Gating Dependence of Inner Pore Access in Inward Rectifier K+ Channels

Neuron, Vol. 37, 953–962, March 27, 2003, Copyright 2003 by Cell Press

Gating Dependence of Inner Pore Access in Inward Rectifier Kⴙ Channels L. Revell Phillips, Decha Enkvetchakul, and Colin G. Nichols* Department of Cell Biology and Physiology Washington University School of Medicine 660 South Euclid Avenue Saint Louis, Missouri 63110

Summary Cation channel gating may occur either at or below the inner vestibule entrance or at the selectivity filter. To differentiate these possibilities in inward rectifier (Kir) channels, we examined cysteine accessibility in the ATP-gated Kir6.2 channel. MTSEA and MTSET both block channels and modify M2 cysteines with identical voltage dependence. If entry is restricted to open channels, modification rates will slow in ATP-closed channels, but because the reagent can be trapped in the pore following brief openings, this may not be apparent until open probability is extremely low (⬍0.01). When these conditions are met, modification does slow significantly, indicating gated access and highlighting an important caveat for interpretation of MTS-accessibility measurements: reagent “trapping” in nominally “closed” channels may obscure gated access. Introduction Many different potassium (K) channels are involved in regulating excitability in all cells and are critical determinants of electrical activity in neurons and glia (Barres et al., 1990; Chiu et al., 1999; Peusner et al., 1998; Pongs, 1999; Raap et al., 2002; Roeper and Pongs, 1996; Rudy and McBain, 2001; Sah and Davies, 2000). All K channels, and indeed all cation channels, are probably built along the same structural pattern as that seen in the bacterial KcsA and MthK channels (Doyle et al., 1998; Jiang et al., 2002a). A p-loop at the outer membrane surface forms a narrow ion selectivity filter, and the transmembrane conduction path is lined by the S6 or M2 helices that follow the p-loop in the primary sequence (Figure 1A). The functional differences between K channels lie in the control of their opening and closing, i.e., their gating. The gating of one major class (Kv channels) is controlled by voltage (Yellen, 1998; Yi and Jan, 2000). The gating of another major class of K channels, the inward rectifiers (Kir channels), is controlled by intracellular ligands, which bind to cytoplasmic domains (Figure 1B; Nichols and Lopatin, 1997; Yi et al., 2001b). Recent studies have begun to define the ligand binding sites on different Kir channels (He et al., 2002; Huang et al., 1995, 1998; Krapivinsky et al., 1995; MacGregor et al., 2002; Mirshahi et al., 2002; Tanabe et al., 1999; Vanoye et al., 2002), but little is known about the opening and closing events themselves. This is not the case for Kv channels, where there is now considerable evidence for *Correspondence: [email protected]

an intracellular gate that opens and closes the channel at the S6 “bundle crossing” by motion of these helices (del Camino et al., 2000; del Camino and Yellen, 2001; Holmgren et al., 1997, 1998; Liu et al., 1997), consistent with the model proposed on the basis of the KcsA (closed) and MthK (open) crystal structures (Figure 1B; Doyle et al., 1998; Jiang et al., 2002a, 2002b). Because of the conservation of pore structures, it might be expected that Kir and other cation channels should also be gated by closure at or below the bundle crossing. Residues in or near the bundle crossing do control ligand gating (Figure 1Bb; Enkvetchakul et al., 2000, 2001; Loussouarn et al., 2001a), and a recent systematic proline-mutagensis of M2 in Kir3 channels indicates a dependence of gating on the position of the M2 helix that is consistent with this notion (Jin et al., 2002). While there is evidence for a similar gating in hyperpolarization-activated (HCN) channels (Shin et al., 2001), recent studies on cyclic nucleotide-gated (CNG) and small Ca2⫹-activated (SK) channels suggest that S6 cysteines may be accessible in the closed channels (BrueningWright et al., 2002; Flynn and Zagotta, 2001). To directly test where ligand gating occurs in Kir channels, we have examined the state dependence of accessibility of cysteines in the inner pore of Kir6.2. We find that access of both MTSEA and MTSET to the inner vestibule is restricted to the open state, but the unconsidered problem of MTS “trapping” in the inner vestibule can obscure this restriction if the degree of closure is not sufficient. We conclude that the S6/M2 gating mechanism proposed for Kv channels (Armstrong, 1966; del Camino and Yellen, 2001; Holmgren et al., 1997; Liu et al., 1997; Rothberg et al., 2002) and inferred from comparison of MthK and KcsA crystal structures (Jiang et al., 2002a, 2002b) is also responsible for gating of Kir channels, and that MTS accessibility measurements should be interpreted with caution. Results and Discussion Pore-Lining Cysteines in M2 Are Modified by MTS Reagents In order to reliably examine the effects of cysteine reagents on substituted cysteines, it is necessary to generate a nonreactive background. With the naturally occurring cysteine (C166) in M2, Kir6.2 channels are blocked by cadmium, but when this residue was mutated to serine, these control (Kir6.2-c) channels were almost completely insensitive to cadmium (Loussouarn et al., 2000) as well as to methane-thiosulfonate-ethyl-ammonium (MTSEA) and methane-thiosulfonate-trimethyl-ammonium (MTSET) (Figure 2A). We previously identified porelining M2 residues based on Cd2⫹ sensitivity of introduced cysteines (Figure 1A; Loussouarn et al., 2000). Current is rapidly reduced following modification of these residues (L157C, A161C, L164C, M169C) by MTSEA or MTSET (Figures 2B and 2C), and there is no significant inhibition of current for residues predicted to face away from the pore (M163C). Equivalent residues

