Two different conformational states of the KirBac3.1 potassium channel revealed by electron crystallography

Potassium channels allow the selective flow of K(+) ions across membranes. In response to external gating signals, the potassium channel can move reversibly through a series of structural conformations from a closed to an open state. 2D crystals of the inwardly rectifying K(+) channel KirBac3.1 from Magnetospirillum magnetotacticum have been captured in two distinct conformations, providing "snap shots" of the gating process. Analysis by electron cryomicroscopy of these KirBac3.1 crystals has resulted in reconstructed images in projection at 9 A resolution. Kir channels are tetramers of four subunits arranged as dimers of dimers. Each subunit has two transmembrane helices (inner and outer). In one crystal form, the pore is blocked; in the other crystal form, the pore appears open. Modeling based on the KirBac1.1 (closed) crystal structure shows that opening of the ion conduction pathway could be achieved by bending of the inner helices and significant movements of the outer helices.     

Exploring models of the influenza A M2 channel: MD simulations in a phospholipid bilayer

The M2 protein of influenza A virus forms homotetrameric helix bundles, which function as proton-selective channels. The native form of the protein is 97 residues long, although peptides representing the transmembrane section display ion channel activity, which (like the native channel) is blocked by the antiviral drug amantadine. As a small ion channel, M2 may provide useful insights into more complex channel systems. Models of tetrameric bundles of helices containing either 18 or 22 residues have been simulated while embedded in a fully hydrated 1-palmitoyl-2-oleoyl-sn-glycerol-3-phosphatidylcholine bilayer. Several different starting models have been used. These suggest that the simulation results, at least on a nanosecond time scale, are sensitive to the exact starting structure. Electrostatics calculations carried out on a ring of four ionizable aspartate residues at the N-terminal mouth of the channel suggest that at any one time, only one will be in a charged state. Helix bundle models were mostly stable over the duration of the simulation, and their helices remained tilted relative to the bilayer normal. The M2 helix bundles form closed channels that undergo breathing motions, alternating between a tetramer and a dimer-of-dimers structure. Under these conditions either the channel forms a pocket of trapped waters or it contains a column of waters broken predominantly at the C-terminal mouth of the pore. These waters exhibit restricted motion in the pore and are effectively "frozen" in a way similar to those seen in previous simulations of a proton channel formed by a four-helix bundle of a synthetic leucine-serine peptide (, Biophys. J. 77:2400-2410).     

Structural models of the MscL gating mechanism.

Three-dimensional structural models of the mechanosensitive channel of large conductance, MscL, from the bacteria Mycobacterium tuberculosis and Escherichia coli were developed for closed, intermediate, and open conformations. The modeling began with the crystal structure of M. tuberculosis MscL, a homopentamer with two transmembrane alpha-helices, M1 and M2, per subunit. The first 12 N-terminal residues, not resolved in the crystal structure, were modeled as an amphipathic alpha-helix, called S1. A bundle of five parallel S1 helices are postulated to form a cytoplasmic gate. As membrane tension induces expansion, the tilts of M1 and M2 are postulated to increase as they move away from the axis of the pore. Substantial expansion is postulated to occur before the increased stress in the S1 to M1 linkers pulls the S1 bundle apart. During the opening transition, the S1 helices and C-terminus amphipathic alpha-helices, S3, are postulated to dock parallel to the membrane surface on the perimeter of the complex. The proposed gating mechanism reveals critical spatial relationships between the expandable transmembrane barrel formed by M1 and M2, the gate formed by S1 helices, and "strings" that link S1s to M1s. These models are consistent with numerous experimental results and modeling criteria.     

A structural model of the acetylcholine receptor channel based on partition energy and helix packing calculations.

