Carvalho-de-Souza JL , Bezanilla F .

R01 GM030376 NIGMS NIH HHS

	Figure 1. Biophysical properties of Shaker dimers when mutated in PDs. a Top: representation of Shaker protein showing wild-type voltage sensor (VSDwt) and pore domains (PDwt); Middle: Top view of Shaker channel in the membrane, showing each subunit (protein) of the tetramer represented by a different color, with N and C representing amino-terminus and carboxy-terminus, respectively. The black circle in the center represents K+ ions within the conductive pathway. Bottom: Typical K+ currents from Shaker channels expressed in Xenopus oocytes with the indicated voltage protocol. Horizontal bar means 2 s. b Top and middle: schematics showing dimerized Shaker, with protomer 1 in blue and protomer 2 in red. These channels do not contain W434F mutations and are called wt-PD Shaker dimers. Bottom: K+ currents recorded from wt-PD Shaker dimers in similar condition as in a. c, d Top and middle: representation of Shaker dimers containing the mutation W434F (W434F-Shaker dimers) in protomer 1 in c and in protomer 2 in d. The mutant domains are marked by a diamond. Bottom: K+ currents recorded with the respective channels in similar conditions as in a. In e–h, open squares are for Shaker, gray circles for wt-PD Shaker dimers, black triangles, and black squares for W434F-Shaker dimers with W434F mutation in protomers 1 and 2, respectively. e Normalized conductance–voltage (G–V) curves for the channels represented in a–d. f Time constants of the K+ currents inactivation in the indicated channels during 10 s depolarizing voltages periods (same protocol shown in a, bottom, and same symbols convention as in e). Fast and slow time constants from two exponentials fit to the data (see “methods” section for details), are shown for all Shaker dimers as in b–d. g Fast component relative amplitude of the exponential components in the channels indicated. h K+ currents 10 s inactivation curves (Inact-V) in the channels indicated. All data points are the average of 4–6 independent experiments. The vertical bars in e–h are the standard error of the mean. VSDwtPDwt: Shaker; VSDwtPDwt-VSDwtPDwt: wt-PD Shaker dimer; VSDwtPDW434F-VSDwtPDwt and VSDwtPDwt-VSDwtPDW434F: W434F-Shaker dimers with W434F in protomer 1 and in protomer 2, respectively
	Figure 2. Schematic representation of near and far configuration Shaker dimers. a Near configuration dimers shown in two possible designs, both with VSDILT and PDW434F in different protomers (VSDwtPDW434F-VSDILTPDwt or VSDILTPDwt-VSDwtPDW434F). Protomer 1 is shown in blue and protomer 2 is shown in red in the top left and right panels with the mutant VSD indicated by circles and mutant PD indicated by diamonds. Note that mutant domains are in different protomers in both cases. b Far configuration dimers shown in their two possible designs, where VSDILT and PDW434F are always in the same protomer (VSDILTPDW434F-VSDwtPDwt or VSDwtPDwt-VSDILTPDW434F). Protomers are color coded as in a. Similarly, the mutant VSD are indicated by circles and mutant PD indicated by diamonds
	Figure 3. Functional characterization of all types of W434F-Shaker dimers bearing VSDILT. K+ currents from dimers with PDW434F mutation in protomer 1 are shown in a and b, configured as near and far channels, respectively. Horizontal bar in a, top, means 2 s. In c and d, K+ currents are shown from dimers with PDW434F mutation in protomer 2, configured as near and far, respectively. All currents in a–d were activated by the same voltage protocol as shown in a, top panel. G–V and Inact-V curves are shown in e for near (orange) and far (green) dimers with W434F in protomer 1 and in f with W434F in protomer 2. Curves in black shown in e and f are from the corresponding control, PDW434F in protomer 1 and in protomer 2, respectively, and with no VSD mutated (see Fig. 1). Inact-V curves in W434F-Shaker dimers, near and far, with the PDW434F in protomer 1 (g) and in protomer 2 (h) are also shown together with the Q–V curve for VSDILT. i