with added Ca²⁺ ('high monovalent' condition). For GluR-A(Q) channels, P_Ca/P_Na derived from all three approaches was concentration independent, and for the reversal potential-based approaches were of comparable magnitude. Their most distinctive property was that P_Ca/P_Na derived from P_f measurements was strongly voltage dependent. For GluR-6(Q) channels, P_Ca/P_Na derived from P_f measurements was weakly voltage dependent. On the other hand, P_Ca/P_Na derived from all three approaches was the most strongly concentration dependent of any GluR subtype and, except for low Ca²⁺ concentrations, the values were of comparable magnitude. Thus, the three Ca²⁺-permeable GluR subtypes showed unique patterns of Ca²⁺ permeability, indicating that distinct biophysical and molecular events underlie Ca²⁺ influx in each subtype.

(Resubmitted 22 June 2001; accepted after revision 28 September 2001)
Corresponding author L. P. Wollmuth: Department of Neurobiology and Behavior, State University of New York at Stony Brook, Stony Brook, NY 11794-5230, USA. Email: lwollmuth{at}notes1.cc.sunysb.edu

	INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

Ionotropic glutamate receptors (GluRs) are ligand-gated ion channels that mediate excitatory neurotransmission in the mammalian brain. They are key elements in numerous physiological processes, including changes in synaptic efficacy and development and maintenance of synaptic connections (Bliss & Collingridge, 1993; Constantine-Paton & Cline, 1998). GluRs also contribute to the acute and chronic cell death associated with numerous neurological diseases (Pellegrini-Giampietro et al. 1997; Lee et al. 1999). Three major subtypes of ionotropic GluRs exist: N-methyl-D-aspartate receptors (NMDARs), -amino-3-hydroxy-5-methyl-4-isoxazolepropionate receptors (AMPARs) and kainate receptors (KARs). Because of differences in their molecular and biophysical properties, these subtypes contribute differentially to synaptic physiology. A major determinant of the specific role of GluRs is the degree to which the channel associated with the receptor, when open, allows Ca²⁺ into the cell. Indeed, NMDAR channels represent the major synaptically controlled pathway for post-synaptic Ca²⁺ influx (e.g. Malinow et al. 1994), and recent evidence has shown that Ca²⁺ influx through Ca²⁺-permeable AMPAR channels has physiological significance (e.g. Mahanty & Sah, 1998; Liu & Cull-Candy, 2000). Despite the importance of Ca²⁺ influx through Ca²⁺-permeable GluR subtypes, many fundamental aspects of this process remain poorly defined.

Under physiological conditions, the current carried through Ca²⁺-permeable GluR channels is a mixture of monovalent cations (K⁺ and Na⁺) and Ca²⁺. Two general approaches have been used to characterize quantitatively this Ca²⁺ influx (Burnashev, 1996). The first, and most common, is the measurement of Ca²⁺ permeability ratios, relative to monovalent cations, using zero-current or reversal potential measurements. This approach is advantageous in that these measurements are relatively easy to make, even on native GluR channels (e.g. Koh et al. 1995). However, to derive Ca²⁺ permeability ratios from reversal potential measurements and to relate these measurements to Ca²⁺ influx under physiological conditions requires certain assumptions about the channels; namely, that they follow the Goldman-Hodgkin-Katz (GHK) or constant field assumptions. An alternative approach to characterize quantitatively Ca²⁺ influx is the combined use of Ca²⁺ photometry and high concentrations of intracellular fura-2 (dye overload) to measure the fraction of the total current carried by Ca²⁺ (see Neher, 1995). This method has the advantage of being model independent and can be used to characterize Ca²⁺ influx over a wide voltage range rather than just at the reversal potential. Further, measuring fractional Ca²⁺ currents directly quantifies the relevant flux carried by Ca²⁺ under physiological conditions, and therefore defines the important physiological parameter.

At present, the relationship, in GluR subtypes, between fractional Ca²⁺ currents and Ca²⁺ permeability, as defined by reversal potential measurements, remains unresolved. Defining this relationship is important for understanding the mechanism of Ca²⁺ influx in GluR channels, especially as it relates to the differences between the subtypes. In addition, measuring Ca²⁺ permeability using reversal potentials is a convenient approach to characterizing Ca²⁺ influx in native channels. However, without a clear relationship to P_f measurements, one cannot quantitatively relate these Ca²⁺ permeability measurements to Ca²⁺ influx under physiological conditions. Some reports have indicated that Ca²⁺ influx in NMDAR channels (Schneggenburger, 1996) and Ca²⁺-permeable AMPAR channels (Wollmuth & Sakmann, 1998) follows GHK assumptions, whereas others have indicated that it does not (NMDAR channels: Burnashev et al. 1995; Wollmuth & Sakmann, 1998; Ca²⁺-permeable AMPAR channels: Burnashev et al. 1995). In part, these differences may reflect that assumptions underlying analysis of the data differed, different monovalent species were used as a reference, a limited range of concentrations were tested and/or unphysiological conditions were used.

To address the relationship between Ca²⁺ permeability measured using changes in reversal potentials and that obtained from fractional Ca²⁺ currents, we characterized and contrasted Ca²⁺ permeability in GluR subtypes (NMDAR, Ca²⁺-permeable AMPAR and KAR channels) over a wide voltage and concentration range. To do so, we took advantage of measuring fractional Ca²⁺ currents and relative Ca²⁺ permeability using various means to quantify changes in reversal potentials. Our overall approach differs from earlier publications in that we contrasted all three Ca²⁺-permeable GluR subtypes simultaneously, under the same ionic conditions, and derived an equation relating P_f to Ca²⁺ permeability that specifically defines the monovalent species. Also, we quantified all P_f measurements relative to the reversal potential, thus correcting for any shifts in potentials due to different Ca²⁺ concentrations. We found, using GHK as a reference, that Ca²⁺ permeability in Ca²⁺-permeable KAR channels deviated from this formalism under all conditions, whereas for NMDAR and Ca²⁺-permeable AMPAR channels such deviations were more limited. However, each of the subtypes was about equally different from the others, indicating that unique biophysical and molecular events underlie the process of Ca²⁺ influx in the GluR subtypes.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Heterologous expression of GluR channels

All experiments were performed with previously described expression constructs for wild-type rat NMDAR (Wollmuth et al. 1996), AMPAR (Burnashev et al. 1992) and KAR (Köhler et al. 1993) subunits. The KAR subunit GluR-6 was fully edited within the M1 segment (V,C) (Köhler et al. 1993). Homomeric channels composed of this subunit are Cl^- impermeable (Burnashev et al. 1996). AMPAR subunits were of the flip-splice variant form. Channels were expressed transiently in HEK 293 cells (human embryonic kidney cell line) using Lipofectamine 2000 (Gibco BRL, Rockville, MD, USA). A vector for green fluorescent protein was co-transfected at a ratio of 1:9. Cells were recorded from 1 to 2 days after transfection.

Current recordings

Currents were recorded at room temperature (20-23 °C) using an EPC-9 amplifier with PULSE software (HEKA Elektronik, Lambrecht, Germany), low-pass filtered at 300-500 Hz, and digitized at 2 kHz. Pipettes had resistances of 1-4 M when filled with the pipette solution and measured in the Cs⁺ or Na⁺ reference solutions. External solutions were applied using a piezo-driven double-barrel application system. For NMDARs, one barrel contained the external solution plus glycine (20 µM) while the other barrel contained the same solution with added glutamate (100 µM). For non-NMDARs, the glutamate concentration was 1 mM. To minimize desensitization, we included in all external solutions for AMPARs 15 µM cyclothiazide (stock solution was 10 mM cyclothiazide in 100 mM NaOH). For KARs, we initially incubated cells for 2-3 min in concanavalin A (0.3 mg ml^-1) and included concanavalin A in all reference solutions. Unless otherwise noted, all chemicals were obtained from Sigma (St Louis, MO, USA) or J. T. Baker (Phillipsburg, NJ, USA).

