Metal bridges to probe membrane ion channel structure and function

Introduction

Membrane proteins, including receptors, transporters, pumps, and ion channels, are involved in almost all aspects of communication between individual cells and their environment. However, because these proteins exist embedded in the lipid membrane, their structure is difficult to study directly using techniques such as crystallography. Furthermore, membrane proteins have to undergo substantial conformational changes as they transmit signals between the extracellular environment and the interior of the cell, and the nature and extent of these dynamic conformational changes are not apparent from static structural images. For these reasons, a wealth of alternative approaches have been developed to study the dynamic aspects of the structure and function of membrane proteins in their native, membrane-embedded environment, using functional readouts that are sensitive to changes in protein conformation. Several of these approaches take advantage of the unique chemistry of the cysteine side-chain among the natural amino acids that make up proteins. The cysteine side-chain contains a thiol group (-SH) that can be readily deprotonated to form a reactive thiolate anion (-S^-) that is capable of nucleophilic attack on electrophilic centers. Under oxidizing conditions, this leads to the rapid formation of disulfide bonds with other nearby sulfur-containing groups. In native proteins this process leads to the redox-sensitive formation of inter- or intra-molecular disulfide bonds between pairs of cysteine side-chains (1, 2). It can also be used to label covalently cysteine side-chains in proteins in situ with designer probes containing reactive sulfur groups such as thiosulfonates. Polarizable, anionic cysteine thiols also bind a number of metal ions, such as Cu²⁺, Zn²⁺, Cd²⁺, Hg⁺, and Ag⁺(3–7). Coordination of the metal ligand resulting from simultaneous interaction with multiple cysteine side-chains results in increased affinity metal binding, effectively forming a high-affinity metal ‘bridge’ between two or more (potentially up to four) cysteine side-chains (3–5).

Cysteine-specific probes are used in combination with site-directed mutagenesis to introduce a cysteine side-chain into any part of a protein. At the two-dimensional level, reactivity of the cysteine side-chain can then be used to identify the location of amino acids in any part of the primary sequence, their relative accessibility from different sides of the membrane, and changes in accessibility during protein conformational changes. At the three-dimensional level, the proximity of different amino acids can be ascertained using substances that can generate cross-links or bridges between two cysteine side-chains introduced into two different parts of the protein. Again, changes in amino acid proximity during conformational changes in the protein can give valuable insights into the nature and extent of the protein movements that are involved in these conformational changes. These techniques have been used most extensively to study the structure and function of ion channels, the function of which is relatively accessible using electrophysiological techniques to measure the tiny current that flows across the membrane carried by these channel proteins; the current review is focused on these channel proteins. However, in principle these techniques can be used to study any membrane protein for which some aspect of structure or function can be monitored experimentally (8–17).

