The conductance of KCNQ2 and its pathogenic variants is determined by individual subunit gating

Figures

• 

  Fig. 1.. The E140R mutation impairs KCNQ2 channel function. (A) Schematic of S4 activation in wt KCNQ2 (Q) channels and when the S4 movement is impeded by the E140R mutation (Qi). (B to D) Current traces from oocytes expressing (B) wt KCNQ2 (Q) and (C) KCNQ2-E140R channels (Qi) or (D) water-injected oocytes using the voltage protocol shown above. Sweep-to-sweep intervals were 15 s. (E) Western blot of surface fraction or total lysate extracted from water-injected oocytes and oocytes expressing Q and Qi channels. The Na/K ATPase and β-actin are shown as controls. (F) Current (black) and fluorescence (cyan) traces from Alexa Fluor 488–labeled KCNQ2-E140R-F192C (Qi*) channels using the voltage protocol in (A). Cartoons depict channel assemblies: wt KCNQ2 (Q, black circles), KCNQ2-E140R (Qi, red circles), and Alexa Fluor 488–labeled KCNQ2-E140R-F192C (Qi*, cyan cartoons attached to red circles). (G) Normalized G(V) (black circles) and F(V) (cyan circles) curves from Qi* channels. Normalized G(V) curves from Q (squares) and unlabeled Qi (triangles) channels are shown for comparison. The midpoint of activation for the wt KCNQ2 (Q) fit is GV1/2 Q=−41.7±1.9 mV.Means ± SEM. n = 8 to 21. (H) TMRM…

• 

  Fig. 2.. KCNQ2 channels bearing two E140R mutations are functional. (A to C) Representative current (upper) and fluorescence (lower) traces from Alexa Fluor 488–labeled F192C (*) QQ* (triangles)–, Q*Qi (squares)–, and QQi* (circles)–linked dimers for voltage steps between −160 and +60 mV in steps of 20 mV from a holding potential of −80 mV, followed by a −40-mV tail potential. In (B) and (C), a prepulse to −120 mV was used to deactivate the S4s more completely. Cartoons depict the channel assembly in the constructs: wt KCNQ2 subunit (Q, black circles), E140R subunit (Qi, red circles), and Alexa Fluor 488 maleimide–labeled F192C (Q*, black or cyan cartoons attached to circles). Black lines between circles indicate tethers between subunits. (D to F) Normalized G(V) (open symbols and solid line from a Boltzmann fit) and F(V) (closed symbols and solid line from a Boltzmann fit) curves from labeled QQ* (triangles), Q*Qi (squares), and QQi* (circles), respectively. In (D) to (F), dotted and dashed-dotted lines represent the G(V) curves of single Q and dimers of QQ, respectively, for comparison. The midpoints of activation for the fits are as follows: GV1/2 QQ*=−74.4±2.8 mV, FV1/2 QQ*=−82.7±2.9…

• 

  Fig. 3.. State-dependent modification of QcQi-linked dimers by external MTSET is consistent with outward S4 motion in the R198C subunit (Qc). KCNQ2-R198C QcQc (top), QcQi (middle), and QQic (bottom) channel currents using the voltage protocol shown above (A) before and (B and C) after application of 100 μM MTSET in the (B) closed and (C) open states. MTSET is first applied [“closed state” in (A′)] at −140 mV or −100 mV for 5 s in between 25-s washouts for 8 to 15 cycles, and the change in current is measured at +20 mV. On the same cell and after MTSET is washed out, MTSET is reapplied [“open state” in (B′)] at +20 mV, as previously described in (21). (D) G(V) curves (lines from a Boltzmann fit) of (top) QcQc, (middle) QcQi, and (bottom) QQic channels normalized to peak conductance before MTSET application (black). The G(V) relationships after MTSET application in the closed (gray) and open (blue) states are obtained from recordings of (B) and (C), respectively. The midpoints of current activation for the fits are as follows: GV1/2 QCQC before=−47.0±1.3 mV (n = 7), GV1/2 QCQC closed=−44.9±0.6 mV (n = 7), GV1/2 QCQC open=−87.5±1.5 mV (n = 7), GV1/2 QCQi before=−46.3±2.5 mV (n = 5), GV1/2…

• 

  Fig. 4.. Tetramers of KCNQ2 subunits containing one, two, or three E140R mutations produce currents. (A to E) Representative current traces from tsA201 cells expressing linked tetrameric channels as labeled for voltage steps between −100 and +60 mV in steps of 20 mV from a holding potential of −80 mV, followed by a −40-mV tail potential. Cartoons depict the channel assembly in the constructs: wt KCNQ2 subunit (Q, black circles) and KCNQ2-E140R subunit (red circles). Black lines in between circles indicate tethers between subunits. (F) Normalized G(V) (solid line from a Boltzmann fit) curves from constructs shown in (A) to (E). The midpoints of activation for the fits are as follows: GV1/2 QQQQ=−24.5±1.4 mV (n = 14), GV1/2 QQiQQ=−26.3±4.0 mV (n = 3), GV1/2 QQiQQi=−29.1±2.1 mV (n = 5), GV1/2 QQQiQi=−20.4±3.3 mV (n = 3), GV1/2 QQiQiQi=−23.0±2.4 mV (n = 13); P > 0.05, one-way ANOVA and Tukey’s post hoc test; see Table 1. Data represent means ± SEM.