Experimental protocols

Fractional Ca²⁺ currents. Fura-2 (1 mM) was loaded into cells via the patch pipette to measure the fraction of the total current carried by Ca²⁺ (Schneggenburger et al. 1993; Neher, 1995). Briefly, cells were illuminated alternately at 360 and 380 nm (2-10 Hz) by a polychromatic illumination system (T.I.L.L. Photonics, München, Germany). Excitation light was coupled to the microscope via a fibre optics light guide. A 425 nm dichroic mirror and a 500-530 nm band-pass emission filter were included in the light path. Fluorescence signals were measured with a photodiode (T.I.L.L. Photonics).

Fractional Ca²⁺ currents (P_f) were quantified using the relationship P_f (%) = Q_Ca/Q_T 100, where Q_Ca and Q_T are the charge carried by Ca²⁺ and the total charge, respectively, during a defined time interval. Q_T was derived from the current integral. Q_Ca was derived from the relationship, Q_Ca = F₃₈₀/f_max, where F₃₈₀ is the change in the fluorescence signal with 380 nm excitation and f_max is the proportionality constant between the charge carried by inward Ca²⁺ and F₃₈₀ (see below). F₃₈₀ was normalized to the fluorescence of 4.5 mm diameter fluoresbrite BB beads (lot no. 481613; Polysciences Inc., Warrington, PA, USA) and expressed in bead units (BU). The bead unit was determined on each experimental day as the mean fluorescence of 5-10 beads at 380 nm excitation. The proportionality constant, f_max, was determined at -100 mV in 10 mM Ca²⁺, 140 mM N-methyl-D-glucamine (NMDG⁺), using NMDAR NR1-NR2A channels and was 0.032 ± 0.001 BU pC^-1 (n = 8) using the KCl-based internal solution (see Wollmuth & Sakmann, 1998).

In measuring fractional Ca²⁺ currents, our intracellular solution consisted of (mM): 140 KCl, 10 Hepes and 1 fura-2 (K₅-fura-2) with the pH adjusted to 7.2 with KOH. The total intracellular K⁺ concentration, [K⁺]_i, was 148.5 mM. The external solution consisted of (mM): 140 NaCl and 10 Hepes with the pH adjusted to 7.2 using NaOH and added CaCl₂ (0.3-50 mM). The total extracellular Na⁺ concentration, [Na²⁺]_o, was 143.5 mM. The low Na⁺ solution consisted of (mM): 20 NaCl, 120 NMDG⁺, 10 Hepes and 1.8 CaCl₂ with the pH adjusted to 7.2 using HCl. All membrane potentials reported for P_f measurements are relative to the reversal potential. Fura-2 was obtained from Molecular Probes (Eugene, OR, USA).

Ca²⁺ permeability. We used three approaches to quantify Ca²⁺ permeability relative to monovalent cations, either Na⁺ (P_Ca/P_Na) or Cs⁺ (P_Ca/P_Cs), in GluR subtypes. The first approach, following that of Schneggenburger et al. (1993) and Burnashev et al. (1995), is based on measurements of fractional Ca²⁺ currents and relates Ca²⁺ permeability to P_f using Goldman-Hodgkin-Katz (GHK) assumptions. Our approach differs from that in earlier publications, however, in that we explicitly defined the intracellular and extracellular monovalent cation composition and their relative permeability. To derive this more general equation, we defined a term, P_Ca' = P_Ca/(1 + exp(VF/RT)), analogous to that in the derivation of the Lewis equation (Lewis, 1979). Assuming that the only permeant ions are K⁺ intracellularly and Na⁺ and Ca²⁺ extracellularly, the form of this equation is:

(1)

I_Ca, I_K and I_Na are the currents carried by Ca²⁺, K⁺ and Na⁺, respectively. R, T and F have their normal thermodynamic meanings, and the quantity RT/F was 25.4 mV (21 °C). For our purposes, V refers to the membrane potential relative to the reversal potential (V = V_test - V_rev) where V_test and V_rev are the test and reversal potentials, respectively. V_rev was determined empirically for each cell and for each Ca²⁺ concentration tested. These potentials were not corrected for junction potentials. Appropriate P_Na/P_K values are found in Table 2 (P_Na/P_K = P_Na/P_Cs P_Cs/P_K).

Equation (1) reduces to the simpler form found in Burnashev et al. (1995) if one uses the same assumption therein (i.e. [K⁺]_i = [Na⁺]_o and P_K/P_Na = 1). The predicted P_f values from eqn (1) and its simpler form - if the P_f values are adjusted relative to the different reversal potentials they predict - differ by -1 to 8 % depending on the membrane potential (eqn (1) predicts a stronger voltage dependence). We used eqn (1) rather than the simpler form because it properly defined P_Ca/P_Na rather than P_Ca/P_M, which can vary greatly depending on which monovalent species (M) is used as a reference (see e.g. Fig. 4). Further, to relate predicted and measured P_f values, especially those around the reversal potential, requires that V_rev be explicitly defined, which can be accomplished only by eqn (1) since P_K/P_Na =/ 1, and that, in our conditions, [K⁺]_i =/ [Na⁺]_o. Finally, to further standardize our measurements, all potentials were set relative to V_rev rather than the absolute potential. This approach properly quantified P_f (and hence P_Ca/P_Na) not only around V_rev, but also under different ionic conditions including when [Ca²⁺]_o was altered.

The two alternative approaches to quantify Ca²⁺ permeability were based on measuring changes in the reversal potential, E_rev, for glutamate-activated currents on replacing a reference solution with a test solution. The first of these reversal potential-based approaches, the 'biionic' approach, involved measuring E_rev on replacing Na⁺ in a Na⁺-based reference solution with Ca²⁺ (e.g. Wollmuth & Sakmann, 1998). The reference solution consisted of (mM): 140 NaCl and 10 Hepes, pH adjusted to 7.2 with NaOH ([Na⁺]_o = 143.5 mM). The high Ca²⁺-containing solution consisted of (mM): 108 Ca²⁺, 2 Ca(OH)₂ and 10 Hepes (final pH was 7.2). For concentrations of 0.5, 1.8 or 10 mM, Ca²⁺ was added to an NMDG⁺ solution (mM: 140 NMDG⁺ and 10 Hepes, pH adjusted to 7.2 with HCl). NMDG⁺ is impermeant in NMDAR channels (Villarroel et al. 1995) but shows a weak permeability in non-NMDAR channels. We assumed this NMDG⁺ permeability (P_NMDG/P_Na) was approximately 0.02 in GluR-A(Q) channels and 0.01 in GluR-6(Q) (Burnashev et al. 1996) (corrected assuming P_NMDG/P_Na = P_NMDG/P_Cs P_Cs/P_Na, see Table 2). This small NMDG⁺ permeability in non-NMDAR channels affects the magnitude of P_Ca/P_Na, especially in 0.5 mM Ca²⁺. It also makes the use of the term 'biionic' not absolutely appropriate, but we continue to use this term since it readily conveys the general ionic conditions. E_rev was converted to P_Ca/P_Na using the Lewis equation (see eqn (7) in Wollmuth & Sakmann, 1998).

The second reversal potential-based approach we used involved measuring E_rev on going from the Na⁺-based reference solution to the same solution but with added Ca²⁺ (0.3-50 mM Ca²⁺). Because extracellular monovalent cations are always present at a high concentration, we refer to this approach as the 'high monovalent' approach. To calculate permeability ratios (P_Ca/P_Na) from E_rev for individual Ca²⁺ concentrations, we started with the Lewis equation. Assuming Na⁺ and Ca²⁺ were the only permeant ions in the external solution, the form of this equation is:

(2)

where P_Ca' is P_Ca/(1 + exp(E_rev,CaF/RT)). Equation (2) can also be rearranged to fit E_rev as a function of Ca²⁺ concentration to derive P_Ca/P_Na. The form of this equation is:

(3)

Note that this equation contains P_Ca rather than P_Ca'. The derivation of this equation requires the assumption that E_rev,Ca and E_rev are the same, that is E_rev,Na = 0. This latter condition occurs only if [K⁺]_i = [Na⁺]_o and P_Na/P_K = 1. Because of this assumption, the use of eqn (3) fitted to E_rev leads to ~20 % error in P_Ca/P_Na (see Fig. 5).

