Relative motion of transmembrane segments S0 and S4 during voltage sensor activation in the human BK_Ca channel

Large-conductance voltage- and Ca²⁺-activated K⁺ (BK_Ca or MaxiK) channels are ubiquitous membrane proteins that potently regulate cellular excitability (Toro et al., 1998; Latorre and Brauchi, 2006; Salkoff et al., 2006; Wu and Marx, 2010). BK_Ca channels are formed by α subunit homotetramers, each comprising a conserved transmembrane voltage-sensing domain (VSD), a pore domain, and a large intracellular ligand-binding domain (Fig. 1), recently visualized by cryoelectron microscopy (Wang and Sigworth, 2009). BK_Ca channels possess several unique features that set them apart from other voltage-gated ion channels. First, as their name suggests, they exhibit an exceptionally high conductance for K⁺, over an order of magnitude greater than that of other voltage-gated K⁺ channels (Latorre and Miller, 1983). Second, they are activated by the allosteric interplay of membrane depolarization and Ca binding (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). Ca is thought to bind with high affinity to the cytosolic domain (Wei et al., 1994; Schreiber and Salkoff, 1997; Bian et al., 2001; Braun and Sy, 2001; Xia et al., 2002; Bao et al., 2004; Sheng et al., 2005; Zeng et al., 2005; Yusifov et al., 2008, 2010; Yang et al., 2010; Yuan et al., 2010), which is also sensitive to other ligands and biological partners that modulate channel activation (Lu et al., 2006; Hou et al., 2009; Cui, 2010).

Another unique aspect of BK_Ca channels is their transmembrane topology: a typical voltage-gated K⁺ channel α subunit crosses the membrane six times; 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) (Fig. 1). BK_Ca channel α subunits possess an additional transmembrane domain, S0, which renders their N-terminal tail extracellular (Wallner et al., 1996; Meera et al., 1997; Morrow et al., 2006). S0 and the extracellular N-terminal flank are required for the functional interaction between channel-forming α subunits and accessory β subunits (Wallner et al., 1996; Morrow et al., 2006; Liu et al., 2008b, 2010; Wu et al., 2009), which modulate the channel activation mechanism and, because of their subtype-restricted expression pattern, confer tissue-specific effects to BK_Ca channel function (Orio et al., 2002, 2006; Bao and Cox, 2005; Orio and Latorre, 2005; Savalli et al., 2007; Sweet and Cox, 2009; Latorre et al., 2010; Wu and Marx, 2010).

Because S0 is a unique transmembrane feature of BK_Ca channels with little homology to the N termini of other K_V channels, its position and orientation relative to other transmembrane segments, as well as its precise boundaries, cannot be inferred from a direct comparison with K_V channel crystal structures available to date. Its hydropathy profile indicates that its hydrophobic transmembrane portion spans residues Met-21 to Trp-43 and is flanked by arginines at positions 20 and 44 (Wallner et al., 1996; Morrow et al., 2006). Tryptophan substitution at specific positions in S0 impairs BK_Ca channel voltage sensing, presumably because they perturb a site of interaction between S0 and the VSD (Koval et al., 2007). Indeed, studies based on disulfide bridge formation efficiency determined that S0 is highly associated with segments S3 and S4 (Liu et al., 2008a, 2010), both of which possess voltage-sensing charged residues and are thus key components of the BK_Ca voltage-sensing apparatus (Stefani et al., 1997; Díaz et al., 1998; Cui and Aldrich, 2000; Ma et al., 2006; Savalli et al., 2006; Pantazis et al., 2010).

If S0 is indeed intimately associated with the BK_Ca voltage-sensing apparatus, then S0 or its immediate proximity should undergo voltage-dependent transitions. We have previously used the voltage clamp fluorometry method (Gandhi and Isacoff, 2005; Claydon and Fedida, 2007; Gandhi and Olcese, 2008; Horne and Fedida, 2009) to optically resolve the voltage-dependent transitions of S2 and S4 in the BK_Ca VSD (Savalli et al., 2006, 2007; Pantazis et al., 2010). In this work, we sought to probe the role of S0 in the BK_Ca VSD by resolving protein rearrangements in its immediate proximity.

