Voltage-dependent conformational changes in

Download PDF
1 downloads 0 Views 1MB Size Report
Comment
Aug 15, 2006 - explaining the unquenching of the fluorescence. However, the time course of these movements is slower than the one expected for voltage ...
Voltage-dependent conformational changes in human Ca2ⴙ- and voltage-activated Kⴙ channel, revealed by voltage-clamp fluorometry Nicoletta Savalli*, Andrei Kondratiev*, Ligia Toro*†‡§, and Riccardo Olcese*‡§¶ Departments of *Anesthesiology–Division of Molecular Medicine and †Molecular Pharmacology, ‡Brain Research Institute, and §Cardiovascular Research Laboratory, David Geffen School of Medicine at University of California, Los Angeles, CA 90095-7115 Edited by Ramo´n Latorre, Center for Scientific Studies, Valdivia, Chile, and approved June 6, 2006 (received for review February 11, 2006)

BK 兩 gating current 兩 MaxiK 兩 fluorescence 兩 voltage sensor

oltage- and Ca2⫹-activated K⫹ channels (BKCa) are membrane proteins that play a fundamental role in controlling smooth muscle tone and neuronal excitability. In most of the tissues, they form a complex consisting of a pore-forming ␣ subunit and regulatory ␤ subunits. The ␣ subunit encodes for the selective pore as well as for the voltage and Ca2⫹-sensing structures. The BKCa channel is a tetramer with each ␣ subunit organized in seven transmembrane domains (S0–S6) (1), a long intracellular C-terminal domain where a high-affinity Ca2⫹binding site has been identified (2, 3), and an extracellular N terminus. The human isoform (hSlo), similarly to other voltagedependent ion channels, possesses a voltage sensor that is mainly represented by the S4 transmembrane domain, containing three positively charged residues (4, 5). Changes in membrane potential displace the voltage sensor and, for adequate depolarizations, the consequent conformational change sets the channel in a conducting state. The movement of the voltage sensor produces a transient current (gating current) that precedes in time and voltage the ionic current activation (6, 7). Thus, gating currents report on the rearrangement of the channel structure in a varying membrane potential but do not provide direct information regarding the motion of regions of the channel outside the voltage field. Structural changes occurring during gating have been elegantly resolved by using site-directed fluorescent labeling, a technique pioneered in the E. Isacoff laboratory (8) and applied to a variety of voltage-gated K⫹ and Na⫹ channels (9–24), ligand-gated channels (25, 26), and transporters (27–31). However, nothing is known regarding the dynamical changes of BKCa channel during gating. Using site-directed fluorescence labeling combined with the cut-open oocyte voltage clamp technique (COVG), we have resolved the conformational changes occurring in hSlo voltage-sensing region, unraveling extremely slow conformational changes not expected from gating current measurements. We have used thiol-reactive fluores-

V

www.pnas.org兾cgi兾doi兾10.1073兾pnas.0601176103

cent probes [tetramethyl rhodamine-5-maleimide (TMRM) or 1-(2-maleimidylethyl)-4-(5-(4-methox yphenyl) oxazol-2yl)pyridinium methansulfonate (PyMPO)] to assess the dynamics of the S4 region conformational changes in BKCa channels. Results BKCa channels possess a functional voltage sensor as demonstrated by direct measurement of ionic and gating currents in the absence of internal Ca2⫹ (6, 7, 32, 33). In addition, Diaz et al. (4) have shown that mutations in the S4 segment alter the voltage dependence of hSlo channel activation. If the S4 transmembrane segment of BKCa channels is part of the voltage-sensing machinery, conformational changes of the S4 region should share some of the features of channel-gating currents. In this study, we have investigated conformational changes of the region between the S3 and S4 transmembrane segments in the hSlo channel. The region of interest and the residues fluorescently labeled in this study are illustrated in a schematic drawing in Fig. 1A. In hSlo, the S3–S4 linker is almost absent, and a sequence alignment of this region shows no homology with KVAP, Shaker, and KV1.2. All of the positions were studied by using two fluorophores sensitive to the environment, TMRM and PyMPO (ref. 34; Fig. 1B). Absence of Voltage-Dependent Fluorescence Changes (⌬F兾F) in the C-Less Channels. We initially investigated whether our background

clone (hSlo C-less R207Q) gives rise to change in voltagedependent fluorescence upon depolarization. We used the R207Q mutant to facilitate channel opening (4). Oocytes injected with cRNA encoding for the C-less clone yielded a high expression level of the channel. Attesting for the appropriateness of the C-less clone, no significant changes in the fluorescence baseline were detected, with either TMRM or PyMPO labeling (see Fig. 7, which is published as supporting information on the PNAS web site). The maximum ⌬F兾F (for a potential step from –160 mV to 100 mV) was 0.0052 ⫾ 0.0011% (TMRM, n ⫽ 6) or 0.0107 ⫾ 0.0013% (PyMPO, n ⫽ 3). Conformational Changes Reported by Positions 199 and 200: Fluorescence Correlates with One Component of Ionic Current Kinetics. K⫹

