A synergistic blocking effect of Mg²⁺ and spermine on the inward rectifier K⁺ (Kir2.1) channel pore

Abstract

Inward rectifier K(+) channels (Kir2.1) exhibit an extraordinary rectifying feature in the current-voltage relationship. We have previously showed that the bundle-crossing region of the transmembrane domain constitutes the crucial segment responsible for the polyamine block. In this study, we demonstrated that the major blocking effect of intracellular Mg(2+) on Kir2.1 channels is also closely correlated with K(+) current flow, and the coupled movements of Mg(2+) and K(+) seem to happen in the same flux-coupling segment of the pore as polyamines. With a preponderant outward K(+) flow, intracellular Mg(2+) would also be pushed to and thus stay at the outermost site of a flux-coupling segment in the bundle-crossing region of Kir2.1 channels to block the pore, although with a much lower apparent affinity than spermine (SPM). However, in contrast to the evident possibilities of outward exit of SPM through the channel pore especially during strong membrane depolarization, intracellular Mg(2+) does not seem to traverse the Kir2.1 channel pore in any case. Intracellular Mg(2+) and SPM therefore may have a synergistic action on the pore-blocking effect, presumably via prohibition of the outward exit of the higher-affinity blocking SPM by the lower-affinity Mg(2+).

Figure 1. The affinity of intracellular Mg²⁺ for WT Kir2.1 channel in symmetrical 100 mM K⁺.

(a)Sample sweeps demonstrating the inhibition of WT Kir2.1 currents in the same patch by intracellular Mg²⁺ in symmetrical 100 mM K⁺. The membrane voltage was repolarized from a holding potential of 0 mV to −100 mV for 20 ms, and then stepped to test pulses between −150 and +100 mV in 10 mV increment for 150 ms. The dashed lines indicate the zero current level. (b) The relative current is defined by the ratio between the steady–state currents in Mg²⁺ and in control at each designated membrane potential, and are plotted against Mg²⁺ concentration (n = 5–8). The data obtained at different membrane voltages (from −60 to +60 mV) are fitted by Eq. 1 (see Methods). (c) The Kd from part b is plotted against membrane voltage on a semi–logarithmic scale. The Kd–voltage relationship is clearly nonlinear, with the most evident change around 0 mV (the reversal potential of K⁺ ion; E_K⁺). The IR index (defined as the ratio between the Kd values at −30 and +30 mV13) is ~95 in this case.

Figure 2. Inhibition of WT Kir2.1 currents by intracellular Mg²⁺ in symmetrical 20 mM K⁺, 300 mM K⁺, and external 20/ internal 100 mM K⁺.

(a)The Kd of the intracellular Mg²⁺ block in symmetrical 20 and 300 mM K⁺ is plotted against voltage on a semi–logarithmic scale (data were obtained with the same approach as that in Fig. 1b). The Kd–voltage relationship of Mg²⁺ block in symmetrical 100 mM K⁺ (solid line) is taken from Fig. 1c for comparison. (b) The Kd–voltage relationship of Mg²⁺ block in external 20/internal 100 mM K⁺ is apparently similar to that in symmetrical 100 mM K⁺, but is shifted only by ~−20 mV in the voltage axis (data were obtained with the same approach as that shown in Fig. 1c).

Figure 3. The emergence of a slow component of intracellular Mg²⁺ unblock in mutant E224 and E299 channels.

(a) The Kd of intracellular Mg²⁺ block in E224G, E224Q, and E299S mutant channels is plotted against voltage on a semi–logarithmic scale (the same approach as that shown in Fig. 1c). The changes in Kd of intracellular Mg²⁺ block around 0 mV are evidently much smaller with specific of E224G, E224Q, and E299S mutations. (b) The patches which expressed E299S, E224G, and E224A mutant channels were first held from 0 mV and stepped twice to −100 mV. The sample sweeps show a “slow tail” when the patch was repolarized from a positive voltage of +100 mV to a negative voltage of −100 mV in the presence but not in the absence of intracellular Mg²⁺, signaling Mg²⁺ block from the channel. The development of the slow tail in the second −100 mV pulse thus is used as a measure of intracellular Mg²⁺ binding to this slow unblocking site at +100 mV (dashed lines, see also part c). The dashed lines indicate the zero current level. Note that when the currents are turning from outward to inward, there is a very fast phase of development of inward current before the slow tail phase (insert figure). (c) Cumulative results were obtained with the same protocol described in part b for the E299S, E224G, and E224A mutant channels (n = 4–7). tau _{on, 1} and tau _{on, 2} denote the time constants of the growing courses of slow tails and the time constants of the decay of outward currents in the presence of 100 μM intracellular Mg²⁺ in part b respectively. (d) The reciprocals of the time constants of development of slow tail currents (on rate at +100 mV) are plotted against the decay time constant of the slow tail (off rate at −100 mV) for each of the E224 mutant (E224A, E224C, and E224G) and E299 mutant (E299A, E299S, and E299C) channels. The lines are linear regressions demonstrating the strong linear correlation between the two parameters.