Two-dimensional information: substituted cysteine accessibility mutagenesis

The technique of substituted cysteine accessibility mutagenesis (SCAM) was first developed to study the structure of ion channels (18, 19) and has since been applied to almost all channel types as well as many other classes of proteins. This technique is based on the covalent reaction between the cysteine thiol and the thiosulfonate group of small, hydrophilic molecules, and rests on the assumption that this covalent modification will lead to some change in protein function that can be easily monitored. The most commonly used probes are small methanethiosulfonate (MTS) reagents such as the negatively charged [2-sulfonatoethyl] methanethiosulfonate (MTSES), the positively charged [2-(trimethylammonium)ethyl] methanethiosulfonate (MTSET), and the neutral/partially positively charged [2-aminoethyl] methanethiosulfonate (MTSEA) (20). Because these thiosulfonates are highly hydrophilic, and also because they react at least 5×10⁹ times faster with the ionized (-S^-) form of the cysteine thiolate than the uncharged (-SH) form, it is assumed that any reaction will be with cysteine side-chains that are on a water-accessible surface of the protein. When the reaction is with a cysteine in the membrane-spanning part of a transmembrane protein such as an ion channel, it is most likely that this water-accessible surface is the lining of the permeation pathway or some other water-filled crevice in the protein (Figure 1B). Cysteine side-chains can be introduced into any part of the channel protein using site-directed mutagenesis. In this way, channel pore-lining residues can be identified by sequentially ‘scanning’ membrane-spanning parts of the protein, which is to say that each amino acid is mutated, one at a time, to cysteine, and its accessibility to MTS reagents tested using functional approaches. Assuming that covalent modification of a pore-lining cysteine side-chains by a small charged reagent will alter channel function in some way, electrophysiological techniques can then be used to discriminate pore-lining (reactive) from non-pore-lining (non-reactive) cysteine side-chains, with the caveat that a negative result could reflect either non-reactivity or reactivity that is without measureable functional consequence. MTS reagents react rapidly with the cysteine thiol; the rate of reaction between these reagents and the thiol of β-mercaptoethanol (β-ME) in free solution is >10 000 m^-1 s^-1 (20). However, the measured reaction rate with cysteine side-chains in ion channels may be considerably less due to restricted access from the bulk solution to the channel pore. Furthermore, where the introduced cysteine is in a part of the protein that undergoes significant conformational changes, the rate of modification may be different in different conformations (Figure 1A,B). Because MTS reagents form a stable disulfide bond with the cysteine thiol, the functional effects of MTS reagents are expected to be sustained even after these reagents have been washed from the experimental chamber; however, they can usually be reversed using reducing agents such as dithiothreitol (DTT) or β-ME to break the disulfide bond. To avoid the possibility of reaction with native cysteines, SCAM studies should be carried out using functional versions of proteins in which all endogenous cysteine residues have been replaced by other amino acids using site-directed mutagenesis (so-called cys-less protein backgrounds). However, in practice, many published ion channel SCAM studies use wild-type backgrounds including some or all native cysteines, so long as the wild-type background is found to be non-responsive to application of MTS reagents. Use of a cys-less protein background simplifies interpretation of the experimental results, since an introduced cysteine is the only possible target of cysteine-reactive reagents, but it also has its own potential pitfalls. Most importantly, removal of endogenous cysteines can impact the function of the protein. For example, in the cystic fibrosis transmembrane conductance regulator (CFTR) Cl^- channel, removal of the 18 endogenous cysteine residues to create a cys-less protein severely compromises protein trafficking to the cell membrane (21), necessitating maneuvers such as additional mutations and growing transfected cells at reduced temperature (~27°C) for functional studies on cys-less CFTR expressed in mammalian cells (22). Cys-less CFTR also has a slightly (15%–20%) elevated single-channel conductance, which was shown to be due to substitution of a single cysteine residue (C343) in a membrane-spanning part of the protein (23). Furthermore, introducing cysteines into different parts of the protein can also alter channel function, although such functional changes are often ignored if they do not prevent SCAM investigation. Again taking CFTR as an example, individual cysteine residues introduced into a cys-less background have been shown to cause complete protein mis-trafficking (even in cells grown in culture at 27°C) (24), to drastically reduce single-channel conductance (25, 26), and to modify channel gating (24). Other practical aspects of SCAM for the study of ion channels and other proteins have been described in detail previously (8–12, 20).

Figure 1:

Conceptual overview of the ways in which cysteine chemistry is used to probe conformational changes in ion channel proteins.

As an example, this overview considers a conformational change that opens and closes the channel pore. (A) Channel opening is controlled by a localized ‘gate’, in this example located in the narrow central region of the pore; although it is noted that in some channel types, the gate is not located in the narrow region. (B) Large probes such as MTS reagents may access wide, outer parts of the pore in one conformation only (in this case, in the closed state) to react covalently with an introduced cysteine side-chain (cys). (C) Smaller probes such as Cd²⁺ are expected to have access to all parts of the pore and therefore should report conformational changes that are relevant to the permeant ion (in this case, opening of the channel gate). (D) Two cysteine side-chains are close enough together to form a disulfide bond (blue line) only in the open state, effectively ‘locking’ the channel in this state. (E) Two cysteine side-chains are close enough together to form a metal bridge only in the open state, effectively stabilizing the channel in this state in the presence of Cd²⁺ ions.

