Structural basis of gating modulation of Kv4 channel complexes

Modulation of voltage-gated potassium (Kv) channels by auxiliary subunits is central to the physiological function of channels in the brain and heart1,2. Native Kv4 tetrameric channels form macromolecular ternary complexes with two auxiliary β-subunits—intracellular Kv channel-interacting proteins (KChIPs) and transmembrane dipeptidyl peptidase-related proteins (DPPs)—to evoke rapidly activating and inactivating A-type currents, which prevent the backpropagation of action potentials1–5. However, the modulatory mechanisms of Kv4 channel complexes remain largely unknown. Here we report cryo-electron microscopy structures of the Kv4.2–DPP6S–KChIP1 dodecamer complex, the Kv4.2–KChIP1 and Kv4.2–DPP6S octamer complexes, and Kv4.2 alone. The structure of the Kv4.2–KChIP1 complex reveals that the intracellular N terminus of Kv4.2 interacts with its C terminus that extends from the S6 gating helix of the neighbouring Kv4.2 subunit. KChIP1 captures both the N and the C terminus of Kv4.2. In consequence, KChIP1 would prevent N-type inactivation and stabilize the S6 conformation to modulate gating of the S6 helices within the tetramer. By contrast, unlike the reported auxiliary subunits of voltage-gated channel complexes, DPP6S interacts with the S1 and S2 helices of the Kv4.2 voltage-sensing domain, which suggests that DPP6S stabilizes the conformation of the S1–S2 helices. DPP6S may therefore accelerate the voltage-dependent movement of the S4 helices. KChIP1 and DPP6S do not directly interact with each other in the Kv4.2–KChIP1–DPP6S ternary complex. Thus, our data suggest that two distinct modes of modulation contribute in an additive manner to evoke A-type currents from the native Kv4 macromolecular complex.


S6
T1-S1 linker  Fig. 5a), which was not observed in previous studies. In addition, the intracellular S6 helix interacts with the T1-S1 linker in the structure of the Kv4.2-KChIP1 complex ( Fig. 1b-d, Extended Data Fig. 5c-e). By contrast, the intracellular S6 helix bends at A419 in the structure of Kv4.2 alone, which results in a partial loss of interaction between the intracellular S6 helix and the T1-S1 linker (Fig. 1d, Extended Data Fig. 5d, e), suggesting a key mechanism of Kv4 gating modulation by KChIPs. The last 130 or so C-terminal amino acid residues of Kv4.2 (residues 496-630) are not resolved and are thus predicted to lack secondary structure (Fig. 1b, Extended Data Fig. 5a, f), suggesting their flexibility. As in the Kv1.2 structure 30,31 , the tetrameric T1 domain of Kv4.2 is located under the tetrameric channel pore domains at a distance of 25 Å-provided by the long T1-S1 linker and the long intracellular S6 helix-thus creating sufficient space for K + ions to laterally enter the channel pore (Fig. 1a, Extended Data Fig. 5a). However, it should be noted that, within the protomer of both Kv4.2 alone and the Kv4.2-KChIP1 complex, the topological relationship between the T1 and transmembrane domains is different from that in Kv1.2, owing to the distinct orientation of the T1-S1 linker following the T1 domain (Extended Data Fig. 5b, g-i). The Kv4-specific topology of the T1 domain would facilitate the proper interaction between the intracellular S6 helix and KChIP1 (Extended Data Fig. 5j).