currents arising from oocytes injected with L199C mutant are reported in Fig. 2A. TMRM labeling of the cysteine inserted at this position (probably located at the C terminus of S3) elicited, upon depolarization, an extremely small (but resolvable) voltage-dependent change of the fluorescence emission (Fig. Conflict of interest statement: No conflicts declared. This paper was submitted directly (Track II) to the PNAS office. Abbreviations: COVG, cut-open oocyte voltage clamp technique; MES, methanesulfonate; NMG, N-methyl-D-glucamine; PyMPO, 1-(2-maleimidylethyl)-4-(5-(4-methoxyphenyl) oxazol2-yl) pyridinium methansulfonate; TMRM, tetramethyl rhodamine-5-maleimide. ¶To

whom correspondence should be addressed. E-mail: [email protected].

© 2006 by The National Academy of Sciences of the USA

PNAS 兩 August 15, 2006 兩 vol. 103 兩 no. 33 兩 12619 –12624

PHYSIOLOGY

Large conductance voltage- and Ca2ⴙ-activated Kⴙ (BKCa) channels regulate important physiological processes such as neurotransmitter release and vascular tone. BKCa channels possess a voltage sensor mainly represented by the S4 transmembrane domain. Changes in membrane potential displace the voltage sensor, producing a conformational change that leads to channel opening. By site-directed fluorescent labeling of residues in the S3–S4 region and by using voltage clamp fluorometry, we have resolved the conformational changes the channel undergoes during activation. The voltage dependence of these conformational changes (detected as changes in fluorescence emission, fluorescence vs. voltage curves) always preceded the channel activation curves, as expected for protein rearrangements associated to the movement of the voltage sensor. Extremely slow conformational changes were revealed by fluorescent labeling of position 202, elicited by a mutual interaction of the fluorophore with the adjacent tryptophan 203.

Fig. 1. Topology of hSlo ␣ subunit and the structure of the fluorophores used in this study. (A) hSlo ␣ subunit has seven transmembrane domains (S0 –S6), an extracellular N terminus, and a C-terminal region that comprises four hydrophobic domains (S7–S10). Sequence alignment of the S3–S4 regions of hSlo, Shaker, KVAP, and KV1.2 K⫹ channels is shown. Underlined residues were mutated into cysteines for site-directed fluorescent labeling (L199C, N200C, R201C, and S202C). Note the very short hSlo S3–S4 linker, similar in length to KVAP. (B) Structures of the membrane impermeable thiol-reactive fluorophores (PyMPO and TMRM) used in this study.

2B). The onset of the fluorescence was properly fit with a single exponential function, whereas the ionic current activation required a double exponential approximation (Fig. 2C). The time course of the conformational changes reported by TMRM labeling position 199 closely follows the fast component of the ionic current activation that accounts for ⬇65% of the total current (Fig. 2C; see Table 1, which is published as supporting information on the PNAS web site). As shown in Fig. 2D, the voltage dependence of the fluorescence signal [F(V)] preceded the activation curve [G(V)] by ⬇80 mV, as expected for protein rearrangement associated to the movement of the channel voltage-sensing region. The half activation potentials (Vhalf) were FVhalf ⫽ ⫺50.10 ⫾ 2.93 mV and GVhalf ⫽ 27.92 ⫾ 3.35 mV (n ⫽ 4). Similar results were obtained by labeling the same position with PyMPO (data not shown). The adjacent position (N200) is located probably at the outmost end of S3. We were able to label a cysteine placed at this position only with TMRM. Positions 200 and 199 were the most resistant to fluorophores labeling, probably because they are still part of S3. The labeling with TMRM elicited a strong ⌬F, suggesting that this position is experiencing a clear change in the surrounding environment. A striking feature of the fluorescence signal arising from N200C was its very slow time course during depolarization. At all potentials tested, fluorescence lagged ionic current activation (Fig. 2 E–G) reporting conformational changes ⬇10 to 20 times slower than the preceding position (199) (see Table 1). The time course of fluorescence activation reported by this mutant seems to follow the slow component of ionic current activation, as shown in the ␶ vs. membrane potential plots (Fig. 2G). Despite these clear kinetics differences between fluorescence arising from L199C and N200C (probably due to different local interactions of the fluorophore with the environment), their steady-state properties were similar (FVhalf ⫽ ⫺77.69 ⫾ 4.96 mV and GVhalf ⫽ ⫺19.72 ⫾ 10.11, n ⫽ 5) giving a separation of F(V) and G(V) of ⬇60 mV (Fig. 2H). The lack of fluorescence signal when using PyMPO at position 200 (data not shown) reveals the importance of the fluorophore structural properties in this type of study. Properties of Charge Movement and Fluorescence Changes at Position R201C. We have labeled the cysteine in position 201 with both