The shift in G–V curves introduced by the presence of VSDILT in Shaker (black bar) and in the four types of W434F-Shaker dimers, as labeled in the graph. j and k show K+ currents inactivation fast and slow time constants taken from two exponential fits (see “methods” section) from near and far configuration dimers, respectively. In both graphs filled and empty circles (orange or green) are for dimers with PDW434F in protomer 1 and in protomer 2, respectively. l Fraction of the fast inactivating component are shown for near and far dimers with the position of the PDW434F indicated by the labels in the graph. All data points are the average of 4–6 independent experiments. The vertical bars in e–l are the standard error of the mean
	Figure 4. Study of the VSDmut//PDW434F interfaces in W434F-Shaker dimers. a Voltage dependences of Shaker containing different VSDmut. Plots show three voltage dependences of Shaker channels with different VSDs: V0 and V1 from Q–V and V1/2 from G–V curves. b, c Inact-V curves from W434F-Shaker dimers, near and far configuration, respectively and as indicated, for each VSDmut case. Mutations in the VSD are color/shape coded and are shown in the insets. d, e G–V curves for the same channels, near and far, shown in b and c with the mutations indicated by same color/shape code as in b and c. All Inact-V curves, except from near dimers containing VSD361A or VSDILT, were fitted by a two-state model: noninactivated and inactivated (see “methods” section for details), generating a voltage-dependent parameter Vinact (Supplementary Tables 1 and 2). In near dimers containing VSD361A or VSDILT, the curves were fitted by two independent two-state models to generate two voltage-dependent parameters: V1st and V2nd. f, g Plots of Vinact, V1st, and V2nd from near and far different dimers as labeled, respectively. Squared residuals from the fittings that generated the voltage-dependent parameters Vinact, V1st, and V2nd are also plotted and connected to them by the vertical dotted lines. Shifts in Vinact (or V1st and V2nd) from near (orange circles for Vinact, squares for V1st and triangles for V2nd, h and j) and far dimers (green circles, i and k) relative to the values in the respective Shaker dimers with VSDwt only plotted against the respective shifts in V0 (h and i) and V1 (h and i) from VSDmut taken from Q–V curves in nonconductive Shaker. l The same shift values of Vinact, V1st and V2nd shown in h–k from near and far dimers were plotted against the shifts in V1/2 in Shaker induced by the respective VSDmut (also relative to channels with VSDwt only). Labels at each data point were omitted for clarity. All data points are the average of 4–6 independent experiments. The vertical bars in b–g are the standard error of the mean. The error bars in plots h–l were omitted for clarity
	Figure 5. wt-PD Shaker dimers (no W434F mutation) and nonconductive Shaker dimers (all PD mutated with W434F) show independent VSD domains. a Q–V curves are shown for all nonconductive dimers studied including dimers with only PDW434F and the ones possessing two VSDmut, indicated in the inset in a color/shape code. The curves were fitted with two independent two-state models for the generation of two voltage dependences values, V0* and V1*. These values are plotted in b and for comparison, V0 and V1 (from nonconductive Shaker) are plotted in c. d G–V curves from wt-PD Shaker dimers possessing only VSDwt and the ones possessing two VSDmut, similarly indicated in the inset in a color/shape code. Data were fitted with a two-state model and the values of the G–V midpoints (V1/2*) were used to calculate the shifts produced by the mutations in the VSD as compared with V1/2* from wt-PD Shaker dimers that only contains VSDwt (e). A horizontal dotted line shows the overall average value of the % shift and its value is 49.8%. All data points are the average of 4–6 independent experiments. The vertical bars in a and d are the standard error of the mean. The plots in b and c were omitted of error bars for clarity. Continuous lines over the data points represent the best fit of two independent two-state models for Q–V curves in a and a two-state model for G–V curves in d
