Voltage-activated proteins can sense, and respond to, changes in the electric field pervading the cell membrane by virtue of a transmembrane helix bundle, the voltage-sensing domain (VSD). Canonical VSDs consist of four transmembrane helices (S1–S4) of which S4 is considered a principal component because it possesses charged residues immersed in the electric field. Membrane depolarization compels the charges, and by extension S4, to rearrange with respect to the field. The VSD of large-conductance voltage- and Ca-activated K+ (BK) channels exhibits two salient inconsistencies from the canonical VSD model: (1) the BK channel VSD possesses an additional nonconserved transmembrane helix (S0); and (2) it exhibits a “decentralized” distribution of voltage-sensing charges, in helices S2 and S3, in addition to S4. Considering these unique features, the voltage-dependent rearrangements of the BK VSD could differ significantly from the standard model of VSD operation. To understand the mode of operation of this unique VSD, we have optically tracked the relative motions of the BK VSD transmembrane helices during activation, by manipulating the quenching environment of site-directed fluorescent labels with native and introduced Trp residues. Having previously reported that S0 and S4 diverge during activation, in this work we demonstrate that S4 also diverges from S1 and S2, whereas S2, compelled by its voltage-sensing charged residues, moves closer to S1. This information contributes spatial constraints for understanding the BK channel voltage-sensing process, revealing the structural rearrangements in a non-canonical VSD.
Large-conductance voltage- and Ca-activated K+ (BK) channels are potent regulators of excitability, broadly expressed in most mammalian cells (Toro et al., 1998; Latorre and Brauchi, 2006; Salkoff et al., 2006). BK channels are formed by α-subunit homotetramers (Shen et al., 1994) (Fig. 1), each comprising a conserved transmembrane voltage-sensing domain (VSD), helices contributing to the ion-conducting pore domain, and large intracellular ligand-binding domains that assemble into the gating ring superstructure in the tetramer (Wang and Sigworth, 2009; Wu et al., 2010; Yuan et al., 2010, 2012; Javaherian et al., 2011). BK channels exhibit an exceptionally high, selective conductance for K+, which is over an order of magnitude greater than that of other voltage-gated K+ channels (Latorre and Miller, 1983), and their activation is regulated by the synergistic action of membrane depolarization and intracellular Ca binding (Stefani et al., 1997; Cui and Aldrich, 2000; Rothberg and Magleby, 2000; Horrigan and Aldrich, 2002; Magleby, 2003; Latorre and Brauchi, 2006; Sweet and Cox, 2008; Cui et al., 2009; Latorre et al., 2010; Lee and Cui, 2010; Savalli et al., 2012). Ca is thought to bind with high affinity to the cytosolic domain (Wei et al., 1994; Schreiber and Salkoff, 1997; Bian et al., 2001; Xia et al., 2002; Bao et al., 2004; Zeng et al., 2005; Yusifov et al., 2008, 2010; Yuan et al., 2010, 2012; Zhang et al., 2010; Javaherian et al., 2011), which is also sensitive to other ligands and biological partners that modulate channel activation (Lu et al., 2006; Hou et al., 2009).
Typical voltage-gated K+ channel α subunits possess six transmembrane helices; the four most N-terminal helical transmembrane segments (S1–S4) comprise the VSD, whereas S5 and S6 contribute to the central ion-selective pore (Armstrong, 2003; Swartz, 2004; Long et al., 2005, 2007). Of the four conserved VSD helices, S4 is considered the principal component of voltage sensing because it has been found to concentrate most voltage-sensing charged residues, which are immersed in the electric field pervading the cell membrane and compel the helix to adopt an active conformation upon membrane depolarization (Tombola et al., 2006; Bezanilla, 2008; Chanda and Bezanilla, 2008; Swartz, 2008).
In contrast, BK channels exhibit a decentralized distribution of voltage-sensing charged residues, whereby most voltage-sensing charge is contributed from segments S2 and S3 (Ma et al., 2006). Specifically, S2 carries two voltage-sensing residues, D153 and R167; S3 bears D186, and S4 only contributes R213 (Ma et al., 2006) (Fig. 1 A). We recently showed that the charged residues borne by S2 confer intrinsic voltage-sensing properties to the S2 helix, causing it to rearrange with respect to the electric field upon depolarization, a process that results in a unique mode of intra-subunit cooperativity between S2 and S4 that could involve electric field focusing by dynamic (state-dependent) water-filled crevice formation (Pantazis et al., 2010a).