TMRM and PyMPO fluorophores and recorded the ionic cur12620 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0601176103

Fig. 2. Conformational changes reported by TMRM labeling positions 199 and 200. Representative K⫹ current traces from oocytes expressing L199C (A) (10 mM ext. K⫹) or N200C (E) (50 mM ext. K⫹) clones, elicited by a 100-ms depolarization. The corresponding TMRM fluorescence traces are shown in B and F. The best fits to a single exponential function are superimposed. The ON fluorescence and the ionic current activation were fit to a mono- and biexponential function, respectively. The time constants of fluorescence (F) and ionic current activation [␶fast (E) and ␶slow (‚)] are shown for L199C (C) and N200C (G) labeled with TMRM. Time course of fluorescence reported by TMRM labeling residue 199 follows the fast component of current activation (C). Fluorescence reported by position 200 displays a time course comparable with the slow component of current activation (G). (D and H) Averaged F(V) (F) and G(V) (E) curves for L199C and N200C, respectively. F(V) is leftward shifted on the membrane potentials axis of ⬇80 mV (D) or ⬇60 mV (H). Single Boltzmann fitting parameters L199C: F(V), Vhalf ⫽ ⫺50.00 mV, z ⫽ 0.83; G(V), Vhalf ⫽ 28.13 mV, z ⫽ 1.05. N200C: F(V), Vhalf ⫽ ⫺79.32 mV, z ⫽ 0.96; G(V), Vhalf ⫽ ⫺18.54 mV, z ⫽ 0.85.

rents (Fig. 3 A and B) and the corresponding simultaneous fluorescence traces (Fig. 3 C and D) from the R201C clone. The kinetics of the fluorescence during depolarization appeared to be almost identical with both TMRM (Fig. 3E) and PyMPO (Fig. 3F) (see Table 1). The fluorescence from R201C labeled with TMRM (but not PyMPO) displays a slow relaxation that becomes more pronounced with large depolarizations. Although the voltage dependence of G(V) curves for channels labeled at position 201 with TMRM (Fig. 3G) or PyMPO (Fig. 3H) was not significantly different (GVhalf ⫽ 3.41 ⫾ 5.04 mV, n ⫽ 6 with TMRM and GVhalf ⫽ 9.34 ⫾ 2.98 mV, n ⫽ 6 with PyMPO), the F(V) curve reported by TMRM was ⬇30 mV more negative than the F(V) curve reported by PyMPO (FVhalf ⫽ ⫺60.77 ⫾ 6.08 mV, n ⫽ 6 with TMRM and FVhalf ⫽ ⫺24.65 ⫾ 4.32 mV, n ⫽ Savalli et al.

Fig. 3. Conformational changes reported by TMRM and PyMPO labeling position 201. (A and B) Representative K⫹ current traces from oocytes expressing R201C clone, labeled with TMRM and PyMPO, respectively. The corresponding fluorescence traces are shown in C and D. The time constants of fluorescence (F) and ionic current activation [␶fast (E) and ␶slow (‚)] are shown for TMRM (E) and PyMPO (F) labeling. Fluorescence reported by TMRM (E) follows a time course comparable with the one reported by PyMPO (F) (fitting as described in Fig. 2). (G and H) Averaged F(V) (F) and G(V) (E) curves for R201C labeled with TMRM and PyMPO, respectively. Single Boltzmann fitting parameters: (G) TMRM: F(V), Vhalf ⫽ ⫺63.00 mV, z ⫽ 0.69; G(V), Vhalf ⫽ 4.91 mV, z ⫽ 0.91; (H) PyMPO: F(V), Vhalf ⫽ ⫺26.99 mV, z ⫽ 0.71; G(V), Vhalf ⫽ 11.25 mV, z ⫽ 1.02.

6 with PyMPO). This discrepancy led us to investigate the voltage dependence of the gating currents in the R201C clone. The gating currents were recorded after complete K⫹ depletion of oocytes expressing R201C clone (see Materials and Methods) labeled with either TMRM (Fig. 4A) or PyMPO (Fig. 4B). The voltage dependence of the time integral of the On-gating current [Q(V)] shares a close voltage dependence with the one of TMRM fluorescence signal (Fig. 4C), suggesting that the fluorophore bound at this position is faithfully reporting conformational changes closely related to the charge displacement of the BKCa voltage-sensing region. On the other hand, the Q(V) curve constructed from gating currents arising from channels labeled with PyMPO (Fig. 4D) was leftward shifted on the voltage axis compared with the F(V), suggesting that the two fluorophores might sense different rearrangements of the channels. Note that the fitting parameters of the two Q(V) curves (TMRM, Vhalf ⫽ ⫺65.60 mV, z ⫽ 0.81; PyMPO, Vhalf ⫽ ⫺65.56 mV, z ⫽ 0.87) are very similar, proving that PyMPO does not affect gating. Superslow Conformational Changes Reported by PyMPO Labeling Position 202. As shown in Fig. 5A, we were able to record large