Neuron 954

Figure 1. Structure of the Kir Pore and Gating (A) Comparative model of Kir6.2 pore (Loussouarn et al., 2000), based on KcsA crystal structure (Doyle et al., 1998). Critical residues examined in this study are indicated as space-filling side chains. (B) Cartoon to illustrate potential ligand-dependent gates in Kir6.2 channels (adapted from Enkvetchakul et al., 2000). The open channel is lined by the P-region at the extracellular side of the membrane and by M2 helices toward the cytoplasm. The open state is stabilized by interaction of the cytoplasmic domain with the membrane phospholipids (Shyng and Nichols, 1998). Conversely, ATP stabilizes the closed state that could arise from (a) steric closure at or below the M2 bundle crossing, or (b) by a blockage of permeation in the selectivity filter.

were predicted to face toward (V169C, A173C, I176C) or away (I175C) from the pore of Kir2.1 channels (Lu et al., 1999). Partial reduction of current following modification (Figure 2B) is consistent with an electrostatic effect on permeation rates, rather than a closing off of the permeation pathway. Complete modification of L157C channels by MTSEA reduces macroscopic currents by ⵑ50% (Figure 2C). This reduction results from a reduction of the single channel current (Figure 3A) and not from an effect on gating (Figure 3B), and the sensitivity of channels to ATP inhibition (Figure 3C) is essentially unaltered after MTSEA modification. Determination of absolute opening and closing rates from concentration-jump experiments such as these is compounded by diffusion limitations

Figure 2. MTS Reagents Modify Cysteines in the Kir6.2 Inner Vestibule (A) No modification occurs in control Kir6.2-c (see Experimental Procedures) channels. ATP, MTSET, or MTSEA were applied to an inside-out patch as indicated (Vm ⫽ ⫺50 mV). (B) L157C mutant channels are irreversibly modified by MTSET, or MTSEA, applied as indicated (Vm ⫽ ⫺50 mV). (C) Fractional decrease of patch current resulting from steady-state modification by MTSEA or MTSET (⫾ SEM, n ⫽ 3–10 in each case). (D) Apparent modification rate constants obtained from exponential fits to currents, as in (B). MTSEA and MTSET were generally applied at 0.1 mM or 1 mM, respectively (mean ⫾ SEM, n indicated).

(Cannell and Nichols, 1991; Nichols et al., 1991), but since diffusion is not altered by MTSEA modification, the similar time courses of opening and closing, before and after modification, indicates that the channel kinetics are also unaltered following modification (Figure 3D).

Gating Mechanism of KATP Channel 955

Figure 3. MTSEA Modification Reduces L157C Single-Channel Current but Does Not Affect Gating (A) Representative record of mutant L157C single channel activity, before and after MTSEA modification (Vm ⫽ ⫺50 mV), shows ⵑ50% reduction of single-channel current. (B) Representative record of mutant L157C channel activity, before and after MTSEA modification (Vm ⫽ ⫺50 mV), shows inhibition by 10 mM ATP. (C) Representative dose-response curve to ATP before and after MTSEA modification, from another patch. (D) Scaled, superimposed, faster time base recordings (from B) before and after MTSEA modification during channel closure in ATP (left) and reopening (right). Both ATP sensitivity (C) and kinetics (D) are essentially unaltered by modification.

Accessibility of the Inner Vestibule Is Voltage-Dependent It is clear from Figures 2B and 2D that the rate of modification of inner vestibule cysteines by MTS reagents is very slow in Kir channels at ⫺50 mV. For both MTSEA and MTSET, the modification rates are orders of magnitude lower than rates reported in open Kv channels (Liu et al., 1997), and similarly low rates have been reported for modification of Kir2.1 channels (Lu et al., 1999). It is also reported that MTSEA accessibility to open Kv channels is only weakly voltage dependent (Liu et al., 1997), but both MTSEA (Figure 4A) and MTSET (not shown) cause steeply voltage-dependent (Z ⵑ1.25) block of nonmodified Kir6.2-c channels (Figure 4B). At the concentrations examined, there is actually very little MTSEA block of Kir6.2-c at ⫺50 mV, and since both block and modification occur from within the inner vestibule, modification may also be voltage dependent. We examined this possibility using experiments like that shown in Figures 4C and 4D. While there is no modification of Kir6.2-c channels even at positive voltages (Figure 4C), modification of L157C is almost complete following just a 2 s step to ⫹75 mV (Figure 4D). As shown

in Figure 4E, there is actually close correlation between the voltage dependence of irreversible modification of L157C and of reversible occupancy of the “blocked” channel. This similar voltage dependence of modification and block implies that the sites for each are positioned at a similar electrical “depth” within the pore. As shown in Figure 4E, the modification rate constant appears to saturate at ⵑ2 s⫺1. This saturation of modification rate, as occupancy of the pore is saturated, indicates that modification is not occurring from MTSEA in the bulk solution, but from a saturable site within the inner vestibule and implies that the modification rate will become concentration independent as occupancy of the site approaches saturation. We tested this further by examining MTSEA modification rates at a series of voltages and concentrations (Figure 5A). The modification is concentration dependent at negative voltages but reaches the same saturating rate constant (ⵑ2 s⫺1) as either voltage or concentration are increased. The simplest kinetic scheme that accounts for this concentration and voltage-dependent behavior requires three states (Scheme 1):

Neuron 956

Figure 4. Both MTSEA Block and Modification of L157C Are Voltage Dependent (A) Representative records of Kir6.2-c current in the absence and presence of 0.1 mM MTSEA, in response to voltage steps from Vh ⫽ ⫺80 mV to voltages between ⫺70 and ⫹80 mV (10 mV steps). (B) Top: Steady-state current-voltage relationships (from A). Bottom: Relative conductance (Grel)-voltage relationships computed from above (Grel ⫽ IMTSEA/Icont), fit with Boltzmann function (Z ⫽ 1.25, V1/2 ⫽ ⫹55 mV). (C and D) Representative inside-out currents recorded from patches containing SUR1 ⫹ Kir6.2-c (C) or L157C (D). Vm was repeatedly stepped from ⫺50 to ⫹50 or ⫹75 mV. MTSEA was applied as indicated. (E) Modification rate of L157C, from records as in (D) (see Experimental Procedures), as a function of membrane potential (symbols, mean ⫾ SEM, n ⫽ 3–7, left scale), and fractional block of Kir6.2-c (1 ⫺ Grel, from Boltzmann fit in B, right scale), versus membrane potential.