	Putative key residue S412 in the VSD//PD interface in Shaker. The graph shows Q–V curves recorded in nonconductive Shaker, as well as G–V curves from Shaker carrying (red circles) or not (black circles) the mutation S412V, supposedly disruptive of the VSD//PD interface. Q–V curves are plotted as filled circles and G–V curves with open circles. Q–V curves were fitted with a three-state model to obtain V0 and V1 voltage dependences. G–V curves were fitted with a two-state model to get a midpoint V1/2. All data points are the average of 4–6 independent experiments. The vertical bars are the standard error of the mean. Continuous lines over the data points represent the best fit of a three-state model for Q–V curves and a two-state model for G–V curves. All V0, V1, and V1/2 voltage-dependent parameters are shown in Supplementary Table 2
	Figure 7. Structural models of a Shaker-like channel (KV1.2, PDB:3LUT), showing the suggested VSD//PD interface. a Top view of the tetrameric channel, with two subunits in blue and red colors and the other two, for clarity, in light gray. Notice that the black square region indicates the VSD//PD interface between the VSD from different blue and red neighboring subunits. b Black square region from a zoomed in and turned by 90°. The cartoon shows with ribbons the S4 segment from one subunit in blue and the S5 segment from the neighbor subunit in red. Colored spheres are the van der Waals volumes of the atoms from the residues indicated (V369 in blue, S412 in green, S411 in orange, F433 in yellow, and finally W434 in magenta (residues numbering as in Shaker)). These residues are proposed to be the physical connection of the noncanonical coupling pathway between VSD and PD. c Top view of the zoomed black square region shown in a. Color code and labels correspond to a. Note that in a and in b the van der Waals surfaces are in contact, giving support to the hypothesis of the VSD//PD interface
	Supplementary figure 1. Possible reason why Shaker dimers with W434F mutation (diamonds) in only one of their PD (all VSD-wt) show different inactivation voltage dependences and kinetics depending on which protomer is mutated in its PD. Shaker dimers are represented in panels a and b with the protomers color coded in blue for protomer 1 and red for protomer 2. a shows Shaker dimer with the protomer 1 containing mutation PD-W434F and b shows dimers with that mutation in protomer 2. Note that since protomer 1 has a C-terminal that is not free, mutations in its PD may introduce slightly different features to the inactivation process as compared with the inactivation in dimers with the mutation in protomer 2 which has a free C-terminal.
	Supplementary figure 2. Gating currents in non-conductive Shaker dimers. A, top panel, voltage protocol for gating currents recording; bottom panel, typical gating currents from non-conductive Shaker dimers bearing only wild-type VSD. B-H, gating currents from non-conductive Shaker dimers containing the indicted different mutations (B: 358E, C: 358R, D: 358W, E: 361E, F: 361R, G: 361A, H: ILT) in the VSD of protomer 2. These currents were recorded in response to the same voltage protocol shown in A, except for the recordings shown in H where the returning (final) voltage was -80 mV. Leak currents from all traces were subtracted offline. Bars indicating amplitude and time base are shown separately for each panel.
	Supplementary figure 3. Speculative diagonally arranged dimerized Shaker proteins. a and b show how the channels would look from the extracellular side if the mutations is VSD and PD are in different protomers as for near channels configuration. c and d show top views of channels configuration. Note that the four channels shown in a-d are very similar in their VSD//PD interfaces, with the same four different interfaces in all of them.