K⫹ currents from the S202C mutant. However, TMRM fluoSavalli et al.

rescent labeling of position 202 produced a very weak but resolvable voltage-dependent ⌬F (Fig. 5B). Fluorescent signal had fast on and off kinetics that does not seem to have a correlation with ionic current activation (see Table 1 and Fig. 5C). The time constant of fluorescence kinetics as a function of membrane potential lies in between the two components of ionic current. For example, at 80 mV, fluorescence rises with ␶ ⫽ 1.42 ⫾ 0.17 ms, whereas for the same potential, the ionic current time constants were ␶fast ⫽ 0.33 ⫾ 0.01 ms and ␶slow ⫽ 6.46 ⫾ 1.66 ms (n ⫽ 3). On the other hand, PyMPO revealed extremely slow conformational changes taking place within the S4 region (see Table 1). Note the different time scales of the recordings in Fig. 5 E and F (PyMPO) compared with Fig. 5 A and B (TMRM). The fluorescence signal rises slowly during 1-s depolarization step (␶ ⫽ 243 ⫾ 16 ms, –160 mV to 80 mV step, n ⫽ 5; Fig. 5F). The return (off fluorescence) to the baseline potential during repolarization to ⫺160 mV followed a relatively faster time constant (␶ ⫽ 43.37 ⫾ 12.12 ms, n ⫽ 3; Fig. 5F). The fluorescence kinetics observed by PyMPO-labeling residue 202 had a voltage dependence opposite to the one of ionic current activation (Fig. 5G). Although both components of the ionic current became faster with larger depolarization, fluorescence time constants increased with the membrane potential. Thus, for potentials ⬎20 mV, the fluorescence significantly displayed kinetics slower than the slow component of the ionic current. For example, at 100 mV, the time constant of the fluorescence and of the slow component of current activation were ␶ ⫽ 267.49 ⫾ 12.03 ms and ␶ ⫽ 37.58 ⫾ 11.53 ms (n ⫽ 5), respectively. These results reveal an unexpected slow rearrangement of the BKCa channel during normal gating. Although the open probability has stabilized after a few milliseconds of the depolarizing step, the fluorescence signal keeps rising. Because it is established that BKCa channels possess multiple open states (7, 35), we speculate that this slow rearrangement represents slow transitions among open states. Despite the slow onset, the change in fluorescence intensity increased with depolarizations displaying a voltage dependence PNAS 兩 August 15, 2006 兩 vol. 103 兩 no. 33 兩 12621

PHYSIOLOGY

Fig. 4. Gating currents from R201C share the same voltage dependence of TMRM fluorescence. (A and B) Gating currents traces from oocytes expressing R201C clone, labeled with TMRM and PyMPO, respectively, are elicited by depolarizations to the indicated potential after K⫹ depletion (see Materials and Methods). Averaged F(V) (F) and G(V) (E) curves for R201C labeled with TMRM (C) or PyMPO (D) are reported for comparison from Fig. 3 G and H, respectively. (C) The Q(V) curve (䊐) constructed from the experiment in A is plotted in the same graph. Single Boltzmann fitting parameters: Q(V) Vhalf ⫽ ⫺65.60 mV, z ⫽ 0.81. (D) Q(V) Vhalf ⫽ ⫺65.56 mV, z ⫽ 0.87. Data points (䊐) are the average of two experiments.

Fig. 5. TMRM and PyMPO labeling position 202 report different conformational changes. (A and E) representative K⫹ current traces from oocytes expressing S202C clone, labeled with TMRM and PyMPO respectively. Note the different time scale. The corresponding fluorescence traces are shown in B and F. The time constants of fluorescence (F) and ionic current activation [␶fast (E) and ␶slow (‚)] are shown for TMRM (C) and PyMPO (G) labeling. Fitting is performed as in Fig. 2. Fluorescence reported by PyMPO (F and G) reveals superslow conformational changes of the S3–S4 region of BKCa channel. (D and H) The averaged F(V) (F) and G(V) (E) curves for S202C labeled with TMRM and PyMPO, respectively, are shown. Single Boltzmann fitting parameters: (D) TMRM: F(V), Vhalf ⫽ ⫺49.00 mV, z ⫽ 0.82; G(V), Vhalf ⫽ ⫺45.94 mV, z ⫽ 1.29. (H) PyMPO: F(V), Vhalf ⫽ ⫺75.00, z ⫽ 1.20; G(V), Vhalf ⫽ ⫺4.10 mV, z ⫽ 1.17.