For both reversal potential-based approaches, the reversal or zero potential was determined by plotting peak current amplitudes, generated by voltage steps in 2, 3 or 5 mV increments, against voltage and fitting them with a fourth-order polynomial (see Wollmuth & Sakmann, 1998). The pipette solution consisted of (mM): 140 KCl, 10 BAPTA and 10 Hepes, pH adjusted to 7.2 with KOH. The control recording was typically an average of the control recordings made before (pre) and after (post) exposure to the test solution. When E_rev was < 5 mV, results were not included in the final analysis unless both the pre- and post-control recordings were made.

In general, we did not correct for ionic activity, since activity coefficients () were not greatly different between the different solutions (see Wollmuth & Sakmann, 1998). In a few examples, we did calculate P_Ca/P_Na using activity coefficients (i.e. concentration terms in eqn (1) were multiplied by ). _Na and _K were assumed to be 0.72 and _Ca (for 1.8 mM Ca²⁺) was assumed to be 0.57 (Burnashev et al. 1995).

Divalent and monovalent cation permeability. The permeability of a variety of divalent (D) ions, in addition to Ca²⁺, and monovalent (M) ions was also measured in GluR subtypes. To quantify P_D/P_Cs and P_M/P_Cs we followed the biionic approach outlined above with several small alterations. First, we used Cs⁺ as the reference ion. In addition, to avoid the problem of not having a hydroxide ion for a particular salt, we used Trizma base rather than Hepes as a buffer. Our reference solution therefore consisted of 143 mM CsCl and 10 mM Trizma base (pH between 7.1 and 7.3). Current amplitudes and reversal potentials were identical when recording in the two Cs⁺-based reference solutions (C. Jatzke & L. P. Wollmuth, unpublished observations). For measuring divalent cation permeability, the high divalent cation solution consisted of (mM): 110 divalent salt (CaCl₂, BaCl₂, MgCl₂, SrCl₂ or CoCl₂) and 10 Trizma base. Lower concentrations were measured using the NMDG⁺-based solution. For measuring monovalent permeability, CsCl in the reference solution was replaced by NaCl, KCl, RbCl or LiCl in the test solution. Permeability ratios were calculated using the Lewis equation.

Data analysis

All curve fitting was done using Igor Pro (WaveMetrics Inc., Lake Oswego, OR, USA). Results are reported as means ± S.E.M. and shown graphically as means ± 2 S.E.M. An analysis of variance was used to test for statistical differences with the Tukey test used for multiple comparisons. Significance was assumed if P < 0.05.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Fractional Ca²⁺ currents in GluR subtypes

Figure 1 illustrates the dye overload approach to measuring the fraction of the total current carried by Ca²⁺. The upper panel shows whole-cell glutamate-activated currents in HEK 293 cells expressing, from left to right, NMDAR NR1-NR2A, AMPAR GluR-A(Q) or KAR GluR-6(Q) channels. In all examples, the external solution contained 1.8 mM Ca²⁺ in 143.5 mM NaCl. The application of glutamate (filled bar) generated an inward current. It also generated a decrement in the fluorescence signal at 380 nm excitation (F₃₈₀), the magnitude of which depended on the specific GluR subtype. With dye overload, changes in the fluorescence signal (F₃₈₀) are proportional to the total Ca²⁺ influx (Q_Ca) with the proportionality constant defined by f_max according to the relationship Q_Ca = I_Cadt = F₃₈₀/f_max. Fractional Ca²⁺ currents were derived from the relationship P_f (%) = Q_Ca/Q_T 100. Q_T, the total charge during the defined time interval, was derived from the current integral (shaded area in current plot). In NR1-NR2A channels, fractional Ca²⁺ currents in 1.8 mM Ca²⁺ at -60 mV were around 13.5 % (Table 1). As expected, the comparable P_f measurement in GluR-A(Q) channels was greatly reduced to around 3.6 % and even further reduced in GluR-6(Q) channels to around 2.4 %. These P_f values are comparable in magnitude, though all slightly higher, than those reported previously for the GluR subtypes (Burnashev et al. 1995). They are also consistent with the general picture that NMDAR channels are more permeable to Ca²⁺ than Ca²⁺-permeable AMPAR or KAR channels.

		View larger version [in this window] [in a new window]
Figure 1. Fractional Ca2+ currents (Pf) in GluR subtypes Simultaneous measurement of whole-cell currents (I, top) and fluorescence intensity with 380 nm excitation (F380, bottom) evoked by glutamate applications (filled bars) in HEK 293 cells expressing NMDAR NR1-NR2A (left panel), AMPAR GluR-A(Q) (middle panel) or KAR GluR-6(Q) (right panel) channels. The potential (V) was -60 mV to the reversal potential (see Methods). In the current records (upper panels), the dashed lines represent zero current, and the shaded regions correspond to the current integral (QT), which was approximately the same for each of the example records. The F380 plot is expressed in bead units (BU). F380 was derived as the difference between the F380 amplitude at the indicated time (arrow) and the baseline F380 signal (continuous line), extrapolated from a linear fit to the F380 amplitudes prior to glutamate application.

Voltage dependence of fractional Ca²⁺ currents in GluR subtypes

Figure 2A shows mean fractional Ca²⁺ currents quantified over a wide voltage range in NMDAR NR1-NR2A, AMPA GluR-A(Q) and KAR GluR-6(Q) channels. As expected, P_f values, on going from negative potentials to V_rev, increased in magnitude and then decreased in magnitude at potentials positive to V_rev. At -20 mV, for example, P_f values were increased compared to those at -60 mV, by about 30 % in NR1-NR2A channels, from around 13.5 % at -60 mV to 17.3 % at -20 mV (Table 1). For GluR-A(Q), the comparable values were nearly doubled, from 3.6 % at -60 mV to 7.1 % at -20 mV, whereas for GluR-6(Q) they were increased by about 50 %, from 2.4 to 3.6 %.

		View larger version [in this window] [in a new window]
Figure 2. Voltage dependence of Pf measurements in GluR subtypes A, mean Pf values measured in GluR subtypes over a wide voltage range (n > 5). Voltages are relative to Vrev. The lines are predicted Pf values, using eqn (1), with PCa/PNa derived from the Pf measurement either at -60 mV (continuous line), or at -20 mV (dashed line; see Table 1). B, average PCa/PNa derived from Pf values measured at different membrane potentials (only Pf values greater than 2 % were included in this plot). The continuous lines are fits of the relationship PCa/PNa(V) = PCa/PNa(Vrev) exp(VzF/RT), where PCa/PNa(Vrev) is the estimated PCa/PNa at the reversal potential and z is the voltage dependence of the process. For non-NMDAR channels, PCa/PNa was fitted only at potentials positive to -60 mV where a clear voltage dependence existed. PCa/PNa(Vrev) and z were: 1.50, 0.41 for GluR-A(Q) (); 1.40, 0.40 for GluR-B(Q) (, lower); 2.10, 0.32 for GluR-B(N) (, upper); and 0.63, 0.18, for GluR-6(Q).

P_f measurements are model independent. Using GHK assumptions, P_f measurements can be converted into relative Ca²⁺ permeability ratios, in our case P_Ca/P_Na, using eqn (1). With this approach, we derived P_Ca/P_Na based on the P_f measurement at single membrane potentials, and then used this derived P_Ca/P_Na to predict P_f values, again using eqn (1), over the entire voltage range. In Fig. 2A, the predicted P_f values were based on P_Ca/P_Na derived from the P_f measurement either at -60 mV (continuous lines) or at -20 mV (dashed lines). For NR1-NR2A channels, the derived P_Ca/P_Na at -60 mV as well as at -20 mV was around 3. The predicted P_f based on this P_Ca/P_Na describes the voltage dependence of P_f over the entire voltage range, even at potentials positive to the reversal potential. In contrast, for GluR-A(Q) the predicted P_f based on the P_Ca/P_Na derived at -60 mV (~0.71) does not describe the results close to the reversal potential, significantly under-estimating the measured values. Indeed, the derived P_Ca/P_Na at -20 mV was significantly greater, around 1.10, than that derived at -60 mV. Correspondingly, the predicted P_f from this derived P_Ca/P_Na (dashed line) over-estimates the measured P_f at potentials negative to -30 mV. For GluR-6(Q) a similar deviation occurs to that for AMPAR channels but its magnitude is considerably less. The derived P_Ca/P_Na at -60 mV was around 0.48 whereas at -20 mV it was around 0.54.

