|
|
|
|
|
|
|
||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
ARTICLE |
Correspondence to Jian Yang: jy160{at}columbia.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 subunit of P/Q type Ca2+ channels. We found that positions above the putative membrane/cytoplasm interface, including two positions below the corresponding S6 bundle crossing in K+ channels, showed pronounced state-dependent accessibility to internal MTSET, reacting 
1,000-fold faster with MTSET in the open state than in the closed state. In contrast, a position at or below the putative membrane/cytoplasm interface was modified equally rapidly in both the open and closed states. Our results suggest that the S6 helices of the 1 subunit of VGCCs undergo conformation changes during gating and the activation gate is located at the intracellular end of the pore.
Abbreviations used in this paper: CNG, cyclic nucleotide–gated; MTSET, methanethiosulfonate ethyltrimethylammonium; VGCC, voltage-gated Ca2+ channel.
 |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
; Larsson et al., 1996; Liu et al., 1997; del Camino et al., 2000; Liu and Siegelbaum, 2000; del Camino and Yellen, 2001; Flynn and Zagotta, 2001), fluorescence measurement (Mannuzzu et al., 1996; Mannuzzu and Isacoff, 2000), FRET spectroscopy (Cha et al., 1999; Glauner et al., 1999), and electron paramagnetic resonance (Perozo et al., 1999; Liu et al., 2001), investigators have begun to unravel the complex conformation changes associated with gating, including the movement of the voltage sensor and the activation gate. The crystal structures of several bacterial K+ channels (KcsA, MthK, KvAP, and KirBac) have provided particularly useful information on the molecular mechanism of gating (Doyle et al., 1998; Jiang et al., 2002a,b, 2003a,b; Kuo et al., 2003). These structures suggest that the M2/S6 transmembrane segments of K+ channels undergo large conformational changes during gating; they form a narrow constriction (5 Å in diameter) at the intracellular end when the channel is closed but widen to 12 Å when the channel is open.
The location of the activation gate for several types of channels has recently been explored. Based on properties of block of voltage-gated K+ (Kv) channels by organic molecules, Armstrong first proposed that the activation gate of these channels was located at the intracellular entrance to the pore (Armstrong, 1966, 1971, 1974). More recent work by studying gating-dependent changes in the rate of chemical modification of cysteines engineered on both sides of the S6 bundle crossing have provided strong support for the intracellular location of the activation gate in Kv channels (Liu et al., 1997; del Camino and Yellen, 2001). The intracellular gate also acts as the activation gate in hyperpolarization-activated cation channels (Shin et al., 2001; Rothberg et al., 2002). On the other hand, another gate, denoted as the "pore gate," located at the selectivity filter, also appears to be involved in the activation gating of Kv channels (Chapman et al., 1997; Zheng and Sigworth, 1997, 1998). Furthermore, the pore gate appears to be the primary activation gate in several other types of ion channels, including cyclic nucleotide–gated (CNG) channels (Liu and Siegelbaum, 2000; Flynn and Zagotta, 2001), SK-type Ca2+-dependent K+ channels (Bruening-Wright et al., 2002), and inward rectifier K+ channels (Xiao et al., 2003; but see Phillips et al., 2003).
Voltage-gated Ca2+ channels (VGCCs) are evolutionarily related to Kv channels (Hille, 2001). Although the pore of VGCCs is formed by four homologous repeats of a single 1 subunit rather than by four subunits as in Kv channels, the general features of the S4 voltage sensor and the S6 transmembrane segment are conserved. It is thus expected that VGCCs share similar gating mechanisms as Kv channels. A recent study using Y3+ block as a tool suggests that the activation gate of the 1G T-type Ca2+ channel is located on the intracellular side of the selectivity filter (Obejero-Paz et al., 2004). However, the exact location of this gate and the molecular movement associated with activation gating has not been determined.
In this work, we examined the location of the activation gate of the P/Q-type (Cav2.1) Ca2+ channel by studying the open/closed state dependence of modification by intracellular methanethiosulfonate ethyltrimethylammonium (MTSET) of engineered pore-lining cysteines in the S6 segments of the 1 subunit. These segments form the inner pore (Zhen et al., 2005). We found that in each repeat, cysteines placed above the putative membrane/cytoplasm interface were modified 1,000-fold faster by MTSET in the open state than in the closed state. Interestingly, two of these cysteines are located below the corresponding S6 bundle crossing in K+ channels. On the other hand, a cysteine placed at the putative membrane/cytoplasm interface was modified equally rapidly in both the open and closed states. These results suggest that the S6 helices of VGCCs undergo conformation changes upon voltage changes and the activation gate is located at or below the membrane/cytoplasm interface.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). cRNA for all constructs (control 1, mutant 1, 2
, and ß2a subunits) were transcribed in vitro using T7 polymerase after linearization. Oocytes preparation and cRNA injection were also as described (Zhen et al., 2005).
