|
FIGURE 1Ethosuximide (ETX) block of GIRK channels is subunit- and Gβγ-dependent. (a) Side view of the GIRK2-Gβγ bound crystal structure (4KFM) (Whorton & MacKinnon, 2013). Two of the four GIRK2 subunits are shown in grey. The geranylgeranyl moiety of Gγ (not in crystal structure), embedded into the plasma membrane, is shown as a yellow serpentine at the end of Gγ's C-terminus. Residue S148 (F137 in GIRK1) is presented in cyan spheres. (b) Record of currents in a representative oocyte expressing hGIRK2 (top), and I-V relationships (bottom). GIRK currents were monitored using two-electrode voltage clamp, at −80-mV holding potential. Oocytes were initially perfused with a 2-mM [K+]out solution ND96, then with high-K+ (24 mM) bath solution, HK24, with incremental concentrations of ETX (0–30 mM, indicated on the trace). For each ETX concentration, an I-V curve was acquired using 2-s-long voltage ramps from −120 to 50 mV. Net GIRK currents were derived by subtracting currents remaining after blocking the GIRK channels with 2.5-mM Ba2+. (c) Same as b for an oocyte expressing hGIRK2 and Gβγ. To express Gβγ, 5-ng Gβ1 and 1-ng Gγ2 RNA were injected. (d) ETX dose response for different GIRK channel combinations across all experiments with all subunit combinations used (symbols present mean ± SEM, number of experiments and cells is shown in e). G1 through G4 stand for GIRK1-GIRK4. For GIRK1/2, mGIRK2 was used. Data were fitted to Hill equation in each cell (except when EC50 was >10 mM); then Kd,app and nH values were averaged across all cells. Solid lines for each subunit combination's ETX dose–response were drawn using Hill equation with these average values. The approximate therapeutic range of ETX is shown by the striped rectangle for illustrative purposes. (e) Mean ± SEM of Kd,app and nH obtained from Hill equation fits. For GIRK2 without Gβγ, ETX dose–response measurements were done only with hGIRK2, because Ibasal of mGIRK2 were too small (<100 nA). n = number of cells, N = number of experiments. Statistical analysis: effects of Gβγ on Kd,app and nH of hGIRK2 and GIRK1/2 were compared by unpaired t test. ****P < 0.0001.
|
|
FIGURE 2Ethosuximide (ETX) block of GIRK2 is allosteric and Gβγ-dependent. (a) to (c) present the results of one experiment, representative of three. (a) hGIRK2 (0.5 ng/oocyte) current amplitudes with increasing amounts of Gβγ. (b) Increasing expression levels of Gβγ incrementally enhance ETX block. ETX dose–response relationships were obtained in individual oocytes; symbols show mean % inhibition (±SEM), number of cells (n) is listed in c. Solid lines were drawn using average Kd,app and nH values (shown in c) from Hill equation fits from individual cells. (d) Two models of ETX block of GIRK2. Black: the basic scheme describing GIRK2 activation by Gβγ (Wang et al., 2016). Dark red: block by ETX binding following channel opening. Blue: the allosteric model of ETX block. Kd and Kd* are dissociation constants of Gβγ to GIRK without and with bound ETX, respectively. Kd,E is ETX dissociation constant (see Figure S3 and Table S1 for details). (e) Simulated ETX dose–response curves of open channel block (left) and allosteric block (right) models. % inhibition represents the calculated decrease in proportion of time spent in open state, which is equivalent to change in channel's open probability. (f) I-V curves at different doses of ETX in a representative cell (left) and ETX dose–response curve (right) of hGIRK2 activated by 2-μM ivermectin. The dose response was fitted with Hill equation (Kd,app = 4.48 ± 1.02 mM, nH = 0.5 ± 0.04) and a two component binding equation (Equation 3; K1 = 0.03 ± 0.01 mM, K2 = 7.59 ± 0.95 mM, c = 0.21 ± 0.03). (g) The open channel N94H mutant is less sensitive to activation by Gβγ compared with GβγWT (left panel; unpaired t test, P = 0.0034). Right panel shows ETX dose response of mGIRK2 + Gβγ (Kd,app = 0.031 ± 0.007 mM, nH = 0.85 ± 0.07) and GIRK2N94H with and without Gβγ (no fitting was done for GIRK2N94H).