The results shown in Fig. 2A (left and centre panel) are qualitatively comparable to those found previously (e.g. Burnashev et al. 1995; Schneggenburger, 1996). However, to convert these P_f measurements to Ca²⁺ permeability, we derived an equation (eqn (1)), that allowed us to specifically define P_Ca/P_Na rather than P_Ca/P_M. With this approach, we derived P_Ca/P_Na for each individual potential and plotted these values against voltage (Fig. 2B). As expected, P_Ca/P_Na was voltage independent in NR1-NR2A channels. In contrast, for GluR-A(Q), P_Ca/P_Na showed little voltage dependence at potentials negative to -40 mV but was strongly voltage dependent at potentials close to the reversal potential. Fitting over the voltage range from -40 to -10 mV (continuous line), yielded a predicted Ca²⁺ permeability at the reversal potential, P_Ca/P_Na(V_rev), of 1.50 and a voltage dependence (z) of around 0.41. A weaker voltage dependence, 0.18, was found in GluR-6(Q) channels with a P_Ca/P_Na(V_rev) around 0.63.

Ca²⁺-permeable, non-NMDAR channels are blocked by intracellular polyamines in a voltage-dependent manner, with the block stronger at potentials around 0 mV (e.g. Koh et al. 1995). This block is absent in mutant GluR-B channels containing an asparagine (N) at the Q/R-site (GluR-B(N)). To test whether this polyamine block underlies the voltage dependence of P_Ca/P_Na, as derived from P_f measurements, we quantified P_f measurements over a wide voltage range in GluR-B(Q) and GluR-B(N) channels, and used these values to derive P_Ca/P_Na (Fig. 2B, middle panel, ). For both GluR-B(Q) (lower set) and GluR-B(N) (upper set), P_Ca/P_Na was strongly voltage dependent with z values of around 0.40 and 0.32, respectively. The results for GluR-B(N) support the idea that the voltage dependence of P_Ca/P_Na is not due to block by intracellular polyamines, and presumably reflects an intrinsic property of the ion conduction pathway.

In summary, the voltage dependence of fractional Ca²⁺ currents, as characterized by P_Ca/P_Na, varies between the different GluR subtypes. Ca²⁺-permeable AMPAR subtypes showed a strong voltage dependence, specifically around the reversal potential, whereas this effect was much weaker in KAR channels and absent in NMDAR channels. Based on these results, Ca²⁺-permeable non-NMDAR but not NMDAR channels deviate from GHK assumptions.

Concentration dependence of fractional Ca²⁺ currents in GluR subtypes

The results shown in Fig. 2 were obtained under approximately physiological concentrations of Ca²⁺. To further contrast Ca²⁺ influx in GluR subtypes, we measured fractional Ca²⁺ currents over a wide range of extracellular Ca²⁺ concentrations (0.1-50 mM Ca²⁺) (Fig. 3). Figure 3A shows the mean P_f values for the various GluR subtypes at -60 mV relative to V_rev, thus correcting for the shifts in V_rev due to the different Ca²⁺ concentrations (cf. Fig. 5). The continuous lines are fitted by Hill equations. Since P_f values showed no clear saturation for any of the GluR subtypes, the derived parameters from these fits can only be viewed as a first approximation. The concentration at which a half-maximal effect was observed was approximately 10 mM for NR1-NR2A at -60 mV. The apparent affinity was significantly lower in GluR-A(Q) channels, around 57 mM, but surprisingly for GluR-6(Q) the affinity was around 28 mM. This higher affinity in GluR-6(Q) channels relative to GluR-A(Q) reflects that for the Hill equation fit, P_f,max was only 34 % for GluR-6(Q) whereas for GluR-A(Q) it was around 100 %.

		View larger version [in this window] [in a new window]
Figure 3. Concentration dependence of Pf measurements in GluR subtypes A, mean Pf values, at -60 mV to the reversal potential, in cells expressing NR1-NR2A, GluR-A(Q) or GluR-6(Q) channels (n > 4). The continuous lines through the points are fitted Hill equations (Pf,max/(1 + (K0.5/[Ca2+])nH)) where Pf,max is the maximal Pf, K0.5 the half-maximal response, and nH the Hill coefficient. For all fits, the Hill coefficient was around 1. The fits yielded Pf,max and K0.5 of approximately: 90 %, 10 mM for NR1-NR2A; 100 %, 57 mM for GluR-A(Q); and 34 %, 28 mM for GluR-6(Q). Measurements at -20 mV were not made for all concentrations (see B) so comparable fits could not be made. B, average PCa/PNa derived using eqn (1) from Pf values measured at -60 mV (; shown in A) or at -20 mV (). The continuous lines are the average of the PCa/PNa values measured around physiological concentrations (1 and 1.8 mM Ca2+) and were, at -60 and -20 mV: 3.1, 3.1 for NR1-NR2A; 0.68, 1.10 for GluR-A(Q); and 0.49, 0.57 for GluR-6(Q).

Figure 3B shows P_Ca/P_Na derived from P_f measurements at -60 mV () or -20 mV () using eqn (1) over a wide concentration range. In the case of NR1-NR2A channels (left panel), P_Ca/P_Na was concentration dependent, being reduced from around 3.56 at 0.1 mM Ca²⁺ to 2.38 at 20 mM Ca²⁺, a reduction of around 33 %. Like that in 1.8 mM Ca²⁺ (Fig. 2), P_Ca/P_Na derived from P_f at -20 mV was not significantly different from that derived at -60 mV for any of the concentrations tested. Hence, the concentration dependence of P_Ca/P_Na derived from P_f measurements appears to occur equally over the entire voltage range.

Compared to NMDARs, P_Ca/P_Na values derived from P_f measurements for GluR-A(Q) showed two distinct differences. First, the values derived at -60 and -20 mV were significantly different over the entire concentration range and, second, both sets of values were independent of concentration. The average values at physiological concentrations of Ca²⁺ (1 and 1.8 mM; continuous lines) were around 0.68 at -60 mV and 1.09 at -20 mV. A comparable concentration independence was found in GluR-B(Q) and GluR-B(N) channels (Table 1).

P_Ca/P_Na derived from P_f measurements in GluR-6(Q) channels was also concentration dependent, as in NR1-NR2A channels. However, this concentration dependence was considerably stronger, with P_Ca/P_Na being reduced from 0.52 at 0.3 mM to 0.19 at 50 mM, a 64 % reduction. Also, the derived P_Ca/P_Na values at -60 and -20 mV diverged at low concentrations (< 1.8 mM), with that at -60 mV levelling off at low concentrations and that at -20 mV continuing to increase in magnitude.

In summary, P_Ca/P_Na derived from P_f measurements is concentration dependent in NMDAR and KAR but not in AMPAR channels, with this concentration dependence strongest in KAR channels. Such a concentration dependence is inconsistent with GHK, but the pattern observed here differs from that for the voltage dependence, where AMPARs and to a lesser extent KARs deviated from GHK whereas NMDARs did not.

Ca²⁺ permeability in GluR subtypes based on changes in reversal potentials

A more traditional approach to measuring Ca²⁺ selectivity is the use of changes in reversal potentials on going from a solution without Ca²⁺ to one with Ca²⁺ to derive a Ca²⁺ permeability ratio. To measure Ca²⁺ permeability ratios based on changes in reversal potentials, we used two different approaches. The first, termed the 'biionic' approach, involved measuring changes in reversal potentials on going from a Na⁺-based reference solution to a solution containing predominantly Ca²⁺ as the permeant ion. Figure 4 illustrates and summarizes the results of the biionic approach. For NR1-NR2A and GluR-6(Q), P_Ca/P_Na () was strongly concentration dependent, whereas for GluR-A(Q) it was concentration independent. To compare this concentration dependence, we looked at the ratio of P_Ca/P_Na in 1.8 mM Ca²⁺ to that in 110 mM Ca²⁺. This ratio was highest for GluR-6(Q), around 4.9, and was about 2 for NR1-NR2A. On the other hand, in 0.5 mM Ca²⁺, P_Ca/P_Na continued to increase in magnitude for GluR-6(Q) whereas for NR1-NR2A it strongly decreased. Thus, NMDAR and KAR channels, but not AMPAR channels, show a concentration dependence in Ca²⁺ permeability when measured under biionic conditions. The pattern of concentration dependence, however, differs with KAR channels, having a minimum at 110 mM Ca²⁺ and increasing monotonically at lower concentrations, whereas NMDAR channels have a maximum at 1.8 mM Ca²⁺.