[1] in which Oin represents the MTS-blocked state, from which modification proceeds. km is given by the saturating modification rate (2 s⫺1). The concentration- and voltage-dependent entry and exit rate constants (kin, kout) can be estimated from relaxation time constants and equilibrium block in the presence of MTS, as shown in Figures 5B and 5C. This fully constrained model allows us to simulate the time course of modification as a function of concentration and voltage (see Experimental Procedures). With the parameters listed in Figure 5, the

model accurately predicts the experimentally observed concentration and voltage dependence of modification rates. Is Accessibility of the Inner Vestibule Reduced in the Closed State? Gating of Kv channels seems to occur by a movement of the S6 (M2) helices that pinches off the permeation pathway at the bundle crossing of the helices (del Camino and Yellen, 2001; Jiang et al., 2002b). Are Kir channels closed in the same way? Preliminary experiments indicated little difference between modification rates in

Gating Mechanism of KATP Channel 957

naive interpretation fails to stand up to further quantitative consideration. The simplest extension of Scheme 1 to include gating requires three additional closed states, limiting cases being Scheme 2a, with the inner vestibule being freely accessible to MTS in the bulk solution in both open and closed states, and Scheme 3a, with the vestibule being accessible only in the open state:

[2a]

[3a]

Figure 5. MTSEA Modification Rates Saturate at High Concentration and Positive Voltages (A) Concentration and voltage dependence of rate of MTSEA modification of L157C. Data points are from experiments (mean ⫾ SEM), lines are predictions of Scheme 1 (see B and C). (B) Expanded time base records of Kir6.2-c current in the presence of 0.1 mM MTSEA, in response to voltage steps from Vh ⫽ ⫺80 mV to voltages between ⫹30 and ⫹80 mV (from Figure 4A), over which range block is increasing. Exponential fits are indicated. (C) MTSEA entry (kin) and exit (kout) rate constants as a function of Vm, estimated from current relaxation and steady-state block (from B), kin(Vm) ⫽ kin(0)·exp(zin·F·Vm), kout(Vm) ⫽ kout(0) ·exp(zout·F·Vm), where F ⫽ Faraday’s constant, kin(0) ⫽ 550 mM⫺1·s⫺1, zin ⫽ 0.6, kout(0) ⫽ 840 s⫺1, zout ⫽ ⫺0.65. The curves in (A) are predictions of this model (with kM ⫽ 2 s⫺1). For modeling MTSET modification, as in Figure 7B, corresponding parameters were kin(0) ⫽ 150 mM⫺1·s⫺1, zin ⫽ 0.9, kout(0) ⫽ 2200 s⫺1, zout ⫽ ⫺0.35, kM ⫽ 0.2 s⫺1 (data not shown).

the presence or absence of ATP (Phillips and Nichols, 2002). Very similar modification rates, and lack of clear dependence on gating, were also preliminarily reported for Kir2.1 channels (Xiao and Yang, 2002). On the face of it, such data provide primary evidence that, because there is no slowing of modification when the channels are closed, there should be no ligand-operated gate to MTS reagents and hence to permeant ions at the level of the bundle crossing in Kir channels. However, this

As shown in Figures 3B–3D, neither the steady-state ATP dependence of channel activity nor the kinetics of channel opening and closing are altered by modification. Thus, the gating process is essentially the same in the unoccupied and modified states (and likely to be the same in the blocked state). If the rate constants for ATP-dependent opening and closing are known, these models are thus also constrained and modification kinetics can be predicted as a function of the open probability. As shown in Figure 3D, the opening and closing rates of L157C channels are approximately exponential with opening ␶ and closing ␶ (in 10 mM in ATP) in the order of 100 msec. The detailed kinetics of ligand gating are actually quite complex (Enkvetchakul et al., 2000), and a two-state model is oversimplified. However, the conclusions we draw below are not critically dependent on the detailed gating scheme nor on the opening/closing rates (see Figure 6A, Scheme 3b), but only on the degree of closure. Figure 6A shows predicted apparent modification rate versus open probability using the parameters for Scheme 1 determined in Figure 5, plus opening (ko) and closing (kc) rate constants given in Figure 6. If there is no gating-dependent accessibility (Scheme 2a), the modification rate is unaffected by closure. However, if the vestibule is only accessible in the open state (Scheme 3a), the apparent modification rate slows as the channel is closed, but by how much depends very nonlinearly on open probability. Strikingly, given the measured MTS accessibility rates, modification rates are not predicted to slow appreciably until the channels are ⬎99% closed. The above predictions assume that the final modification rate (km) is the same in open and closed channels. We also consider the additional cases (Schemes 2b and 3b) in which the modification rate constant is different in the closed (km,c) and open (km,o) states:

Neuron 958

Figure 6. MTSEA Modification Rate Depends Nonlinearly on Open Probability (A) Theoretical dependence of modification rate on open probability (see text). Heavy lines correspond to predictions of limiting Schemes 2 and 3 described in Results, with parameters for 0.1 mM MTSEA modification at ⫹50 mV from Figure 5, and with ko ⫽ 100 s⫺1, kc ⫽ 100 mM⫺1·s⫺1. Also shown (dashed) are the predictions for Scheme 3b with 10fold higher (i, ko ⫽ 1000 s⫺1, kc ⫽ 1000 mM⫺1·s⫺1) and 10-fold lower (ii, ko ⫽ 10 s⫺1, kc ⫽ 10 mM⫺1·s⫺1) opening and closing rates. (B) Representative consecutive records of poly-L-lysine-treated L157C currents. The patch was exposed to 0.1 mM MTSEA (plus ATP as indicated) immediately before a brief pulse (0.5 s) from ⫺50 to ⫹50 mV, then returned to control solution immediately after. Zero current (solid), prehit current (dashed), and fully modified current (dotted) are indicated. (C) Higher resolution of the currents in 1 or 10 mM ATP from times indicated in (B). From these records, the mean channel current in the presence of ATP was estimated, and the relative open probability (rPo) was calculated (rPo ⫽ IATP/IControl).