|
|
FIGURE 3Loss-of-function mutants of Gβ lose the ability to regulate ethosuximide (ETX) blocking effect in GIRK1/2 channels. Results of one experiment are shown, representative of two. Amounts of RNA were as follows: 50-ng RNA of both GIRK1 and mGIRK2 subunits, 5-ng Gβ RNA and 1-ng Gγ RNA. (a) GIRK1/2 channels are similarly activated by GβWTγ and the Gβ1 mutants K78R, I80N, and I80T. The P values above bars relate to the difference from control (Ibasal, no Gβγ). There was no difference among Gβγ-activated groups (P > 0.05 in all pairwise comparisons; one-way ANOVA followed by Tukey's test). (b) Comparison of Kd,app values extracted from data shown in c (one-way ANOVA followed by Tukey's test). The values of P versus control (no Gβγ) are shown. (c) ETX dose response of GIRK1/2 without or with WT or mutant Gβγ (left), and the Hill plot parameters from fits in individual oocytes (right). Mean ± SEM values are listed.
|
|
FIGURE 4GIRK3 gives clues for the ethosuximide (ETX) binding site. (a) The GIRK2-GIRK3 chimeras used. Note that in GIRKs the CSDs consist of a short N-terminal domain and a long C-terminal domain. (b) ETX dose response of the heterotetramers of GIRK2, GIRK3, and GIRK2/3 chimeras coexpressed with GIRK1. Details on RNA doses used, current amplitudes and Rβγ found in these experiments are shown in Figure S5B and Table S2.
|
|
FIGURE 5Ethosuximide (ETX) binds to the GIRK2 channel close to the activator PIP2 and leads to closing of the HBC gate. Twenty ETX molecules were placed randomly in the solvent, resulting in a concentration of 18.2 mM. (a) Aligned time-averaged density maps of ETX molecules binding to GIRK2 near the activator PIP2. The maps are shown as blue, orange, and olive mesh for run1, run2, and run3, respectively. PIP2 is coloured green, and one subunit of GIRK2 is highlighted in grey. (b) Distances between the ETX molecules binding to GIRK2 near the activator PIP2 and the GIRK2 residue R92. (c) ETX binding site identified in run1. On the right, a close-up of the ETX binding site is displayed. The GIRK2 channel, PIP2, and ETX are coloured in grey, green, and blue, respectively. (d) Minimum distances and Cα distances between the HBC gate-forming F192 of opposing subunits in run1. Pairs of opposing subunits are coloured in different shades of red (see inset). ETX binds to GIRK subunit C.
|
|
FIGURE 6Shortest pathways between the ethosuximide (ETX) binding site and the HBC gate based on the protein network analysis from the trajectory of run1. (a) Schematic illustration of the shortest pathways between the ETX binding residues R92 and F93, and the HBC gate residue F192. ETX and interactions of ETX with GIRK2 are shown in blue. Residues conserved between GIRK2 and GIRK3 are presented as white circles, while nonconserved residues are shown as grey circles. Conformational changes and interactions between residues are presented in different colours. Arrows show conformational changes, while dotted lines indicate nonbonded interactions between residues. (b) Overtime plots showing stable interactions between ETX and GIRK in blue. The resulting conformational changes and alterations of nonbonded interaction energies (nbIE) of channel residues are plotted in different colours, reflecting the colours of the arrows in a.
|
|
FIGURE 7Mutagenesis in the PIP2 binding region reveals amino acid residues crucial for ethosuximide (ETX) inhibition. (a) Sequence alignment of the PIP2-binding region, generated using ClustalW, reveals disparities between all GIRK subunits. (b,c) ETX dose responses of heterotetramers of GIRK1 with K90-N94 segment mutants of GIRK2. Data are from two separate experimental series detailed in Figure S8. (d) ETX dose response of GIRK1/3WT and GIRK1/3KFN with or without Gβγ. (e) ETX dose responses of homotetrameric GIRK2WT and mutants. Data are from two separate experiments detailed in Figure S9. Number of cells and Hill equation fit parameters for all experiments are shown in Table S3.
