Potassium-selective block of barium permeation through single KcsA channels

Abstract

Ba(2+), a doubly charged analogue of K(+), specifically blocks K(+) channels by virtue of electrostatic stabilization in the permeation pathway. Ba(2+) block is used here as a tool to determine the equilibrium binding affinity for various monovalent cations at specific sites in the selectivity filter of a noninactivating mutant of KcsA. At high concentrations of external K(+), the block-time distribution is double exponential, marking at least two Ba(2+) sites in the selectivity filter, in accord with a Ba(2+)-containing crystal structure of KcsA. By analyzing block as a function of extracellular K(+), we determined the equilibrium dissociation constant of K(+) and of other monovalent cations at an extracellular site, presumably S1, to arrive at a selectivity sequence for binding at this site: Rb(+) (3 µM) > Cs(+) (23 µM) > K(+) (29 µM) > NH(4)(+) (440 µM) >> Na(+) and Li(+) (>1 M). This represents an unusually high selectivity for K(+) over Na(+), with |ΔΔG(0)| of at least 7 kcal mol(-1). These results fit well with other kinetic measurements of selectivity as well as with the many crystal structures of KcsA in various ionic conditions.

Figure 1.

K⁺-binding sites in KcsA selectivity filter. Small purple spheres mark locations of sites where K⁺ ions appear in crystal structures of the KcsA selectivity filter, shown as stick models of T74-Y78 of two opposing subunits of the tetramer. Sites are numbered S1–S4 from external to internal sides of the filter (deposited in the Protein Data Bank under accession no. 1K4C).

Figure 2.

Personality quirks in KcsA-E71A. Illustrative recordings are shown for three clearly distinguishable single-channel personalities observed here, with names indicated. Dashed lines mark zero-current level, here and throughout.

Figure 3.

Ba²⁺ block of KcsA-E71. Single-channel recordings were collected at 50 mV with 200 mM of internal K⁺/20 mM of internal K⁺, with or without Ba²⁺ as indicated. (A) Ba²⁺ blocks from either side of the membrane. Nonconducting dwell-time histograms (log-binned) are shown in each condition, along with exponential fits. Time constant for zero Ba²⁺ is 0.4 ms; time constants of triple-exponential fits for internal Ba²⁺ are 0.9, 6, and 370 ms, and for external Ba²⁺ are 0.3, 5, and 350 ms, respectively. (B and C) Bimolecular behavior of Ba²⁺ block. Representative recordings at two concentrations of internal Ba²⁺ are shown along with plots of inverse time constants for unblocked times (1/τ_o) and fast and slow block times. Second-order rate constant for Ba²⁺ association derived from slope of unblocked-time data is 2.9 × 10⁵ M⁻¹s⁻¹.

Figure 4.

Two-site sequential blocking scheme. The three pertinent states in our kinetic blocking scheme are shown, along with cartoons illustrating Ba²⁺ (purple circles) in positions observed in KcsA and K⁺ sites as small black circles (Protein Data Bank accession no. 2ITD). “O” represents the unblocked conducting channel, and B₁ and B₂ are the two blocked states postulated from the two Ba²⁺-induced components of the nonconducting dwell-time distributions. The scheme refers to internal Ba²⁺conditions, with z representing irreversible Ba²⁺ exit to the Ba²⁺-free external side.

Figure 5.

Suppression of Ba²⁺ escape by external K⁺. (A) Recordings are shown in the presence of 60 µM of internal Ba²⁺ at zero or 5 mM of external K⁺. The presence of external K⁺ lengthens block times ∼30-fold. (B) Nonconducting time distribution at 5 mM K⁺. Solid curve represents double-exponential block-time distribution fit by: a_f and a_s = 75 and 1.5 s⁻¹ ; τ_f and τ_s = 6 and 365 ms (time constants for Ba²⁺-induced blocks are marked by arrows). Dashed curve represents block-time distribution predicted by rate constants derived from Bayesian analysis, using maximum entropy prior.