	Fig. 1. Biophysical properties of Shaker dimers when mutated in PDs. a Top: representation of Shaker protein showing wild-type voltage sensor (VSDwt) and pore domains (PDwt); Middle: Top view of Shaker channel in the membrane, showing each subunit (protein) of the tetramer represented by a different color, with N and C representing amino-terminus and carboxy-terminus, respectively. The black circle in the center represents K+ ions within the conductive pathway. Bottom: Typical K+ currents from Shaker channels expressed in Xenopus oocytes with the indicated voltage protocol. Horizontal bar means 2 s. b Top and middle: schematics showing dimerized Shaker, with protomer 1 in blue and protomer 2 in red. These channels do not contain W434F mutations and are called wt-PD Shaker dimers. Bottom: K+ currents recorded from wt-PD Shaker dimers in similar condition as in a. c, d Top and middle: representation of Shaker dimers containing the mutation W434F (W434F-Shaker dimers) in protomer 1 in c and in protomer 2 in d. The mutant domains are marked by a diamond. Bottom: K+ currents recorded with the respective channels in similar conditions as in a. In e–h, open squares are for Shaker, gray circles for wt-PD Shaker dimers, black triangles, and black squares for W434F-Shaker dimers with W434F mutation in protomers 1 and 2, respectively. e Normalized conductance–voltage (G–V) curves for the channels represented in a–d. f Time constants of the K+ currents inactivation in the indicated channels during 10 s depolarizing voltages periods (same protocol shown in a, bottom, and same symbols convention as in e). Fast and slow time constants from two exponentials fit to the data (see “methods” section for details), are shown for all Shaker dimers as in b–d. g Fast component relative amplitude of the exponential components in the channels indicated. h K+ currents 10 s inactivation curves (Inact-V) in the channels indicated. All data points are the average of 4–6 independent experiments. The vertical bars in e–h are the standard error of the mean. VSDwtPDwt: Shaker; VSDwtPDwt-VSDwtPDwt: wt-PD Shaker dimer; VSDwtPDW434F-VSDwtPDwt and VSDwtPDwt-VSDwtPDW434F: W434F-Shaker dimers with W434F in protomer 1 and in protomer 2, respectively
	Fig. 2. Schematic representation of near and far configuration Shaker dimers. a Near configuration dimers shown in two possible designs, both with VSDILT and PDW434F in different protomers (VSDwtPDW434F-VSDILTPDwt or VSDILTPDwt-VSDwtPDW434F). Protomer 1 is shown in blue and protomer 2 is shown in red in the top left and right panels with the mutant VSD indicated by circles and mutant PD indicated by diamonds. Note that mutant domains are in different protomers in both cases. b Far configuration dimers shown in their two possible designs, where VSDILT and PDW434F are always in the same protomer (VSDILTPDW434F-VSDwtPDwt or VSDwtPDwt-VSDILTPDW434F). Protomers are color coded as in a. Similarly, the mutant VSD are indicated by circles and mutant PD indicated by diamonds
	Fig. 3. Functional characterization of all types of W434F-Shaker dimers bearing VSDILT. K+ currents from dimers with PDW434F mutation in protomer 1 are shown in a and b, configured as near and far channels, respectively. Horizontal bar in a, top, means 2 s. In c and d, K+ currents are shown from dimers with PDW434F mutation in protomer 2, configured as near and far, respectively. All currents in a–d were activated by the same voltage protocol as shown in a, top panel. G–V and Inact-V curves are shown in e for near (orange) and far (green) dimers with W434F in protomer 1 and in f with W434F in protomer 2. Curves in black shown in e and f are from the corresponding control, PDW434F in protomer 1 and in protomer 2, respectively, and with no VSD mutated (see Fig. 1). Inact-V curves in W434F-Shaker dimers, near and far, with the PDW434F in protomer 1 (g) and in protomer 2 (h) are also shown together with the Q–V curve for VSDILT. i The shift in G–V curves introduced by the presence of VSDILT in Shaker (black bar) and in the four types of W434F-Shaker dimers, as labeled in the graph. j and k show K+ currents inactivation fast and slow time constants taken from two exponential fits (see “methods” section) from near and far configuration dimers, respectively. In both graphs filled and empty circles (orange or green) are for dimers with PDW434F in protomer 1 and in protomer 2, respectively. l Fraction of the fast inactivating component are shown for near and far dimers with the position of the PDW434F indicated by the labels in the graph. All data points are the average of 4–6 independent experiments. The vertical bars in e–l are the standard error of the mean