|
|
FIGURE 8Ethosuximide (ETX) uncovers the role of F137 of GIRK1 in the allosteric pathway of channel gating. (a) ETX strongly blocks GIRK1*, with or without coexpressed Gβγ. The dotted lines show the ETX dose–response curves of hGIRK2, hGIRK2 + Gβγ, and mGIRK2 from Figure 1d. (b) ETX dose–response of GIRK1/2, GIRK1*/2, GIRK1ΔdCT/2, and GIRK1*ΔdCT/2 with and without coexpressed Gβγ, and GIRK1*/3 with coexpressed Gβγ. Notation: G1/2 stands for GIRK1/2, G1*/2 stands for GIRK1*/2, and so on. (c) Summary of Kd,app and nH extracted from ETX dose–response experiments of a and b.
|
|
Fig. S1. Current amplitudes of different subunit combinations of GIRK channels. When Gβγ was
present, 5 ng Gβ and 1 ng Gγ RNAs were injected per oocyte. G1 through G4 stands for GIRK1
through GIRK4. A, an exemplary recording of the dose-dependent effect of ETX on mGIRK2
coexpressed with Gβγ (mGIRK2+Gβγ). B, current amplitudes of hGIRK2, hGIRK2+Gβγ, and
mGIRK2+Gβγ. Numbers above bars denote number of cells (n), with number of experiments (N)
in brackets. **** p<0.0001, one-way ANOVA followed by Tukey’s test. C, RNA amounts of
channel subunits, current amplitudes and fold activation by Gβγ (Rβγ; mean±SEM) of GIRK1/2,
GIRK1/3, GIRK1/4 with and without Gβγ, and GIRK4+Gβγ. This is a summary of all experiments
shown in Fig. 1. D, comparison of basal and Gβγ-activated currents in a subset of experiments
with the indicated amounts of channels’ RNA (bottom). Numbers above bars show number of
cells with number of experiments (in brackets). RNA doses of GIRK subunits varied depending on
their combination, thus only unpaired t-test was performed, to compare currents without and
with Gβγ. E, Rβγ of GIRK1/2, GIRK1/3, GIRK1/4 and hGIRK2 from the experiments shown in A
and D. Statistical comparison (bottom) was done by one-way ANOVA followed by Tukey’s test.
|
|
Fig. S2. T-type and L-type Ca2+ channels are blocked by ETX with low affinity. A, examples of
current records of α1H (CaV3.2), α1H + a2δ like subunit, Cachd1, and α1C + a2δ + β2b (CaV1.2),
respectively, with increasing doses of ETX. Currents were measured by voltage steps from -80
mV to -10 mV for CaV3.2, and from -80 to +20 mV for CaV1.2, in a solution containing 40 mM
Ba2+
. B, dose dependence of ETX inhibition shows that the apparent affinity of ETX to T-type and
L-type VGCCs is low.
|
|
Fig. S3. Kinetic modeling of ETX block of GIRK2. A, GIRK2 channel activity was recorded in
oocytes injected with 5 ng/oocyte RNA of Gβγ. Openings are shown as upward deflections. B,
open times histogram with a one-exponential fit (τ = 0.62 ms). C, the tested schemes of GIRK2-
Gβγ-ETX interaction, explained in Table S1. α and β are the opening and closing rate constants
for the transition C4↔O. KD and KD,E are the dissociation constant of binding of Gβγ and ETX,
respectively. KD
*
is the KD of Gβγ binding to ETX-bound channel. 1 – GIRK2 channel activation by
Gβγ, 2 – open channel blocker scheme, 3 – allosteric closed channel blocker scheme, 4 – state-
7
dependent closed channel blocker scheme (ETX interacts only with C4), 5 – state-independent
closed channel blocker scheme (ETX binds to all closed sates with the same affinity and no
assumption of change in ETX-occupied channel affinity to Gβγ is made). D, simulation of the
dose-response to ETX with different level of expression of Gβγ utilizing scheme #4. E, simulation
of the dose-response to ETX with different level of expression of Gβγ utilizing scheme #5.
|
|
Fig. S4. Effects of IVM and N94H mutation on GIRK2 current amplitudes. A, 2 µM ivermectin
(IVM) activates hGIRK2 (2 ng RNA) similarly to Gβγ (5 ng Gβ RNA, 1 ng Gγ RNA) (Kruskal-Wallis
test followed by Dunn’s test). The difference between IVM- or Gβγ-activated current was not
significant (p>0.999). B, mGIRK2N94H has a high basal current and is additionally activated by Gβγ,
while mGIRK2WT has a small basal current and a much greater extent of activation by Gβγ than
mGIRK2N94H (unpaired t-test for all comparisons). Note that the high basal current of mGIRK2N94H
was observed with only 0.5 ng RNA/oocyte, which is 4 fold less than for mGIRK2WT.