Kv4.2-KChIP1 interaction
KChIP1s are laterally anchored next to the T1 domains of Kv4.2, consistent with the previous crystal structures of the Kv4.3 T1 domain-KChIP1 complex 26,27 (Fig. 2a-c). The N-terminal hydrophobic segment (A2-R35) of Kv4.2, referred to as the inactivation ball, was captured by KChIP1 (Fig. 2b), which may explain why Kv4.2 exhibits a closed inactivated (CSI) mechanism, rather than an open inactivated (OSI) mechanism like Kv1.2, as previously discussed for Kv4.3 26,27 . The present structure of the full-length Kv4.2-KChIP1 complex reveals that the C terminus of Kv4.2 tightly interacts with both KChIP1 and the N terminus of Kv4.2 (Figs. 1a, c, 2b, d-f). The C-terminal cytoplasmic S6 helix continuously extends from the transmembrane S6 helix and terminates at S450, which is localized at the bottom of the complex (Fig. 2a, b, Extended Data Fig. 5a). Although the residues from G451 to G471 are disordered, the following second cytoplasmic helix with a short loop (C-terminal segment: S472-D495) fits into the hydrophobic crevice formed by KChIP1 and the Kv4.2 N-terminal segment (A2-R35) from the neighbouring Kv4.2 subunit (Fig. 2b). In addition, the cytoplasmic S6 helix (around S450) is captured by KChIP1 directly and indirectly, through the hydrophobic interactions between Kv4.2 (Y444), Kv4.2 (H479-L482-F493) and KChIP1 (H84) and the electrostatic interactions of Kv4.2 (Y444-K448) with Kv4.2 (H483-E486), respectively (Fig. 2f). Together, these interactions suggest that KChIPs modulate the inactivation and recovery of the Kv4 channel by directly regulating S6 gating, and are consistent with a previous study that suggested that the Kv4 C-terminal region is involved in modulation by KChIPs 33 . The amino acid sequence of the C-terminal helix segment (S473-T489) perfectly matches the dendritic targeting motif that is conserved in the Shal family of potassium channels-including Kv4-suggesting that this motif has a dual function as a KChIP-binding site and a dendrite localization signal 34 .
To examine how the interaction of KChIP1 with the C terminus of Kv4 (S472-D495) affects Kv4 modulation, four alanine-substituted mutant versions of Kv4 were generated (F474A/H478A, H480A, L482A/L485A and H491A/F493A/V494A) on the basis of the hydrophobic interactions with KChIP1 (Fig. 2e). Using two-electrode voltage clamp (TEVC) recording in Xenopus oocytes, we assessed the effects of these mutations on activation, inactivation and recovery (Fig. 3 (Fig. 3a, Extended Data Fig. 7a). When expressed alone, all of the Kv4.2 mutants exhibited similar current-time traces to those of the wild type, and the H480A and H491A/F493A/V494A mutants exhibited slightly faster inactivation than the wild type (Fig. 3a, Extended Data Fig. 8 Supplementary Fig. 5c).
Finally, we assessed the effects of the mutations on recovery from inactivation, as KChIP1 reportedly accelerates the recovery from inactivation of Kv4s 24 . In the absence of KChIP1, all of the Kv4.2 C-terminal mutants exhibited quite similar recovery rates to that of the wild type ( Fig. 3b-e, Supplementary Fig. 5d, Extended Data Fig. 8). However, each mutant received a different modulatory effect on the recovery rate by KChIP1 (Fig. 3b-e, Supplementary Fig. 5d, Extended Data Fig. 8). KChIP1 accelerated the recovery rate of the L482A/L485A mutant, but more weakly compared to the wild type (Fig. 3d, Extended Data Fig. 8), whereas it did not affect the recovery rate of the H491A/F493A/V494A mutant (Fig. 3e, Extended Data Fig. 8). KChIP1 accelerated the recovery rate of the F474A/H478A and H480A mutants even more strongly than the wild type, together with an 'overshoot' current 36,37 (Fig. 3b, c, Extended Data Fig. 8). Altogether, these results indicate that the interaction of the Kv4.2 C-terminal segment with KChIP1 affects the gating modulation of Kv4.2.

Structures of Kv4.2-DPP6S and Kv4.2-DPP6S-KChIP1
DPP6 and DPP10 are single-pass transmembrane proteins with a large extracellular domain and a short intracellular segment 38,39 . DPPs reportedly accelerate the activation, inactivation and recovery of Kv4s 38,39 . DPPs modulate Kv4s through their single transmembrane helices and short intracellular segments 40,41 , suggesting that they have modulatory mechanisms that are distinct from those of KChIPs. To investigate how DPPs modulate the properties of Kv4, we solved the structures of the human Kv4.2-DPP6S binary and Kv4.2-DPP6S-KChIP1 ternary complexes (Fig. 1a, Supplementary Fig. 2c, d, Extended Data Table 1). During 3D classification with C1 symmetry, two different classes of structures were obtained, with two or four DPP6S molecules integrated in the complex (Extended Data Figs. 9, 10), which is consistent with the previous stoichiometric analysis of the Kv4-DPP complex 42 . The 3D classes that contained four DPP6S molecules were selected for further 3D refinement with C2 symmetry imposed, because two DPP6S dimers were integrated with C2 symmetry in the complexes (Extended Data Figs