[2b]

[3b]

For scheme 2b, 5-fold increase in the closed state modification rate (km,c) will cause the overall modification rate to increase asymptotically as the open probability decreases (Figure 6A). For the limiting scheme 3b, with zero access in the closed state, the same (5-fold) increase in km,c leads to a bell-shaped dependence of modification rate on open probability (Figure 6A). At low degrees of closure, the modification rate will be enhanced; at ⫹50 mV the MTSEA modification rate is not expected to slow until Po ⬍ 0.01. At more negative voltages, and for the slower modifying MTSET, considerably lower Po must be attained before appreciable slowing is observed (see Figure 7).

Gating Mechanism of KATP Channel 959

posures to MTSEA. In the presence of 10 mM ATP, a high degree of closure (rPo ⫽ 0.0004) was now attained (Figure 6C), and MTSEA failed to significantly modify the current during a 500 msec pulse to ⫹50 mV (Figure 6Bi). However, in 1 mM ATP (rPo ⫽ 0.0056), there was significant modification during the 500 msec pulse to ⫹50 mV (Figure 6Bii), similar to the extent of modification in the absence of ATP (Figure 6Biii). In Figure 7, apparent modification rates for individual L157C patches are plotted as a function of open probability, estimated in the absence or presence of various concentrations of ATP. For both MTSEA and MTSET, the data follow a bell-shaped dependence on the degree of closure, with modification rates slowing at very low open probabilities. Also shown in Figure 7A are the predicted relationships for zero (Scheme 2b) or equal accessibility (Scheme 3b) as well as for “partially gated” models in which the accessibility in the closed state is a fraction of that in the open state. The data indicate that accessibility is clearly at least 100-fold lower in the closed state (i.e., the scaling constant ⬍ 0.01). When modification rates are examined for MTSEA modification at ⫺50 mV (Figure 7A) or for MTSET modification at ⫹75 mV (Figure 7B), both the predicted and observed relationships shift on the y axis and change “curvature” with rPo. Together, the data provide strong evidence for a gradual slowing of modification rates as high degrees of closure are attained, and argue against significant MTS access in the closed state.

Figure 7. MTS Accessibility Is Gated below L157C (A and B) Relationship between modification rate of L157C and open probability for MTSEA at ⫹50 and ⫺50 mV (A) or MTSET (B) at ⫹75 mV, from records as in Figure 6B. Dashed lines connect individual measurements from single patches. The heavy lines indicate the expected relationship from Scheme 3b, with parameters as in Figures 5 and 6. Also shown for ⫹50 mV in (A) are the relationships for Scheme 3a with partially gated access, in which the entry and exit rates of MTSEA in the closed state are scaled fractions of the entry and exit rates in the open state (scaling constants indicated).

The ATP sensitivity of KATP channels varies from patch to patch, depending on the ambient phospholipids in the patch (Loussouarn et al., 2001b). For L157C channels, 10 mM ATP typically reduces channel current to only about 1%–5%, i.e., rPo of 0.01–0.05. However, ATP sensitivity can be increased, and much more complete channel closure can be achieved, by treatment with poly-L-lysine, which reduces the effective PIP2 concentration in the membrane (Enkvetchakul et al., 2000; Koster et al., 1999; Shyng and Nichols, 1998). Figure 6B shows modification of a poly-L-lysine-treated L157C patch by repeated ex-

A Common Structural Basis of Gating in Kv and Kir Channels? In the present study, we have observed that MTS reagents modify essentially the same group of residues that coordinate Cd2⫹ ions (Loussouarn et al., 2000). These residues are equivalent to those that are MTS accessible in Kir2.1 (Lu et al., 1999) and align with pore lining residues in KcsA (Doyle et al., 1998). Given that chimeras in which the KcsA pore is substituted into a Kir (Kir2.1) channel or into a Kv (Kv2.1) channel express functional inward rectifiers and voltage-gated K channels, respectively (Lu et al., 2001c), we may place confidence in the likelihood that the overall architecture of KcsA (Doyle et al., 1998) is replicated in Kir channels, and that M2 helices form an “inverted tepee” lining the inner vestibule of the channel. The KcsA structure appears to be closed to permeation at the lower end of the inner vestibule, where the M2 helices cross (the bundle crossing). By contrast, the M2 helices are splayed open by ⵑ30⬚ (at the residue equivalent to Gly156 in Kir6.2) in the MthK structure. This splaying of the helices opens the inner vestibule to a wide, almost cylindrical, tunnel (Jiang et al., 2002a). Small ions should freely permeate the bundle crossing in such a structure, and hence it is proposed that gating of K channels reflects transitions between a KcsA-like and MthK-like structure (Jiang et al., 2002b). Significant evidence has been marshaled in support of the idea that a very tight steric closure of Kv and HCN channels occurs at the level of the S6 (del Camino et al., 2000; del Camino and Yellen, 2001; Liu et al., 1997; Rothberg et al., 2002; Shin et al., 2001) bundle crossing