The central principle of voltage clamp fluorometry is that many fluorophores are sensitive to nearby environmental factors, such as hydropathy, quenching groups, etc. (Lakowicz, 2006). By engineering such fluorophores with a sulfhydryl-reactive moiety, such as maleimide, it is possible to covalently attach them to a specific position of the protein where a unique cysteine has been introduced. If this area of the protein then undergoes movement upon depolarization that causes a change of the fluorophore environment, a deflection in the fluorescence intensity will be observed. This technique was pioneered in Shaker K⁺ channels (Mannuzzu et al., 1996; Cha and Bezanilla, 1997) and has since been used to probe the voltage-dependent operation of other K_V channels (Smith and Yellen, 2002; Bannister et al., 2005; Vaid et al., 2008; Horne et al., 2010), Na_V channels (Chanda and Bezanilla, 2002), HCN channels, (Bruening-Wright et al., 2007), H_V channels (Tombola et al., 2010), and the VSD-activated phosphatase, Ci-VSP (Kohout et al., 2008; Villalba-Galea et al., 2008).

By combining this approach with the quenching properties of a tryptophan residue in the extracellular tip of S4, we optically resolved a voltage-dependent divergence of the extracellular portions of S0 and S4: a direct demonstration of protein segments rearranging relative to each other during activation. In this context, we speculate that S0 could act as a pivot element, closely associated with the voltage-sensing S4 at rest; however, upon membrane depolarization, S4 moves against and away from S0 to facilitate channel activation.

For site-directed fluorescence labeling with a thiol-reactive fluorophore, an hSlo clone (no. U11058; provided by L. Toro, University of California, Los Angeles, CA) (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 P_O (Díaz et al., 1998; Ma et al., 2006). A single cysteine was substituted at positions 17, 18, 19, or 20 (outside S0), 145 (outside S2), or 202 (outside S4) to track voltage-dependent conformational transitions from these segments. Single point mutations were generated with QuikChange Site-Directed Mutagenesis kit (Agilent Technologies) and confirmed by sequencing. The cDNAs were translated to cRNAs in vitro (mMESSAGE MACHINE; Applied Biosystems) 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 (0.2–0.5 µg/µl) using a nano-injector (Drummond Scientific Company). Injected oocytes were maintained at 18°C in an amphibian saline solution supplemented with 50 µg/ml gentamycin (Invitrogen), 200 µM DTT, and 10 µM EDTA. 3–6 d after injection, oocytes were stained for 30–45 min with 10 µM of a membrane-impermeable, thiol-reactive fluorophore, tetramethylrhodamine-5′-maleimide (TMRM; Invitrogen), in a depolarizing K⁺ solution (in mM: 120 K-methanesulfonate [MES], 2 Ca(MES)₂, and 10 HEPES, pH 7) at room temperature. TMRM stock (100 mM) was dissolved in DMSO and stored at −20°C. The oocytes were then thoroughly rinsed in a dye-free solution before being mounted in the recording chamber. Changes in fluorescence emission were a result of environmental differences sensed by the fluorophores.

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 rings. 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 (Gandhi and Olcese, 2008). The optical setup consists of a microscope (Axioscope FS; Carl Zeiss, Inc.) with filters (Omega Optical) appropriate for TMRM. 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 (in mM): 110 Na-MES, 10 K-MES, 2 Ca(MES)₂, and 10 Na-HEPES, pH 7.0. The internal solution contained (in 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 analyzed with a customized program developed in our division and using fitting routines run in Microsoft Excel. The G(V) curves were calculated by dividing the current–voltage relationships (I-V curves) by the driving force (V_m−E_K), where V_m is the membrane potential and E_K the equilibrium potential for K⁺, estimated using the Nernst equation. Data for the membrane conductance (G(V)) and the fluorescence (F(V)) curves were fitted to one or two Boltzmann distributions of the form:

where G_max and F_max are the maximal G and F, F_min is the minimal F, z is the effective valence of the distribution, V_half is the half-activating potential, V_m is the membrane potential, and F, R, and T are the usual thermodynamic values.

The fluorescence of TMRM in solution (1 µM dissolved in external solution, pH 7.0) was measured using a spectrofluorometer (Fluorolog-3; HORIBA): λ_ex = 545 nm, 5-nm slit; λ_em = 550–650 nm, 1-nm slit. A stock of l-tryptophan analogue N-acetyl-l-tryptophanamide (NATA; Sigma-Aldrich) was prepared at 40 mM in external solution containing 1 µM TMRM, pH 7.0, and its concentration was ascertained by its absorbance at 280 nm according to its extinction coefficient (5,690 M⁻¹). NATA solution was progressively added to measure its quenching effect on TMRM fluorescence emission. This was quantified according to the Stern-Volmer equation (Lakowicz, 2006):

where F₀ is peak TMRM fluorescence emission (at 573 nm) without NATA, F is the peak of TMRM fluorescence emission peak (573 nm) with NATA, K_SV is the Stern-Volmer quenching constant, and [NATA] is NATA concentration.