Furthermore, BK α subunits possess a distinct transmembrane topology. In addition to broadly conserved helices S1–S6, BK channels have additional transmembrane helix S0 (Wallner et al., 1996; Meera et al., 1997), which is involved in the association of auxiliary β subunits (Wallner et al., 1996; Morrow et al., 2006; Liu et al., 2008, 2010; Wu et al., 2009) and could be directly involved in the BK voltage-sensing process (Koval et al., 2007). The position of S0 with respect to other VSD and pore helices has been inferred by investigating disulfide cross-linking efficiency by Liu et al. (2010) (Fig. 1 D). The close proximity of S0 and S4 was further supported by the identification of a Trp side chain in S4 as the collisional quencher of fluorophores labeling the extracellular flank of S0 (Pantazis et al., 2010b).
Given these salient structural and functional features of the BK VSD, it is likely that BK channels use a distinct mode of sensing the membrane potential. How does S4 rearrange with respect to other VSD helices during activation? As S2 exhibits intrinsic voltage dependence, what is the direction of its depolarization-induced motion? We sought to understand how BK channels sense, and respond to, changes in the membrane potential by optically tracking voltage-dependent rearrangements using voltage-clamp fluorometry. The latter involves labeling unique cysteines introduced at specific protein positions with small thiol-reactive, environment-sensitive fluorophores (Mannuzzu et al., 1996; Cha and Bezanilla, 1997; Claydon and Fedida, 2007; Gandhi and Olcese, 2008). In addition, we exploited the photochemical properties of the tryptophan side chain, which can quench the fluorescence of small organic dyes by photo-induced e− transfer upon van der Waals collision (Mansoor et al., 2002, 2010; Doose et al., 2005, 2009; Islas and Zagotta, 2006; Pantazis et al., 2010b) to gain structural insight to the source of the observed fluorescence deflections. Thus, we combined site-directed fluorescent labeling under voltage clamp with manipulation of fluorescence quenching by the removal of a native Trp, or the strategic introduction of additional Trp residues, to reveal the relative voltage-dependent rearrangements of assigned protein positions. We found that the extracellular portion of S4 diverges from S1 upon VSD activation, similar to the relative rearrangement of S4 and S0 (Pantazis et al., 2010b). On the other hand, membrane depolarization causes S2 to approach S1, whereas the principal voltage-sensing helices of BK, S2 and S4, are within collisional distance at rest and diverge upon activation.
MATERIALS AND METHODS
For site-directed fluorescence labeling with a thiol-reactive fluorophore, an hSlo clone (provided by L. Toro, University of California, Los Angeles, Los Angeles, CA; NCBI Protein database accession no. U11058) (Wallner et al., 1995) transcribed from the fourth methionine without extracellular cysteines (C14S, C141S, and C277S), was used. Background mutation R207Q was introduced to increase PO at low [Ca2+]i (Díaz et al., 1998; Ma et al., 2006) and to be consistent with the previous fluorometric investigations of the human BK channel voltage sensor (Savalli et al., 2006, 2007, 2012; Pantazis et al., 2010a,b). A single cysteine was substituted at positions at the extracellular flank of S1 (S135C) or S2 (Y145C) for subsequent modification by thiol-reactive fluorescent labels, and Trp residues were substituted at positions along the S1–S2 extracellular linker. Single-point mutations were generated with QuikChange Site-Directed Mutagenesis kit (Agilent Technologies) and confirmed by sequencing. The cDNAs were transcribed to cRNAs in vitro (mMESSAGE MACHINE; Ambion) and stored at −80°C.