	Fig. 4. Study of the VSDmut//PDW434F interfaces in W434F-Shaker dimers. a Voltage dependences of Shaker containing different VSDmut. Plots show three voltage dependences of Shaker channels with different VSDs: V0 and V1 from Q–V and V1/2 from G–V curves. b, c Inact-V curves from W434F-Shaker dimers, near and far configuration, respectively and as indicated, for each VSDmut case. Mutations in the VSD are color/shape coded and are shown in the insets. d, e G–V curves for the same channels, near and far, shown in b and c with the mutations indicated by same color/shape code as in b and c. All Inact-V curves, except from near dimers containing VSD361A or VSDILT, were fitted by a two-state model: noninactivated and inactivated (see “methods” section for details), generating a voltage-dependent parameter Vinact (Supplementary Tables 1 and 2). In near dimers containing VSD361A or VSDILT, the curves were fitted by two independent two-state models to generate two voltage-dependent parameters: V1st and V2nd. f, g Plots of Vinact, V1st, and V2nd from near and far different dimers as labeled, respectively. Squared residuals from the fittings that generated the voltage-dependent parameters Vinact, V1st, and V2nd are also plotted and connected to them by the vertical dotted lines. Shifts in Vinact (or V1st and V2nd) from near (orange circles for Vinact, squares for V1st and triangles for V2nd, h and j) and far dimers (green circles, i and k) relative to the values in the respective Shaker dimers with VSDwt only plotted against the respective shifts in V0 (h and i) and V1 (h and i) from VSDmut taken from Q–V curves in nonconductive Shaker. l The same shift values of Vinact, V1st and V2nd shown in h–k from near and far dimers were plotted against the shifts in V1/2 in Shaker induced by the respective VSDmut (also relative to channels with VSDwt only). Labels at each data point were omitted for clarity. All data points are the average of 4–6 independent experiments. The vertical bars in b–g are the standard error of the mean. The error bars in plots h–l were omitted for clarity
	Fig. 5. wt-PD Shaker dimers (no W434F mutation) and nonconductive Shaker dimers (all PD mutated with W434F) show independent VSD domains. a Q–V curves are shown for all nonconductive dimers studied including dimers with only PDW434F and the ones possessing two VSDmut, indicated in the inset in a color/shape code. The curves were fitted with two independent two-state models for the generation of two voltage dependences values, V0* and V1*. These values are plotted in b and for comparison, V0 and V1 (from nonconductive Shaker) are plotted in c. d G–V curves from wt-PD Shaker dimers possessing only VSDwt and the ones possessing two VSDmut, similarly indicated in the inset in a color/shape code. Data were fitted with a two-state model and the values of the G–V midpoints (V1/2*) were used to calculate the shifts produced by the mutations in the VSD as compared with V1/2* from wt-PD Shaker dimers that only contains VSDwt (e). A horizontal dotted line shows the overall average value of the % shift and its value is 49.8%. All data points are the average of 4–6 independent experiments. The vertical bars in a and d are the standard error of the mean. The plots in b and c were omitted of error bars for clarity. Continuous lines over the data points represent the best fit of two independent two-state models for Q–V curves in a and a two-state model for G–V curves in d
	Fig. 6. Putative key residue S412 in the VSD//PD interface in Shaker. The graph shows Q–V curves recorded in nonconductive Shaker, as well as G–V curves from Shaker carrying (red circles) or not (black circles) the mutation S412V, supposedly disruptive of the VSD//PD interface. Q–V curves are plotted as filled circles and G–V curves with open circles. Q–V curves were fitted with a three-state model to obtain V0 and V1 voltage dependences. G–V curves were fitted with a two-state model to get a midpoint V1/2. All data points are the average of 4–6 independent experiments. The vertical bars are the standard error of the mean. Continuous lines over the data points represent the best fit of a three-state model for Q–V curves and a two-state model for G–V curves. All V0, V1, and V1/2 voltage-dependent parameters are shown in Supplementary Table 2