Kv4.2-DPP6S interaction
In the structures of the Kv4.2-DPP6S and Kv4.2-DPP6S-KChIP1 complexes, the DPP6S transmembrane helix hydrophobically interacts with the voltage-sensing domain of Kv4.2, specifically at the lower half of S1 and the upper half of S2 (Fig. 4a, b). This is consistent with a previous domain-swapping study, which suggested that DPP10 interacts with S1 and/or S2 of Kv4.3 40 . Recently, two potassium channel structures (Kv7.1 and Slo1) in complex with a modulatory transmembrane β-subunit have been reported 44,45 . In both the Kv7.1-KCNE3 and Slo1-β4 complexes, the β-subunit associates with the transmembrane interface between neighbouring α-subunits (Extended Data Fig. 12). The structure of the Kv4.2-DPP6S complex therefore represents a distinct interaction mode among the potassium channel complexes reported thus far. The interaction of the Kv4.2-DPP6S complex somewhat resembles that of voltage-gated sodium channels, such as the Nav1.4-β1 and Nav1.7-β1 complexes, in which the β1 transmembrane helix interacts with S0 and S2 of Nav (Extended Data Fig. 12), and therefore their modulation mechanisms could be similar 46,47 . However, the specific involvement of S1 in the Kv4.2-DPP6S interaction suggests the unique modulatory mechanisms of Kv4. Although the side chains of the DPP6S transmembrane helix could not be easily assigned owing to the lack of characteristic density (Supplementary Figs. 6-8), the Kv4.2-DPP6S structure revealed seven hydrophobic residues in S1 and S2 of Kv4.2 that face and potentially interact with DPP6S (Fig. 4b). To examine the importance of these residues in the modulation of Kv4.2 by DPP6S, we generated a series of Kv4.2-DPP6S interface mutants by substituting each residue in S1-S2 with tryptophan residue to physically interfere with their potential interaction. When expressed alone, the wild type and all Kv4.2 S1-S2 mutants exhibited similar current-time traces and voltage-dependent activation curves. (Fig. 4c-e, Supplementary Fig. 9). As reported previously 38 , DPP6S accelerates activation and inactivation and also shifts the voltage-dependent activation curve to more negative membrane potentials ( Fig. 4c-e, Extended Data Figs. 8, 13a-g, Supplementary  Fig. 9). Although the quite rapid activation mediated by DPP6S made it difficult to evaluate the effects of the mutations on the activation kinetics, three mutants (V190W in S1; and A228W and C231W in S2) were inactivated more slowly than the wild type in the presence of DPP6S (Fig. 4c-e, Extended Data Fig. 8, Supplementary Fig. 9, Supplementary Table 1). In addition, in the presence of DPP6S these three mutants exhibited smaller negative voltage shifts for channel activation, as compared to the wild type (Extended Data Figs. 8, 13c, e, f). Kv4 . We next assessed the mutational effects on voltage-dependent inactivation (Extended Data Figs. 8, 13h-n, Supplementary Fig. 10). DPP6S shifted the inactivation curves of the wild type to the negative direction with the steeper voltage dependence, indicating relative stabilization of the inactivated state 38 (Extended Data Figs. 8, 13h-n). Five mutants (T182W, V186W, F194W, A228W and C231W) showed a similar negative voltage shift in the presence of DPP6S to that of the wild type (Extended Data Figs. 8, 13h, i, k-m). By contrast, DPP6S shifted the inactivation curves of V190W in S1 and A235W in S2 mutants to the positive direction (Extended Data Figs. 8, 13j, n), suggesting that the S1 and S2 helices of Kv4.2 are important for the modulation of steady-state inactivation by DPP6S.
DPP6S reportedly accelerates the recovery of Kv4.2 from inactivation 38 (Fig. 4f-h, Extended Data Figs. 8, 13o-u, Supplementary Fig. 11). However, the V190W and C231W mutants exhibited slower recovery rates than the wild type in the presence of DPP6S, even though the V190W mutant alone recovered faster than the wild type in the absence of DPP6S (Fig. 4g, h, Extended Data Fig. 8). The V186W mutant alone recovered faster than the wild type, which made it difficult to evaluate the effect of DPP6S on this mutant (Fig. 4f, Extended Data Fig. 8). Together, all these results indicate that DPP6S modulates the activation, inactivation and recovery of Kv4.2 through interactions with the S1 and S2 helices of the Kv4.2 voltage-sensing domain.