MTS reagents are excellent probes for cysteines located on the surface of proteins, including relatively accessible regions close to the wide entrances to aqueous ion channel pores (Figure 1B). However, channel pores usually narrow to a region that is not much wider than the permeant ion (usually a few angstroms), and these narrow regions may be inaccessible to large MTS reagents (Figure 1B): for example, MTSES, MTSET, and MTSEA fit into a cylinder approximately 6 Å in diameter and 10 Å in length (20). Indeed, MTS reagents are usually considered to be impermeant in ion channel pores, implying that there are parts of the permeation pathway that cannot be accessed by these reagents. For these reasons, smaller probes have also been used extensively to test the accessibility of cysteine side-chains introduced into ion channel pores. Small transition metal ions also bind to cysteine side-chains, albeit reversibly, a fact that is reflected in nature in the use of cysteine side-chains in the binding sites for metals such as Cu²⁺, Zn²⁺, and Cd²⁺ in many metalloproteins (3–5, 27). The ionic radii of some of these metal ions (e.g. Cd²⁺, 1.09 Å; Zn²⁺, 0.88 Å; Ag⁺, 1.29 Å) are similar to those of biologically relevant ion channel cationic substrates (K⁺, 1.52 Å; Na⁺, 1.16 Å; Ca²⁺, 1.14 Å) (28). As such, these thiophilic metal ions are expected to have access to all parts of ion channel pores that are visited by the natural permeant ion (Figure 1C), and Cd²⁺, Zn²⁺, and Ag⁺have all been used as probes of cysteine side-chains introduced into different cation channel pores (29–32). One major difference between the use of metal ions and MTS reagents is that the interaction between metal ions and cysteine side-chains is expected to be reversible, and so the functional effects of metal ions should be reversed once these ions have been washed from the experimental chamber.

Because the dimensions of these thiophilic metal ion probes closely match those of the natural substrate, it is to be expected that they will be subject to similar restrictions on accessibility to different parts of the pore as the channel transitions between different open and closed conformational states (Figure 1C). As a result, these small probes have been used successfully to identify the location of ion channel ‘gates’ that open and close to allow or prevent the flow of ionic current and so control the overall function of K⁺channels (33–36), hyperpolarization-activated (HCN) channels (37), cyclic nucleotide-gated channels (38), TRPV1 channels (39), and purinergic P2X receptor channels (40, 41). State-dependent interactions of metal ions with introduced cysteines may also probe other conformational changes occurring within ion channel pores, such as those that control channel inactivation (42–45).

Thiophilic metal ions such as Cd²⁺ have also been used to identify accessible cysteine side-chains introduced into narrow regions of anion-selective channel pores (46). In these channels, however, cationic metal ions cannot be described as appropriate substitutes for the normal anionic permeant anions such as Cl^- and HCO₃^-. Small anions Au(CN)₂^- and Ag(CN)₂^- have been shown to interact covalently with cysteine side-chains introduced into the pore of the CFTR Cl^- channel by a process of ligand exchange (47, 48). Because these small monovalent anions are highly permeant in CFTR (49, 50), they have been used to probe cysteines introduced throughout the pore region (48, 51–55). These small anions have also been used to probe the location of the channel ‘gate’ that opens and closes the CFTR Cl^- channel (53, 55, 56) as well as other conformational changes in the CFTR pore (51, 54). Because these reactive anions are able to make a covalent reaction with cysteine side-chains, their effects should be irreversible on washing; however, the covalent reaction can be reversed using DTT or excess CN^- (47).

Three-dimensional information: cysteine cross-linking

The reactivity of the cysteine thiol group allows individual cysteine side-chains to make covalent attachments to other, nearby cysteine side-chains via reversible disulfide bond formation (1, 2). Because disulfide bond formation and breakage is highly sensitive to redox conditions, reversible disulfide bond formation is used in nature to produce oxidant-sensing proteins (2). Again, researchers have taken extensive advantage of the natural process of cysteine-cysteine disulfide bond formation to investigate the structure and function of membrane proteins. Pairs of cysteine side-chains introduced into different regions of channel proteins using site-directed mutagenesis can be induced to form disulfide bonds under oxidizing conditions provided that the two cysteine side-chains are in close proximity, potentially providing a great deal of three-dimensional information on protein structure and dynamics.

In practical terms, the use of disulfide chemistry to probe membrane proteins involves introduction of two cysteines into different parts of the protein that are hypothesized to be located close together in three-dimensional space, at least at some point in the range of conformational states in which the protein can exist. The formation and breakage of disulfide bonds between the two cysteine side-chains can then be ascertained by some change in protein function (for example, a change in ion channel activity monitored using electrophysiological recording techniques) (57–60), or some other parameter, such as a change in protein mobility on SDS-PAGE gels (11). Because of the potential to form disulfide bonds between an introduced cysteine and any native cysteine side-chains that may exist in the protein under investigation, it is highly desirable to carry out such experiments using cys-less versions of the protein (see above).