Electrophysiology
All patch-clamp recordings were performed using the inside-out configuration. Procedures and protocols for most of the recordings were as described (Zhen et al., 2005). In some experiments where rapid solution exchange was required, solutions were delivered to the intracellular side of the membrane through a pressurized fast perfusion system (SF-77B, Warner Instruments), which under our experimental conditions, could achieve a complete exchange of solution in 150 ms.
Macroscopic currents were evoked from a holding potential of –80 mV every 6 s by a 500-ms depolarization to +20 or +30 mV. In experiments to determine the voltage dependence of modification of K+ currents, the holding potential was set at –100 mV and currents were evoked by depolarization to different voltages, from –80 to +40 mV. For closed-state modification using regular perfusion, current was evoked by a 20-ms depolarization to +20 or +30 mV before and after application of MTSET, which was applied for 1–2 min and subsequently washed out for 1 min. When pressurized fast perfusion was used, current was evoked every 20 s by a 20-ms depolarization to +20 mV. MTSET was applied for 10 s in between the test pulses. For closed-state modification of IIS6-D0C, current was evoked every 10 s by a 20-ms depolarization to +10 mV and MTSET was applied for 4 s in between the test pulses. Currents were filtered at 2 kHz, digitized at 10 kHz. Data acquisition and analysis were performed using pClamp8 (Axon Instruments) on a PC through a Digidata 1200 interface. Experiments were performed at 21–23°C.
No corrections were made for leakage current, which was negligible in macropatch recordings, or for channel rundown, which was greatly reduced by inclusion of 2 mM Mg-ATP and 3 µM PIP2 in the intracellular solution. The time constant of modification was obtained by fitting the time course of modification with a single exponential function. The apparent second-order rate constant of MTSET reaction with a cysteine mutant channel was then calculated as the reciprocal of the time constant divided by the MTSET concentration. To determine the appropriate concentration of MTSET to use, we examined the relationship between MTSET concentration and its modification rate on the IIS6-A4C channel. At the voltage that opens 80% of the channels, the relationship was linear between 0.5 and 2 mM but reached a plateau at 3 mM (not depicted). Therefore, 1 mM MTSET was used for most of the open- and closed-state modification experiments. In experiments to determine the voltage dependence of MTSET modification (Fig. 5), different concentrations (0.5, 1, or 2 mM) of MTSET were used at different voltages in order to obtain a more accurate estimate of the second-order rate constants. MTSET (Toronto Research Chemicals) was stored at –20°C and was dissolved in the control solution before each experiment, generally <5 min before application.
To determine the closed-state modification rate for IIS6-D0C, we used the fast perfusion system mentioned above to apply MTSET only in the closed state. The time course of modification was plotted against the cumulative MTSET exposure time in the closed state, which was then fitted with a single exponential to get the time constant and the closed-state rate. The open-state modification rate was calculated by subtracting the closed-state rate from the rate obtained in both the open and closed states.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). These positions are highlighted in the amino acid sequence alignment of the four S6 segments of Cav2.1 (Fig. 1), where position 0 is presumed to lie at the membrane/cytoplasm interface and residues in the membrane are defined as 1, 2, 3, etc., from the intracellular side to the extracellular side. Fig. 1 also aligns the Cav2.1 S6 segments with the M2/S6 segment of two different types of K+ channels (adopted from Lipkind and Fozzard, 2003). According to this alignment, the S6 bundle crossing of the K+ channels is located three amino acids above the membrane/cytoplasm interface.
|
), we selected several cysteine mutant channels (including IS6-L5C, IIS6-V1C, IIS6-A2C, IIS6-A4C, and IIIS6-V9C) that were rapidly and almost completely modified by internal MTSET when it was applied continuously in both the open and closed states. We first examined modification by internal MTSET of IIIS6-V9C in the closed state (Fig. 2). Channel activity was recorded in inside-out membrane patches and was monitored by a brief (2 ms) depolarization to +30 mV from a holding potential of –80 mV. MTSET (1 mM) was applied for 2 min to the intracellular side only when channels were closed. The decrease in current, if any, was considered to be caused solely by MTSET modification in the closed state. There was only a slight decrease in current amplitude after MTSET application (Fig. 2 B), indicating that there was little modification of IIIS6-V9C in the closed state. By contrast, IIIS6-V9C was modified irreversibly and rapidly when MTSET was applied continuously in both the open and closed states (Fig. 3). Together, these results indicate that IIIS6-V9C can be modified only when the channel is open. The open-state modification could also be seen during the depolarizing test pulse, as reflected by the faster current decay (Fig. 3 A).
|
|
To calculate the modification rate in the closed state, we assume that current decay during MTSET application follows a single exponential. Using the current amplitude obtained before and after MTSET application, we calculated the time constant of the current decay and then the second-order rate constant, which was less or 1 M–1s–1 for all the positions situated above the putative membrane/cytoplasm interface.