Figure 6.

Zero-K⁺ block-time distribution with rising phase. (A) Nonconducting dwell-time histogram is shown for a typical dataset with zero external K⁺. This histogram is linearly binned to reveal the rising phase of the Ba²⁺-induced part of the distribution (solid curve, after removal of the Ba²⁺-independent gating). Parameters of the block-time distribution are: a_f and a_s = −41 and +78 s⁻¹ ; τ_f and τ_s = 4 and 15 ms (marked by arrows). Dotted curve is predicted from the mean values of rate constants derived from Bayesian analysis using a maximum entropy prior, as in Appendix A (b, c, d, and z = 41, 88, 42, and 132 s⁻¹). Dashed curve is predicted by Bayesian analysis using the prior distribution for z shown in Fig. 7 B (b, c, d, and z = 46, 29, 7, and 204 s⁻¹). (B) Distributions of rate constants derived by Bayesian analysis of zero-K⁺ dwell times. Parameters from all data are reported in Table I.

Figure 7.

Block by external Ba²⁺ at zero external K⁺. (A) Block-time histogram with zero external K⁺ and 1 mM of external Ba²⁺. Solid curve is a triple-exponential fit with a 1-ms component representing heavily filtered spontaneous closings, an 80-ms component (gray arrow) representing rare inactivation events (see Materials and methods), and only a single component for blocks (black arrow), with τ_f constrained by internal Ba²⁺ data to a value of 5.8 ms. (B) Distributions of the rate constants derived from Bayesian analysis of the internal Ba²⁺ data of Fig. 6, now combined with external Ba²⁺ block data. Parameters are reported in Table I.

Figure 8.

Block of Ba²⁺ permeation by K⁺. Ba²⁺ block datasets were recorded as in Fig. 5, at different concentrations of external K⁺. (A) Nonconducting time histograms are shown (log-binned except for 10 µM K⁺ data, which is binned linearly to show its rising phase). (B) Behavior of fast/slow time constants (top) and amplitudes (bottom) of block-time distributions. Curves have no theoretical meaning, except for the 1/τ_s curve, which plots Eq. 7. Gray shading on the amplitude plot indicates the range of K⁺ concentrations where the fast amplitude, in reversing sign, becomes too small to observe and thereby produces a single-exponential block-time distribution.

Figure 9.

Ion selectivity of external Ba²⁺-trapping site. Suppression of Ba²⁺ escape by external monovalent cations was performed as in Fig. 8 for K⁺. (A) Representative recordings of Ba²⁺ block with external addition of the indicated cations. (B) Titration of inverse slow time constant of block for all cations tested, with data fit by single-site inhibition curves (Eq. 7) constrained to the same zero-K⁺ value (solid curves), with half-inhibition constants, in order of affinity: Rb⁺ (3 µM), Cs⁺ (23 µM), K⁺ (29 µM), and NH₄⁺ (440 µM).

Figure B1.

Single KcsA-E71A channels were recorded as in the main text, with 200 mM K⁺ in/10 mM K⁺ and 190 mM Na⁺ out. 10 µM Ba²⁺ was also included in the internal solution to produce block events. Immediately after recording a single-channel I-V curve, 25 µM valinomycin was added to the bilayer to raise the conductance ∼100-fold, and the macroscopic I-V curve was measured within a minute. (Top) Illustrative records (filtered at 50 Hz) at voltages on either side of the reversal potential before the addition of valinomycin. Dashed line marks zero-current level. (Bottom) I-V curves for open channel (squares), blocked channel (circles), and valinomycin (diamonds). Curves (dashed for channel, solid for valinomycin) are third-order polynomial fits to the data points. Reversal potential is the voltage at which the open-channel or valinomycin curve crosses the blocked-channel curve. Right panel is a blown-up view of left panel.