	Fig. 7. Structural models of a Shaker-like channel (KV1.2, PDB:3LUT), showing the suggested VSD//PD interface. a Top view of the tetrameric channel, with two subunits in blue and red colors and the other two, for clarity, in light gray. Notice that the black square region indicates the VSD//PD interface between the VSD from different blue and red neighboring subunits. b Black square region from a zoomed in and turned by 90°. The cartoon shows with ribbons the S4 segment from one subunit in blue and the S5 segment from the neighbor subunit in red. Colored spheres are the van der Waals volumes of the atoms from the residues indicated (V369 in blue, S412 in green, S411 in orange, F433 in yellow, and finally W434 in magenta (residues numbering as in Shaker)). These residues are proposed to be the physical connection of the noncanonical coupling pathway between VSD and PD. c Top view of the zoomed black square region shown in a. Color code and labels correspond to a. Note that in a and in b the van der Waals surfaces are in contact, giving support to the hypothesis of the VSD//PD interface

Bezanilla, Inactivation of the sodium channel. I. Sodium current experiments. 1977, Pubmed

Bezanilla, Inactivation of the sodium channel. I. Sodium current experiments. 1977, Pubmed
Brohawn, Crystal structure of the human K2P TRAAK, a lipid- and mechano-sensitive K+ ion channel. 2012, Pubmed
Carvalho-de-Souza, Nonsensing residues in S3-S4 linker's C terminus affect the voltage sensor set point in K+ channels. 2018, Pubmed , Xenbase
Chen, Structure of the full-length Shaker potassium channel Kv1.2 by normal-mode-based X-ray crystallographic refinement. 2010, Pubmed
Conti, Reciprocal voltage sensor-to-pore coupling leads to potassium channel C-type inactivation. 2016, Pubmed , Xenbase
Cuello, Structural mechanism of C-type inactivation in K(+) channels. 2010, Pubmed
Dedek, Neonatal convulsions and epileptic encephalopathy in an Italian family with a missense mutation in the fifth transmembrane region of KCNQ2. 2003, Pubmed , Xenbase
Dong, K2P channel gating mechanisms revealed by structures of TREK-2 and a complex with Prozac. 2015, Pubmed
Fernández-Mariño, Gating interaction maps reveal a noncanonical electromechanical coupling mode in the Shaker K+ channel. 2018, Pubmed , Xenbase
Grabe, Structure prediction for the down state of a potassium channel voltage sensor. 2007, Pubmed
HODGKIN, Currents carried by sodium and potassium ions through the membrane of the giant axon of Loligo. 1952, Pubmed
Hoshi, Biophysical and molecular mechanisms of Shaker potassium channel inactivation. 1990, Pubmed , Xenbase
Hou, Inactivation of KCNQ1 potassium channels reveals dynamic coupling between voltage sensing and pore opening. 2017, Pubmed , Xenbase
Kalstrup, S4-S5 linker movement during activation and inactivation in voltage-gated K+ channels. 2018, Pubmed , Xenbase
Kanevsky, Determinants of voltage-dependent gating and open-state stability in the S5 segment of Shaker potassium channels. 1999, Pubmed , Xenbase
Lacroix, Intermediate state trapping of a voltage sensor. 2012, Pubmed , Xenbase
Ledwell, Mutations in the S4 region isolate the final voltage-dependent cooperative step in potassium channel activation. 1999, Pubmed , Xenbase
Lee, Two separate interfaces between the voltage sensor and pore are required for the function of voltage-dependent K(+) channels. 2009, Pubmed , Xenbase
Lee, A novel mutation in KCNA1 causes episodic ataxia without myokymia. 2004, Pubmed
Li-Smerin, A localized interaction surface for voltage-sensing domains on the pore domain of a K+ channel. 2000, Pubmed , Xenbase
Liu, KCNQ1 and KCNH2 mutations associated with long QT syndrome in a Chinese population. 2002, Pubmed
Long, Crystal structure of a mammalian voltage-dependent Shaker family K+ channel. 2005, Pubmed
Lu, Coupling between voltage sensors and activation gate in voltage-gated K+ channels. 2002, Pubmed , Xenbase
Lu, Ion conduction pore is conserved among potassium channels. 2001, Pubmed
Ostmeyer, Recovery from slow inactivation in K+ channels is controlled by water molecules. 2013, Pubmed
Pan, Structure of the human voltage-gated sodium channel Nav1.4 in complex with β1. 2018, Pubmed
Perozo, Gating currents from a nonconducting mutant reveal open-closed conformations in Shaker K+ channels. 1993, Pubmed
She, Structural insights into the voltage and phospholipid activation of the mammalian TPC1 channel. 2018, Pubmed
Stefani, Cut-open oocyte voltage-clamp technique. 1998, Pubmed , Xenbase
Sun, Cryo-EM Structure of a KCNQ1/CaM Complex Reveals Insights into Congenital Long QT Syndrome. 2017, Pubmed , Xenbase
Wu, Structure of the voltage-gated calcium channel Ca(v)1.1 at 3.6 Å resolution. 2016, Pubmed
Yang, How does the W434F mutation block current in Shaker potassium channels? 1997, Pubmed , Xenbase