|
|
Fig. S5. Design and Gβγ activation of the GIRK2/GIRK3 chimeras. A, sequence alignment of
TMDs of GIRK2 and GIRK3. Dots indicate identical a.a. Note that the TMDs of GIRK2 and GIRK3
differ by 24 amino acid residues. B, comparison of the extent of Gβγ activation, Rβγ. One-way
ANOVA (F (3, 46) = 75.17, p<0.0001) followed by Tukey test.
|
|
Fig. S6. ETX binding-induced conformational changes of the GIRK channel observed in run1. A,
GIRK2 subunits are colored based on an alignment between the subunit conformations at the
beginning of the simulations (t=0 µs) and the end of the simulation (t=1.5 µs). The F192 of the
HBC gate is shown in stick representation. B, Alignment of the starting conformation (t= 0 µs,
not bound to ETX; white) and the end conformation (t=1.5 µs, ETX-bound; blue) of subunit C.
The HBC gate, the ETX-binding residues R92 and F93, and the residues in proximity to the ETX
binding site are shown in stick representation. On the right, a close up of the alignment is
shown, focusing on the ETX binding region at the TM1 and the HBC gating region.
|
|
Fig. S7. The effect of ETX binding on the PIP2 conformation in run1. A, the structures show the
PIP2 conformation at the beginning (t= 0 μs) and at 1.5 μs of the MD simulation. PIP2 is shown in
green and GIRK2 in grey. PIP2 and PIP2 binding residues of GIRK2 are shown as sticks. Red dashed
lines represent hydrogen bonds. B, RMSDs and hydrogen bonds formation of ETX (blue) and PIP2
(green) over time. The dashed lines indicate the time ETX binds to GIRK2 (t= 0.26 µs), and where
PIP2 is stabilized in a displaced conformation, 5.5 Å from its original position (t= 0.49 µs).
|
|
Fig. S8. The effect of mutations in the PIP2 binding region in GIRK1/2 and GIRK1/3
heterotetramers. The amount of injected RNA of Gβγ was 5 ng Gβ and 1 ng Gγ. The amounts of
RNA for channel subunits are shown in panels. A-F, results from 2 experimental series with
GIRK1/2 heterotetrameric constructs: 5 experiments in series 1 with the QLS and K90Q mutants
of GIRK2, and 1 experiment in series 2 tested the F93L and N94S mutants. In A and D, the left
panels show original recordings of ETX effect on the indicated constructs, and the right panels
show the same traces normalized to the initial amplitude in each cell, without ETX. B and E,
current amplitudes of GIRK1/2 and GIRK1/2 mutants with and without Gβγ (unpaired t-test for
all comparisons). Currents of GIRK1/2QLS and GIRK1/2K90Q were similar to GIRK1/2WT (P>0.05;
11
Kruskal-Wallis followed by Dunn’s multiple comparison test). For GIRK1/2F93L and GIRK1/2N94S a
direct comparison with WT is not possible because of the different amounts of RNAs injected. C
and F, Rβγ was similar to GIRK1/2WT for all GIRK2 mutants. G-I, comparison of GIRK1/3WT and
GIRK1/3KFN. G, representative recordings of GIRK1/3WT and GIRK1/3KFN, without (left) and with
(right) amplitude normalization. H, current amplitudes of GIRK1/3WT and GIRK1/3KFN. I, Rβγ of
GIRK1/3WT and GIRK1/3KFN is similar (p>0.05).
|
|
Fig. S8. The effect of mutations in the PIP2 binding region in GIRK1/2 and GIRK1/3
heterotetramers. The amount of injected RNA of Gβγ was 5 ng Gβ and 1 ng Gγ. The amounts of
RNA for channel subunits are shown in panels. A-F, results from 2 experimental series with
GIRK1/2 heterotetrameric constructs: 5 experiments in series 1 with the QLS and K90Q mutants
of GIRK2, and 1 experiment in series 2 tested the F93L and N94S mutants. In A and D, the left
panels show original recordings of ETX effect on the indicated constructs, and the right panels
show the same traces normalized to the initial amplitude in each cell, without ETX. B and E,
current amplitudes of GIRK1/2 and GIRK1/2 mutants with and without Gβγ (unpaired t-test for
all comparisons). Currents of GIRK1/2QLS and GIRK1/2K90Q were similar to GIRK1/2WT (P>0.05;
11
Kruskal-Wallis followed by Dunn’s multiple comparison test). For GIRK1/2F93L and GIRK1/2N94S a
direct comparison with WT is not possible because of the different amounts of RNAs injected. C
and F, Rβγ was similar to GIRK1/2WT for all GIRK2 mutants. G-I, comparison of GIRK1/3WT and
GIRK1/3KFN. G, representative recordings of GIRK1/3WT and GIRK1/3KFN, without (left) and with
(right) amplitude normalization. H, current amplitudes of GIRK1/3WT and GIRK1/3KFN. I, Rβγ of
GIRK1/3WT and GIRK1/3KFN is similar (p>0.05).