Xenopus laevis (Nasco) oocytes (stage V–VI) were prepared as described previously (Haug et al., 2004) and then injected with 50 nl of total cRNA (0.1–0.5 µg/µl) using a nanoinjector (Drummond Scientific). Injected oocytes were maintained at 18°C in an amphibian saline solution supplemented with 50 µg/ml gentamycin (Gibco), 200 µM DTT, and 10 µM EDTA. 3–6 d after injection, oocytes were stained for 40 min with 10 µM of membrane-impermeable, thiol-reactive fluorophore, tetramethylrhodamine-5′-maleimide (TMRM; AnaSpec), in a depolarizing solution (in mM: 120 K-methanesulfonate [MES], 2 Ca(MES)2, and 10 HEPES, pH 7.0) at 18°C in the dark. Alternatively, oocytes were incubated with 20 µM MTS-TAMRA (Santa Cruz Biotechnology, Inc.) for 40 s on ice. 100 mM of fluorescent probe stock was dissolved in DMSO and stored at −20°C. After fluorescent labeling, the oocytes were rinsed in dye-free saline before being mounted in the recording chamber. Both TMRM and MTS-TAMRA are rhodamine-based fluorescent labels that can be collisionally quenched by photoinduced e− transfer by Trp residues (Doose et al., 2005, 2009; Mansoor et al., 2010; Pantazis et al., 2010b).
The cut-open oocyte voltage clamp.
The cut-open oocyte Vaseline gap technique is a low-noise, fast-clamp technique (Stefani and Bezanilla, 1998). The oocyte is placed in a triple-compartment Perspex chamber, with a diameter of 600 µm for the top and bottom apertures. The top chamber isolates the oocyte’s upper domus and maintains it under clamp. The middle chamber provides a guard shield by clamping the middle part of the oocyte to the same potential as the top chamber. The bottom chamber injects current intracellularly through the saponin-permeabilized part of the oocyte. Fluorescence emission and ionic current were simultaneously measured from the same area of membrane isolated by the top chamber (Cha and Bezanilla, 1997; Gandhi and Olcese, 2008). The optical setup consists of a microscope (Axioscope FS; Carl Zeiss) with filters (Semrock) appropriate for rhodamine excitation and emission wavelengths. The light source is a 100-W microscope halogen lamp. A TTL-triggered shutter (Uniblitz VS 25; Vincent Associates) is mounted on the excitation light path. The objective (40×, water immersion; LUMPlanFl; Olympus) has a numerical aperture of 0.8 and a working distance of 3.3 mm (Olympus), which leaves enough room for the insertion of the microelectrode. The emission light is focused on a photodiode (PIN-08-GL; UDT Technologies). An amplifier (Photomax 200; Dagan) is used for the amplification of the photocurrent and background fluorescence subtraction. The external solution contained (mM): 110 Na-MES, 10 K-MES, 2 Ca(MES)2, and 10 Na-HEPES, pH 7.0. The internal solution contained (mM): 120 K-glutamate and 10 HEPES, pH 7.0. Standard solution for the intracellular recording micropipette is (in mM): 2,700 Na-MES and 10 NaCl. Low access resistance to the oocyte interior was obtained by permeabilizing the oocyte with 0.1% saponin carried by the internal solution.
Experimental data were recorded and analyzed with a customized program developed in our division. The G-V curves were calculated by dividing the I-V relationships (I-V curves) by the driving force (Vm−EK), where Vm is the membrane potential and EK is the equilibrium potential for K+, estimated using the Nernst equation. Data for the membrane conductance (G-V) and the fluorescence (ΔF(V)) curves were fitted to one or two Boltzmann distributions of the form:
where Gmax and ΔFmax are the maximal G and ΔF; ΔFmin is the minimal ΔF; z is the effective valence of the distribution; V0.5 is the half-activating potential; Vm is the membrane potential; and F, R, and T are the usual thermodynamic values. Fitting was performed by least squares in Microsoft Excel. To characterize the overall voltage-dependent ΔF reported from a label, we normalized the total fluorescence change (ΔFtotal = ΔFmax − ΔFmin) and normalized by background fluorescence (F0) and maximal conductance (Gmax):
Online supplemental material
In Fig. S1, the Trp side chain can quench a thiol-reactive fluorophore, PyMPO maleimide, in solution. In a previous investigation (Savalli et al., 2006), PyMPO labeling position 202, outside S4, reported slow voltage-dependent ΔF in the presence of W203. As PyMPO can be quenched by the Trp side chain, the reported ΔF can be interpreted as the rearrangement of S4, with respect to the S3–S4 linker, over a slow (200-ms) time course. Fig. S1 is available at http://www.jgp.org/cgi/content/full/jgp.201210807/DC1.