We also examined closed-state modification of IIS6-A4C and IIIS6-V9C by applying MTSET in the closed state between brief test pulses (see voltage and MTSET application protocol in Fig. 4 A) through a pressurized fast perfusion system. The time course of modification was very slow (Fig. 4 B). By fitting the current decay phase during MTSET application with a single exponential (Fig. 4 C), we obtained a second-order rate constant of 0.5 M–1s–1, the same as that obtained with the method described above.
|
1,000-fold faster than those in the closed state, suggesting that there is a physical barrier for MTSET diffusion at or below the putative membrane/cytoplasm interface.
|
120-fold faster than at –80 mV. This large increase in the modification rate is apparently a consequence of channel gating. On the other hand, the modification rate increased by only 1.3-fold between –20 and +40 mV, when channel open probability reached maximum. This small increase indicates that MTSET diffusion into the inner pore has little intrinsic voltage dependence.
|
1,000-fold faster in the open state than in the closed state. In contrast, IIS6-D0C was modified rapidly in both the open and closed states, with only a threefold difference between the modification rates. More important, the closed-state modification rate of IIS6-V1C and IIS6-A2C is >1,000-fold slower than that of IIS6-D0C, suggesting that the gate for MTSET diffusion is situated between position 0 and 1 of IIS6, right at the putative membrane/cytoplasm interface.
Trapping of MTSET Also Suggests an Intracellular Gate
In many channels, blocking molecules are often trapped in the pore when the channel is closed (e.g., Armstrong, 1966, 1971, 1974; Holmgren et al., 1997; Shin et al., 2001; for review see Hille 2001), a phenomenon indicative of an intracellular gate. Does trapping occur to MTSET in VGCCs? To address this question, we compared MTSET modification of IIS6-A4C using two different voltage protocols, in which the magnitude of the test pulse (+30 mV) and the pulse interval (6 s) were maintained the same but the duration of the test pulse was set at 500 ms in one case and 10 ms in another (Fig. 7 A). IIS6-A4C was chosen because it exhibited little modification in the closed state (Fig. 5 A). This mutant channel was modified with both voltage protocols, but the modification was much faster with the 500-ms test pulse than with the 10-ms test pulse for the same number of test pulses (Fig. 7 B). This is expected if modification occurs predominantly in the open state. However, when the time course of modification was plotted using the cumulative channel open time, it became apparent that the modification was significantly faster (approximately fourfold) with the 10-ms test pulse (Fig. 7 C). One potential factor contributing to this apparent faster modification is the time (2–3 ms) taken for the open channels to completely close, which would increase the actual cumulative channel open time for the 10-ms test pulse by 20–30% (2–3 ms adds little to the 500-ms test pulse). However, this addition would increase the modification rate by only 20–30%. Furthermore, since 2–3 ms was also needed for all the channels to open, the real open time for the 10-ms test pulse should remain more or less 10 ms. A more likely explanation is trapping: MTSET enters the inner pore when the channel is open; it then gets trapped in the inner pore when an intracellular gate is closed and reacts with the cysteine while the channel remains in the closed state. Since more trapping events occur with the 10-ms test pulse for the same cumulative channel open time, the apparent modification rate becomes faster. Trapping in another way supports the existence of an intracellular gate.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
; Flynn and Zagotta, 2001): cysteines situated above the gate would be modified much more rapidly by an internally applied modifying reagent in the open state than in the closed state, but cysteines situated below the gate would be modified equally rapidly in both the open and closed states. If there is no intracellular gate, all the cysteines along the inner pore would be modified similarly in both the open and closed states. A critical assumption for using this method to locate the activation gate is that the intrinsic reactivity of the engineered cysteine does not change with gating. This assumption is difficult to test directly, but the strong state-dependent modification of cysteines above the putative membrane/cytoplasm interface in three different S6 segments and the clear-cut difference between those cysteines and IIS6-D0C (Fig. 5 A) suggest that such an assumption is a reasonable one.