|
|
Fig. S9. The effect of mutations in the PIP2 binding region in GIRK2 homotetramers
coexpressed with Gβγ. Very small or no currents were observed without Gβγ. In all experiments
the amount of injected RNA was 5 ng channel, 5 ng Gβ and 1 ng Gγ. Results are from 2
experiments with GIRK2 mutants. Experiment 1 tested the QLS and K90Q mutants, and
experiment 2 tested the F93L and N94S mutants. A shows the original recordings (not
normalized) of ETX effect on the indicated constructs, and B shows the same traces normalized
to the initial amplitude in each cell, without ETX. Color coding is shown in B. C, current
amplitudes of GIRK2WT and mutants with Gβγ (Kruskal-Wallis followed by Dunn’s multiple
comparison test). There was not difference in currents of GIRK2WT, GIRK2F93L and GIRK2N94S
(P>0.05).
|
|
Fig. S10. Single channel properties of GIRK1*. A, representative cell-attached, single channel
recordings of GIRK1* expressed alone (left, in black); GIRK1* coexpressed with m2 receptor and
activated by acetylcholine (ACh) (middle, in blue; 1-5 ng m2 receptor RNA); and GIRK1*
coexpressed with m2 receptor and Gβγ (right, in red; 5 ng Gβ and 1 or 2 ng Gγ RNA). c denotes
the closed state and o denotes the open state of the channel. B, Po of GIRK1*, GIRK1* activated
by ACh, and GIRK1* activated by Gβγ. Gβγ had significantly increased the Po of GIRK1*. The
average Po of GIRK1* in the presence of Gβγ was 0.043±0.01 (n=8). C, a representative all-points
histogram from a single channel recording in an oocyte expressing GIRK1* with Gβγ, used to
assess the single channel amplitude, isingle. Histogram (red line) was fitted to a two-component
Gaussian (blue line). D, the single channel current amplitude of GIRK1* at -80 mV was similar in
all conditions (p=0.21, Kruskal-Wallis test). The average (±SEM) isingle for GIRK1* in the presence
of Gβγ, at -80 mV, was 1.18±0.03 pA (n=17). For comparison, for Gβγ-activated GIRK1/2
recorded under identical conditions, Po is ~0.105 and isingle is ~2.8 pA (Yakubovich et al., 2015).
|
|
Fig. S11. The reversal potential (Vrev) of GIRK1* is depolarized compared to GIRK1/2. A, current
amplitudes at -80 mV of GIRK1/2 and GIRK1* (n=10 and n=8, respectively). In this experiment,
the channels were expressed with Gβγ. The RNA doses were (in ng/oocyte): GIRK1* 5, GIRK1
0.05, GIRK2-YFP 0.05, Gβ 5, Gγ 1. B, the reversal potential (Vrev) in cells shown in A. Vrev of
GIRK1* is more positive than of GIRK1/2WT in HK24 solution. In HK24 solution, according to
Nernst equation the expected EK = -36 mV, assuming Kin = 100 mM (Dascal, 1987). C, I-V curves
in representative oocytes.
|
|
Figure S12. S148 is not solvent-accessible but is part of the allosteric network coupling ETXand Gβγ-binding sites, along with N94, in GIRK2. A, S148 in GIRK2 structure is not accessible to
ETX because it is not solvent-accessible. S148 and the surrounding residues are represented as
spheres. The left figure shows the upper part of the TMD, with S148 colored red. On the right, a
close up of S148 and its surrounding residues can be seen. The structures are colored according
to their local solvent accessibility, with white and purple meaning no and full solvent
accessibility, respectively. B, the shortest pathway linking S148 and the Gβ binding residue R337
passes through the GIRK2 residue N94, important for channel inhibition by ETX. GIRK2 is shown
as white cartoon, the residues of the pathway are represented as red sticks. The network was
calculated based on the dynamics of subunit C in run1 that binds ETX in the course of the MD
simulation.