Figure 4. The blocking and unblocking kinetics of internal Mg²⁺ in E224G mutant channels.

(a) The inverses of the time constants for the development of the slow tail in Fig. 3c are plotted against the Mg²⁺ concentration (the depolarization pulses were set between +70 and +110 mV). The lines are linear fits with slope and Y–intercept of 1.8 × 10⁶ M⁻¹s⁻¹ and 29 s⁻¹ (+70 mV); 2 × 10⁶ M⁻¹s⁻¹ and 35 s⁻¹ (+80 mV); 2.1 × 10⁶M⁻¹s⁻¹ and 49 s⁻¹ (+90 mV); 2.3 × 10⁶M⁻¹s⁻¹and 52 s⁻¹ (+100 mV); and 2.3 × 10⁶M⁻¹s⁻¹ and 63 s⁻¹ (+110 mV), respectively. (b) The slopes in part a are plotted against voltage and fitted with the equation: k_on = 1.2 × 10⁶ exp (0.15V/25 mV) M⁻¹s⁻¹, where V is the membrane voltage potential in mV. The regression line indicates an equivalent electrical distance (δ) of ~0.08 for the Mg²⁺ blocking rate (assuming +2 effective charges on Mg²⁺).(c) The inverses of the relaxation time constants of the slow tail currents are plotted against membrane potential on a semi–logarithmic scale for the E224G mutant channels (n = 5). The second negative pulse in the same two–pulse protocol was set between −20 and −100 mV to examine the voltage dependence of Mg²⁺ unblock from the blocking site. The data are fitted with the equation: off rate _(V) = off rate ₍₀₎ × exp(–ZδV/25 mV), where V is the membrane potential in mV, and Z and δ are the charges on the blocker and the equivalent electric distance of the blocking site from inside, respectively. The off rate₍₀₎ and Zδ are ~280 s⁻¹ and ~0.67 for the E224G mutant channels.

Figure 5. Similar voltage dependence of the intracellular Mg²⁺ unblocking kinetics in the WT, E224G, and M183W mutant channels.

(a) The two–pulse protocol is basically similar to that in Fig. 3b, except that the voltage across the patch membrane was first depolarized from the holding potential −100 mV to +100mV for ~50 ms, and then stepped to different test voltages between −100 and 0 mV for ~300 ms in 10 mV increment for M183W single, and E224G/M183W double mutant channels. Representative current traces were recorded in control or in internal 100 μM Mg²⁺. The slow tail is present only in the presence of intracellular Mg²⁺ at appropriate membrane potentials. (b) The reverses of the relaxation time constants of the slow tail currents are plotted against voltage in semi–logarithmic scale for the WT, and single as well as corresponding double mutant channels. The negative pulse in the same protocol was set to −20 ~ −100 mV to examine the voltage dependence of Mg²⁺ unblock from the binding site. The E224G mutant channel data are taken from Fig. 4c for comparison (the blue line). The off rate₍₀₎ and Zδ are ~4300 s⁻¹and 0.48 for the WT channel, ~370 s⁻¹ and 0.56 for the E224G, ~2800 s⁻¹ and 0.46 for the M183W, ~230 s⁻¹ and 0.62 for the E224G/M183W mutant channels, respectively.

Figure 6. Elimination of the extremely slow component of Mg²⁺ unblock in E224Q mutant channel by concomitant specific mutations at M183 and A178.