Fig. 6. Mutual interaction between W203 and PyMPO labeling position 202. Representative K⫹ current traces from oocytes expressing S202C W203V mutant, labeled with PyMPO (A) and TMRM (C). The corresponding fluorescence traces are reported in B and D, respectively. Note the absence of PyMPO ⌬F when W203 is mutated into a valine (B). TMRM fluorescence remains unchanged after W203V; the best fits to a single exponential function are shown superimposed (D). (E) Averaged F(V) (F) and G(V) (E) curves for S202C W203V labeled with TMRM. Single Boltzmann fitting parameters: F(V), Vhalf ⫽ ⫺80.87 mV, z ⫽ 0.83; G(V), Vhalf ⫽ ⫺4.69 mV, z ⫽ 1.05. (F) The time constants of TMRM fluorescence (F) and ionic current activation [␶fast (E) and ␶slow (‚)] are shown in the same plot. Fitting is performed as in Fig. 2.

Tryptophan 203 in the Upper S4 Reports the Slow Conformational Changes by Quenching PyMPO Fluorescence. Aromatic residues are

S202C W203V clone still could be labeled with TMRM and ionic currents and fluorescence changes could be efficiently recorded during depolarization (Fig. 6 C and D). In addition, this view has been confirmed by previous incubation with PyMPO that prevented TMRM labeling. The fluorescence signals acquired with both the TMRM and PyMPO filter set did not display any voltage dependence (see Fig. 9, which is published as supporting information on the PNAS web site). In this condition, the lack of TMRM fluorescence (⌬F兾F ⫽ 0.004 ⫾ 0.003%, n ⫽ 7) indicates that the cysteines in position 202 are already labeled with PyMPO (Fig. 9E). Therefore, the accessibility to the cysteine, and steady-state voltage dependence and kinetics of both ionic currents and fluorescence signals, were not compromised by the W203V mutation (Fig. 6 E and F). The specificity of the quenching effect of W203 on PyMPO labeling 202 is supported by the fact that W203V substitution did not affect ionic current and fluorescence signal in the R201C W203V clone after labeling with either PyMPO or TMRM (data not shown).

excellent fluorescence quenchers. We tested the hypothesis that the slow conformational changes observed with PyMPO labeling position 202 were reported by a mutual interaction between the fluorophore and the adjacent tryptophan (W203). When W203 was mutated into a valine (W203V), despite the fact that current properties and expression levels were normal (Fig. 6A), the labeling of S202C with PyMPO did not give rise to voltagedependent ⌬F (Fig. 6B). The absence of PyMPO ⌬F in S202C W203V is not a consequence of lack of labeling (Fig. 6B). In fact,

Discussion The data presented here revealed the motion of the region between S3 and S4 transmembrane segments comprising a stretch of amino acids facing the extracellular side of the plasma membrane. Each position reports distinctive kinetics information and for some residues the fluorescence signal lags the ionic current. Nevertheless, for all of the positions tested, the F(V) curve preceded the G(V) curve. This result is not surprising if

that shares the same features of the F(V) curves constructed for other positions: F(V) precedes on the voltage axis the ionic current activation (Fig. 5H). These slow fluorescence signals seem to report true conformational changes of the protein, because they do not depend on the ionic flux. Replacing the external solution with 120 mM tetraethylammonium (TEA) completely blocked the ionic flux, leaving the fluorescence signal practically unchanged (see Fig. 8, which is published as supporting information on the PNAS web site).

12622 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0601176103

Savalli et al.

Savalli et al.