Voltage-dependent ΔF signals reported from the extracellular flank of the S1 helix track voltage sensor activation and require W203 at the extracellular flank of S4
S1 is a hydrophobic helix, possessing no charges that contribute to the BK channel voltage-sensing process (Ma et al., 2006). However, considering the close packing of the VSD helix bundle in the membrane (Long et al., 2007), the BK S1 is likely intimately associated with segments S2, S3, and S4 (Fig. 1 D), which do carry voltage-sensing charges (Ma et al., 2006) and undergo voltage-dependent conformational changes (Savalli et al., 2006; Pantazis et al., 2010a). To probe for voltage-dependent conformational rearrangements in the vicinity of the extracellular flank of S1, we substituted a unique cysteine at position 135 (S135C) and labeled the channel with MTS-TAMRA (Fig. 2 A). Positive depolarization-evoked ΔF was observed with a voltage dependence that preceded that of ionic conductance (Fig. 2 B), indicating that the ΔF reported structural rearrangements of the voltage sensor. The voltage dependence of S1-reported ΔF was very similar to that of the ΔF of TMRM labeling the extracellular flank of S4 (Savalli et al., 2006, 2007, 2012; Pantazis et al., 2010a) and S0 (Pantazis et al., 2010b). The ΔF reported by the S0 label was found to be caused by its state-dependent quenching by a Trp residue at the extracellular tip of S4, W203 (Pantazis et al., 2010b).
To test whether W203 is also responsible for the quenching of S1-conjugated fluorophores, we removed it by mutation W203V (Fig. 2 C), as previously (Savalli et al., 2006, 2007, 2012; Pantazis et al., 2010a,b). MTS-TAMRA ΔF reported from position 135 in W203V channels was strongly attenuated (Fig. 2, D and E). The voltage dependence of this residual ΔF signal was similar to that from channels without the W203V mutation (Fig. 2 F), indicating that the observed decrease in ΔF amplitude was caused by attenuation of the fluorophore quenching process that gave rise to the observed fluorescence deflections, rather than a shift in voltage dependence.
W203, at the extracellular end of S4, is the principal quencher of fluorescence reported from S2
To investigate how S2 and S4 rearrange with respect to each other, we labeled S2 (position 145) with TMRM (Fig. 3 A), resolving positive fluorescence deflections (Fig. 3 B) indicating protein rearrangements with a voltage dependence similar to those reported by S4-, S0-, and S1-labeled channels. Substitution of W203, extracellular to S4 (W203V) (Fig. 3 C), resulted in an attenuation of the reported ΔF (Fig. 3, D and E). The ΔF reported from S2 recapitulates those reported from S0 (Pantazis et al., 2010b) and S1 (Fig. 2), suggesting that the side chain of W203, at S4, interacts with fluorophore-labeling segments S0, S1, and S2 in a voltage-dependent manner. The voltage dependence of the residual ΔF reported from S2 in W203V channels was right-shifted toward more depolarized potentials (Fig. 3, D and F). In a previous work, we reported that ΔF from TMRM-labeling S2 in W203V channels differs significantly in terms of voltage dependence and kinetics from the ΔF reported from S4, in BK channels with intact voltage sensing as well as those with voltage-sensing charge neutralizations, postulating that they reflect the intrinsic voltage dependence of helix S2 (Pantazis et al., 2010a).
In the next experiments, we provide evidence that, even though the fluorescence deflections reported from the extracellular flanks of S1 and S2 are similar in voltage dependence and dependence on W203, the two segments do no behave as a single structural unit, but they undergo a relative rearrangement upon membrane depolarization.
Resolving the relative voltage-dependent rearrangement of S1 and S2 by manipulating the quenching environment of the fluorescent label with site-directed Trp substitution
The S2 helix of the BK VSD bears two charged residues that contribute to the voltage-sensing charge: D153 and R167 (Ma et al., 2006). As such, S2 would be expected to rearrange with respect to the electric field upon depolarization. Accordingly, TMRM labels attached to S2 (in W203V channels, as in Fig. 3 D) exhibit ΔF with distinct kinetics and voltage dependence to the ΔF of S4 labels, strongly suggesting that S2 undergoes conformational changes upon depolarization (Pantazis et al., 2010a). On the other hand, S1 bears no voltage-sensing charges and is therefore unlikely to undergo large conformational changes upon membrane depolarization. Because of the physical proximity of the two helices and difference in their expected behavior upon depolarization, we reasoned that the two segments also undergo a voltage-dependent relative rearrangement, and sought to resolve it by directly manipulating the environment of the S1-conjugated fluorescent label by the introduction of a Trp residue extracellular to S2.