		View larger version [in this window] [in a new window]
Figure 4. Ca2+ permeability in GluR subtypes under biionic conditions Average PCa/PNa () or PCa/PCs () derived from Erev for individual Ca2+ concentrations using the Lewis equation (n > 4). Erev was measured on replacing Na+ (or Cs+) in a reference solution with 0.5, 1.8, 10 or 110 mM Ca2+ (see Wollmuth & Sakmann, 1998). The lines through the points have no theoretical meaning. The weak NMDG+ permeability in non-NMDAR channels (see Methods) strongly alters the magnitude of PCa/PNa, especially in 0.5 mM Ca2+, and these values should be viewed cautiously. Nevertheless, assuming a PNMDG/PNa value of between 0 and 0.02 for GluR-6(Q) does not alter the overall pattern of concentration dependence (the values shown were derived using PNMDG/PNa = 0.01). PCa/PCs was not measured in 0.5 mM Ca2+.

Figure 4 also shows Ca²⁺ permeability ratios measured relative to Cs⁺ (), for the various GluR subtypes. As expected from the relative monovalent cation ratio, P_Na/P_Cs (Table 2), which was less than unity for all GluR channels, P_Ca/P_Cs was smaller than P_Ca/P_Na for all subtypes, being reduced by slightly more than 20 % in NR1-NR2A channels, and by about 15 % in non-NMDAR channels. This plot directly illustrates that in order to quantify the relationship between P_Ca and P_f measurements, the reference monovalent cation needs to be explicitly defined.

A second and alternative reversal potential-based approach to quantify Ca²⁺ permeability involves measuring E_rev on going from a Na⁺-based reference solution to the same solution but with added Ca²⁺. The results of this 'high monovalent' approach for the GluR subtypes are shown in Fig. 5. Figure 5A shows the mean E_rev over a wide concentration range of added Ca²⁺ (0.3-20 mM Ca²⁺ for NMDAR, and 5-50 mM for Ca²⁺-permeable non-NMDARs). For NR1-NR2A, a single P_Ca/P_Na (~4.35, continuous line) describes E_rev over the entire concentration range, consistent with previous reports (Mayer & Westbrook, 1987; Zarei & Dani, 1994; Schneggenburger, 1996). We also converted E_rev at individual Ca²⁺ concentrations to P_Ca/P_Na using eqn (2) (Fig. 5B). Both approaches reveal that P_Ca/P_Na is concentration independent for NR1-NR2A, yet they yield quantitatively different results. In particular, the fit of E_rev (Fig. 5A) gave a P_Ca/P_Na of around 4.35 (continuous line in Fig. 5B), whereas the average of the individually derived P_Ca/P_Na values (dashed line) was around 3.60. This difference reflects the fact that to derive the equation to fit E_rev in Fig. 5A (i.e. eqn (3)) the assumption must be made that in the reference solution the reversal potential occurs at 0 mV, a condition arising only if [K⁺]_i = [Na⁺]_o and P_Na/P_K = 1 (see Methods). Thus, to quantitatively describe P_Ca/P_Na in the high monovalent approach, we used the values derived from individual Ca²⁺ concentrations.

		View larger version [in this window] [in a new window]
Figure 5. Ca2+ permeability in GluR subtypes under high monovalent conditions A, mean Erev on going from the Na+-based reference solution to the same solution but with added Ca2+ (0.3-50 mM Ca2+; n > 4). The continuous lines are fits of eqn (3), yielding PCa/PNa values of 4.35 for NR1-NR2A, 1.74 for GluR-A(Q) and 0.32 for GluR-6(Q). B, average PCa/PNa derived from Erev (shown in A) for individual Ca2+ concentrations using eqn (2). The continuous lines are from the fits in A. The dashed lines are the average of the individually derived PCa/PNa values (3.6 for NR1-NR2A and 1.55 for GluR-A(Q)).

For GluR-A(Q) (Fig. 5, middle panels), a single P_Ca/P_Na (~1.74, continuous line in Fig. 5A) described E_rev over the entire concentration range. Thus, like NMDAR channels, Ca²⁺-permeable AMPAR channels show a concentration-independent P_Ca/P_Na derived under the high monovalent condition. The average of the individually derived P_Ca/P_Na values was around 1.55 (dashed line in Fig. 5B). In contrast, for GluR-6(Q) (right panels), a single P_Ca/P_Na could not describe E_rev over the entire concentration range. Correspondingly, P_Ca/P_Na was strongly concentration dependent, ranging from about 0.52 at 5 mM Ca²⁺ down to 0.16 at 50 mM Ca²⁺.

Deviation from GHK is not due to competition between Na⁺ and Ca²⁺

The magnitude of Ca²⁺ permeability in NMDAR channels depended strongly on the approach used to measure it, with the biionic approach showing the highest P_Ca/P_Na at 1.8 mM Ca²⁺. For the biionic approach, the Ca²⁺ reversal potential was measured in the absence of external Na⁺ with the impermeant NMDG⁺ representing the monovalent ion. Ca²⁺ and Na⁺ ions compete for the pore (see Discussion). Hence, the high P_Ca/P_Na measured in the biionic approach may arise because Ca²⁺ has better access to the pore in the absence of Na⁺. To test this idea, we measured P_f at -60 mV relative to the reversal potential in low external Na⁺ (Fig. 6A). As expected given the higher proportion of Ca²⁺ (1.8 mM) relative to Na⁺ (20 mM), P_f values were considerably higher than those in physiological conditions (1.8 mM Ca²⁺, 140 mM Na⁺). Nevertheless, P_Ca/P_Na values derived from these P_f measurements were either unchanged or less than those derived under physiological concentrations (Fig. 6B). These experiments indicate that competition between Ca²⁺ and Na⁺ for the pore does not explain the high P_Ca/P_Na found in the biionic approach.

		View larger version [in this window] [in a new window]
Figure 6. Pf measurements in low extracellular Na+ A, mean Pf values measured in low extracellular Na+. Cells were bathed in a solution containing 1.8 mM Ca2+, 20 mM Na+ and 120 mM NMDG+. Values were measured at -60 mV. The number of cells recorded was, from left to right, 4, 3, 3. B, comparison of PCa/PNa values derived from Pf measurements at -60 mV either in 143.5 mM Na+ or in 20 mM Na+.

Divalent permeability in GluR subtypes

To further contrast the process of Ca²⁺ selectivity in these GluR subtypes, we examined their permeability to different divalent cations (D) (see also Dingledine et al. 1992; Egebjerg & Heinemann, 1993; Tsuzuki et al. 1994). To measure P_D/P_Cs, we quantified E_rev on going from the Cs⁺-based reference solution to a solution containing either 110 mM divalent cation or 10 mM divalent cation plus 140 mM NMDG⁺. The divalent cations tested included CaCl₂, BaCl₂, SrCl₂, MgCl₂, and CoCl₂. Figure 7 summarizes P_D/P_Cs, measured with either 10 or 110 mM divalent cation, for NR1-NR2A, GluR-A(Q) and GluR-6(Q). NR1-NR2A channels were highly permeable to three divalent cations, Ca²⁺ (), Ba²⁺ () and Sr²⁺ (). In contrast, Co²⁺ () and Mg²⁺ () were poorly permeable. The selectivity sequence, in 110 mM divalent cations, was: Ca²⁺ (2.55) > Ba²⁺ (2.10) > Sr²⁺ (1.60) >> Co²⁺ (0.04) Mg²⁺ (0.04). GluR-A(Q) channels differed from NMDAR channels in that any selectivity among the divalent cations was very weak. The magnitude of this permeability, compared to NR1-NR2A, was either considerably reduced (Ca²⁺, Ba²⁺ and Sr²⁺) or considerably greater (Co²⁺ and Mg²⁺) with the selectivity sequence: Ca²⁺ (1.43) Ba²⁺ (1.43) > Sr²⁺ (1.10) > Mg²⁺ (0.86) Co²⁺ (0.57). As for GluR-A(Q), GluR-6(Q) channels showed a weak selectivity amongst divalent cations, but the magnitude of this permeability was greatly reduced for all divalent cations. The selectivity sequence was: Mg²⁺ (0.16) Ca²⁺ (0.12) > Sr²⁺ (0.08) Ba²⁺ (0.07) Co²⁺ (0.06).