(a) Sample sweeps of the extremely slow tail currents in the E224Q mutant channel in symmetrical 100 mM K⁺ (see Fig. 3c for the pulse protocol). The slow tails are considerably slower than that in the E224G mutant channel. Also, much higher concentrations of Mg²⁺ are required for the emergence of slow tails. In addition, the extremely slow tails signaling Mg²⁺ unblocking in the E224Q single mutant channel are evidently accelerated in specific double–mutant (e.g., E224Q/M183W, E224Q/M183N, and E224Q/A178C) channels. The dashed lines indicate the zero current level. (b) Cumulative results were obtained for the E224Q, E224Q/M183W, E224Q/M183N, and E224Q/A178C double mutant channels (each n = 5). tau _{on, 1} and tau _{on, 2} denote the time constant of the growing course of the slow tails (at −100 mV) and the time constant of the decay of the macroscopic outward currents (at +100 mV) in 3 mM Mg²⁺ from the experiments in part a respectively (the same approach as that shown in Fig. 3c).

Figure 7. Negative cooperative effects of E224G and the TM2 bundle–crossing region mutations on the apparent affinity of Mg²⁺.

(a) Comparison of the ratio between the apparent Kd of intracellular Mg²⁺ at −30 and +30 mV among the WT and mutant channels. (b) Double–mutant cycle analysis of the E224G+G177N and E224G+M183N single– and double–mutant channels at +40 mV. The Kd of Mg²⁺ block is derived from the same approaches as that described in Fig. 1c. The coupling coefficients (Ω) are defined as the ratio between the product of the Kd changes in each single mutant channel (versus the WT channel) and the Kd change in the double–mutant channel, and are ~0.08 and ~0.14 for the E224G/G177N and the E224G/M183N double mutations, respectively. (c) The coupling coefficient is plotted for different pairs of double mutations. The dashed line denotes a coupling coefficient of 1.0, which signals a purely additive effect of the paired mutations. Double mutations involving E224 (E224G) and most of the residues between I176 and M183 in the bundle–crossing region of TM2 show a prominent negative cooperative effect, except for the E224G/A181T double mutant channels. E224G/D172N and E224G/A184Q double mutations involving E224 and presumably either the outer or the inner boundaries of the central cavity, also have coupling coefficients close to ~1.0.

Figure 8. Inhibition of outward currents in the WT Kir2.1 channel by internal 1 mM Mg²⁺ and 10 μM SPM.

(a) Sample sweeps demonstrating inhibition of WT Kir2.1 currents in the same patch by internal 1 mM Mg²⁺, 10 μM SPM, and concomitant 1 mM Mg²⁺ and 10 μM SPM in symmetrical 100 mM K⁺ (see Fig. 1a for the pulse protocol). The dotted lines indicate the zero current level. (b) The experiments were performed in symmetrical 100 mM extracellular K⁺ for the WT Kir2.1 channel. The relative currents are defined in the same way as that in Fig. 1. Note the stronger inhibitory effect (smaller residual currents) with 10 μM SPM than 1 mM Mg²⁺, and the even stronger effect (even smaller residual currents) with the concomitant presence of both 1 mM Mg²⁺ and 10 μM SPM. * and **p < 0.05 and <0.005, respectively, by student’s independent t–test. (c) The experiments were performed in external 20/ internal 100 mM K⁺ conditions for the WT Kir2.1 channel. The relative currents are defined in the same way as that in Fig. 1. The protocols were the same as that in Fig. 1a. The blocking effect of 1 mM Mg²⁺ is gradually increased, whereas the blocking effect of 10 μM SPM is gradually decreased, from +20 to +100 mV.

Figure 9. Inhibition of outward currents in the WT Kir2.1 channel by internal 0.1 μM Mg²⁺ and 0.01 μM SPM.

(a) Sample sweeps demonstrating the inhibition of WT Kir2.1 currents in the same patch by intracellular 0.1 μM Mg²⁺, 0.01 μM SPM, and concomitant 0.1 μM Mg²⁺ and 0.01 μM SPM in symmetrical 100 mM K⁺ (the same pulse protocol as that in Fig. 1a). The dashed lines indicate the zero current level. (b) The experiment was conducted in symmetrical 100 mM extracellular K⁺ for the WT Kir2.1 channel. Note the stronger blocking effect in the concomitant presence of both 0.1 μM Mg²⁺ and 0.01 μM SPM (e.g., +20, +40, and +60 mV) than either blocker. * and **p < 0.05 and <0.005, respectively, by student’s independent t–test. (c) Cumulative results were obtained from experiments with the same protocols described in part a (n = 7). Note that the macroscopic binding rates of concomitant intracellular 0.1 μM Mg²⁺ and 0.01 μM SPM are roughly the sum of the binding rates of each blocker alone. * and **p < 0.05 and <0.005, respectively, by student’s independent t–test.