Conclusions
The structures we present here, combined with complementary electrophysiological analyses, suggest that KChIP1 stabilizes the S6 conformation to modulate synchronized and accelerated gating of the S6 helices within the tetramer, preventing N-type inactivation but promoting fast CSI and recovery. On the other hand, DPP6S may accelerate the voltage-dependent movement of the S4 helices by stabilizing the S1-S2 conformation. KChIP1 and DPP6S do not directly interact with each other, and they interact with distinct structures of Kv4.2 to modulate its gating kinetics in different manners. Therefore, our results suggest that these two distinct modes of modulation additively contribute to evoking A-type currents from the native Kv4 macromolecular complex by eliminating OSI, and accelerating CSI and fast recovery from CSI ('Discussion' in Methods).

Online content
Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41586-021-03935-z.

Article
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Data reporting
No statistical methods were used to predetermine sample size. The experiments were not randomized and the investigators were not blinded to allocation during experiments and outcome assessment.

Protein expression in Xenopus laevis oocytes
The human Kv4.2 (NP_036413.1; wild type and mutants), human KChIP1 (NP_055407.1; wild type), and human DPP6S (NP_001927.3; wild type) genes were cloned into the pGEMHE expression vector 52 . The cRNAs were transcribed using a mMESSAGE mMACHINE T7 Transcription Kit (Thermo Fisher Scientific). Oocytes were surgically taken from female Xenopus laevis anaesthetized in water containing 0.15% tricaine (Sigma-Aldrich, E10521) for 15-30 min. They were treated with collagenase (Sigma-Aldrich, C0130) for 6-7 h at room temperature to remove the follicular cell layer. Defolliculated oocytes of a similar size at stage V or VI were selected and microinjected with 50 nl of cRNA solution. They were then incubated for 1-2 days at 18 °C in MBSH buffer, containing 88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO 3 , 10 mM HEPES, 0.3 mM Ca(NO 3 ) 2 , 0.41 mM CaCl 2 , and 0.82 mM MgSO 4 , pH 7.6, supplemented with 0.1% penicillin-streptomycin solution (Sigma-Aldrich, P4333) 25,42 . All experiments were approved by the Animal Care Committee of Jichi Medical University and were performed following the institutional guidelines.

Two-electrode voltage clamp recordings
Ionic currents were recorded under two-electrode voltage clamp with an OC-725C amplifier (Warner Instruments) at room temperature. The bath chamber was perfused with ND-96 buffer, containing 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 and 5 mM HEPES, pH 7.5. The microelectrodes were drawn from borosilicate glass capillaries (Harvard Apparatus, GC150TF-10) using a P-1000 micropipette puller (Sutter Instrument) to a resistance of 0.2-0.5 MΩ and filled with 3 M KCl. Generation of voltage-clamp protocols and data acquisition were performed using a Digidata 1550 interface (Molecular Devices) controlled by the pClampex 10.7 software (Molecular Devices). Data were sampled at 10 kHz and filtered at 1 kHz by the pClampfit 10.7 software (Molecular Devices).

Data analysis
For the voltage-dependent activation. The holding potential was −80 mV. After 500 ms of hyperpolarization at −110 mV to remove inactivation, currents were elicited by 400-ms test pulses to membrane potentials from −80 to 40 mV with 10-mV increments. Conductance values were calculated from peak current amplitudes by normalizing to the maximum current amplitude obtained in the experiment, assuming a linear open channel current-voltage relationship and a reversal potential of −98 mV (normalized chord conductance). Normalized peak conductance was plotted versus voltage and fitted with single Boltzmann functions to estimate the half-activation voltage (V 1/2,act ) and the effective charge (z act ) in Extended Data Fig. 8.
Recovery from inactivation. The currents were elicited by a two-pulse protocol using the prepulse (500 ms) and the test pulses (100 ms) at 40 mV with an interpulse interval of the duration from 10 to 490 ms at −100 mV. The fractional recovery at each point was determined by normalizing the peak current amplitude of the test pulse by the amplitude of the prepulse and fitted with single exponential functions to estimate the recovery time constant (τ rec ) in Extended Data Fig. 8. For the Kv4.2 (F474A/H478A) with KChIP1 and Kv4.2 (H480A) with KChIP1 conditions, only data obtained using prepulses from 10 ms to 90 ms were used for single-exponential fits owing to reduced fractional recovery at longer prepulses.
Voltage-dependent prepulse inactivation. The holding potential was −100 mV. After 5 s of prepulses from −120 mV to 0 mV with 10-mV increments, currents were elicited by 250-ms test pulses at 60 mV. The fractional recovery at each point was determined by normalizing the peak current amplitude of the test pulse by the test pulse after the prepulse of −120 mV and fitted with single Boltzmann functions to estimate the half-inactivation voltage (V 1/2,inact ) and the effective charge (z inact ) in Extended Data Fig. 8.