Fig. 4 summarizes the results of substituting Trp in two positions along the S1–S2 linker, in channels fluorescently labeled with MTS-TAMRA at the extracellular portion of S1 (position 135). Fig. 4 (A–D) demonstrates a condition similar to that shown in Fig. 2 (A and B), whereby the label reports positive ΔF upon depolarization, which was subsequently shown to be caused by voltage-dependent interaction of the S1 fluorophore with the side chain of the S4 tryptophan, W203 (Fig. 2). When a Trp is substituted near S2 in the S1–S2 linker (Fig. 4 E, C141W), MTS-TAMRA–labeling position 135 (outside S1) reports a complex fluorescence signal upon depolarization (Fig. 4 G). In addition to the dequenching component, which is caused by the rearrangement of S1 and S4 (Figs. 2 and 4 C), a voltage-dependent quenching component is evident, which demonstrates the state-dependent interaction of the S1 label with W141, near S2. The unquenching component responsible for the fluorescence transients was absent in channels lacking the native S4 tryptophan (Fig. 4, I and K, W203V). However, the concurrent removal of native W203 and the introduction of W141 resulted in a highly shifted voltage dependence of activation (Fig. 4 J).
To further probe the photochemistry of Trp introduction in the S1–S2 linker, we introduced a Trp closer to the fluorescent label (Fig. 4 M, I138W). This resulted in the almost complete absence of fluorescence deflections (Fig. 4 O), likely caused by static, voltage-independent quenching of fluorophore by W138 (Doose et al., 2005; Mansoor et al., 2010). A weak fluorescence transient is consistently observed upon sufficient depolarizations (Fig. 4 P). Although the molecular events underlying these fluorescence dynamics are unclear, their presence confirms that MTS-TAMRA did modify C135, and the lack of ΔF reported from I138W channels is caused by the static photochemical interaction of W138 and the C135-conjugated label.
Model-independent inference of dynamic structural information from voltage-clamp fluorometry data
The BK channel VSD is an ideal system to try and test the principles of structurally assigned voltage-clamp fluorometry by optically tracking the photochemical interaction between a fluorophore and Trp attached to specific positions in the protein: first, the property of the W203 side chain, at S4, to collide with and quench small fluorophores attached to positions around it according to the state of the VSD, revealing the relative movement of S4 with respect to nearby fluorescently labeled helices (Fig. 6) (Pantazis et al., 2010b); second, the S1–S2 system, which consists of the charge-bearing S2 (Ma et al., 2006), which rearranges with respect to the electric field upon depolarization (Pantazis et al., 2010a), and voltage-insensitive S1 (Ma et al., 2006), connected with S2 via a short extracellular linker. This provided the ideal platform to detect the relative voltage-dependent rearrangement that would be expected to occur between a voltage-sensitive and a voltage-insensitive transmembrane helix, by combining site-specific fluorescent labeling at an extracellularly accessible position near S1, and Trp introduction near S2, to induce Trp-mediated quenching (Figs. 4, A–L, and 5, B and C). We have combined the structural information from this work with the findings of previous investigations on BK channels and the KV1.2–2.1 atomic structure (Long et al., 2007) to propose the rearrangements of the human BK channel VSD shown in Fig. 6, without applying a computational model. Nevertheless, because the fluorescence deflections reported by Trp-induced quenching reflect the probability of Trp–fluorophore collisions, the fluorescence data can be used as important constraints for, and to reduce the a priori assumptions of, the elaborate computational modeling of molecular rearrangements involved in voltage-dependent activation.