In the simplest case, disulfide bonds might form spontaneously between two introduced cysteine side-chains. In this case, channel function may be altered upon exposure to a reducing agent (usually DTT) to break the disulfide bond. However, in mammalian cells, this spontaneous disulfide formation usually only applies to external cysteines, due to the reducing conditions of the cytoplasm. In most cases, an oxidizing agent is applied to induce disulfide bond formation (preferably after treatment with DTT to reduce any disulfide bonds that might have formed spontaneously). A common choice for cysteine side-chains that are exposed on the surface of the protein is copper(II)-o-phenanthroline (CuPhe) (61), which catalyzes oxidation by ambient molecular oxygen to promote effective disulfide bond formation, although other strong oxidizing agents such as H₂O₂ or I₂ may also be effective.

The ability of two identified cysteine side-chains to form a disulfide bond is used to provide evidence for the proximity of those side-chains in the three-dimensional structure of the protein. For a disulfide bond to form, it is thought that the S-S distance must be ~2 Å, with the Cβ-Cβ distance ~3.5–4.6 Å and the Cα-Cα distance ~3.8–6.8 Å, along with strict limits on each of the dihedral angles Cα-Cβ, Cβ-S, and S-S (62–66) (Figure 2). However, it has been found that cysteine side-chains with Cα-Cα distances >15 Å in the equilibrium structure can form disulfide bonds, which is presumed to reflect the effects of protein thermal fluctuations (11, 63).

More distant cysteine side-chains can be connected experimentally using chemical cross-linkers that contain two cysteine-reactive groups. For example, homobifunctional cross-linkers containing two thiol-reactive MTS groups connected by spacer arms of different lengths have been used as ‘molecular rulers’ to gauge the distances between pairs of cysteine residues introduced into membrane proteins including P-glycoprotein (67), CFTR Cl^- channel (68, 69), lactose permease (70), cyclic nucleotide gated channels (71), KCNQ1/KCNE1 voltage-gated K⁺ channels (72), and ENaC Na⁺ channels (73). One caveat to the use of these reagents is that long spacer arms result in considerable flexibility of the cross-linker molecule (70), such that cysteine-cysteine distances judged using these cross-linkers should be considered maximal distances and may in fact be considerably less.

Figure 2:

Structure of the cysteine side-chain and experimentally derived proximity of cysteine pairs.

(A) Chemical composition of the cysteine side-chain within a peptide chain. (B) Stick model of the cysteine side-chain. (C) As described in the text, to form either disulfide bonds or metal bridges, the two cysteine side-chains are subject to strict spatial constraints (S-S, Cβ-Cβ, Cα-Cα), as well as limited regarding the dihedral angles Cα-Cβ, Cβ-S, and S-S.

Conformational movements are necessary for many aspects of membrane protein function, and disulfide bond formation can be used to obtain important information on the relative proximity of different parts of the protein in different conformational states that may not be accessible using other techniques (Figure 1D). The principle here is that, if the physical proximity of two reporter cysteine side-chains changes concomitant with a conformational change, then (i) disulfide bond formation may be possible only in the conformation in which the two side-chains are close together and/or (ii) disulfide bond formation may ‘trap’ or ‘lock’ the protein in the native conformation in which the two cysteine side-chains are close together [and conversely, disulfide bond breakage using DTT will liberate the protein to transition into other conformation(s)]. As shown by the selected examples listed in Table 1, the use of engineered disulfide bond formation has shed light on many different kinds of conformational changes that are crucial to the normal activity of many different classes of ion channels.

Three-dimensional information: cysteine-metal bridging

Cysteine cross-linking has provided a great deal of information on the three-dimensional structure of ion channel proteins and on the three-dimensional nature of conformational changes in these proteins (Table 1). However, the essentially irreversible nature of the disulfide bond under non-reducing conditions limits the information that can be obtained in this way. As described above, whereas the formation of the disulfide bond is subject to strict spatial constraints, it has been observed that relatively distant cysteine side-chains can form disulfide bonds under rare conditions due to protein flexibility. This potential caveat is particularly acute when the aim is to study protein conformational changes. The overall rate at which disulfide bonds will form is dependent on both the relative occupancy of different protein conformations and the proximity of cysteine side-chains in these different conformations. Thus, if disulfide bonds form rapidly in an uncommon (or even unnatural) conformation, cross-linked proteins will tend to accumulate in that conformation due to the high strength of the disulfide bond (often referred to as ‘disulfide trapping’).