Statistical analysis
The electrophysiological data were expressed as mean ± s.e.m. (n = 8). Differences between wild type and mutants, between wild type with KChIP1 and mutants with KChIP1, and between wild type with DPP6S and mutants with DPP6S were evaluated by Dunnett's test with EZR software 53 .

Discussion
Modulation by KChIP1. KChIPs reportedly prevent OSI and accelerate CSI and recovery from inactivation 11,24,35 (Fig. 3, Extended Data Fig. 7a, Supplementary Fig. 5). The structural comparison between Kv4.2 alone and the Kv4.2-KChIP1 complex provides insight into how KChIPs modulate the gating of Kv4s. In the Kv4.2-KChIP1 complex, KChIP1s bind and sequester the both N-terminal inactivation ball and the C terminus (amino acids 472-495) of Kv4.2, which would therefore result in preventing N-type inactivation. Moreover, while S6 gating helices adopt a more flexible conformation with weaker interaction with T1-S1 linkers in the structure of Kv4.2 alone, KChIP1 stabilizes these structures and enhances their interactions in the structure of Kv4.2-KChIP1. These structural changes mediated by KChIPs, together with the following three observations and reports, might explain how KChIPs accelerate the S6 gating of Kv4s, including CSI and recovery from inactivation. First, one KChIP1 stabilizes the S6 conformation as well as the N terminus from the neighbouring subunit of Kv4.2. Second, one KChIP1 also interacts with two T1 domains from neighbouring subunits 26,27 (Fig. 2b). Third, previous functional studies have suggested that the T1-S1 linker of Kv4 dodecameric channels undergoes major conformational shifts tightly coupled to movements of the S6 tail 54,55 , although we do not know what the T1 conformational change is. Together, these structural features mediated by KChIP1 may allow synchronized and accelerated S6 gating to enable fast CSI and recovery (Extended Data Fig. 14a).
Modulation by DPP6. DPP6S reportedly accelerates the activation, inactivation, and recovery of K4 channels 38 . In the Kv4.2-DPP6S complexes, the single-spanning transmembrane helix of DPP6S apparently stabilizes the structure of S1 and S2 helices because it simultaneously interacts with the lower half of S1 and the upper half of S2 (Fig. 4a, b). DPP6S reportedly accelerates both the outward and the inward movements of the Kv4.2 gating charge after depolarization and repolarization, respectively 28 . Among the hypotheses to explain the voltage dependency in voltage-gated channels, the hypothesis that S4 slides on the surface formed by S1 and S2 depending on the membrane potential might be most likely 13 .Therefore, the stabilization of the S1-S2 conformation may facilitate the movement of the S4 helices upon depolarization and repolarization, which could explain the fast kinetics of activation and recovery from the closed inactivated state (Extended Data Fig. 14b).
Previous studies suggest that DPP6S accelerates both OSI and CSI of Kv4s 39,56 (Extended Data Fig. 14b). The acceleration of OSI by DPP6S could involve the N-terminal intracellular domain of DPP6S and the N terminus of Kv4s 39 ; however, both regions are disordered in the structure of Kv4.2-DPP6S and further investigations are required. Previous studies suggest that the dynamic interaction of the S4-S5 linker and the S6 gate is the molecular basis of CSI 12,22 . Therefore, the acceleration of CSI by DPP6S could be, at least in part, attributed to the accelerated conformational change of S4 as discussed above (Extended Data Fig. 14b).