Fig. S1 shows a proposed mechanism to account for the composite fluorescence signal reported from position 20. This scheme explains how two distinct quenching processes can generate the observed fluorescence signal. It is available.

Recent studies have proposed that S0 closely associates with the voltage-sensing apparatus (Koval et al., 2007; Liu et al., 2008a, 2010), raising the prospect of its involvement in VSD operation. To investigate this hypothesis we sought to optically track voltage-dependent conformational rearrangements from its extracellular flank. A cysteine residue was substituted into position 17, 18, or 19 in a human BK_Ca channel α subunit clone (hSlo) devoid of native extracellular cysteines (Cless). The mutation R207Q was also introduced to increase P_O at low [Ca] (Díaz et al., 1998). Xenopus oocytes expressing these channels were incubated with the environment-sensitive fluorophore TMRM, which covalently bound to the engineered cysteine (Fig. 2, A, E, and I). The labeled oocytes were mounted on a cut-open oocyte Vaseline gap voltage clamp setup modified for epifluorescence measurement to simultaneously record ionic current (Fig. 2, B, F, and J) and fluorescence emission. Pronounced voltage-dependent fluorescence deflections (ΔF/F), which underlie protein rearrangements, were reported from all tested positions (17, 18, and 19) on the extracellular flank of S0 (Fig. 2, C, G, and K). The similarity in time course, amplitude, and voltage dependence (Table I) of the fluorescence deflections resolved from position 17, 18, or 19 suggests that fluorophores at these three sites track the same molecular process. Their voltage dependence is indeed similar to that of protein rearrangements reported from voltage-sensing segments S2 and S4 (Savalli et al., 2006; Pantazis et al., 2010) (Table I), suggesting a process related to the operation of the BK_Ca voltage sensor.

The conformational rearrangements reported from positions 17, 18, and 19 could imply a voltage-dependent movement of S0. To undergo voltage-dependent motion, a transmembrane segment needs to either possess voltage-sensing charge(s) (i.e., charged residues experiencing the membrane electric field; Tombola et al., 2006; Bezanilla, 2008; Chanda and Bezanilla, 2008; Swartz, 2008) or interact with another voltage-sensing segment. According to the hydropathy profile of S0, a positively charged arginine at position 20 maps at the interface between the membrane and the extracellular solution, whereas another arginine, R44, defines the putative cytosolic flank of S0 (Wallner et al., 1996; Meera et al., 1997).

Accordingly, R20 could either experience a fraction of the membrane electric field or be involved in electrostatic interactions with charged residues in another voltage-sensing segment. We have investigated the voltage dependence of protein rearrangements and ionic conductance in channels bearing mutation R20A and labeled at S0 (Fig. 3 A), S2 (Fig. 3 F), or S4 (Fig. 3 K). R20A mutants exhibited a positive shift in the voltage dependence of ionic conductance by ≈30 mV (Fig. 3, D, I, and N, and Table I). The voltage dependence of conformational rearrangements reported from S0, S2, and S4 exhibited a shift in the same direction (Fig. 3 E, J, and O, and Table I). However, although the R20A mutation appeared to perturb the voltage-dependent activation of BK_Ca channels, it failed to significantly reduce the voltage dependence of the conformational rearrangements reported from S0. That is, the effective valence (z) of ΔF/F resolved from S0-labeled channels was not significantly reduced by mutation R20A (Table I). In contrast, voltage-sensing charge neutralization within BK_Ca S2 and S4 domains reduced the effective valence of these segments by ≈70% (Pantazis et al., 2010). These results suggest that R20 is not responsible for the voltage-dependent motions reported from the extracellular flank of S0; it is therefore unlikely to have a role as a voltage-sensing charge.

We did not achieve functional channel expression in a channel bearing the mutation R44A or R44Q at the cytosolic tip of S0, preventing a further characterization of the role of this charge in the wild-type (wt) channel.

To better characterize the nature of the conformational rearrangements detected from S0, we labeled position 20 (R20C) with TMRM (Fig. 4 A). The fluorescence deflections resolved from this position are intriguing in that they exhibit two distinct components: upon depolarization, the fluorescence is initially quenched (Fig. 4 F, state 1), followed by unquenching (state 2). Upon repolarization, the fluorescence is transiently unquenched further (state 3), followed by return to resting state (state 0). Thus, TMRM-labeled position 20 reports two kinetically distinct processes with opposite effects on the fluorescence emission. The two processes also possess different voltage dependence at steady state: for voltage steps up to −60 mV, only the quenching component is evident, as shown by the downward fluorescence deflections (Fig. 4 C). However, for higher test potentials, the unquenching process becomes prominent, resulting in an overall positive fluorescence deflection (Fig. 4, C and F). As such, the steady-state F-V relationship is a convolution of two processes with opposite effects on fluorophore emission and distinct voltage dependence (Fig. 4 E). An implication of the fluorescence at position 20 tracking two processes with distinct kinetics and voltage dependence is that the fluorophore could be affected by molecular events outside the S0 helix.