		View larger version [in this window] [in a new window]
Figure 7. Divalent cation permeability in GluR subtypes Mean permeability ratios, PD/PCs, for different divalent ions in the GluR subtypes (n > 3). The divalent ions tested included Ca2+ (), Ba2+ (), Sr2+ (), Mg2+ () and Co2+ (). For GluR-6(Q), only Ca2+ and Mg2+ could be measured at 10 mM because of the very negative reversal potential and the small current size in the test solution.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The use of dye overload to measure the fraction of the total current carried by Ca²⁺ in channels with a mixed Ca²⁺/ monovalent cation permeability directly quantifies Ca²⁺ influx under physiological conditions (Schneggenburger et al. 1993; Vernino et al. 1994; Neher, 1995). Using this approach, as well as those based on reversal potential measurements, we characterized Ca²⁺ permeability in the three major GluR subtypes over a wide voltage and Ca²⁺ concentration range. This comparison indicates that the various GluR subtypes show similarities and differences in the underlying process of Ca²⁺ permeation. Surprisingly, except for the absolute magnitude, NMDAR channels frequently displayed Ca²⁺ permeability properties intermediate between those for Ca²⁺-permeable AMPAR and KAR channels. Thus, despite the fact that AMPAR and KAR subunits show a much higher sequence similarity to each other than to NMDAR subunits, they do not share common mechanisms of Ca²⁺ influx.

Comparison of Ca²⁺ permeability derived from fractional Ca²⁺ currents and reversal potential measurements in GluR subtypes

We used three different experimental approaches to characterize the voltage and concentration dependence of P_Ca/P_Na in GluR subtypes. The first approach was based on deriving P_Ca/P_Na from P_f measurements. The two other, more traditional approaches involved deriving P_Ca/P_Na from changes in reversal potential (E_rev) on going from a Na⁺-based reference solution to either a pure Ca²⁺-containing solution (the biionic approach) or a solution containing Na⁺ plus added Ca²⁺ (the high monovalent approach). The derivation of P_Ca/P_Na for all three approaches is based on the GHK equation.

Figure 8 summarizes and contrasts P_Ca/P_Na derived from P_f measurements (at -60 mV, and at -20 mV, ) to those derived from reversal potentials either using the biionic (, left panel) or the high monovalent (, right panel) approach. Across panels, one can also compare the magnitude of P_Ca/P_Na derived from the two reversal potential-based approaches.

		View larger version [in this window] [in a new window]
Figure 8. Comparison of PCa/PNa derived from Pf measurements and from changes in reversal potentials Comparison of PCa/PNa derived from Pf measurements at -60 mV () or at -20 mV () (from Fig. 3) to PCa/PNa derived from changes in reversal potentials using either the biionic (left panel, ; from Fig. 4) or the high monovalent (right panel, ; from Fig. 5) approach. The dashed lines in all panels show the average PCa/PNa derived from Pf measurements in 1 and 1.8 mM Ca2+ either at -60 or at -20 mV. A, NMDAR NR1-NR2A. Dashed line, 3.1. The continuous line in the left panel has no theoretical meaning, whereas that in the right panel is the average PCa/PNa (3.60). B, AMPAR GluR-A(Q). Dashed lines, 0.68 (-60 mV) or 1.10 (-20 mV). The continuous lines are the average of the individually derived PCa/PNa (1.73, left panel and 1.55, right panel). The '' in the left panel represents the estimated PCa/PNa at 0 mV (1.50) from Fig. 2B. C, KAR GluR-6(Q). Dashed lines, 0.49 (-60 mV). The line at -20 mV is not shown. The continuous lines have no theoretical meaning. PCa/PNa in 0.5 mM Ca2+ under the biionic conditions is not shown in order to compare more readily the values at higher Ca2+ concentrations.

NMDAR channels. For NR1-NR2A, P_Ca/P_Na derived from P_f measurements or from reversal potentials under biionic conditions was concentration dependent (Fig. 8A). Similarly, P_Ca/P_Na values derived from the two reversal potential-based approaches were significantly different. These results argue against the validity of GHK assumptions for Ca²⁺ influx in NMDAR channels. However, P_Ca/P_Na derived from P_f measurements was voltage independent (see also Burnashev et al. 1995; Schneggenburger, 1996), and P_Ca/P_Na using the high monovalent approach was concentration independent (see also Mayer & Westbrook, 1987; Zarei & Dani, 1994; Schneggenburger, 1996). Also, P_Ca/P_Na values derived from these two approaches were not greatly different, especially at physiological and lower concentrations of Ca²⁺ (Fig. 8A, right panel), as has been found previously (Schneggenburger, 1996). Hence, under physiological conditions - specifically for Ca²⁺ and Na⁺ (see below) - NMDAR channels display properties that deviate only weakly from GHK. Nevertheless, other evidence indicates that NMDAR channels have properties incompatible with GHK, including concentration-dependent P_Ca/P_Na values derived from P_f values (Fig. 3B), multiple sites in the pore for monovalent cations (Antonov et al. 1998) and Ca²⁺ (Premkumar & Auerbach, 1996; Sharma & Stevens, 1996), and a strong interaction between Ca²⁺ and intracellular monovalent cations (Wollmuth & Sakmann, 1998).

For NMDARs, P_Ca/P_Na derived using the biionic approach differed greatly from that derived using the other approaches. In 1.8 mM Ca²⁺, for example, P_Ca/P_Na using the biionic approach was around 7.2. Such a P_Ca/P_Na would predict a P_f at -60 mV and in physiological conditions of around 27 %, a value twice as large as that observed (~13.5 %). In the biionic approach, NMDG⁺ completely replaced Na⁺ in the test solution. One possible explanation for this high P_Ca/P_Na is that NMDG⁺, our presumed inert ion, alters channel function. Although this alternative cannot be completely ruled out, we have observed comparable results for P_Ca/P_Na using tetramethylammonium (TMA) as the inert ion (Wollmuth & Sakmann, 1998). Alternatively, the high P_Ca/P_Na could reflect competition or lack thereof between Ca²⁺ and Na⁺ for the pore (and hence P_Ca/P_Na is lower when Na⁺ is present). Recent evidence has shown that NMDAR channels have sites in the pore for Na⁺ (Antonov et al. 1998; Zhu & Auerbach, 2001), and monovalent currents in NMDAR channels are blocked by Ca²⁺ (Ascher & Nowak, 1988). Nevertheless, the interaction between Ca²⁺ and Na⁺ apparently does not account for the high P_Ca/P_Na with the biionic approach (Fig. 6B).

Ca²⁺-permeable AMPAR channels. For GluR-A(Q), as well as GluR-B(Q) and GluR-B(N), all three approaches to measuring Ca²⁺ permeability yielded P_Ca/P_Na values that were concentration independent (Fig. 8B, Table 1). Similarly, the two reversal potential-based approaches yielded comparable values, though the average value for the biionic condition (~1.73) was slightly higher than that for the high monovalent condition (~1.55). The major deviation from GHK for AMPAR channels was that P_Ca/P_Na was strongly voltage dependent (Fig. 2; see also Burnashev et al. 1995). Accordingly, P_Ca/P_Na values derived from P_f measurements differed from those derived from the two reversal potential-based approaches in a voltage-dependent manner with the deviation increasing at more negative potentials. Indeed, the P_Ca/P_Na estimated to occur at 0 mV based on the voltage dependence of P_Ca/P_Na ( 1.50, Fig. 2B) was comparable to that derived from the reversal potential measurements. Thus, P_f measurements in AMPAR channels deviate more strongly from GHK at more negative potentials.