Figure 10. Molecular dynamics simulation of coexistence Mg²⁺ and SPM action in the WT Kir2.1 channel pore.

(a) In the presence of approximately forty water molecules, one Mg²⁺, one SPM molecules, and seven K⁺ ions are applied to the central cavity region of the WT Kir2.1 channel. The Mg²⁺ (colored are yellow) and SPM molecules (hydrogen atoms are colored grayish white; carbon, black; and nitrogen, blue) of the blocking sites are presented in a CPK model. The representative structure of the cytoplasmic and transmembrane domains of one subunit of the WT Kir2.1 channel is shown. A regional view of the residues T141, S165, and D172 is shown in the boxed panel. The side chains of T141, S165, and D172 are shown in CPK model, and the hydrogen atoms are colored grayish white; nitrogen, blue; oxygen, red; and carbon, black. All of the other atoms in the channel protein are colored green. (b) With the same approaches as that in part a two subunits of the channel are presented in a solid–ribbon model. The distance between diagonal residues D172 from tip to tip is profoundly decreased (~9.7 Å) in the coexistence presence of Mg²⁺ and SPM, but not in presence of Mg²⁺ (~17.3 Å) or SPM alone (~15.5 Å). Note that Mg²⁺ is located close to S165 whereas SPM is located close to D172. (c) A summary plot of the tip–to–tip diagonal distance between sites D172 in control, and in the presence of SPM, or Mg²⁺, or both, in the WT Kir2.1 channel pore. Note that an evident narrowing occurs only in the concomitant presence of Mg²⁺ and SPM. (d) A summary plot of the diagonal (tip–to–tip) distances between residues T141, S165, and D172 in control (no blockers) and in the concomitant presence of Mg²⁺ and SPM in the Kir2.1 channel pore. Note that an evident narrowing occurs only at the level of D172, but not at T141 and S165.

Figure 11. A Schematic model illustrating the mechanism underlying Mg²⁺ block of the Kir2.1 channel pore.

(a) The WT Kir2.1 channel comprises cytoplasmic and transmembrane domains. The ion conduction pathway in the transmembrane domain could be further divided into the selectivity filter and the central cavity. SPM probably binds to a blocking site involving D172 with a curly form. Moreover, SPM may go through the selectivity filter of the Kir2.1 channel, entering the extracellular side of the channel if there is a large driving force. (b) With strong outward K⁺ currents, intracellular Mg²⁺ is pushed to the outermost site of the flux–coupling region (probably involving S165) in the Kir2.1 channel pore. Mg²⁺ is more weakly bound to its site than SPM. However, unlike SPM, intracellular Mg²⁺ could not traverse the selectivity filter of Kir2.1 channel pore and exit to the outside even with a large driving force. (c) We illustrate that concomitant intracellular Mg²⁺ and SPM are synergistic in the blocking action of the Kir2.1 channel pore. SPM has a strong flow–dependent movement in the central cavity, and thus a strong blocking effect on the outward K⁺ currents when SPM gets “stuck” at the junction between central cavity and selectivity filter. SPM, however, can not completely block the outward K⁺ currents even with highly preponderant outward K⁺ flux (e.g., very strong depolarization) because of the possibility of outward exit of the blocking SPM through the selectivity filter. On the other hand, Mg²⁺ has a much weaker flow–dependent movement in the central cavity, and thus millimolar Mg²⁺ has an even weaker overall blocking effect on the outward K⁺ currents than micromolar SPM. Concomitant Mg²⁺ and SPM, however, would have a synergistic rather than additive blocking effect on the outward K⁺ current especially at strong depolarization. This is because the apparently weaker blocker Mg²⁺ could prohibit the outward exit of SPM and thus effectively reduce the residual outward K⁺ currents, either by direct physical hindrance or by the possible narrowing conformational changes at ~D172 residues induced by coexisting Mg²⁺ and SPM (Fig. 10), or by both.