A structural model of the transmembrane portion of the acetylcholine receptor was developed from sequences of all its subunits by using transfer energy calculations to locate transmembrane alpha-helices and to calculate which helical side chains should be in contact with water inside the channel, with portions of other transmembrane helices, or with lipid hydrocarbon chains. "Knobs-into-holes" side chain packing calculations were used with other factors to stack the transmembrane alpha-helices together. In the model each subunit has the following structures in order along the sequence from the NH2 terminus: a large extracellular domain of undetermined structure, a short apolar alpha-helix that lies on the extracellular lipid surface of the membrane; three apolar transmembrane alpha-helices (I, II, and III), a cytoplasmic domain of undetermined structure, an amphipathic transmembrane alpha-helix (L) that forms the channel lining, a short extracellular alpha-helix, another apolar transmembrane alpha-helix (IV), and a small cytoplasmic domain formed by the COOH-terminal end of the chain. Three concentric layers form the pore. A bundle of five amphipathic L helices forms the channel lining. This bundle is surrounded by a bundle of 10 alternating II and III helices. Helices I and IV cover portions of the outer surface of the bundle formed by helices II and III. Positions of disulfide bridges are predicted and a mechanism for opening and closing conformational changes is proposed that requires tilting transmembrane helices and possibly a thiol-disulfide interchange reaction.     

The peptaibol antiamoebin as a model ion channel: similarities to bacterial potassium channels.

Antiamoebin (AAM) is a polypeptide antibiotic that is capable of forming ion channels in phospholipid membranes: planar bilayer studies have suggested the channels are octamers. The crystal structure of a monomeric form of AAM has provided the basis for molecular modelling of an octameric helical bundle channel. The channel model is funnel-shaped due to a substantial bend in the middle of the polypeptide chain caused by the presence of several imino acids. Inter-monomer hydrogen bonds orientate a ring of glutamine side chains to form a constriction in the pore lumen. The channel lumen is lined both by side chains of Gln11 and by polypeptide backbone carbonyl groups. Electrostatic calculations on the model are compatible with a channel that transports cations across membranes. The AAM channel model is compared with the crystal structures of two bacterial (KcsA andMthK) potassium channels. AAM and the potassium channels exhibit common functional features, such as cation-selectivity and similar single channel conductances. Common structural features include being multimers, each formed from a bundle of eight transmembrane helices, with lengths roughly comparable to the thickness of lipid bilayers. In addition, they all have aromatic amino acids that lie at the bilayer interfaces and which may aid in the stabilization of the transmembrane helices, as well as narrower constrictions that define the ion binding sites or selectivity filters in the pore lumen. The commonality of structural and functional features in these channels thus suggests that antiamoebin is a good, simple model for more complex bacterial and eukaryotic ion channels, capable of providing insight into details of the mechanisms of ion transport and multimeric channel stability.     

Structural Model of the CaV1.2 Pore

Understanding the structure and functional mechanisms of voltage-gated calcium channels remains a major task in membrane biophysics. In the absence of three dimensional structures, homology modelling techniques are the method of choice, to address questions concerning the structure of these channels. We have developed models of the open Cav1.2 pore, based on the crystal structure of the mammalian voltage-gated potassium channel Kv1.2 and a model of the bacterial sodium channel NaChBac. Our models are developed to be consistent with experimental data and modelling criteria. The models highlight major differences between voltage-gated potassium and calcium channels, in the P segments, as well as the inner pore helices. Molecular dynamics simulations support the hypothesis of a clockwise domain arrangement and experimental observations of asymmetric calcium channel behaviour. In the accompanying paper these models were used to study structural effects of a channelopathy mutation.     

Structural model of the Ca V 1.2 pore

Understanding the structure and functional mechanisms of voltage-gated calcium channels remains a major task in membrane biophysics. In the absence of three dimensional structures, homology modeling techniques are the method of choice, to address questions concerning the structure of these channels. We have developed models of the open Ca(V)1.2 pore, based on the crystal structure of the mammalian voltage-gated potassium channel K(V)1.2 and a model of the bacterial sodium channel NaChBac. Our models are developed to be consistent with experimental data and modeling criteria. The models highlight major differences between voltage-gated potassium and calcium channels in the P segments, as well as the inner pore helices. Molecular dynamics simulations support the hypothesis of a clockwise domain arrangement and experimental observations of asymmetric calcium channel behavior. In the accompanying paper these models were used to study structural effects of a channelopathy mutation.     