• 

  Fig. 5.. Increasing the number of E140R-containing subunits in the channel complex reduces single-channel conductance. (A) Representative single-channel recordings of various KCNQ2 constructs as labeled above each set of records. Records are from the 0-mV test pulse portion of a protocol in which the patch was held at a holding potential of −80 mV and pulsed to 0 mV for 4 or 10 s (QQiQiQi only) and then to −40 mV for 0.9 s, with an interpulse interval of 10 or 15 s. (B) All-points amplitude histograms of the traces shown in (A). The bin size for the histograms was 0.01 pA, and the y axis is shown as a log scale. The nonzero peaks were 0.57 pA for QQ, 0.34 pA for QQiQQ, 0.21 pA for QQQiQi, and 0.12 pA for QQiQiQi. Arrows highlight peaks representing subconducting states.

• 

  Fig. 6.. Having two E140R-containing subunits across from one another reduces conductance to a level of construct containing three E140R subunits. (A) Representative single-channel records of QQiQQi. Records are from the 0-mV test pulse portion of a protocol in which the patch was held at a holding potential of −80 mV and pulsed to 0 mV for 4 s and then to −40 mV for 0.9 s, with an interpulse interval of 10 s. (B) All-points amplitude histograms of the traces shown in (A). The bin size for the histogram was 0.01 pA, and the y axis is shown as a log scale. The nonzero peak was 0.11 pA. (C) Unitary current amplitude-voltage curves for each of the dimer and tetramer constructs studied. Conductances were 6.0 ± 0.2 pS for QQ (n = 3 to 6), 6.0 ± 0.4 pS for Q-F192S (n = 3 to 7), 4.5 ± 0.3 pS for QQiQQ (n = 3 to 7), 2.6 ± 0.1 pS for QQQiQi (n = 3 to 5), 1.7 ± 0.2 pS for QQiQQi (n = 3 to 7), 1.3 ± 0.1 pS for QiQ* (n = 3 to 4), 1.5 ± 0.2 pS for Qi*Q (n = 3 to 7), and 1.6 ± 0.2 pS for QQiQiQi (n = 3 to 5); see Table 1.

• 

  Fig. 7.. Models of KCNQ2 activation. (A) Model schemes for labeled KCNQ2-F192C (Q*) and for channels with two E140R (Q*Qi) and three E140R-containing (QQiQiQi) nonactivatable VSs. (B) G(V) and F(V) relations for the three models as indicated. Model G(V)s as solid lines overlay each other; F(V)s are dashed lines. Experimental Q* G(V) data are shown as filled circles and F(V) data as open circles. (C) For current and fluorescence tracings, the model protocol was as follows: Vh −120, 2 s to between −180 and +40 mV, step to −40 mV for 2 s. (D) G(V) and F(V) relations on a semilog plot, legend as for (B). (E) KCNQ2-F192C total fluorescence-voltage relation (Q*) and components F1(V) and F2(V). The V1/2 and k values for the overall model Q* F(V) are −80.5 and 22.9 mV, respectively. For F1(V), V1/2 and k values are −82.9 and 25.2 mV, and for F2(V), V1/2 and k values are −73.4 and 14.4 mV, respectively. See Materials and Methods for rate constants.

• 

  Fig. 8.. The opening of heterozygous channels containing GOF or LOF mutations is controlled by the subunit displaying the strongest hyperpolarized V1/2. (A to D) Representative current traces from cells expressing tetrameric constructs as labeled for voltage steps between −120 and +80 mV in steps of 20 mV from a holding potential of −80 mV, followed by a −40-mV tail potential. Cartoons depict the channel assembly in the constructs: wt KCNQ2 subunit (Q, black circles), KCNQ2-R201C subunit (green circles), KCNQ2-R214W subunit (blue circles), and KCNQ2-E140R subunit (red circles). Black lines in between circles indicate tethers between subunits. (E) Normalized G(V) (solid line from a Boltzmann fit) curves from constructs shown in (A) to (D). (F) Summary data for G(V) midpoints using Boltzmann fits from (E). The midpoints of activation for the fits are as follows: GV1/2 QQQQ=−24.4±4.1 mV (n = 5), GV1/2 QQQQR201C=−37.5±4.8 mV (n = 6), GV1/2 QQQQR214W=−18.0±1.3 mV (n = 5), GV1/2 QiQiQiQR214W=−6.5±5.8 mV (n = 4); see Table 1. Data represent means ± SEM.