The second-order rate constant for MTSET modification in the open state is two to three orders of magnitude lower than that for MTSET modification of free thiols (Stauffer and Karlin, 1994) and of cysteines in S6 of the Shaker K+ channel (Liu et al., 1997; del Camino and Yellen, 2001). Multiple reasons could account for this low rate, including the low intrinsic reactivity of the engineered cysteines and the low open probability of P/Q-type Ca2+ channels. Similar low rates have also been reported for cysteines in M2 of Kir channels (Phillips et al., 2003; Xiao et al., 2003), S6 of SK-type Ca2+-dependent K+ channels (Bruening-Wright et al., 2002), and S6 of CNG channels (Flynn and Zagotta, 2001). For the purpose of locating the activation gate, the critical measure is not the absolute rate of modification but the difference of this rate between the open and closed states. For cysteines above the membrane/cytoplasm interface, we observed an 1,000-fold difference in the modification rate between the open and closed states. The actual difference could be significantly bigger than what we estimated. In our calculation, we assumed that the current decay in the closed state following MTSET application was totally caused by modification in closed state. However, this small decrease of current (e.g., Fig. 2 B and Fig. 4 B) could be largely due to channel rundown. Thus the closed-state modification rate we estimated is an upper limit. For positions below the putative activation gate, we suppose that MTSET access is not gated but that there is a gating-associated movement of the S6 helices, which results in a small change in the modification rate (e.g., threefold difference between the open and closed states for IIS6-D0C). This small change in cysteine reactivity with gating has been seen in the lower S6 of Kv channels (Liu et al., 1997).
The strong state-dependent modification by MTSET of cysteines above the membrane/cytoplasm interface suggests that the S6 helices form a narrow constriction at the intracellular end of pore. Where is this constriction? In KcsA, the M2 bundle crossing forms the narrowest part of the inner pore (Doyle et al., 1998). This bundle crossing also forms the activation gate in the Shaker K+ channel: I477C, which lies at the bundle crossing (Fig. 1), shows strong state-dependent modification by internal MTSET but V478C, a position immediately below the bundle crossing, does not (Liu et al., 1997; del Camino and Yellen, 2001). Notably, in the P/Q-type Ca2+ channel, IIS6-A2C and IIS6-V1C both show strong state-dependent modification (Fig. 5 A). These two positions are located below the corresponding M2/S6 bundle crossing in K+ channels, according to the alignment in Fig. 1. This means that either this alignment is inappropriate or the location of the activation gate in VGCCs is different from that in the Shaker K+ channel. Although the details of the alignment in Fig. 1 may need revision, placing position 0 at the membrane/cytoplasm interface or even slightly below it seems appropriate, based on the hydrophobicity and the effect and pattern of MTSET modification of the amino acids below (i.e., more intracellular to) this position (see Fig. 3 of Zhen et al., 2005). Thus, we propose that the activation gate in VGCCs is located more intracellularly than that in the Shaker K+ channel.
Does the constriction that gates MTSET diffusion also gate Ca2+ conduction? An ideal and direct experiment to answer this question is to study the open/closed state dependence of block by internal Cd2+. Cd2+ binds to the EEEE locus in the selectivity filter with a high affinity (Yang et al., 1993), but since this effect is rapidly reversible, we cannot capitalize on this binding. However, Cd2+ interacts strongly with sulfhydryl groups, preferably organized in a tetrahedral geometry, and this interaction is generally extremely slow to reverse. This property has been exploited to study gating in other types of channels (Holmgren et al., 1998; del Camino and Yellen, 2001; Enkvetchakul et al., 2001; Loussouarn et al., 2001; Rothberg et al., 2002, 2003; Xiao et al., 2003; Webster et al., 2004). We also attempted to create a slowly reversible Cd2+ binding site in the Ca2+ channel pore by engineering two or four pore-lining cysteines at several analogous positions of the four S6 segments or by substituting one or two glutamates in the selectivity filter with cysteines. Unfortunately, our attempts were not successful due to several difficulties. First of all, because of the low membrane expression of Ca2+ channels in mammalian cell lines, all of our experiments were performed in Xenopus oocytes, which contain high levels of Ca2+-activated Cl– channels on the surface membrane. To avoid contamination from these channels, which can be activated by Ba2+ as well, albeit to a lesser degree, we used K+ as the charge carrier. But the contaminating Ca2+ in the intracellular solution (presumably mainly from our deionized water) was sufficient to activate the endogenous Ca2+-activated Cl– channels. Use of a Ca2+ chelator would not alleviate this problem since common chelators such as EGTA, EDTA, and BAPTA bind Cd2+ much more tightly than Ca2+, leaving any contaminating Ca2+ unbuffered. We thus resorted to use a "Ca2+ sponge" (Yang et al., 1993) to remove the contaminating Ca2+ from the intracellular solution. However, even with this maneuver, we still encountered large and variable Cl– currents because, surprisingly, Cd2+ itself could activate the Cl– current. We were therefore left with a low-probability chance of encountering oocytes with no or a very low level of endogenous Ca2+-activated Cl– channels. Furthermore, all quadruple cysteine mutant channels failed to produce a large enough macroscopic current. Although some single and double mutant channels expressed well, we were still unable to carry out the Cd2+ block experiments due to the aforementioned technical difficulty.
