Structure of the KvAP voltage-dependent K+ channel and its dependence on the lipid membrane

Abstract

Voltage-dependent ion channels gate open in response to changes in cell membrane voltage. This form of gating permits the propagation of action potentials. We present two structures of the voltage-dependent K(+) channel KvAP, in complex with monoclonal Fv fragments (3.9 A) and without antibody fragments (8 A). We also studied KvAP with disulfide cross-bridges in lipid membranes. Analyzing these data in the context of the crystal structure of Kv1.2 and EPR data on KvAP we reach the following conclusions: (i) KvAP is similar in structure to Kv1.2 with a very modest difference in the orientation of its voltage sensor; (ii) mAb fragments are not the source of non-native conformations of KvAP in crystal structures; (iii) because KvAP contains separate loosely adherent domains, a lipid membrane is required to maintain their correct relative orientations, and (iv) the model of KvAP is consistent with the proposal of voltage sensing through the movement of an arginine-containing helix-turn-helix element at the protein-lipid interface.

Fig. 1.

KvAP structures are in similar conformations with and without Fv fragments. (a) Four-fold averaged electron density map (1.0 σ) of KvAP at 8 Å calculated with (2 F_o - F_c) Fourier coefficients with the voltage sensor omitted. Two KvAP pores were used to calculate the map. The α-carbon traces of the KvAP structure (dark brown) were generated by superposition of the KvAP-Fv complex structure onto the KvAP pore molecular replacement solutions. (b) Contacts in the KvAP crystal. A layer of channel molecules is formed by lateral packing of voltage sensors with neighboring channel molecules within a layer and end-to-end packing between layers. Two channels (red), surrounded by black lines, define the unit cell. (c) Native sharpened 2-fold averaged electron density map (1.0 σ) of the KvAP-33H1 Fv complex at 3.9 Å calculated with F_o Fourier coefficients. Combined phases (single isomorphous replacement with anomalous scattering and partial molecular replacement phases) were used for map calculation. The α-carbon traces of the KvAP-Fv structure were colored dark brown. (d) Crystal packing in the KvAP-33H1 Fv crystal as viewed down the crystallographic 4-fold axis. Channels are colored blue, and Fv fragments are colored green. Four channels and 16 Fvs, surrounded by black lines, comprise a unit cell.

Fig. 2.

Structure of the KvAP-33HI complex and comparison with previous KvAP structures. (a) Stereo view of a single KvAP subunit from the side with extracellular solution above. S1-S4 helices are colored blue. (b) Stereo view of a single KvAP subunit of the KvAP-6E1 Fab complex (PDB ID code 1ORQ). (c) Comparison with the isolated voltage sensor structure. Stereo view of a superposition of the isolated voltage sensor (gold, PDB ID code 1ORS) and the KvAP-Fv complex (blue). The S2 helix (Leu-55 to Tyr-75) was used for superposition, and the S1 helix is not shown. Residues Asp-62, Asp-72, Arg-76, Glu-93, and Arg-133, which are important for channel function, are shown in ball-and-stick representation.

Fig. 3.

The structure of the S2 to S3 turn. (a) The turn connecting the S2 to S3 helices (green, Tyr-63 to Pro-95) from two different viewpoints. Tyr residues are shown as ball-and-stick representation. (b) Possible transformation of the voltage sensor upon detergent extraction. (Left) The structure of KvAP in a detergent micelle. S2, S3, and S4 helices are colored blue, and the turn connecting S2 to S3 helices is colored green. (Right) The putative conformation of KvAP in the membrane. Asterisk marks the pivot point of proposed voltage sensor rotation.

Fig. 4.

A model of KvAP based on the structure of Kv1.2. (a) A subunit of the Kv1.2 (PDB ID code 2A79) viewed from the side. The voltage sensor region (S1-S4) is colored blue, and the pore region (S5-S6) is colored gray. A queue of K⁺ ions (green spheres) is shown as a reference point for comparison with the KvAP model. (b) The KvAP model. The model was constructed by tilting S2-S4 helices of the KvAP structure (PDB ID code 2A0L), adjusting the S4-S5 linker to resemble the linker in the Kv1.2 structure (PDB ID code 2A79), folding the S1 helix similar to its position in the isolated voltage sensor (PDB ID code 1ORS), and repositioning the sensor slightly to account for EPR O₂ (lipid) accessibility (9).

Fig. 5.

Mapping EPR O₂ (lipid) accessibility data onto the voltage sensor region of the KvAP. (a) The S4-S5 linker (colored blue) of the KvAP model. The side chains of the S4-S5 linker shown as a ball-and-stick representation indicate the amphipathic nature of this linker region. (b and c) The EPR O₂ (lipid) accessibility data (9) mapped onto the voltage sensor structure. O₂ accessibility ranges from white (low accessibility) to red (high accessibility). Numbers in c indicate amino acid position, circles show positions where data are not available. (d) Western blot analysis of cross-linking. Membrane vesicles containing single or double Cys mutations (at numbered residues) in KvAP were subjected to air oxidation. The size of the covalent monomer, dimer, trimer, and tetramer are indicated. Residues 222 and 232 on the S6 helix were chosen as a control for cross-linking based on the crystal structure of KvAP. The 222C/232C cross-linked tetramer migrates slightly faster than the other cross-linked tetramers.

Fig. 6.

A model of the KvAP tetramer in the open conformation. The top-down view (a) and the side view (b) of the proposed model of the KvAP tetramer in the open conformation. Each subunit is colored blue, green, gold, and red. This model is the same as in Fig. 4b but it is shown as a tetramer to show the position of the voltage sensor relative to the pore.