An inherent limitation of this type of studies is the possibility that in some cases the fluorophores could report their own adaptation to a different conformation (e.g., a change in the dihedral angle between the aromatic rings), instead of protein rearrangements. Slow docking of the dye onto a part of the channel that was exposed previously during gating is another possibility. However, we still favor our main conclusion, because we also have observed rather slow conformational changes with the other fluorophore (TMRM) labeling a different position (C200, ␶ ⬇30–40 ms) (Fig. 2), strongly implying that BKCa channels can undergo relatively slow rearrangements that do not depend on the nature of the fluorophore used and on the residue labeled. We found that fluorescence signal arising from PyMPO labeling position 202 was abolished when the adjacent tryptophan (W203) was mutated into a valine. The ability of aromatic rings to quench fluorescence is well established (34). Our interpretation of these results is that at hyperpolarized potentials, W203 exerts a strong quenching effect on PyMPO bound to C202. The mutual interaction between PyMPO in position 202 and W203 decreases at depolarized potentials, unquenching PyMPO fluorescence emission and revealing the slow kinetics of this process. If C202 is part of the hinge point between S3 and S4, a tilting or a rotation of the S4 helix during channel gating (similar to the one proposed for Shaker channels; refs. 9 and 19) can increase in the relative distance between PyMPO and W203, explaining the unquenching of the fluorescence. However, the time course of these movements is slower than the one expected for voltage sensor activation. The quenching interaction between W203 and PyMPO likely occurs only within the same subunit. In fact, taking as a reference the KV1.2 structure (38), PyMPO labeling the N-terminal side of S4 does not seem able to interact with W203 in the neighboring subunits (see Fig. 10, which is published as supporting information on the PNAS web site). The successful PyMPO labeling of this position has been confirmed by the observation that preincubation with PyMPO prevented TMRM labeling. In summary, voltage clamp fluorometry has revealed voltagedependent conformational changes of the S3–S4 region of BKCa channel revealing extremely slow rearrangements. The F(V) curves always preceded the channel activation curves, as expected for conformational changes associated with the movement of the voltage sensor. Materials and Methods Molecular Biology. We used the human (hSlo) BKCa clone (GenBank accession no. U11058) starting from methionine 4 (hSloM4) (39). The three native, extracellularly exposed cysteines have been mutated (C14S-C141S-C277S). Unique cysteines were introduced at positions 199, 200, 201, and 202, together with the R207Q mutation (4) to increase channel open probability. Site-directed mutagenesis by overlap extension with PCR (40) was adopted for all of the single-point mutations required by this study. All of the mutations were confirmed by sequence analysis (Laragen, Los Angeles, CA). The clones were in vitro transcribed (T7 mMessage; Ambion, Austin, TX), and the cRNA was injected into Xenopus oocytes for electrophysiological studies. Oocyte Preparation, cRNA Injection, and Fluorescent Labeling. Xenopus laevis (NASCO, Modesto, CA) oocytes (stages V–VI) were prepared as described in ref. 41. Oocytes were injected with 50 nl of cRNA at 0.01–0.2 ␮g兾␮l depending on the clone and batch of RNA. Injected oocytes were maintained at 18°C in an amphibian saline solution supplemented with 50 mg/ml gentamycin (Gibco BRL, Carlsbad, CA)兾200 ␮M DTT兾10 ␮M EDTA for 2–7 days before experiments. In prior fluorescence recordings, oocytes were incubated 30–40 min at room temperature in PNAS 兩 August 15, 2006 兩 vol. 103 兩 no. 33 兩 12623

PHYSIOLOGY

one considers that, as in the other voltage-dependent K⫹ channels, the movement of the S4 segment is a necessary condition for the voltage-dependent opening of the channel. Although L199C seems to track the fast component of ionic current activation, the labeling of the adjacent residue, N200C, reported a completely different pattern of fluorescence kinetics. In this case, the rising and falling of fluorescence at the beginning and at the end of the depolarizing step followed a time course similar to the slow component of the ionic current activation. Interestingly, voltage-dependent fluorescence changes were observed only by labeling this position with TMRM. Although it is reasonable that the slimmer molecular structure of PyMPO (see Fig. 1B) can allow the labeling of less accessible sites in the channel, it seems that, at this position (N200C), only the bulkier TMRM can interact with the surrounding environment, a necessary condition to detect conformational changes by a varying quenching interaction. The fluorescence changes we have described are most likely related to the movement of the charged S3–S4 region. In fact, the charge movement recorded for the R201C mutant shares a very similar voltage dependence of the fluorescence signal reported by TMRM labeling this position (Fig. 4C). The study of position 201 provided two important pieces of information. First, the neutralization of the positively charged arginine at this position (R201C) did not alter the voltage dependence of the channel: The effective valence (z) of the G(V) curve was 0.87 ⫾ 0.06 (n ⫽ 4) in the C-less clone and 1.06 ⫾ 0.05 (n ⫽ 6) and 0.98 ⫾ 0.07 (n ⫽ 6) in the R201C clone, labeled with PyMPO and TMRM, respectively. This result, in agreement with the work of Diaz et al. (4) and Ma et al. (5), confirms that R201 does not contribute to voltage sensing. Second, this position freely experiences changes in the local environment, as measured with both TMRM and PyMPO being the most accessible of the residues scanned. This result is consistent with an extremely short S3–S4 linker possibly formed only by R201 facing the extracellular solution, making this loop significantly shorter than the one usually found in other voltage-dependent K⫹ channels. On the other hand, residue 202 is probably at the interface between the lipid bilayer and the extracellular solution. In fact, PyMPO labeling this position reported the largest ⌬F兾F among the residues studied (⌬F兾F ⫽ 3.46 ⫾ 0.63%, n ⫽ 6), suggesting that the fluorophore is experiencing a considerable environmental change. These fluorescent signals display an extremely slow kinetics, which does not follow the time course of any of the ionic current components. Time constants of several hundreds of milliseconds are typical of slow inactivation processes. Although BKCa channels do not inactivate, they still possess slow protein motions, slower than any other observable macroscopic process. The BKCa channels changes in fluorescence intensity could report conformational changes occurring during transitions among open states. This result fits with the view that the channel possesses several interconnected open states with similar kinetic properties (7, 35). During activation, the channel can make transitions between different open states maintaining a steady macroscopic current. An analogous situation was observed in Shaker K⫹ channel, in which a double mutation abolished C type inactivation without removing the molecular rearrangements that underlie transitions among multiple open states (36). An emerging view is that BKCa channels are much more than structures engineered for ion conduction. Instead, they are part of multimeric complexes participating in cell signaling (37). An intriguing alternative explanation is that these very slow conformational changes (that have no correlation with kinetics observed in ionic currents) are not directly related to the ‘‘conduction duties’’ of the channel, rather to the exposure of parts of the protein involved in other functions, such as interaction with modulatory partners via binding sites that become unmasked by changes in membrane potentials.