Neuron 960

(i.e., the narrowest part of the pore, defined by the helices in KcsA) (Doyle et al., 1998). In contrast, several studies have used reagent accessibility measurements to argue against such a gate in other cation channels, including Kir channels (Bruening-Wright et al., 2002; Flynn and Zagotta, 2001; Phillips and Nichols, 2002; Xiao and Yang, 2002; Proks et al., 2003). Is it really conceivable that the fundamental structural changes that underlie voltage-gating of Kv channels are different than those that underlie ligand-gating in Kir and other cation channels? Kir channels exhibit multiple gating states, but the kinetics are generally readily separable into fast and slow gating events (Choe et al., 1998, 1999; Drain et al., 1998; Enkvetchakul et al., 2000, 2001). The former are clearly voltage and permeant ion dependent, and are therefore likely to involve the selectivity filter. Recent studies on various Kir channels have demonstrated the involvement of selectivity filter residues and permeant ions in fast gating processes (Choe et al., 1999; Lu et al., 2001a, 2001b; Proks et al., 2001), but slower ligand (e.g., PIP2, ATP, G protein) gating is not obviously dependent on voltage or permeant ions. Importantly, mutation of residues in M2, particularly residues near the expected bundle crossing (Enkvetchakul et al., 2000, 2001; Sadja et al., 2001; Yi et al., 2001a) (and moreover at equivalent residues to those that lead to voltage-independent opening of Kv channels [Espinosa et al., 2001; Hackos et al., 2002]), can have very profound effects on ligand gating in Kir channels, indicating a clear involvement of M2 in at least coupling the gating sensors to the gates themselves. This has led us to implicitly suppose that M2 helices, by analogy with Kv channels and the bacterial KcsA and MthK channels (Doyle et al., 1998; Jiang et al., 2002a, 2002b) (see below), act as the slow gate to permeation and form a tight steric seal in the closed state (Enkvetchakul et al., 2000; Jin et al., 2002; Loussouarn et al., 2001a). The quantitative analysis of the present data does indicate a gated access of MTS reagents to the Kir6.2 inner vestibule, providing evidence to support this hypothesis. So what are the implications of the present findings for the nature of the gating mechanism in other Kir channels and in other ligand-gated channels? The data illustrate a critical caveat to interpretation of experiments using accessibility measurements to assess the position of the channel gate. With rapid gating and reagent access, relative to modification rates, the problem of reagent trapping can obscure a gated access, unless the degree of closure is very high. It is significant that in other studies using MTS accessibility to assess accessibility in other ligand-gated channels, the actual modification rates are uniformly slow (Bruening-Wright et al., 2002; Flynn and Zagotta, 2001; Simoes et al., 2002) and the degree of closure in the (nominally) closed state is not addressed. The possibility that reagent trapping is obscuring a gated access in these channels should perhaps be considered. Conversely, we can turn the question around and ask why no such caveat seems to exist for analysis of Kv channels (Liu et al., 1997). In this case, limiting modification rates are considerably higher (⬎105 M⫺1·s⫺1, i.e., ⬎10 s⫺1 at 100 ␮M MTSET) than in the present experiments. Under these conditions, trapping can become irrelevant, since modification occurs essentially instantaneously once the reagent enters the vesti-

bule, and the measured modification rates will depend only on the accessibility (see Supplemental Data at http:// www.neuron.org/cgi/content/full/37/6/953/DC1). In conclusion, the present data indicate a gated access of MTS reagents to the inner vestibule of Kir channels, consistent with the previously proposed model in which movements of the gating sensor (i.e., the cytoplasmic C-terminal domain (Cukras et al., 2002; Shyng et al., 2000) are directly coupled to the M2 segment (Enkvetchakul et al., 2000). The ligand-operated gating of Kir channels is thus proposed to be fundamentally similar to the voltage gating of Kv channels (Liu et al., 1997) and to the archetypal gating proposed for MthK and KcsA (Jiang et al., 2002b). The analysis highlights a critical caveat for interpretation of MTS accessibility measurements and suggests reconsidering the possibility that gating at or below the bundle crossing is a feature of other, if not all, cation channel superfamily members. Experimental Procedures Expression of KATP Channels in COSm6 Cells In all experiments, Kir6.2-c or cysteine-substituted mutants were coexpressed with SUR1. COSm6 cells were plated at a density of ⵑ2.5 ⫻ 105 cells per well (30 mm 6-well dishes) and cultured in Dulbecco’s Modified Eagle Medium plus 10 mM glucose (DMEMHG), supplemented with fetal calf serum (FCS, 10%). The first day after plating, cells were transfected using 1–2 ␮g each of pCMV6bKir6.2, pECE-SUR1, and pGreenlantern (BRL Co) and 5 ␮l FuGene6 (Roche Co, Indianapolis, IN). Two days after plating, transfected cells were replated onto glass cover slips for patch-clamping. Generation of a Nonreactive Background and Introduction of Cysteines Constructs containing point mutations were prepared by overlap extension at the junctions of the relevant residues by sequential polymerase chain reaction (PCR). Resulting PCR products were subcloned into pCMV6B and sequenced to verify the correct mutant construct before transfection. The starting “control” construct had a deletion of 36 amino acids from the C terminus (Tucker et al., 1997). This ⌬C36 truncation had no effect on the function of channels coexpressed with SUR1 (as in all of the present experiments). The Kir6.2-c construct also contained a N160D substitution. This generates strongly rectifiying channels in the presence of spermine, which permits estimation of zero current for mutants that are insensitive to ATP. In addition, the Control-Kir6.2 construct contained the mutation C166S. Cysteine substitutions were made on this background (Loussouarn et al., 2000). Patch-Clamp Measurements All experiments were performed at room temperature in a chamber that allowed the bathing solution exposed to the patch surface to be rapidly changed using an oil gate (Lederer and Nichols, 1989). Pipettes were pulled from Kimble 73813 soda lime glass using a horizontal puller (Sutter Instrument Co., Novato, CA). Electrode resistances ranged from .5 to 2.5 M⍀. Microelectrodes were sealed onto green fluorescing cells by applying light suction to the rear of the pipette. Inside-out patches were obtained by lifting the electrode and then passing the tip through the oil gate. Pipette and bath solutions contained a modified KINT (in mM 140 KCl, 1 EGTA, 1 EDTA, 2 K2HPO4, 2 KH2PO4 [pH 7.4]), with additions as described. Membrane patches were voltage-clamped using an Axopatch 1D amplifier. Data was normally filtered at 1 kHz and recorded directly onto a computer at 5 kHz using P-Clamp 7 and a digidata board (Axon, Foster City, CA). MTS reagents were purchased from Toronto Research Chemicals (Drownview, Ontario) and were stored at ⫺20⬚C as a powder. Each day a 10 mM stock was made up and stored on ice and diluted to final concentration immediately before use. Dipotassium ATP was