What is the mechanism underlying the voltage dependence of P_f values in AMPAR channels? This voltage dependence was also found in mutant GluR-B channels not blocked by intracellular polyamines (GluR-B(N); Fig. 2B), indicating that it is a property of the ion conduction pathway. In GluR channels, the magnitude of fractional Ca²⁺ currents is defined by the relative magnitude of I_Ca, I_Na and I_K (P_f = I_Ca/(I_Ca + I_Na + I_K)). Thus, assuming currents follow GHK around V_rev, three general possibilities exist to account for the reduced amplitude of P_f (or P_Ca/P_Na) at more negative potentials: (i) the relative rate of Ca²⁺ influx increases less rapidly than expected (that is I_Ca is less than expected, relative to monovalent cations); (ii) the relative rate of Na⁺ influx increases more rapidly than expected (that is I_Na is greater than expected, relative to Ca²⁺); and/or (iii) the relative rate of K⁺ efflux decreases more rapidly than expected (that is I_K is less than expected, relative to Ca²⁺). The third alternative seems unlikely given that the relationship between Ca²⁺ and intracellular monovalent cations (in this case Cs⁺) follows GHK over a wide voltage range (Wollmuth & Sakmann, 1998). Further, P_Ca/P_Na does not change in low extracellular Na⁺ (Fig. 6), suggesting that Na⁺ currents are not greatly different and arguing against the second alternative. Thus the first alternative seems most likely, that is Ca²⁺ influx does not increase as rapidly as Na⁺ influx with negative potentials, but additional experiments will be needed to directly test this idea.

Fractional Ca²⁺ currents in NMDAR channels were not voltage dependent, as expected from GHK. GluR-B(N) channels, which like NMDAR channels have an asparagine at the Q/R/N-site, had a significantly higher fractional Ca²⁺ current (5.2 % at -60 mV) than that in unedited GluR-B(Q) channels (~3.6 %), but P_Ca/P_Na remained strongly voltage dependent in these channels. Thus, at a molecular level, the difference in the voltage dependence between NMDAR and AMPAR channels is not due to the identity of the residue occupying the Q/R/N-site, and may reflect differences in the overall structure of the M2 loop and/or in the composition of the extracellular vestibule.

Ca²⁺-permeable KAR channels. For GluR-6(Q), all three approaches to measuring Ca²⁺ permeability yielded strongly concentration-dependent P_Ca/P_Na values (Fig. 8C). In addition, the magnitudes of these values were comparable, at least at concentrations greater than 1.8 mM Ca²⁺. The strong concentration dependence of P_Ca/P_Na accounts for the disparity between the measured Ca²⁺ permeability found in previous publications (e.g. Köhler et al. 1993; Burnashev et al. 1995) and the relatively high P_f values in KAR under physiological conditions (Fig. 2). Indeed, in previous publications Ca²⁺ permeability was typically measured under biionic conditions using high concentrations of extracellular Ca²⁺ (~110 mM), yielding a Ca²⁺ permeability (P_Ca/P_M) of approximately 0.15, comparable to that in our hands (Fig. 4). Such a Ca²⁺ permeability predicts a fractional Ca²⁺ current at -60 mV and in 1.8 mM Ca²⁺ of around 0.8 %, a value considerably less than the measured P_f value under these conditions (~2.3 %; Fig. 2).

Compared to the value in 110 mM Ca²⁺, P_Ca/P_Na in KAR channels under biionic conditions increased monotonically at lower concentrations. This pattern of concentration dependence is expected if diffuse, non-specific surface charges contribute to the process of Ca²⁺ influx. On the other hand, for NMDAR channels P_Ca/P_Na showed a maximum in 1.8 mM Ca²⁺, but decreased at the lower concentration (0.5 mM Ca²⁺), a response inconsistent with simple diffuse surface charges (see Wollmuth & Sakmann, 1998). Thus, while NMDAR and KAR channels both show concentration dependent Ca²⁺ permeability ratios under biionic conditions, the mechanism for it appears different, at least at very low concentrations of Ca²⁺.

Ca²⁺ selectivity in GluR subtypes

Consider a channel that selects neither for nor against Ca²⁺ (or monovalent ions; i.e. P_Na/P_K = P_Ca/P_M = 1). For such a non-selective channel, predicted P_f values in physiological concentrations of Ca²⁺ (1.03 mM Ca²⁺; 1.8 mM Ca²⁺ corrected for activity) using eqn (1) would be around 3.9 % at -60 mV and around 4.9 % at -20 mV. From this perspective, Ca²⁺ selectivity in GluR channels can be classified into three types: NMDAR channels are highly Ca²⁺ selective over the entire voltage range, displaying a nearly three- to fourfold higher P_f value than expected for a purely non-selective channel (see Table 1). In contrast, Ca²⁺-permeable AMPAR channels are essentially non-selective for Ca²⁺. At -60 mV, the measured P_f values in AMPAR subtypes (~3.6 %) are slightly below that for a non-selective channel whereas those at -20 mV (~7.1 %) are slightly above it. GluR-B(N) channels are weakly Ca²⁺ selective over the entire voltage range. Finally, KAR channels select against Ca²⁺ over the entire voltage range. (This presumably also holds true for 2 channels (Wollmuth et al. 2000)). Thus, understanding the molecular mechanism of Ca²⁺ influx in Ca²⁺-permeable GluR channels will require, to some extent, addressing different issues. For NMDAR channels, it will require defining those sites in the channel responsible for the high Ca²⁺ selectivity, that is for the 9.5 % fractional Ca²⁺ currents (at -60 mV), above and beyond that expected for a non-selective channel. Some of this difference may be due to differences in the structure of the M2-loop, but most appears to be due to elements within the extracellular vestibule (Premkumar & Auerbach, 1996). On the other hand, for AMPAR channels it will require defining the voltage dependence of P_f measurements. Finally, for KAR channels, it will require defining those structures of the channel that select against Ca²⁺. The molecular basis for these different Ca²⁺ permeation properties in Ca²⁺-permeable GluR channels remains unidentified.

Divalent permeability in GluR subtypes

The pattern of divalent permeability in GluR subtypes was consistent with the general idea that NMDARs are highly selective for Ca²⁺ (or divalent cations), AMPARs are non-selective, and KARs select against divalent cations (e.g. Dingledine et al. 1992; Egebjerg & Heinemann, 1993; Tsuzuki et al. 1994). Indeed, NMDAR channels showed considerable selectivity in terms of both the sequence (Ca²⁺ Ba²⁺ > Sr²⁺ >> Co²⁺ Mg²⁺) and magnitude with the highly permeable divalent cations being three to four times more permeable than monovalent cations. In contrast, non-NMDAR subtypes showed only a poor discrimination (GluR-A(Q): Ca²⁺ Ba²⁺ > Sr²⁺ Mg²⁺ > Co²⁺ and GluR-6(Q): Mg²⁺ Ca²⁺ > Sr²⁺ Ba²⁺ Co²⁺). Nevertheless, in terms of the magnitude, AMPAR channels were poorly selective, whereas KAR subtypes strongly selected against all divalent cations.

Conclusion

The comparison of P_Ca/P_Na derived from the different approaches yields three major conclusions. First, using GHK as a reference, Ca²⁺ influx in none of the three GluR subtypes followed this formalism, showing a voltage or concentration dependence and/or a strong divergence between the approaches used to measure P_Ca/P_Na. Second, the pattern of divergence for each of the subtypes was unique, suggesting that while the various subtypes may share common features in terms of their molecular and biophysical basis of Ca²⁺ influx, they all display fundamental differences. Finally, as a reversal potential-based approach to characterize Ca²⁺ flux under physiological conditions, measurement of P_Ca/P_Na using the high monovalent approach is most appropriate. Indeed, for NMDAR and KAR channels, the high monovalent but not the biionic approach quantitatively describes P_f values under physiological conditions. On the other hand, because of the strong voltage dependence of P_f in AMPAR channels, both reversal potential-based approaches over-estimate P_f.