a labeling solutions (120 mM K-methanesulfonate (MES) or NaMES兾2 mM CaCl2兾10 mM Hepes) containing 10 ␮M SHreactive f luorescent dye, TMRM, or PyMPO (Invitrogen– Molecular Probes, Eugene, OR). COVG Fluorometry. Fluorescence, ionic, and gating currents were recorded in voltage clamp condition by using the COVG (42) implemented for epifluorescence measurement (ref. 12; see Supporting Text, which is published as supporting information on the PNAS web site). COVG recording (ionic current). MES was the only anion in the external recording solutions. External solution was 60 mM NaMES, 10 or 50 mM K-MES, 2 mM Ca(MES)2, and 10 mM Na-Hepes. Internal solution is 110 mM K-glutamate and 10 mM K-Hepes. Micropipette solution is 2.7 mM NaMES and 10 mM NaCl. Ionic current was blocked by replacing the extracellular solution with 120 mM tetraethylammonium MES兾2 mM Ca(MES)2兾10 mM Hepes. For all solutions, pH ⫽ 7.0. All recordings were performed at 22–24°C. Holding potential ⫽ –90 mV. Gating currents. Currents were recorded after oocyte internal K⫹ depletion. Sustained depolarizations (⬎30 min) associated with a continuous perfusion of K⫹-free solution of impermeant ions [N-methyl-D-glucamine (NMG)] in the top middle and lower chamber of the COVG. This procedure, when associated with high expression level, causes an efficient K⫹ depletion, allowing the recording of gating current in isolation without the application of blockers. External solution is 110 mM NMG-MES, 2 mM 1. Wallner, M., Meera, P. & Toro, L. (1996) Proc. Natl. Acad. Sci. USA 93, 14922–14927. 2. Schreiber, M. & Salkoff, L. (1997) Biophys. J. 73, 1355–1363. 3. Bian, S., Favre, I. & Moczydlowski, E. (2001) Proc. Natl. Acad. Sci. USA 98, 4776–4781. 4. Diaz, F., Meera, P., Amigo, J., Stefani, E., Alvarez, O., Toro, L. & Latorre, R. (1998) J. Biol. Chem. 273, 32430–32436. 5. Ma, Z., Lou, X. J. & Horrigan, F. T. (2006) J. Gen. Physiol. 127, 309–328. 6. Stefani, E., Ottolia, M., Noceti, F., Olcese, R., Wallner, M., Latorre, R. & Toro, L. (1997) Proc. Natl. Acad. Sci. USA 94, 5427–5431. 7. Horrigan, F. T. & Aldrich, R. W. (1999) J. Gen. Physiol. 114, 305–336. 8. Mannuzzu, L. M., Moronne, M. M. & Isacoff, E. Y. (1996) Science 271, 213–216. 9. Cha, A., Snyder, G. E., Selvin, P. R. & Bezanilla, F. (1999) Nature 402, 809–813. 10. Cha, A., Ruben, P. C., George, A. L., Jr., Fujimoto, E. & Bezanilla, F. (1999) Neuron 22, 73–87. 11. Cha, A., Zerangue, N., Kavanaugh, M. & Bezanilla, F. (1998) Methods Enzymol. 296, 566–578. 12. Cha, A. & Bezanilla, F. (1998) J. Gen. Physiol. 112, 391–408. 13. Cha, A. & Bezanilla, F. (1997) Neuron 19, 1127–1140. 14. Sorensen, J. B., Cha, A., Latorre, R., Rosenman, E. & Bezanilla, F. (2000) J. Gen. Physiol. 115, 209–222. 15. Smith, P. L. & Yellen, G. (2002) J. Gen. Physiol. 119, 275–293. 16. Mannuzzu, L. M. & Isacoff, E. Y. (2000) J. Gen. Physiol. 115, 257–268. 17. Gandhi, C. S., Loots, E. & Isacoff, E. Y. (2000) Neuron 27, 585–595. 18. Glauner, K. S., Mannuzzu, L. M., Gandhi, C. S. & Isacoff, E. Y. (1999) Nature 402, 813–817. 19. Chanda, B., Asamoah, O. K., Blunck, R., Roux, B. & Bezanilla, F. (2005) Nature 436, 852–856. 20. Chanda, B., Asamoah, O. K. & Bezanilla, F. (2004) J. Gen. Physiol. 123, 217–230. 21. Loots, E. & Isacoff, E. Y. (2000) J. Gen. Physiol. 116, 623–636. 22. Schonherr, R., Mannuzzu, L. M., Isacoff, E. Y. & Heinemann, S. H. (2002) Neuron 35, 935–949. 23. Chanda, B. & Bezanilla, F. (2002) J. Gen. Physiol. 120, 629–645.