A potential limitation of manipulating the quenching of small protein-labeling fluorophores with targeted Trp introduction (or the removal of native Trp residues) is the risk of perturbing channel operation and gating. We have previously reported that substitution of the native W203 in BK channels results in a depolarizing shift of channel activation (Savalli et al., 2006; Pantazis et al., 2010b), which is also observed in this work (Figs. 2 and 3). The addition of Trp residues in the S1–S2 linker resulted in a less pronounced effect on channel activation (Fig. 4, F and N), which appeared to be cumulative to W203 substitution (Fig. 4 J). However, the characteristics of the observed fluorescence deflections of the clones studied in this work cannot be readily accounted for by changes of the voltage dependence of channel activation, whereas the photochemical interpretation of the data (Fig. 5) is consistent with the fluorescence characteristics of all the clones studied in this and previous (Pantazis et al., 2010b) work.
Converging toward a dynamic map of voltage-dependent structural rearrangements in the BK voltage sensor
The fluorescence data from S1-conjugated MTS-TAMRA in channels with or without the native Trp outside S4 (Fig. 2, W203) recapitulate the fluorescence of S0-labeling TMRM, in that the positive depolarization-induced ΔF is strongly attenuated by removal of the S4 Trp (W203V) (Pantazis et al., 2010b). As such, the same structural interpretation can account for these data: the extracellular portions of S1 and S4 appear to be within collisional range at rest, so that the S1 label is efficiently quenched by the W203 side chain in S4; however, upon depolarization, the distance between the two helices increases, reducing the efficiency of W203-induced quenching, so that a positive fluorescence deflection is observed (Fig. 5 A). According to Liu et al. (2010), S1 and S0 are almost diametrically opposite S4 on the plane of the membrane (Fig. 1 D). S4 could move away from S1 and S0, as the positively charged, voltage-sensing R213 (Ma et al., 2006) moves outward with respect to the electric field. Such a motion was previously suggested to explain the fluorescence from S0 labels (Pantazis et al., 2010b) and can be used to account for the fluorescence deflections reported from S1 as well (Fig. 6).
Introduction of a Trp near the extracellular flank of S2 (C141W) imposed a depolarization-dependent quenching component (negative ΔF; Fig. 4, G and K) to the fluorescence of MTS-TAMRA label conjugated to S1 (position 135), revealing that the fluorophore and W141 converge upon activation (Fig. 5, B and C). S2 bears two residues that contribute to the voltage-sensing charge of BK: negatively charged D153 at the extracellular half of the helix and positively charged R167 at the intracellular end (Ma et al., 2006), exhibiting intrinsic voltage-sensing properties (Pantazis et al., 2010a). Upon membrane depolarization, D153 would be compelled to move inward and R167 outward, inducing S2 to tilt. Because S1 does not possess sensing charges (Ma et al., 2006), it is likely that S2 tilting toward S1 gives rise to the relative rearrangement of the two segments. Such a movement is included in the proposed structural interpretation of the results in Fig. 6.
A large component of the ΔF signal from labels on S2 arises from its state-dependent interaction with W203 in S4 (Fig. 3), so that the molecular quenching interpretation in Fig. 5 A can also account for the ΔF reported from S2 labels. The data suggest that S2 and S4 are within collisional range at rest but follow different paths upon activation, so that their extracellular flanks diverge. This interpretation is also consistent with the findings that S2 approaches S1 upon activation, but S4 and S1 diverge, and has been considered in constructing the proposed voltage-dependent rearrangements of the BK channel VSD shown in Fig. 6.
How does S3 rearrange with respect to other helices?
We did not directly probe for the motions of S3 relative to the other VSD helices in this work. The extracellular linker between S3 and S4 in the BK VSD is very short, consisting of three residues (N200–R201–S202; Fig. 1 B) and exhibiting secondary structure (Semenova et al., 2009; Unnerståle et al., 2009). Because S4 and the extracellular portion of S3 (S3b) are connected by a very short and structurally robust tether in BK, we propose that the collisional events between the S4 Trp (W203) and S0-, S1-, and S2-labeling fluorophores represent, to the best of our current knowledge, the rearrangements of an S4–S3b unit. When PyMPO maleimide is used to label position 202, a positive voltage-dependent ΔF is reported that is abolished by W203V (Savalli et al., 2006). Because the Trp side chain can also quench PyMPO fluorescence (Stern–Volmer bimolecular quenching constant: 129 M−1; Fig. S1), this result is consistent with a conformational change between these two positions. This rearrangement occurs over a longer time course (≈200 ms) than those reported in this work and may represent reorganizations of the BK VSD occurring while the channel gate is open (Savalli et al., 2006).