Despite the lack of direct supporting data, we propose that the intracellular gate that hinders MTSET diffusion is also a gate for Ca2+, because Ca2+ ions are most likely hydrated, at least partially, when they pass through the bundle crossing, and the size of a partially hydrated Ca2+ ion is similar to or larger than that of MTSET.
In the accompanying study, we concluded that the inner pore of VGCCs can open to >10 Å (Zhen et al., 2005). The closed channel is likely much narrower than this. Thus the S6 helices must undergo significant conformational changes during gating, as in K+ channels. But do Ca2+ channels open the pore like K+ channels do? A highly conserved glycine residue is present in the M2/S6 helix of various K+ channels and closely related CNG channels. It has been proposed (Jiang et al., 2002b) that this glycine acts as a flexible gating hinge, allowing the M2/S6 inner-pore helices to form a narrow constriction as in the KcsA structure and to splay wide open as in the MthK structure. However, it has been suggested recently (Webster et al., 2004) that unlike predictions from the MthK structure, the inner-pore helices of the Shaker K+ channel maintain the KcsA-like bundle-crossing motif in both open and closed states, with a bend at the Pro-X-Pro sequence, which is located above the bundle crossing and is conserved in Kv channels but not in inward rectifier K+ channels and the bacterial K+ channels. The implication of these studies is that the M2/S6 inner-pore helices may undergo different kinds of gating conformation changes in different types of K+ channels. In VGCCs, a conserved glycine is present in IS6 and IIS6 but not in IIIS6 and IVS6, at a position close or analogous to the conserved glycine in K+ channels. There is no Pro-X-Pro sequence in any of the S6 helices. It is thus expected that the gating motions of the S6 helices of VGCCs would be different from those of K+ channels.
|
ACKNOWLEDGMENTS |
|---|
2 cDNAs. This work was supported by National Institutes of Health grant NS45383.
Lawrence G. Palmer served as editor.
Submitted: 28 March 2005
Accepted: 2 August 2005
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1G T-type Ca2+ channel. J. Gen. Physiol. 124:631–640.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Related Article
This article has been cited by other articles:
|
|
 |
|
  H.-R. Chang and C.-C. Kuo The Activation Gate and Gating Mechanism of the NMDA Receptor J. Neurosci., February 13, 2008; 28(7): 1546 - 1556. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Bruening-Wright, W.-S. Lee, J. P. Adelman, and J. Maylie Evidence for a Deep Pore Activation Gate in Small Conductance Ca2+-activated K+ Channels J. Gen. Physiol., November 26, 2007; 130(6): 601 - 610. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Raybaud, E.-E. Baspinar, F. Dionne, Y. Dodier, R. Sauve, and L. Parent The Role of Distal S6 Hydrophobic Residues in the Voltage-dependent Gating of CaV2.3 Channels J. Biol. Chem., September 21, 2007; 282(38): 27944 - 27952. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Berkefeld, C. A. Sailer, W. Bildl, V. Rohde, J.-O. Thumfart, S. Eble, N. Klugbauer, E. Reisinger, J. Bischofberger, D. Oliver, et al. BKCa-Cav channel complexes mediate rapid and localized Ca2+-activated K+ signaling. Science, October 27, 2006; 314(5799): 615 - 620. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Erxleben, Y. Liao, S. Gentile, D. Chin, C. Gomez-Alegria, Y. Mori, L. Birnbaumer, and D. L. Armstrong Cyclosporin and Timothy syndrome increase mode 2 gating of CaV1.2 calcium channels through aberrant phosphorylation of S6 helices. PNAS, March 7, 2006; 103(10): 3932 - 3937. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Highlights from the Literature Physiology, December 1, 2005; 20(6): 369 - 373. [Full Text] [PDF]
|
|
|
|
|
|
X.-g. Zhen, C. Xie, A. Fitzmaurice, C. E. Schoonover, E. T. Orenstein, and J. Yang Functional Architecture of the Inner Pore of a Voltage-gated Ca2+ Channel J. Gen. Physiol., August 29, 2005; 126(3): 193 - 204. [Abstract] [Full Text] [PDF]
|
|