12624 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0601176103

Ca(MES)2, and 10 mM NMG-Hepes. Internal solution is 110 mM NMG-MES and 10 mM NMG-Hepes. Gating currents were recorded unsubtracted after analog compensation of the membrane linear capacity and resistive components at potential positive to 20-mV subtracting protocols. Holding potential ⫽ –90 mV. Data analysis. A customized program and fitting routines constructed with Microsoft Excel (Redmond, WA) were used for data analysis. Membrane conductance G(V), fluorescence F(V), and gating charge Q(V) curves were fitted to a Boltzmann distribution of the form: G(V) ⫽ Gmax兾(1 ⫹ exp(z(Vhalf ⫺ Vm)(F兾RT))), F(V)⫽((Fmax⫺Fmin)兾(1⫹exp(z(Vhalf ⫺Vm)兾(F兾RT)))⫹Fmin, and Q(V) ⫽ ((Qmax ⫺ Qmin)兾(1 ⫹ exp(z(Vhalf ⫺ Vm)兾(F兾RT)))) ⫹ Qmin, where Gmax, Fmax, and Qmax are the maximum G, F, and Q; Fmin and Qmin are the minima F and Q; z is the effective valence of the distribution, Vhalf is the half-activating potential, Vm is the membrane potential, and F, R, and T are the usual thermodynamic constants. We thank Chris Gandhi for insightful discussions. This work was supported by National Institutes of Health (NIH)兾National Institute of Neurological Disorders and Stroke Grant R01NS043240, American Heart Association Grant-in-Aid 0250170N (to R.O.), and NIH Grant HL054970 (to L.T.). 24. Bannister, J. P. A., Chanda, B., Bezanilla, F. & Papazian, D. M. (2005) Proc. Natl. Acad. Sci. USA 102, 18718–18723. 25. Zheng, J. & Zagotta, W. N. (2000) Neuron 28, 369–374. 26. Dahan, D. S., Dibas, M. I., Petersson, E. J., Auyeung, V. C., Chanda, B., Bezanilla, F., Dougherty, D. A. & Lester, H. A. (2004) Proc. Natl. Acad. Sci. USA 101, 10195–10200. 27. Geibel, S., Zimmermann, D., Zifarelli, G., Becker, A., Koenderink, J. B., Hu, Y. K., Kaplan, J. H., Friedrich, T. & Bamberg, E. (2003) Ann. N.Y. Acad. Sci. 986, 31–38. 28. Geibel, S., Kaplan, J. H., Bamberg, E. & Friedrich, T. (2003) Proc. Natl. Acad. Sci. USA 100, 964–969. 29. Larsson, H. P., Tzingounis, A. V., Koch, H. P. & Kavanaugh, M. P. (2004) Proc. Natl. Acad. Sci. USA 101, 3951–3956. 30. Loo, D. D., Hirayama, B. A., Cha, A., Bezanilla, F. & Wright, E. M. (2005) J. Gen. Physiol. 125, 13–36. 31. Meinild, A. K., Hirayama, B. A., Wright, E. M. & Loo, D. D. (2002) Biochemistry 41, 1250–1258. 32. Cui, J., Cox, D. H. & Aldrich, R. W. (1997) J. Gen. Physiol. 109, 647–673. 33. Horrigan, F. T., Cui, J. & Aldrich, R. W. (1999) J. Gen. Physiol. 114, 277–304. 34. Lakowicz, J. R. (1983) Principles of Fluorescence Spectroscopy (Plenum, New York). 35. McManus, O. B. & Magleby, K. L. (1991) J. Physiol. 443, 739–777. 36. Olcese, R., Sigg, D., Latorre, R., Bezanilla, F. & Stefani, E. (2001) J. Gen. Physiol. 117, 149–163. 37. Lu, R., Alioua, A., Kumar, Y., Eghbali, M., Stefani, E. & Toro, L. (2006) J. Physiol. 570, 65–72. 38. Long, S. B., Campbell, E. B. & MacKinnon, R. (2005) Science 309, 897–903. 39. Wallner, M., Meera, P., Ottolia, M., Kaczorowski, G. J., Latorre, R., Garcia, M. L., Stefani, E. & Toro, L. (1995) Receptors Channels 3, 185–199. 40. Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K. & Pease, L. R. (1989) Gene 77, 51–59. 41. Haug, T., Sigg, D., Ciani, S., Toro, L., Stefani, E. & Olcese, R. (2004) J. Gen. Physiol. 124, 173–184. 42. Stefani, E. & Bezanilla, F. (1998) Methods Enzymol. 293, 300–318.

Savalli et al.