The relationship between amide proton chemical shifts and secondary structure in proteins

Summary The parameters for HN chemical shift calculations of proteins have been determined using data from high-resolution crystal structures of 15 proteins. Employing these chemical shift calculations for HN protons, the observed secondary structure chemical shift trends of HN protons, i.e., upfield shifts on helix formation and downfield shifts on -sheet formation, are discussed. Our calculations suggest that the main reason for the difference in NH chemical shifts in helices and sheets is not an effect from the directly hydrogen-bonded carbonyl, which gives rise to downfield shifts in both cases, but arises from an additional upfield shift predicted in helices and originating in residues i-2 and i-3. The calculations also explain the well-known relationship between amide proton shifts and hydrogen-bond lengths. In addition, the HN chemical shifts of the distorted amphipathic helices of the GCN4 leucine zipper are calculated and used to characterise the solution structure of the helices. By comparing the calculated and experimental shifts, it is shown that in general the agreement is good between residues 15 and 28. The most interesting observation is that in the N-terminal half of the zipper, although both calculated and experimental shifts show clear periodicity, they are no longer in phase. This suggests that for the N-terminal half, in the true average solution structure the period of the helix coil is longer by roughly one residue compared to the NMR structures.     

Closing in on the resting state of the Shaker K(+) channel

Membrane depolarization causes voltage-gated ion channels to transition from a resting/closed conformation to an activated/open conformation. We used voltage-clamp fluorometry to measure protein motion at specific regions of the Shaker Kv channel. This enabled us to construct new structural models of the resting/closed and activated/open states based on the Kv1.2 crystal structure using the Rosetta-Membrane method and molecular dynamics simulations. Our models account for the measured gating charge displacement and suggest a molecular mechanism of activation in which the primary voltage sensors, S4s, rotate by approximately 180 degrees as they move "outward" by 6-8 A. A subsequent tilting motion of the S4s and the pore domain helices, S5s, of all four subunits induces a concerted movement of the channel's S4-S5 linkers and S6 helices, allowing ion conduction. Our models are compatible with a wide body of data and resolve apparent contradictions that previously led to several distinct models of voltage sensing.     

Analysis and evaluation of channel models: simulations of alamethicin

Alamethicin is an antimicrobial peptide that forms stable channels with well-defined conductance levels. We have used extended molecular dynamics simulations of alamethicin bundles consisting of 4, 5, 6, 7, and 8 helices in a palmitoyl-oleolyl-phosphatidylcholine bilayer to evaluate and analyze channel models and to link the models to the experimentally measured conductance levels. Our results suggest that four helices do not form a stable water-filled channel and might not even form a stable intermediate. The lowest measurable conductance level is likely to correspond to the pentamer. At higher aggregation numbers the bundles become less symmetrical. Water properties inside the different-sized bundles are similar. The hexamer is the most stable model with a stability comparable with simulations based on crystal structures. The simulation was extended from 4 to 20 ns or several times the mean passage time of an ion. Essential dynamics analyses were used to test the hypothesis that correlated motions of the helical bundles account for high-frequency noise observed in open channel measurements. In a 20-ns simulation of a hexameric alamethicin bundle, the main motions are those of individual helices, not of the bundle as a whole. A detailed comparison of simulations using different methods to treat long-range electrostatic interactions (a twin range cutoff, Particle Mesh Ewald, and a twin range cutoff combined with a reaction field correction) shows that water orientation inside the alamethicin channels is sensitive to the algorithms used. In all cases, water ordering due to the protein structure is strong, although the exact profile changes somewhat. Adding an extra 4-nm layer of water only changes the water ordering slightly in the case of particle mesh Ewald, suggesting that periodicity artifacts for this system are not serious.     