Given the close association of S0 and S4 (Liu et al., 2008a, 2010), and the complex ΔF/F profile of channels labeled at position 20 (see above), we hypothesized that the fluorescence deflections observed from the S0 flank could be a result of the voltage-dependent interaction between the fluorophore and the quenching group in another VSD segment, specifically W203 outside S4, as tryptophan residues are able to quench small fluorophores. The quenching effect is based on electron transport from the tryptophan side chain to the excited state of the fluorophore, and it has been exploited to investigate structural properties in other proteins (Mansoor et al., 2002; Islas and Zagotta, 2006), as well as to resolve the structure of the short S3–S4 linker in BK_Ca channels (Semenova et al., 2009). A previous fluorometric investigation on the voltage dependence of S4 showed that W203 affects the emission of fluorophores (TMRM or PyMPO), labeling the adjacent position 202 (Savalli et al., 2006). We therefore investigated the effects of mutation W203V on the fluorescence deflections resolved from positions 17, 18, 19, and 20 at the extracellular flank of S0.

When position 18 at the S0 extracellular flank was labeled with TMRM in channels bearing mutation W203V (Fig. 5 A), the amplitude of the depolarization-evoked ΔF/F was diminished by an order of magnitude (Fig. 5, B and G). The same was observed when labeling positions 17 and 19 (Fig. 5 G). This result suggests that, at rest, W203 acts as a fluorescence quencher for the fluorophore, labeling the region extracellular to S0. Depolarization evokes a relative motion between the two segments so that the quenching effect is lifted, and a positive ΔF/F is observed. Accordingly, this effect is abolished by mutation W203V. Moreover, in W203V channels labeled at position 20 (Fig. 5 D), the unquenching component is abolished, unmasking the residual quenching process (Fig. 5, E and F).

To confirm tryptophan’s ability to quench TMRM fluorescence, we assessed the quenching efficiency of an uncharged l-tryptophan derivative that resembles the structure of nonterminal Trp in proteins (NATA; Fig. 6 A) on TMRM in solution. The concentration-dependent quenching effect of NATA on 1 µM TMRM is shown in Fig. 6 C. The estimated Stern-Volmer quenching constant is 41.3 M⁻¹ (Fig. 6 D), within the same order of magnitude as the tryptophan quenching efficiency for bimane fluorescence: 83 M⁻¹ (Islas and Zagotta, 2006).

The nonconserved BK_Ca transmembrane domain S0 and the short extracellular N-terminal tail that precedes it are important for the functional interaction between channel-forming α subunits and regulatory β subunits (Wallner et al., 1996; Morrow et al., 2006). More recently, it was reported that an interaction site could exist between S0 and the voltage-sensing apparatus of BK_Ca channels (Koval et al., 2007), whereas S0 was found to associate closely with VSDs S3 and S4, so that disulfide bridge formation between their extracellular regions is highly efficient (Liu et al., 2008a, 2010). Intrigued by this intimate association, we optically investigated voltage-dependent conformational rearrangements from the proximity of S0 (Figs. 2 and 4), testing the hypothesis that S0 undergoes voltage-dependent motions during activation.

The TMRM fluorescence deflections observed from positions immediately extracellular to S0 (17, 18, and 19; Fig. 2) are exceptionally large compared with those resolved from transmembrane segments with intrinsic voltage dependence, S2 and S4 (see ΔF/F amplitude scales in Fig. 3 and Savalli et al., 2006, 2007; Pantazis et al., 2010). When large fluorescence deflections are observed from a contiguous series of labeling positions in a protein segment, it is reasonable to assume that they are a result of an extensive movement of the entire labeled region, as suggested by Pathak et al. (2007). However, we were unable to experimentally establish whether S0 undergoes large movements owing to intrinsic voltage dependence because neutralization of R20 did not perturb the large ΔF/F signal, and R44 mutants could not be investigated. Nevertheless, a clue for the source of the large ΔF/F signal came from the fluorescence signature of position 20: the composite ΔF/F reported from this location (Fig. 4) indicates that at least two different voltage-dependent processes, with distinct time- and voltage-dependent properties, influence the emission of the attached fluorophore, raising the possibility that the fluorescence emission is perturbed by events outside S0.