Gating Mechanism of KATP Channel 961

from Sigma (St. Louis, MO) and was stored as a powder in the freezer and made fresh each day. Data Analysis and Modeling Offline analysis was performed using Axon Clampfit and Microsoft Excel programs. Wherever possible, data are presented as mean ⫾ SEM (standard error of the mean). Microsoft Solver was used to fit data by a least-square algorithm. Modification rates were obtained either from exponential fits to continuous recordings (as in Figure 2B), or from assumed exponential decay during voltage steps (as in Figure 4D). Markov models were assumed to be representative of actual KATP channel kinetics and were used for simulations. Rate constants between states were constrained as described in the Results section and microreversibility was observed. Time courses of occupancy of each state were calculated as described by Colquhoun and Hawkes (1995a, 1995b). Specifically, the occupancy of each state at time t is given by p(t) ⫽ p(0)eQt, where p(t) is a row vector with one element for each state in the model. p(0) contains the occupancy probabilities of each state at t ⫽ 0 and was calculated from the equilibrium occupancy in the indicated [ATP] prior to MTS application. Matrix Q was constructed with each element (i,j) equal to the rate constant from state i to state j, and elements (i,i) were set equal to the negative sum of all the other values in that row. The tasks of calculating rate constants, building the Q matrix, and solving for the above equations were performed either with MathCad6 (MathSoft, Cambridge, MA) or with original programs written in Visual Basic or Visual Basic for Applications (Microsoft, Redmond, WA). The matrix exponential was calculated using either spectral expansion of a matrix or by the Taylor series expansion of an exponential (Colquhoun and Hawkes, 1995a, 1995b), allowing the time course of modification by MTS reagents to be predicted. From these time courses, time to 63% modification was taken as apparent time constant (␶app) and the reciprocal of ␶app taken as the apparent time constant of modification, as indicated in Figures 6 and 7 and Supplemental Data at http:// www.neuron.org/cgi/content/full/37/6/953/DC1. Acknowledgments This work was supported by grants HL54171 (to C.G.N.) and DK60086 (to D.E.) from the NIH and by the NIH Biophysical Training Grant at Washington University. We are also grateful to the Washington University Diabetes Research and Training Center for reagent support. Received: December 5, 2002 Revised: January 27, 2003 References Armstrong, C.M. (1966). Time course of TEA(⫹)-induced anomalous rectification in squid giant axons. J. Gen. Physiol. 50, 491–503. Barres, B.A., Chun, L.L., and Corey, D.P. (1990). Ion channels in vertebrate glia. Annu. Rev. Neurosci. 13, 441–474. Bruening-Wright, A., Schumacher, M.A., Adelman, J.P., and Maylie, J. (2002). Localization of the activation gate for small conductance Ca2⫹- activated K⫹ channels. J. Neurosci. 22, 6499–6506. Cannell, M.B., and Nichols, C.G. (1991). Effects of pipette geometry on the time course of solution change in patch clamp experiments. Biophys. J. 60, 1156–1163. Chiu, S.Y., Zhou, L., Zhang, C.L., and Messing, A. (1999). Analysis of potassium channel functions in mammalian axons by gene knockouts. J. Neurocytol. 28, 349–364. Choe, H., Sackin, H., and Palmer, L.G. (1998). Permeation and gating of an inwardly rectifying potassium channel. Evidence for a variable energy well. J. Gen. Physiol. 112, 433–446. Choe, H., Palmer, L.G., and Sackin, H. (1999). Structural determinants of gating in inward-rectifier K⫹ channels. Biophys. J. 76, 1988– 2003. Colquhoun, D., and Hawkes, A.G. (1995a). The principles of the stochastic interpretation of ion-channel mechanisms. In Single-