Computational modelling of the open-state Kv1.5 ion channel block by bupivacaine

Binding of R(+)-bupivacaine to open-state homology models of the mammalian K_v1.5 membrane ion channel is studied using automated docking and molecular dynamics (MD) methods. Homology models of K_v1.5 are built using the 3D structures of the KcsA and MthK channels as a template. The packing of transmembrane (TM) α-helices in the KcsA structure corresponds to a closed channel state. Opening of the channel may be reached by a conformational transition yielding a bent structure of the internal S6 helices. Our first model of the K_v open state involves a PVP-type of bending hinge in the internal helices, while the second model corresponds to a Gly-type of bending hinge as found in the MthK channel. Ligand binding to these models is probed using the common local anaesthetic bupivacaine, where blocker binding from the intracellular side of the channel is considered. Conformational properties and partial atomic charges of bupivacaine are determined from quantum mechanical HF/6-31G* calculations with inclusion of solvent effects. The automated docking and MD calculations for the PVP-bend model predict that bupivacaine could bind either in the central cavity or in the PVP region of the channel pore. Linear interaction energy (LIE) estimates of the binding free energies for bupivacaine predict strongest binding to the PVP region. Surprisingly, no binding is predicted for the Gly-bend model. These results are discussed in light of electrophysiological data which show that the K_v1.5 channel is unable to close when bupivacaine is bound.     

Computational modelling of the open-state Kv 1.5 ion channel block by bupivacaine

Binding of R(+)-bupivacaine to open-state homology models of the mammalian K(v)1.5 membrane ion channel is studied using automated docking and molecular dynamics (MD) methods. Homology models of K(v)1.5 are built using the 3D structures of the KcsA and MthK channels as a template. The packing of transmembrane (TM) alpha-helices in the KcsA structure corresponds to a closed channel state. Opening of the channel may be reached by a conformational transition yielding a bent structure of the internal S6 helices. Our first model of the K(v) open state involves a PVP-type of bending hinge in the internal helices, while the second model corresponds to a Gly-type of bending hinge as found in the MthK channel. Ligand binding to these models is probed using the common local anaesthetic bupivacaine, where blocker binding from the intracellular side of the channel is considered. Conformational properties and partial atomic charges of bupivacaine are determined from quantum mechanical HF/6-31G* calculations with inclusion of solvent effects. The automated docking and MD calculations for the PVP-bend model predict that bupivacaine could bind either in the central cavity or in the PVP region of the channel pore. Linear interaction energy (LIE) estimates of the binding free energies for bupivacaine predict strongest binding to the PVP region. Surprisingly, no binding is predicted for the Gly-bend model. These results are discussed in light of electrophysiological data which show that the K(v)1.5 channel is unable to close when bupivacaine is bound.     

Modeling the ion channel structure of cecropin.

Atomic-scale computer models were developed for how cecropin peptides may assemble in membranes to form two types of ion channels. The models are based on experimental data and physiochemical principles. Initially, cecropin peptides, in a helix-bend-helix motif, were arranged as antiparallel dimers to position conserved residues of adjacent monomers in contact. The dimers were postulated to bind to the membrane with the NH2-terminal helices sunken into the head-group layer and the COOH-terminal helices spanning the hydrophobic core. This causes a thinning of the top lipid layer of the membrane. A collection of the membrane bound dimers were then used to form the type I channel structure, with the pore formed by the transmembrane COOH-terminal helices. Type I channels were then assembled into a hexagonal lattice to explain the large number of peptides that bind to the bacterium. A concerted conformational change of a type I channel leads to the larger type II channel, in which the pore is formed by the NH2-terminal helices. By having the dimers move together, the NH2-terminal helices are inserted into the hydrophobic core without having to desolvate the charged residues. It is also shown how this could bring lipid head-groups into the pore lining.     