Indeed, the tryptophan residue at the extracellular tip of S4 (W203) was found to be the principal molecular source of the fluorescence deflections observed from positions 17–20, as mutation W203V abolished most of the depolarization-dependent positive ΔF/F component, while apparently not perturbing the quenching component observed from position 20 (Fig. 5). Therefore, the fluorescence deflections resolved must be primarily a result of the relative motion between W203 (S4) and the fluorophore attached to the extracellular flank of S0. The structural interpretation of these results is that the two segments are within collisional proximity at rest, in agreement with the conclusions of previous investigations (Koval et al., 2007; Liu et al., 2008a, 2010). However, upon VSD reorganization after membrane depolarization, the two segments move apart, relieving the fluorophore of the quenching effect of W203. Based on this interpretation, we present a model accounting for the experimental observations in Fig. 7.

We propose a mechanism to account for this complex fluorescence signal, whereupon W203 and another quencher influence TMRM fluorescence, giving rise to the observed positive and negative ΔF/F, respectively (Fig. S1). The nature of the second quencher remains unknown. However, because position 20 is thought to be near the membrane, we speculate that the labeling fluorophore is influenced by transitions at the interface of the lipid bilayer and the polar solution. Indeed, this quenching process (negative ΔF/F) is fully unmasked by mutation W203V (Fig. 5 E).

The positions investigated in this work (17–20 in the S0 extracellular flank) are located in the N-terminal tail of the channel (Wallner et al., 1996; Morrow et al., 2006) and are not thought to be involved in a helical structure (Liu et al., 2008a). Lack of structure in the N-terminal region of the α subunit is compatible with this work. A fluorophore labeling the substituted cysteine in a protein samples a volume of space that is influenced by the flexibility of the protein backbone it is attached to. If the protein backbone spanning positions 17–19 (in the BK_Ca N-terminal tail) were part of an α helix, the residue side chains, and consequently the fluorophore space sampled at each position, would be facing 100° away from each other. However, TMRM labeling positions 17–19 is quenched by W203 with comparable efficiency at rest, so that its depolarization-evoked departure gives rise to fluorescence deflections with similar amplitude and kinetics (Figs. 2 and 5), supporting the view that the protein backbone spanning positions 17–19 is unstructured. In contrast, the W203- and depolarization-dependent unquenching process reported from position 20 is much smaller in amplitude (Fig. 5 G). This result could be explained by a more rigid geometry between positions 19 and 20, thus defining the N-terminal boundary of the S0 helix. Alternatively, the expected proximity of position 20 to the membrane could affect its accessibility for labeling, or the fluorophore at position 20 could simply be further away from W203 at rest, so it is less efficiently quenched. As such, these experiments define a most N-terminal limit for a helical S0 structure at position 19.

BK_Ca channels possess a “decentralized” distribution of voltage-sensing charged residues across the VSD: two residues in S2 (D153 and R167), one in S3 (D186), and one in S4 (R213) (Ma et al., 2006); indeed, these segments have been demonstrated to undergo voltage-dependent conformational transitions (Savalli et al., 2006, 2007; Pantazis et al., 2010). S1 lacks intrinsic voltage dependence (Ma et al., 2006), whereas the voltage-sensing ability of S0 is unknown as, according to its hydropathy profile, it does not bear membrane-immersed charged residues. Nevertheless, S0 is thought to be a critical component of the BK_Ca voltage–sensing apparatus: ΔS0 channels are translated (Wallner et al., 1996; Morrow et al., 2006) but do not produce current—a condition that can be rescued by the coinjection of cRNA encoding S0 (Wallner et al., 1996). Moreover, tryptophan substitution into specific positions in S0 can impair voltage-dependent activation (Koval et al., 2007). By considering the relative voltage-dependent motion between S4 and S0 demonstrated in this work, we propose that S0 could act as a pivot component in the voltage-sensing apparatus of BK_Ca channels, against which S4 moves to facilitate channel activation. Furthermore, the transmembrane segments of auxiliary β subunits were recently discovered to localize near segments S0 and S1 (Liu et al., 2008b, 2010; Wu et al., 2009). Association of β subunits with pivot components could in turn affect the structural transitions of voltage-sensitive segments and thus provide a mechanism for channel modulation.

We are grateful to Michela Ottolia and members of the Olcese laboratory for critical comments on the manuscript. We thank Ligia Toro (University of California, Los Angeles, CA) for the hSlo clone.

This work was supported by research grants from National Institutes of Health (NIGMS R01GM082289 to R. Olcese) and American Heart Association (Western States Affiliate; postdoctoral fellowship 09POST2250648 to A. Pantazis).

Christopher Miller served as editor.