Channel Recording, B. Sakmann and E. Neher, eds. (New York: Plenum Press), pp. 397–482. Colquhoun, D., and Hawkes, A.G. (1995b). A Q-matrix cookbook: how to write only one program to calculate the single-channel and macroscopic predictions for any kinetic mechanism. In Single-Channel Recording, B. Sakmann and E. Neher, eds. (New York: Plenum Press), pp. 589–636. Cukras, C.A., Jeliazkova, I., and Nichols, C.G. (2002). Structural and functional determinants of conserved lipid interaction domains of inward rectifying kir6.2 channels. J. Gen. Physiol. 119, 581–591. del Camino, D., and Yellen, G. (2001). Tight steric closure at the intracellular activation gate of a voltage- gated K(⫹) channel. Neuron 32, 649–656. del Camino, D., Holmgren, M., Liu, Y., and Yellen, G. (2000). Blocker protection in the pore of a voltage-gated K⫹ channel and its structural implications. Nature 403, 321–325. Doyle, D.A., Morais Cabral, J., Pfuetzner, R.A., Kuo, A., Gulbis, J.M., Cohen, S.L., Chait, B.T., and MacKinnon, R. (1998). The structure of the potassium channel: molecular basis of K⫹ conduction and selectivity. Science 280, 69–77. Drain, P., Li, L., and Wang, J. (1998). KATP channel inhibition by ATP requires distinct functional domains of the cytoplasmic C terminus of the pore-forming subunit. Proc. Natl. Acad. Sci. USA 95, 13953– 13958. Enkvetchakul, D., Loussouarn, G., Makhina, E., Shyng, S.L., and Nichols, C.G. (2000). The kinetic and physical basis of K(ATP) channel gating: toward a unified molecular understanding. Biophys. J. 78, 2334–2348. Enkvetchakul, D., Loussouarn, G., Makhina, E., and Nichols, C.G. (2001). ATP interaction with the open state of the K(ATP) channel. Biophys. J. 80, 719–728. Espinosa, F., Fleischhauer, R., McMahon, A., and Joho, R.H. (2001). Dynamic interaction of S5 and S6 during voltage-controlled gating in a potassium channel. J. Gen. Physiol. 118, 157–170. Flynn, G.E., and Zagotta, W.N. (2001). Conformational changes in S6 coupled to the opening of cyclic nucleotide-gated channels. Neuron 30, 689–698. Hackos, D.H., Chang, T.H., and Swartz, K.J. (2002). Scanning the intracellular s6 activation gate in the shaker k(⫹) channel. J. Gen. Physiol. 119, 521–532. He, C., Yan, X., Zhang, H., Mirshahi, T., Jin, T., Huang, A., and Logothetis, D.E. (2002). Identification of critical residues controlling G protein-gated inwardly rectifying K(⫹) channel activity through interactions with the beta gamma subunits of G proteins. J. Biol. Chem. 277, 6088–6096. Holmgren, M., Smith, P.L., and Yellen, G. (1997). Trapping of organic blockers by closing of voltage-dependent K⫹ channels—evidence for a trap door mechanism of activation gating. J. Gen. Physiol. 109, 527–535. Holmgren, M., Shin, K.S., and Yellen, G. (1998). The activation gate of a voltage-gated K⫹ channel can be trapped in the open state by an intersubunit metal bridge. Neuron 21, 617–621. Huang, C.L., Slesinger, P.A., Casey, P.J., Jan, Y.N., and Jan, L.Y. (1995). Evidence that direct binding of G beta gamma to the GIRK1 G protein-gated inwardly rectifying K⫹ channel is important for channel activation. Neuron 15, 1133–1143. Huang, C.L., Feng, S., and Hilgemann, D.W. (1998). Direct activation of inward rectifier potassium channels by PIP2 and its stabilization by Gbetagamma. Nature 391, 803–806. Jiang, Y., Lee, A., Chen, J., Cadene, M., Chait, B.T., and MacKinnon, R. (2002a). Crystal structure and mechanism of a calcium-gated potassium channel. Nature 417, 515–522. Jiang, Y., Lee, A., Chen, J., Cadene, M., Chait, B.T., and MacKinnon, R. (2002b). The open pore conformation of potassium channels. Nature 417, 523–526. Jin, T., Peng, L., Mirshahi, T., Rohacs, T., Chan, K.W., Sanchez, R., and Logothetis, D.E. (2002). The betagamma subunits of G proteins gate a K(⫹) channel by pivoted bending of a transmembrane segment. Mol. Cell 10, 469–481.

Neuron 962

Koster, J.C., Sha, Q., Shyng, S., and Nichols, C.G. (1999). ATP inhibition of KATP channels: control of nucleotide sensitivity by the N-terminal domain of the Kir6.2 subunit. J. Physiol. 515, 19–30. Krapivinsky, G., Krapivinsky, L., Wickman, K., and Clapham, D.E. (1995). G beta gamma binds directly to the G protein-gated K⫹ channel, IKACh. J. Biol. Chem. 270, 29059–29062. Lederer, W.J., and Nichols, C.G. (1989). Nucleotide modulation of the activity of rat heart ATP-sensitive K⫹ channels in isolated membrane patches. J. Physiol. 419, 193–211. Liu, Y., Holmgren, M., Jurman, M.E., and Yellen, G. (1997). Gated access to the pore of a voltage-dependent K⫹ channel. Neuron 19, 175–184.

controlled gating at the intracellular entrance to a hyperpolarizationactivated cation channel. J. Gen. Physiol. 119, 83–91. Rudy, B., and McBain, C.J. (2001). Kv3 channels: voltage-gated K⫹ channels designed for high-frequency repetitive firing. Trends Neurosci. 24, 517–526. Sadja, R., Smadja, K., Alagem, N., and Reuveny, E. (2001). Coupling Gbetagamma-dependent activation to channel opening via pore elements in inwardly rectifying potassium channels. Neuron 29, 669–680. Sah, P., and Davies, P. (2000). Calcium-activated potassium currents in mammalian neurons. Clin. Exp. Pharmacol. Physiol. 27, 657–663.

Loussouarn, G., Makhina, E.N., Rose, T., and Nichols, C.G. (2000). Structure and dynamics of the pore of inwardly rectifying K(ATP) channels. J. Biol. Chem. 275, 1137–1144.

Shin, K.S., Rothberg, B.S., and Yellen, G. (2001). Blocker state dependence and trapping in hyperpolarization-activated cation channels: evidence for an intracellular activation gate. J. Gen. Physiol. 117, 91–101.

Loussouarn, G., Phillips, L.R., Masia, R., Rose, T., and Nichols, C.G. (2001a). Flexibility of the Kir6.2 inward rectifier K(⫹) channel pore. Proc. Natl. Acad. Sci. USA 98, 4227–4232.

Shyng, S.L., and Nichols, C.G. (1998). Membrane phospholipid control of nucleotide sensitivity of KATP channels. Science 282, 1138– 1141.