Identification of Channel-lining Amino Acid Residues in the Hydrophobic Segment of Colicin Ia

Colicin Ia is a bactericidal protein of 626 amino acid residues that kills its target cell by forming a channel in the inner membrane; it can also form voltage-dependent channels in planar lipid bilayer membranes. The channel-forming activity resides in the carboxy-terminal domain of ~177 residues. In the crystal structure of the water-soluble conformation, this domain consists of a bundle of 10 α-helices, with eight mostly amphipathic helices surrounding a hydrophobic helical hairpin (helices H8-H9). We wish to know how this structure changes to form a channel in a lipid bilayer. Although there is evidence that the open channel has four transmembrane segments (H8, H9, and parts of H1 and H6-H7), their arrangement relative to the pore is largely unknown. Given the lack of a detailed structural model, it is imperative to better characterize the channel-lining protein segments. Here, we focus on a segment of 44 residues (573–616), which in the crystal structure comprises the H8-H9 hairpin and flanking regions. We mutated each of these residues to a unique cysteine, added the mutant colicins to the cis side of planar bilayers to form channels, and determined whether sulfhydryl-specific methanethiosulfonate reagents could alter the conduction of ions through the open channel. We found a pattern of reactivity consistent with parts of H8 and H9 lining the channel as α-helices, albeit rather short ones for spanning a lipid bilayer (12 residues). The effects of the reactions on channel conductance and selectivity tend to be greater for residues near the amino terminus of H8 and the carboxy terminus of H9, with particularly large effects for G577C, T581C, and G609C, suggesting that these residues may occupy a relatively constricted region near the cis end of the channel.     

A large iris-like expansion of a mechanosensitive channel protein induced by membrane tension

MscL, a bacterial mechanosensitive channel of large conductance, is the first structurally characterized mechanosensor protein. Molecular models of its gating mechanisms are tested here. Disulfide crosslinking shows that M1 transmembrane alpha-helices in MscL of resting Escherichia coli are arranged similarly to those in the crystal structure of MscL from Mycobacterium tuberculosis. An expanded conformation was trapped in osmotically shocked cells by the specific bridging between Cys 20 and Cys 36 of adjacent M1 helices. These bridges stabilized the open channel. Disulfide bonds engineered between the M1 and M2 helices of adjacent subunits (Cys 32-Cys 81) do not prevent channel gating. These findings support gating models in which interactions between M1 and M2 of adjacent subunits remain unaltered while their tilts simultaneously increase. The MscL barrel, therefore, undergoes a large concerted iris-like expansion and flattening when perturbed by membrane tension.     

Coupled motions between pore and voltage-sensor domains: a model for Shaker B, a voltage-gated potassium channel

A high-resolution crystal structure of KvAP, an archeabacterial voltage-gated potassium (Kv) channel, complexed with a monoclonal Fab fragment has been recently determined. Based on this structure, a mechanism for the activation (opening) of Kv channels has been put forward. This mechanism has since been criticized, suggesting that the resolved structure is not representative of the family of voltage-gated potassium channels. Here, we propose a model of the transmembrane domain of Shaker B, a well-characterized Kv channel, built by homology modeling and docking calculations. In this model, the positively charged S4 helices are oriented perpendicular to the membrane and localized in the groove between segments S5 and S6 of adjacent subunits. The structure and the dynamics of the full atomistic model embedded in a hydrated lipid bilayer were investigated by means of two large-scale molecular dynamics simulations under transmembrane-voltage conditions known to induce, respectively, the resting state (closed) and the activation (opening) of voltage-gated channels. Upon activation, the model undergoes conformational changes that lead to an increase of the hydration of the charged S4 helices, correlated with an upward translation and a tilting of the latter, concurrently with movements of the S5 helices and the activation gate. Although small, these conformational changes ultimately result in an alteration of the ion-conduction pathway. Our findings support the transporter model devised by Bezanilla and collaborators, and further underline the crucial role played by internal hydration in the activation of the channel.     