How do mechanical interactions contribute to the BK voltage-sensing process?
In a previous work on the functional interaction of S2 and S4 in the BK VSD (Pantazis et al., 2010a), we proposed two mechanisms of cooperativity between the two voltage-sensing helices: in the first, S2 and S4 are mechanically coupled, so that the voltage-sensing charge of one helix contributes to that of the other; the second mechanism involves mutual state-dependent focusing of the electric field by dynamic aqueous crevice formation, inspired by previous hypotheses in the literature (Cha and Bezanilla, 1997; Chanda and Bezanilla, 2008). Although the two scenarios are not mutually exclusive, the work presented here contributes more weight to the idea of dynamic focusing of the electric field, which is more consistent with S2 and S4 undergoing different and divergent motions, whereas mechanical interactions might be expected to result in movements of uniform direction.
On the other hand, the Trp-induced quenching mechanism used to resolve the relative helical rearrangements, including those of S2 and S4, relies on electron waveform overlap, i.e., a collision event (Doose et al., 2005, 2009; Mansoor et al., 2010). Because the S4 Trp (W203) quenches S0-, S1-, and S2-labeling fluorophores more efficiently as the membrane potential repolarizes, it would appear that S4 “nudges” S0, S1, and S2 upon VSD deactivation, whereas S2 collides with S1 more frequently in the active state.
The collisions, or mechanical interactions, between VSD segments resolved previously (Pantazis et al., 2010b) and in this work (Fig. 6) could be an important aspect of voltage sensing, which, after all, is an electromechanical process. The mechanical interaction of voltage-sensing helices S2 and S4 may enhance their voltage-sensing properties (Pantazis et al., 2010a). Furthermore, the interactions of voltage-sensing S2 and S4 with voltage-insensitive S0 and S1 could also contribute to the efficiency with which the VSD actuates voltage-dependent channel gating: S0 and S1 likely pose steric constraints or even mechanical guidance and support to the movements of the voltage-sensing S2 and S4. Thus, the interaction of auxiliary β subunits with S0 and S1 (Wallner et al., 1996; Meera et al., 1997; Morrow et al., 2006; Wu et al., 2009; Liu et al., 2010; Morera et al., 2012) may contribute to the modulating effect of β subunits to the BK voltage-sensing process (Orio et al., 2002; Bao and Cox, 2005; Orio and Latorre, 2005; Savalli et al., 2007; Lee et al., 2010; Wu and Marx, 2010; Sun et al., 2012). Moreover, the mutual coordination of Mg2+ between the intracellular ligand-sensing gating ring and D99, in the intracellular S0–S1 linker (Yang et al., 2008), could contribute to the functional coupling between the BK voltage- and ligand-sensing domains (Horrigan and Aldrich, 2002; Horrigan and Ma, 2008; Savalli et al., 2012). Finally, the palmitoylation site in the S0–S1 linker (Jeffries et al., 2010) suggests that the anchoring of helices S0 and S1 to the membrane could be important for their structural role in VSD operation.
By combining site-specific fluorescence labeling and quenching, we have resolved structural rearrangements occurring between assigned loci in the voltage sensor of human BK channels. We propose a possible structural interpretation of how this “non-canonical” VSD operates in Fig. 6. Specifically, we have resolved the direction of movement of the crucial voltage-sensing helices S4 and S2, and revealed their state-dependent mechanical interaction with S0, S1, and each other. Because the VSDs of a multitude of voltage-dependent proteins are structurally conserved (Nelson et al., 1999; Chanda and Bezanilla, 2008), we are confident that site-directed, Trp-induced collisional quenching of fluorescent labels can be used to resolve the relative motions of assigned positions in other VSD-activated proteins and provide crucial constraints for the elaborate computational modeling of voltage-dependent rearrangements.
We are grateful to members of the Olcese laboratory for critical comments on the manuscript. The hSlo clone was a kind gift from Ligia Toro (Dept. of Anesthesiology, UCLA).
This work was supported by research grants from National Institutes of Health/National Institute of General Medical Sciences (grant R01GM082289 to R. Olcese) and American Heart Association (Western States Affiliate) Postdoctoral Fellowship (grant 11POST7140046 to A. Pantazis).
Sharona E. Gordon served as editor.