An alternative approach is to use the propensity of cysteine (or other) side-chains to coordinate the binding of metal ions (so-called metal bridge formation). Two or more cysteine side-chains in close proximity can tightly bind metal ions such as Cd²⁺ or Zn²⁺, resulting in a metal bridge that stabilizes the protein in the conformation in which the side-chains in question are retained in the proximal, high-metal-affinity position. Zinc bridges between introduced histidine side-chains have also been used in the study of some ion channels (see Table 2). Because metal binding is reversible, and its affinity is strongly dependent on protein conformation (via conformation-dependent changes in the proximity of amino acid side-chains), this metal bridge approach is considered more likely to reflect normal, natural protein conformations and to be far less sensitive to the effects of rare protein conformations in which amino acid side-chains briefly come into close proximity. Furthermore, reversible metal binding means that protein conformational changes can continue to occur as the metal ion binds and unbinds from the introduced binding site. In effect, metal bridges act to stabilize natural protein conformations rather than to freeze the protein in a conformation that may or may not occur during normal activity.

In most cases, Cd²⁺ bridges between introduced cysteine side-chains are used to identify pairs of amino acids that are located in close proximity (see Table 2). To ensure that the Cd²⁺ bridge is formed between the introduced cysteine side-chains, use of a cys-less protein background is strongly advisable (see above). Cadmium ions also bind to other amino acid side-chains (histidine, aspartate, glutamate), and so it is possible that Cd²⁺ bridges may form between an introduced cysteine and a nearby non-cysteine side-chain (74–76). In electrophysiological recordings, application of Cd²⁺ may, via formation of a metal bridge between introduced cysteines, ‘lock’ or stabilize the channel in the open conformation (suggesting proximity of two or more cysteine side-chains in the open state) (Figure 1E) or in the closed conformation (suggesting proximity of cysteines in the closed state) (74, 77–80). This locking effect allows strict spatial constraints to be placed on the proximity of different protein regions in different conformational states (Figure 1E). For example, in voltage-gated K⁺ channels, evidence from metal bridges formed between identified amino acid side-chains has been used to identify the conformational changes that open the gate in the pore (74), the movements of the voltage-sensor domain that result in voltage-dependent activation (78, 81–87), and relative movements of the voltage-sensor and pore domains during channel opening (77, 83) (Table 2). Furthermore, manipulations that alter the proportion of time that the channel spends in the open state may influence the observed affinity of Cd²⁺ binding, by altering the likelihood that two cysteines that show state-dependent changes in proximity spend in a position that is amenable to metal bridge formation. For example, our own work on the CFTR Cl^- channel has shown that a mutation that keeps the channel in a permanently open state strengthens the apparent affinity of Cd²⁺ binding coordinated by certain pairs of cysteines introduced into the channel pore (80, 88), suggesting that these cysteines are close enough to form metal bridges only in the open state. Meanwhile, Cd²⁺ binding was practically abolished in permanently open channels for other cysteine pairs (80, 88), suggesting that metal bridges form here only when the channel is closed and that the cysteines in question move apart when the channel opens. Structural implications of these experimental findings are described in Figure 3.

Figure 3:

Examples of the use of metal bridges to determine conformation-dependent distances in the CFTR Cl^- channel pore.

(A) Overall atomic homology structure of the CFTR protein within the cell membrane, according to coordinates provided by Mornon et al. (133). The location of pore-lining amino acids that have been studied in our laboratory using cysteine mutagenesis and Cd²⁺ bridge formation are indicated, close to the outer mouth of the pore (R334, green; T1122, cyan; G1127, blue), and the inner vestibule of the pore (K95, red; M348, purple; S1141, orange). (B, C) Location of identified introduced cysteine side-chains that could form conformation-dependent Cd²⁺ bridges at these levels of the pore, viewed from the extracellular side of the membrane. (B) R334C was reported to form a Cd²⁺ bridge with T1122C in open channels and with G1127C in closed channels (88). The measured S-S distances (red dotted lines) are 22.2 Å for R334C-T1122C and 21.1 Å for R334C-G1127C, inconsistent with metal bridge formation, which suggests that this model does not accurately represent the geometry of this part of either the open or the closed channel. (C) S1141C was reported to form a Cd²⁺ bridge with K95C in open channels and with M348C in closed channels (80). The measured S-S distances (red dotted lines) are 8.6 Å for K95C-S1141C and 8.2 Å for M348C-S1141C, close to distances required for Cd²⁺ coordination, suggesting that only small conformational changes at this level of the pore may occur during pore opening and closing.