Ion channels formed by amphipathic helical peptides

Channel forming peptides (CFPs) are amphipathic peptides, of length ca. 20 residues, which adopt an -helical conformation in the presence of lipid bilayers and form ion channels with electrophysiological properties comparable to those of ion channel proteins. We have modelled CFP channels as bundles of parallel trans-bilayer helices surrounding a central ion-permeable pore. Ion-channel interactions have been explored via accessible surface area calculations, and via evaluation of changes in van der Waals and electrostatic energies as a K⁺ ion is translated along the length of the pore. Two CFPs have been modelled: (a) zervamicin-A1-16, a synthetic apolar peptaibol related to alamethicin, and (b) -toxin from Staphylococcus aureus. Both of these CFPs have previously been shown to form ion channels in planar lipid bilayers, and have been shown to have predominantly helical conformations. Zervamicin-A1-16 channels were modelled as bundles of 4 to 8 parallel helices. Two related helix bundle geometries were explored. K⁺channel interactions have been shown to involve exposed backbone carbonyl oxygen atoms. -Toxin channels were modelled as bundles of 6 parallel helices. Residues Q3, D11 and D18 generate favourable K⁺-channel interactions. Rotation of W15 about its C-C bond has been shown to be capable of occluding the central pore, and is discussed as a possible model for sidechain conformational changes in relation to ion channel gating.     

The closed structure of the MscS mechanosensitive channel. Cross-linking of single cysteine mutants

Mechanosensitive channels must make a large conformational change during the transition from the closed to the open state. The crystal structure of the open form of the Escherichia coli MscS channel was recently solved and depicts a homoheptamer (1). In this study, cross-linking of site-specific cysteine substitutions demonstrates that residues up to 10-33 A apart in the crystal structure readily form disulfide bridges in the closed form and can also be cross-linked by a 10-A linker. Cross-linking between adjacent subunits stabilizes the heptameric form of the channel providing biochemical evidence to support the crystal structure. The data are consistent with the published model (1) in that the membrane domain is highly flexible and that the closed to open transition may involve a significant displacement of transmembrane helices 1 and 2, possibly by as much as 30 A. The data are also consistent with significant flexibility of the cytoplasmic domain.     

Three-dimensional models of non-NMDA glutamate receptors.

Structural models have been produced for three types of non-NMDA inotropic glutamate receptors: an AMPA receptor, GluR1, a kainate receptor, GluR6; and a low-molecular-weight kainate receptor from goldfish, GFKAR alpha. Modeling was restricted to the domains of the proteins that bind the neurotransmitter glutamate and that form the ion channel. Model building combined homology modeling, distance geometry, molecular mechanics, interactive modeling, and known constraints. The models indicate new potential interactions in the extracellular domain between protein and agonists, and suggest that the transition from the "closed" to the "open" state involves the movement of a conserved positive residue away from, and two conserved negative residues into, the extracellular entrance to the pore upon binding. As a first approximation, the ion channel domain was modeled with a structure comprising a central antiparallel beta-barrel that partially crosses the membrane, and against which alpha-helices from each subunit are packed; a third alpha-helix packs against these two helices in each subunit. Much, but not all, of the available data were consistent with this structure. Modifying the beta-barrel to a loop-like topology produced a model consistent with available data.     

Structural studies on delta(3)-delta(2)-enoyl-CoA isomerase: the variable mode of assembly of the trimeric disks of the crotonase superfamily

Subunits of the enzymes in the crotonase superfamily form tight trimeric disks. In most members of this protein superfamily these disks assemble further into hexamers. Here we report on the 2.1 A structure of a tight hexameric crystal form of the yeast peroxisomal delta(3)-delta(2)-enoyl-CoA isomerase (Eci1p). A comparison of this structure to a previously solved crystal form of Eci1p and other structures of this superfamily shows that there is much variability with respect to the relative distance between the disks and their relative orientations. In particular helices H2 and H9 are involved in the inter-trimer contacts and there are considerable structural differences in these helices in this superfamily. Helices H2 and H9 are near the catalytic cavity and it is postulated that the observed structural variability of these helices, stabilized by the different modes of assembly, has allowed the evolution of the wide range of substrate and catalytic specificity within this enzyme superfamily.