Modulation in the Kv4 macromolecular ternary complex. Native
Kv4s form macromolecular ternary complex with KChIPs and DPPs. The structure of the Kv4.2-DPP6S-KChIP1 dodecameric complex (Fig. 1a) supports the additive contribution of KChIPs and DPPs to the modulation of Kv4s in the ternary complex. KChIP1 and DPP6S interact with distinct structures of Kv4.2 to modulate its gating kinetics in different manners (Figs. 1a, 2, 4a). In addition, KChIP1 and DPP6S do not interact with each other. Overall, the modulatory mechanisms of Kv4.2 by KChIP1 and DPP6S are different, and therefore, native Kv4s form ternary macromolecular complexes with both KChIPs and DPPs to exhibit eliminated OSI, accelerated CSI and fast recovery rate from CSI 5 (Extended Data Fig. 14c). Structurally mechanistic elucidations of CSI will further clarify the mechanisms of modulation by KChIPs and DPPs.

Insight into closed-state inactivation of Kv4.2.
The structural correlates of Kv4 in closed-state inactivation (CSI) remain unknown. Previous studies have proposed that the interaction between the S4-S5 linker and S6 in Kv4s, which couples the S4 movement to S6 gating in Kv1, might be lost following the upshifted movement of S4 during depolarization 12,21,22 (Extended Data Fig. 1b). Indeed, the amino acid sequences of Kv4 around the S4-S5 linker and S6 on the intracellular side are unique among the Shaker-related Kv subfamilies (Kv1-Kv4) ( Supplementary Fig. 12a), and mutations of these regions affect the CSI kinetics of Kv4 21,22 (Supplementary Fig. 12b). In addition, the open conformation of Kv4.2 complexes revealed several Kv4-specific residues involved in the intra-subunit interactions between the S4-S5 linker and S6, as well as the inter-subunit interactions between the S4-S5 linker and S5 (Supplementary Fig. 12a, c, d). Further study of this 'pre-closing' conformation may lead to elucidating the mechanism of CSI. Together, future structural studies of the resting and closed inactivated states will provide more mechanistic insights into Kv4 channel gating, CSI and modulation by auxiliary subunits.

Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this paper.   Figs. 5a, b). The previous electrophysiological studies reported that upon depolarization, Kv4s adopt the closed conformation (i.e. CSI) at all physiologically relevant membrane potentials within a cell [11][12][13][14][15][16][17][18] (Extended Data Fig. 1). This discrepancy could be attributed to the micelle which is likely to facilitate the open conformation. Similar inconsistent example was observed in the cryo-EM structure of the HCN channel in a hyperpolarized conformation in which the pore is closed while it is open within a cell 57 .   Note that a single DPP6S interacts with S1-S2 of a single voltage-sensing domain (VSD), whereas KCNE3 and β4 interact with the interface between two neighbouring α subunits in the Kv7.1 and Slo1 complexes, respectively. The interaction of Nav1.7 and β1 is rather similar to that of Kv4.2-DPP6S, in that a single β subunit interacts with a single VSD. However, the interaction of Kv4.2 and DPP6S is unique, because S1 of Nav1.7 is not involved in the interaction with β1. Fig. 14

Data
Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A description of any restrictions on data availability -For clinical datasets or third party data, please ensure that the statement adheres to our policy  Table 1, Extended Data Fig. 3, 4 , 8, 9. Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design
All studies must disclose on these points even when the disclosure is negative.

Sample size
No statistical method was used to determine the sample size. For cryo-EM analyses, sample sizes were determined by the availability of microscope time and the number of particles on electron microscopy grids enough to obtain a structure at the reported resolution. For electrophysiological analyses, sample sizes were determined based on the previous reports of this type of study and the reproducibility of results across independent experiments.
Data exclusions For cryo-EM analyses, particles that did not contribute to improving map quality were excluded following the standard classification procedures in RELION. This is standard practice for structure determination by cryo-EM. For electrophysiological analyses, recordings that contain leak or endogenous currents were excluded. This is standard practice in electrophysiology.

Replication
For cryo-EM analyses, related experiments including FSEC, purification, and SDS-PAGE were reproduced at least two times and structure determination was completed once. For electrophysiological analyses, all data sets were pooled from at least two independent oocyte batches.
Randomization For cryo-EM analyses, particles were randomly assigned to half-maps for resolution determination following the standard procedures in RELION. For electrophysiological analyses, randomization was not performed since samples were not divided into two or more groups.

Blinding
For cryo-EM analyses, blinding was not applicable since this type of studies does not use group allocation. For electrophysiological analyses, blinding was not applied since it was not technically or practically feasible to do so.

Reporting for specific materials, systems and methods
We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response. HEK cells are listed in the register but it does not specify which type of HEK strains. Our secondary HEL293S GnTI-cell lines was purchased by from ATCC, where they validated.