Tuning of CFTR chloride channel function by location of positive charges within the pore

Abstract

High unitary Cl(-) conductance in the cystic fibrosis transmembrane conductance regulator Cl(-) channel requires a functionally unique, positively charged lysine residue (K95) in the inner vestibule of the channel pore. Here we used a mutagenic approach to investigate the ability of other sites in the pore to host this important positive charge. The loss of conductance observed in the K95Q mutation was >50% rescued by substituting a lysine for each of five different pore-lining amino acids, suggesting that the exact location of the fixed positive charge is not crucial to support high conductance. Moving the positive charge also restored open-channel blocker interactions that are lost in K95Q. Introducing a second positive charge in addition to that at K95 did not increase conductance at any site, but did result in a striking increase in the strength of block by divalent Pt(NO(2))(4)(2-) ions. Based on the site dependence of these effects, we propose that although the exact location of the positive charge is not crucial for normal pore properties, transplanting this charge to other sites results in a diminution of its effectiveness that appears to depend on its location along the axis of the pore.

Figure 1

Simple cartoon alignment of TM1 and TM6 in the CFTR pore. The amino-acid side chains in TM1 and TM6 that line the lumen of the inner vestibule were identified by substituted cysteine accessibility mutagenesis (11,13). The two TMs were aligned on the basis of functional and disulfide cross-linking analysis, as described by Wang et al. (13). In this study, the endogenous positive charge of K95 (TM1) was transplanted to Q98 (TM1) and to I344, V345, M348, and A349 (TM6); in previous work we also investigated the effect of moving this positive charge to S341 (TM6) (8).

Figure 2

Single-channel conductance is restored by moving the positive charge from K95. (A) Example single-channel currents carried by the indicated CFTR variants at a membrane potential of −50 mV. The line to the left represents the closed channel level. Note that the reduced single-channel current amplitude of K95Q is rescued by concurrent mutagenesis of other amino acids (Q98, I344, V345, M348, and A349) to lysine. (B) Mean single-channel i/V relationships for different CFTR variants under these conditions. The leftmost panel shows WT (green) and K95Q (red). The other five panels show the effects of the indicated lysine-introducing mutations in either a WT (open green symbols) or K95Q (open red symbols) background. In each case, the lines fitted to WT and K95Q data are indicated in green and red, respectively, as a reference. (C) Mean unitary slope conductance measured from individual patches. ^∗Significant difference from WT (p < 0.05); ^#significant difference from K95Q (p < 10⁻¹⁰); ^†significant difference from the same lysine mutation in a K95Q background (p < 0.05). (D) Mean effect on conductance of removing the positive charge associated with K95 in different backgrounds, quantified as the conductance with a glutamine at this position relative to that with a lysine. In each case, the effect of the K95Q mutation was significantly reduced by the presence of an introduced lysine at other positions (^∗p < 10⁻¹⁰ compared with WT). Mean of data from six to 12 patches in C and D.

Figure 3

Open-channel block by NPPB is restored by moving the positive charge from K95. (A) Example leak-subtracted macroscopic I/V relationships for the four indicated CFTR variants after maximization of channel open probability by treatment with PPi (2 mM). In each case, currents were recorded before (control) and after the addition of 50 μM NPPB to the intracellular (bath) solution. (B) Mean fraction of control current remaining after the addition of NPPB at a membrane potential of −100 mV. ^∗Significant difference from WT (p < 0.005); ^#significant difference from K95Q (p < 0.002). Mean of data from four to five patches.

Figure 4

Additional fixed positive charges strengthen block by cytoplasmic Pt(NO₂)₄²⁻ ions. (A) Example leak-subtracted macroscopic I/V relationships for the three indicated CFTR variants. In each case, currents were recorded before (control) and after the addition of 30 μM Pt(NO₂)₄²⁻ to the intracellular (bath) solution. (B and C) Mean concentration-inhibition relationships for the indicated CFTR variants at a membrane potential of −100 mV. Each was fitted according to Eq. 1. In C, the dotted line is the fit to WT data shown in B. (D) Mean K_d for Pt(NO₂)₄²⁻ as a function of voltage for each channel variant, estimated by fitting data from individual patches as shown in B and C. Symbols represent the same channel variants as in B and C. Fitted line is to Eq. 2. (E and F) Mean K_d at 0 mV membrane potential (E) and apparent blocker valence (zδ) (F), estimated by fitting data from individual patches with Eq. 2. ^∗Significant difference from WT (p < 10⁻⁶(E) and p < 0.05 (F)). Mean of data from five to nine patches.

Figure 5

Block by cytoplasmic Pt(NO₂)₄²⁻ ions is modulated by toggling the positive charge in the pore. (A) Example leak-subtracted macroscopic I/V relationships for V345H at intracellular pH of 5.5 (left) or 9.0 (right, different patch). In both cases, currents were recorded before (control) and after the addition of 100 μM Pt(NO₂)₄²⁻ to the intracellular (bath) solution. (B) Mean concentration-inhibition relationships for WT (circles), I344H (squares), and V345H (triangles) at intracellular pH 5.5 (left, solid symbols) or 9.0 (right, open symbols) at a membrane potential of −100 mV. Each relationship was fitted according to Eq. 1. Note that three overlapping data sets are shown at pH 9.0. (C) Mean K_d for Pt(NO₂)₄²⁻ as a function of voltage for each condition in B, estimated by fitting data from individual patches. Symbols represent the same conditions as in B. (D) Mean K_d at 0 mV membrane potential estimated by fitting data from individual patches with Eq. 2. ^∗Significant difference from WT under corresponding conditions (p < 0.002); ^†significant difference from the same variant at pH 5.5 (^†p < 0.05; ^††p < 0.00002). Mean of data from four to six patches.

Figure 6

Single-channel conductance is unaffected by toggling the positive charge in the pore. (A) Example single-channel currents carried by V345H at a membrane potential of −50 mV, at a bath pH of 5.5 (left) or 9.0 (right, different patch). The line to the left represents the closed-channel level. (B) Mean single-channel i/V relationships for WT and V345H under these two pH conditions. (C) Mean unitary slope conductance measured from individual patches for the indicated CFTR variants at pH 5.5 and 9.0. ^∗Significant difference from WT at the same pH (p < 10⁻⁴). None of the channel variants had a significantly different conductance at these two pH values (p > 0.15). Mean of data from four to 13 patches in B and C.