Loussouarn, G., Pike, L.J., Ashcroft, F.M., Makhina, E.N., and Nichols, C.G. (2001b). Dynamic sensitivity of ATP-sensitive K(⫹) channels to ATP. J. Biol. Chem. 276, 29098–29103.

Shyng, S.L., Cukras, C.A., Harwood, J., and Nichols, C.G. (2000). Structural determinants of PIP(2) regulation of inward rectifier K(ATP) channels. J. Gen. Physiol. 116, 599–608.

Lu, T., Nguyen, B., Zhang, X., and Yang, J. (1999). Architecture of a K⫹ channel inner pore revealed by stoichiometric covalent modification. Neuron 22, 571–580.

Simoes, M., Garneau, L., Klein, H., Banderali, U., Hobeila, F., Roux, B., Parent, L., and Sauve, R. (2002). Cysteine mutagenesis and computer modeling of the S6 region of an intermediate conductance IKCa channel. J. Gen. Physiol. 120, 99–116.

Lu, T., Ting, A.Y., Mainland, J., Jan, L.Y., Schultz, P.G., and Yang, J. (2001a). Probing ion permeation and gating in a K⫹ channel with backbone mutations in the selectivity filter. Nat. Neurosci. 4, 239–246.

Tanabe, K., Tucker, S.J., Matsuo, M., Proks, P., Ashcroft, F.M., Seino, S., Amachi, T., and Ueda, K. (1999). Direct photoaffinity labeling of the Kir6.2 subunit of the ATP-sensitive K⫹ channel by 8-azidoATP. J. Biol. Chem. 274, 3931–3933.

Lu, T., Wu, L., Xiao, J., and Yang, J. (2001b). Permeant ion-dependent changes in gating of Kir2.1 inward rectifier potassium channels. J. Gen. Physiol. 118, 509–522.

Tucker, S.J., Gribble, F.M., Zhao, C., Trapp, S., and Ashcroft, F.M. (1997). Truncation of Kir6.2 produces ATP-sensitive K⫹ channels in the absence of the sulphonylurea receptor. Nature 387, 179–183.

Lu, Z., Klem, A.M., and Ramu, Y. (2001c). Ion conduction pore is conserved among potassium channels. Nature 413, 809–813.

Vanoye, C.G., MacGregor, G.G., Dong, K., Tang, L., Buschmann, A.S., Hall, A.E., Lu, M., Giebisch, G., and Hebert, S.C. (2002). The carboxyl termini of K(ATP) channels bind nucleotides. J. Biol. Chem. 277, 23260–23270.

MacGregor, G.G., Dong, K., Vanoye, C.G., Tang, L., Giebisch, G., and Hebert, S.C. (2002). Nucleotides and phospholipids compete for binding to the C terminus of KATP channels. Proc. Natl. Acad. Sci. USA 99, 2726–2731. Mirshahi, T., Mittal, V., Zhang, H., Linder, M.E., and Logothetis, D.E. (2002). Distinct sites on G protein beta gamma subunits regulate different effector functions. J. Biol. Chem. 277, 36345–36350. Nichols, C.G., and Lopatin, A.N. (1997). Inward rectifier potassium channels. Annu. Rev. Physiol. 59, 171–191. Nichols, C.G., Lederer, W.J., and Cannell, M.B. (1991). ATP dependence of KATP channel kinetics in isolated membrane patches from rat ventricle. Biophys. J. 60, 1164–1177. Peusner, K.D., Gamkrelidze, G., and Giaume, C. (1998). Potassium currents and excitability in second-order auditory and vestibular neurons. J. Neurosci. Res. 53, 511–520. Phillips, L.R., and Nichols, C.G. (2002). Cold draft through the teepee: MTSEA access to the inner vestibule of ATP closed Kir6.2 channels implies gating does not occur at the smokehole. Biophys. J. 82, 590a. Pongs, O. (1999). Voltage-gated potassium channels: from hyperexcitability to excitement. FEBS Lett. 452, 31–35. Proks, P., Capener, C.E., Jones, P., and Ashcroft, F.M. (2001). Mutations within the P-loop of Kir6.2 modulate the intraburst kinetics of the ATP-sensitive potassium channel. J. Gen. Physiol. 118, 341–353. Proks, P., Antcliff, J.F., and Ashcroft, F.M. (2003). The ligand-sensitive gate of a potassium channel lies close to the selectivity filter. EMBO Rep. 4, 70–75. Raap, M., Biedermann, B., Braun, P., Milenkovic, I., Skatchkov, S.N., Bringmann, A., and Reichenbach, A. (2002). Diversity of Kir channel subunit mRNA expressed by retinal glial cells of the guinea-pig. Neuroreport 13, 1037–1040. Roeper, J., and Pongs, O. (1996). Presynaptic potassium channels. Curr. Opin. Neurobiol. 6, 338–341. Rothberg, B.S., Shin, K.S., Phale, P.S., and Yellen, G. (2002). Voltage-

Xiao, J., and Yang, J. (2002). Where is the activation gate for PIP2 gating of Kir2.1 channels? Biophys. J. 82, 180a. Yellen, G. (1998). The moving parts of voltage-gated ion channels. Q. Rev. Biophys. 31, 239–295. Yi, B.A., and Jan, L.Y. (2000). Taking apart the gating of voltagegated K⫹ channels. Neuron 27, 423–425. Yi, B.A., Lin, Y.F., Jan, Y.N., and Jan, L.Y. (2001a). Yeast screen for constitutively active mutant G protein-activated potassium channels. Neuron 29, 657–667. Yi, B.A., Minor, D.L., Jr., Lin, Y.F., Jan, Y.N., and Jan, L.Y. (2001b). Controlling potassium channel activities: interplay between the membrane and intracellular factors. Proc. Natl. Acad. Sci. USA 98, 11016–11023.