|
|
Fig. S13. Representative current recordings and mean current amplitudes of GIRK1*- and
GIRK2S148F-containing channels. Representative recordings without (top) and with Gβγ (bottom):
16
A, GIRK1/2WT. B, GIRK1/2S148F. C, GIRK1*/2WT. D, GIRK1*/3WT+Gβγ. E, GIRK1Δ121/2WT. F,
GIRK1Δ121/2s148f. G, GIRK1*Δ121/2WT. H, GIRK1*Δ121/2s148f. I, current amplitudes at -80 mV of the
channels presented in A-H (left), and the amounts of injected RNA (right). Oocytes were injected
with 5 ng of Gβ and 1 ng of Gγ in all cases. Note that the amounts of channel RNA without and
with Gβγ were often different, therefore a direct comparison of amplitudes with/without Gβγ
and an estimation of extent of Gβγ activation was not possible.
|
|
Fig. S13. Representative current recordings and mean current amplitudes of GIRK1*- and
GIRK2S148F-containing channels. Representative recordings without (top) and with Gβγ (bottom):
16
A, GIRK1/2WT. B, GIRK1/2S148F. C, GIRK1*/2WT. D, GIRK1*/3WT+Gβγ. E, GIRK1Δ121/2WT. F,
GIRK1Δ121/2s148f. G, GIRK1*Δ121/2WT. H, GIRK1*Δ121/2s148f. I, current amplitudes at -80 mV of the
channels presented in A-H (left), and the amounts of injected RNA (right). Oocytes were injected
with 5 ng of Gβ and 1 ng of Gγ in all cases. Note that the amounts of channel RNA without and
with Gβγ were often different, therefore a direct comparison of amplitudes with/without Gβγ
and an estimation of extent of Gβγ activation was not possible.
|
|
Fig. S14. Voltage dependence of ETX block in GIRK1-containing channels. Currents at different
voltages were determined from the I-V curves obtained by voltage ramps. A, GIRK1/2 channels
show different apparent affinity towards ETX at different voltages. B, ETX block of GIRK1/2
remained sensitive to voltage with coexpressed Gβγ. C-F, ETX block of GIRK1*/2 channels (with
or without Gβγ), GIRK1* with Gβγ, and GIRK2 with Gβγ did not exhibit voltage dependence. G,
comparison of Kd,app from all GIRK combinations shown in A-F shows that GIRK1/2 and
GIRK1/2+Gβγ are blocked differently at voltages between -100 to -60 mV, while GIRK1*/2,
GIRK1*/2+Gβγ, GIRK1*+Gβγ, and GIRK2+Gβγ had similar Kd,app at all voltages. Statistical analysis:
two-way ANOVA (mixed-effects model (REML)) followed by Tukey's multiple comparisons test.
H, statistical significance (p values) for the comparison of Kd,app at -100 and -80 mV vs -60 mV
(same data as in g) from two-way ANOVA test.
|
|
Fig. S14. Voltage dependence of ETX block in GIRK1-containing channels. Currents at different
voltages were determined from the I-V curves obtained by voltage ramps. A, GIRK1/2 channels
show different apparent affinity towards ETX at different voltages. B, ETX block of GIRK1/2
remained sensitive to voltage with coexpressed Gβγ. C-F, ETX block of GIRK1*/2 channels (with
or without Gβγ), GIRK1* with Gβγ, and GIRK2 with Gβγ did not exhibit voltage dependence. G,
comparison of Kd,app from all GIRK combinations shown in A-F shows that GIRK1/2 and
GIRK1/2+Gβγ are blocked differently at voltages between -100 to -60 mV, while GIRK1*/2,
GIRK1*/2+Gβγ, GIRK1*+Gβγ, and GIRK2+Gβγ had similar Kd,app at all voltages. Statistical analysis:
two-way ANOVA (mixed-effects model (REML)) followed by Tukey's multiple comparisons test.
H, statistical significance (p values) for the comparison of Kd,app at -100 and -80 mV vs -60 mV
(same data as in g) from two-way ANOVA test.
|