The binding affinity of Cd²⁺ for cysteines introduced into a protein is expected to be influenced by the number of cysteine side-chains that can coordinate the Cd²⁺ ion. In metalloproteins Cd²⁺is almost always coordinated by binding to four cysteine side-chains in a tetrahedral coordination geometry (3–6, 89, 90). The Cd²⁺ binding affinity of such well-coordinated tetra-cysteine binding sites is estimated to be on the order of 10^-14m (10 fm) (4). Zinc binding to cysteine or histidine side-chains in proteins also usually occurs in a tetrahedral geometry (3, 5, 90). Electrophysiological investigation of ion channels often suggests Cd²⁺ binding to a single cysteine side-chain with a measured affinity of >100 μm (32, 80, 91), whereas measured affinities of <1 μm have been reported with two proximal cysteine side-chains (32, 57, 77, 80, 91). These quantitative reports are consistent with the suggestion that each additional well-positioned, coordinating cysteine ligand increases Cd²⁺ binding affinity by a factor of ~1000-fold (79). However, it is unlikely that two or more cysteine side-chains introduced experimentally into an ion channel could match the exact coordination geometry found in metalloproteins, and so considerably lower apparent affinities of Cd²⁺ binding are not unexpected experimentally. Furthermore, in some cases, it may not be known how many introduced cysteine side-chains can interact simultaneously with a bound Cd²⁺ ion. In the case of monomeric ion channel proteins, the number of introduced cysteines is known, and in these cases, the measured affinity of Cd²⁺ binding with two cysteines can be compared directly with the affinity when only one cysteine is present (32, 57, 80, 91). In the case of multimeric proteins, however, introduction of only a single cysteine into each subunit may result in the formation of inter-subunit metal bridges of unknown coordination geometry. For example, in tetrameric K⁺ channels (92, 93) and trimeric P2X receptor channels (40), it has been observed that a single cysteine substitution results in very high affinity Cd²⁺ binding, which has been used to suggest that more than two cysteine side-chains are able to interact simultaneously with the bound Cd²⁺ ion to form a high-affinity Cd²⁺binding site. Besides the number of interacting cysteine side-chains, other factors may influence the apparent affinity of Cd²⁺ binding measured in electrophysiological experiments, including ease of access of the Cd²⁺ ion from the bulk solution on one side of the membrane to its binding site, the exact proximity and relative orientation of introduced cysteine side-chains, local structural features such as the presence of other potentially interacting side-chains such as histidines, aspartates, and glutamates, conformational changes and dynamic protein flexibility that influence the proportion of the time that cysteine side-chains are in a permissive position for Cd²⁺ bridge formation, strain on the metal bridge due to sub-optimal coordination, and the propensity of the channel to undergo conformational changes that can break conformationally sensitive Cd²⁺ bridges. As illustrated in Table 2, the range of Cd²⁺ concentrations that have been used experimentally to provide evidence for metal bridge formation between introduced cysteine side-chains in ion channel proteins span several orders of magnitude.

As with disulfide bond formation, identification of metal bridges is used to suggest strict spatial constraints on the structure of channel proteins. Structural studies with a number of model systems suggest a Cd²⁺-S distance of ~2.5 Å (3, 6, 89, 90), implying an S-S distance between coordinating cysteine side-chains of ~5 Å (Figure 2). The Zn²⁺-S distance for binding to cysteine is similar, ~2.3 Å, while the Zn²⁺-N distance for binding to histidine is reportedly ~2.1 Å (3, 90).

Conclusions

Techniques that take advantage of the reactivity of introduced cysteine side-chains have been used to probe the structure and function of practically all known types of ion channels; the examples cited in Tables 1 and 2 represent just a small representative sample of the kinds of important dynamic rearrangements that have been identified in different channel types. Metal bridges positively identified as being formed between different regions of ion channel proteins can give invaluable structural constraints that are otherwise inaccessible in the study of these integral membrane proteins, while at the same time allowing normal conformational rearrangements to proceed. These functional techniques therefore provide a powerful complement and constraint to structural techniques such as X-ray crystallography.

List of abbreviations

β-ME

β-mercaptoethanol

CFTR

cystic fibrosis transmembrane conductance regulator

CuPhe

copper(II)-o-phenanthroline

DTT

dithiothreitol

LBD

ligand binding domain

MTS

methanethiosulfonate

MTSEA

[2-aminoethyl] methanethiosulfonate

MTSES

[2-sulfonatoethyl] methanethiosulfonate

MTSET

[2-(trimethylammonium)ethyl] methanethiosulfonate

SCAM

substituted cysteine accessibility mutagenesis

TM

transmembrane segment