Estimating the pKa values of basic and acidic side chains in ion channels using electrophysiological recordings: a robust approach to an elusive problem

Abstract

As a step toward gaining a better understanding of the physicochemical bases of pK(a)-value shifts in ion channels, we have previously proposed a method for estimating the proton affinities of systematically engineered ionizable side chains from the kinetic analysis of single-channel current recordings. We reported that the open-channel current flowing through mutants of the (cation-selective) muscle nicotinic acetylcholine receptor (AChR) engineered to bear single basic residues in the transmembrane portion of the pore domain fluctuates between two levels of conductance. Our observations were consistent with the idea that these fluctuations track directly the alternate protonation-deprotonation of basic side chains: protonation of the introduced basic group would attenuate the single-channel conductance, whereas its deprotonation would restore the wild-type-like level. Thus, analysis of the kinetics of these transitions was interpreted to yield the pK(a) values of the substituted side chains. However, other mechanisms can be postulated that would also be consistent with some of our findings but according to which the kinetic analysis of the fluctuations would not yield true pK(a)s. Such mechanisms include the pH-dependent interconversion between two conformations of the channel that, while both ion permeable, would support different cation-conduction rates. In this article, we present experimental evidence for the notion that the fluctuations of the open-channel current observed for the muscle AChR result from the electrostatic interaction between fixed charges and the passing cations rather than from a change in conformation. Hence, we conclude that bona fide pK(a) values can be obtained from single-channel recordings.

Figure 1.. The AChR channel.

(A) Membrane-threading pattern common to all Cys-loop receptors, a superfamily of pentameric ligand-gated ion channels that, in vertebrates, includes receptors to acetylcholine, serotonin, γ-aminobutyric acid and glycine. (B) Ribbon representation of the transmembrane portion of the PDB file 2BG9 (4-Å resolution; ref. 22), a model of the AChR from the electric organ of Torpedo marmorata (a muscle-type AChR) in the absence of activating ligands as viewed from the extracellular side. The color code is the same as in (A). In adult muscle, the γ subunit is replaced by the ε subunit; thus, the stoichiometry of the adult-type AChR is (α1)₂β1δε. The molecular image was made with VMD (ref. 23).

Figure 2.. pH-dependent fluctuations of the open-channel current.

Upon engineering single basic residues in the pore lining of the (cation-selective) AChR, the open-channel current fluctuates between two levels of different conductance: a wild-type-like level (the “main level”) and a level of lower conductance (the “sublevel”). (A) Effect of pH. Single-channel inward currents recorded from the δA266K AChR mutant using pipette solutions of different pH, at ~− 100 mV (negative inside the cell). One representative burst of channel openings is shown at each pH. Residue δA266 is in the M2 segment of the δ subunit, approximately halfway across the membrane width. As the concentration of protons increases so does the probability of the open-channel current signal dwelling in the sublevel. The pK_a of a lysine substituted at this position (which, using the prime-numbering system, is referred to as position 10′) was estimated to be 7.53 ± 0.01 (n = 8), at pH 7.4 (ref. 13). (B) Effect of type of basic side chain. Inward currents recorded at ~−100 mV; one representative burst of openings is shown for each mutant. The probability of the open-channel current signal dwelling in the sublevel is highest for the arginine-substituted, and lowest for the histidine-substituted, δS268 mutant AChR. Residue 268 of the δ subunit corresponds to position 12′ of M2. At pH 7.4, fluctuations of the open-channel current signal between the main level and a sublevel could only be detected (and hence, a pK_a could only be estimated) for the lysine mutant; the pK_a was found to be 8.87 ± 0.01 (n = 4) (ref. 13). All current traces are displayed at a cut-off frequency of ~6 kHz.

Figure 3.. Two possible mechanisms.

Different mechanisms can account for pH-dependent fluctuations of the open-channel current. (A) The channel can adopt ion-conductive (“open”) and non-conductive (“shut”) conformations, which, in turn, can have the substituted ionizable residue in a protonated or deprotonated state. The alternate protonation and deprotonation of the engineered titratable group while the channel is open (see gray-shaded boxes) may manifest as fluctuations of the open-channel current; proton-transfer events occurring while the channel is shut would go undetected. The kinetic scheme in (A) illustrates this mechanism for the particular case of substituted basic side chains (hence, the positive charge on the protonated forms) but the same type of mechanism applies to acidic side chains. (B) While open, the channel can adopt a high-conductance or a low-conductance conformation (denoted as “Open_H” and “Open_L”, respectively), which, in turn, can have the substituted ionizable residue in a protonated or deprotonated state. The alternate occupancy of either conformation, irrespective of the protonation state of the substituted side chain (see gray-shaded boxes), would give rise to fluctuations in the open-channel current. If the pK_a values of the side chain in question differed in these two conformations, then the probability of the channel occupying either conformation would be a function of pH. For example, if the pK_a were higher in the low-conductance conformation, then the probability of the channel adopting this conformation would increase as the pH decreases. The equilibrium constant corresponding to the interconversion between the two conformations of different conductance when the substituted side chain is protonated can be expressed in terms of the other three equilibrium constants around the cycle. For clarity, neither the proton donors/acceptors nor the shut states are shown.