	REFERENCES

Top Abstract Introduction Methods Results Discussion References

ANTONOV, S. M., GMIRO, V. E. & JOHNSON, J. W. (1998). Binding sites for permeant ions in the channel of NMDA receptors and their effects on channel block. Nature Neuroscience 1, 451-461	[Medline]
ASCHER, P. & NOWAK, L. (1988). The role of divalent cations in the N-methyl-D-aspartate responses of mouse central neurones in culture. Journal of Physiology 399, 247-266	[Abstract]
BLISS, T. V. P. & COLLINGRIDGE, G. L. (1993). A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361, 31-39	[Medline]
BURNASHEV, N. (1996). Calcium permeability of glutamate-gated channels in the central nervous system. Current Opinion in Neurobiology 6, 311-317	[Medline]
BURNASHEV, N., MONYER, H., SEEBURG, P. H. & SAKMANN, B. (1992). Divalent ion permeability of AMPA receptor channels is dominated by the edited form of a single subunit. Neuron 8, 189-198	[Medline]
BURNASHEV, N., VILLARROEL, A. & SAKMANN, B. (1996). Dimensions and ion selectivity of recombinant AMPA and kainate receptor channels and their dependence on Q/R site residues. Journal of Physiology 496, 165-173	[Abstract]
BURNASHEV, N., ZHOU, Z., NEHER, E. & SAKMANN, B. (1995). Fractional calcium currents through recombinant GluR channels of the NMDA, AMPA and kainate receptor subtypes. Journal of Physiology 485, 403-418	[Abstract]
CONSTANTINE-PATON, M. & CLINE, H. T. (1998). LTP and activity-dependent synaptogenesis: the more alike they are, the more different they become. Current Opinion in Neurobiology 8, 139-148	[Medline]
DINGLEDINE, R., HUME, R. I. & HEINEMANN, S. F. (1992). Structural determinants of barium permeation and rectification in non-NMDA glutamate receptor channels. Journal of Neuroscience 12, 4080-4087	[Abstract]
EGEBJERG, J. & HEINEMANN, S. F. (1993). Ca2+ permeability of unedited and edited versions of the kainate selective glutamate receptor GluR6. Proceedings of the National Academy of Sciences of the USA 90, 755-759	[Abstract]
KOH, D.-S., BURNASHEV, N. & JONAS, P. (1995). Block of native Ca2+-permeable AMPA receptors in rat brain by intracellular polyamines generates double rectification. Journal of Physiology 486, 308-312
KOH, D.-S., GEIGER, J. R., JONAS, P. & SAKMANN, B. (1995). Ca2+-permeable AMPA and NMDA receptor channels in basket cells of rat hippocampal dentate gyrus. Journal of Physiology 485, 383-402	[Abstract]
KÖHLER, M., BURNASHEV, N., SAKMANN, B. & SEEBURG, P. H. (1993). Determinants of Ca2+ permeability in both TM1 and TM2 of high-affinity kainate receptor channels: diversity by RNA editing. Neuron 10, 491-500	[Medline]
LEE, J. M., ZIPFEL, G. J. & CHOI, D. W. (1999). The changing landscape of ischaemic brain injury mechanisms. Nature 399, A7-14	[Medline]
LEWIS, C. A. (1979). Ion-concentration dependence of the reversal potential and the single channel conductance of ion channels at the frog neuromuscular junction. Journal of Physiology 286, 417-445	[Abstract]
LIU, S. Q. & CULL-CANDY, S. G. (2000). Synaptic activity at calcium-permeable AMPA receptors induces a switch in receptor subtype. Nature 405, 454-458	[Medline]
MAHANTY, N. K. & SAH, P. (1998). Calcium-permeable AMPA receptors mediate long-term potentiation in interneurons in the amygdala. Nature 394, 683-687	[Medline]
MALINOW, R., OTMAKHOV, N., BLUM, K. I. & LISMAN, J. (1994). Visualizing hippocampal synaptic function by optical detection of Ca2+ entry through N-methyl-D-aspartate channel. Proceedings of the National Academy of Sciences of the USA 91, 8170-8174	[Abstract]
MAYER, M. L. & WESTBROOK, G. L. (1987). Permeation and block of N-methyl-D-aspartic acid receptor channels by divalent cations in mouse cultured central neurones. Journal of Physiology 394, 501-527	[Abstract]
NEHER, E. (1995). The use of fura-2 for estimating Ca buffers and Ca fluxes. Neuropharmacology 34, 1423-1442	[Medline]
PELLEGRINI-GIAMPIETRO, D. E., GORTER, J. A., BENNETT, M. V. & ZUKIN, R. S. (1997). The GluR2 (GluR-B) hypothesis: Ca2+-permeable AMPA receptors in neurological disorders. Trends in Neurosciences 20, 464-470	[Medline]
PREMKUMAR, L. S. & AUERBACH, A. (1996). Identification of a high affinity divalent cation binding site near the entrance of the NMDA receptor channel. Neuron 16, 869-880	[Medline]
SCHNEGGENBURGER, R. (1996). Simultaneous measurement of Ca2+ influx and reversal potentials in recombinant N-methyl-D-aspartate receptor channels. Biophysical Journal 70, 2165-2174	[Abstract]
SCHNEGGENBURGER, R., ZHOU, Z., KONNERTH, A. & NEHER, E. (1993). Fractional contribution of calcium to the cation current through glutamate receptor channels. Neuron 11, 133-143	[Medline]
SHARMA, G. & STEVENS, C. F. (1996). Interactions between two divalent ion binding sites in N-methyl-D-aspartate receptor channels. Proceedings of the National Academy of Sciences of the USA 93, 14170-14175	[Abstract/Full Text]
TSUZUKI, K., MOCHIZUKI, S., IINO, M., MORI, H., MISHINA, M. & OZAWA, S. (1994). Ion permeation properties of the cloned mouse epsilon 2/zeta 1 NMDA receptor channel. Molecular Brain Research 26, 37-46	[Medline]
VERNINO, S., ROGERS, M., RADCLIFFE, K. A. & DANI, J. A. (1994). Quantitative measurement of calcium flux through muscle and neuronal nicotinic acetylcholine receptors. Journal of Neuroscience 14, 5514-5524	[Abstract]
VILLARROEL, A., BURNASHEV, N. & SAKMANN, B. (1995). Dimensions of the narrow portion of a recombinant NMDA receptor channel. Biophysical Journal 68, 866-875	[Abstract]
WOLLMUTH, L. P., KUNER, T., JATZKE, C., SEEBURG, P. H., HEINTZ, N. & ZUO, J. (2000). The Lurcher mutation identifies delta 2 as an AMPA/kainate receptor-like channel that is potentiated by Ca2+. Journal of Neuroscience 20, 5973-5980	[Abstract/Full Text]
WOLLMUTH, L. P., KUNER, T., SEEBURG, P. H. & SAKMANN, B. (1996). Differential contribution of the NR1- and NR2A-subunits to the selectivity filter of recombinant NMDA receptor channels. Journal of Physiology 491, 779-797	[Abstract]
WOLLMUTH, L. P. & SAKMANN, B. (1998). Different mechanisms of Ca2+ transport in NMDA and Ca2+-permeable AMPA glutamate receptor channels. Journal of General Physiology 112, 623-636	[Abstract/Full Text]
ZAREI, M. M. & DANI, J. A. (1994). Ionic permeability characteristics of the N-methyl-D-aspartate receptor channel. Journal of General Physiology 103, 231-248	[Abstract]
ZHU, Y. & AUERBACH, A. (2001). Na+ occupancy and Mg2+ block of the N-methyl-D-aspartate receptor channel. Journal of General Physiology 117, 275-286	[Abstract/Full Text]

Acknowledgements

We thank Dr A. Sobolevsky for his comments on the manuscript, L. Rooney and S. Metz for technical assistance, and Dr P. H. Seeburg for the wild-type glutamate receptor cDNAs. This work was supported by NIH RO1 grant NS39102 and a Sinsheimer Scholars Award (L.P.W.).

This article has been cited by other articles:

D. S. K. Samways, B. S. Khakh, and T. M. Egan Tunable Calcium Current through TRPV1 Receptor Channels J. Biol. Chem., November 14, 2008; 283(46): 31274 - 31278.