Figure 4.. Different mechanisms, similar pH dependences.

Whether reflecting the fluctuating protonation state of a pore’s ionizable side chain or the alternative conformations of the channel’s pore domain, the interconversion between different levels of conductance may be sensitive to pH. In blue, the plot shows the ratio of deprotonated versus protonated forms of a hypothetical side chain whose alternate charging and uncharging leads to two different conductance levels according to the “electrostatic-block mechanism” shown in Figure 3(A); in this example, pK_a = 6.0. In red, the plot shows the ratio of high-conductance versus low-conductance conformations of a hypothetical channel that can switch between these forms in a pH-dependent manner according to the “conformational-change mechanism” in Figure 3(B); in this example, pK_{a, H} = 6.0, pK_{a, L} = 10.0 and K_H→L = 0.1. Note that, regardless of the underlying mechanism, a pH range exists within which the computed ratio is predicted to vary with pH in a remarkably similar fashion.

Figure 5.. Experimental evidence for the electrostatic-block mechanism.

(A) Extent of channel block. The extent to which the single-channel conductance is attenuated in the sublevel relative to the main level as a result of engineering single lysines in the AChR pore domain is plotted as a function of the mutated position along M1 (open red circles), M2 (including the M1–M2 linker and part of the M2–M3 linker; filled gray circles) and M3 (including the rest of the M2–M3 linker; filled black circles). Most of the plotted extent-of-block values correspond to the effect of substitutions in the δ subunit. Positions in and flanking M2 are denoted using the prime-numbering system according to which, in the δ subunit, proline 250 corresponds to position −7′ whereas threonine 281 corresponds to 25′ (we omit the −3′ label in the δ subunit). Positions in and flanking M1 or M3 are denoted using the δ-subunit residue numbers. The unbroken line joining the gray symbols is a cubic-spline interpolation. Proposed membrane boundaries and the relative depths of the three helices in the membrane are tentative. For most positions, the horizontal error bars (standard errors) are smaller than the symbols. The lumen of the pore would be to the right of the plot; see also ref. . (B) Single-channel inward currents recorded from the δS262E AChR mutant using pipette solutions of different pH, at ~−100 mV. This construct also included the T264P mutation in the ε subunit (see Methods). As the concentration of protons increases so does the probability of the open-channel current signal dwelling in the (wild-type-like) main level. The high-conductance level (154 ± 2 pS; main-level conductance = 114 ± 1 pS) is denoted as a “superlevel”. Residue 262 of the δ subunit corresponds to position 6′ of M2. The pK_a of a lysine substituted at this position could not be estimated because the open-channel current signal appeared to remain fixed at the low-conductance level even at pH 9.0. Thus, in order to estimate the pK_a of a basic side chain at this position, we engineered a histidine; its pK_a was estimated to be 7.30 ± 0.01 (n = 3), at pH 7.4 (ref. 13). (C) Single-channel inward currents recorded from the δV269E AChR mutant using pipette solutions of different pH, at ~−100 mV. The conductance of the superlevel was found to be 166 ± 1 pS (main-level conductance = 127 ± 1 pS). Residue 269 of the δ subunit corresponds to position 13′ of M2. As was the case for position 6′, the pK_a of a lysine substituted at position 13′ could not be estimated even at pH 9.0. The pK_a of a histidine at this position was estimated to be 6.68 ± 0.03 (n = 3), at pH 7.4 (ref. 13). (D) Single-channel inward currents recorded from the α1F225K AChR mutant using pipette solutions of different pH, at ~−100 mV. The two α1 subunits of the channel contained the substituted lysine. Note that the presence of two lysines (in this case, in M1) gives rise to two conductance sublevels. For panels (B), (C) and (D), one representative burst of channel openings is shown at each pH (display cut-off frequency ≈ 6 kHz). The larger amplitude of the main-level current through acidic residue-substituted AChRs [this figure, panels (B) and (C)] compared to that through the main level of basic residue-substituted channels [this figure, panel (D), and Fig. 2] is simply due to the different pipette solutions employed in each case. Currents through acidic residue-substituted AChRs were recorded in the nominal absence of Ca²⁺ or Mg²⁺ in the pipette solution (see Methods). Even at low millimolar concentrations, these divalent cations reduce the amplitude of single-channel channel currents through the AChR.