Previous Article |
Table of Contents
| Next Article 
BIOLOGICAL SCIENCES / NEUROSCIENCE
Mg2+ mediates interaction between the voltage sensor and cytosolic domain to activate BK channels
,
Department of Biomedical Engineering and Cardiac Bioelectricity and Arrhythmia Center, Washington University, St. Louis, MO 63130; and Department of Physiology, University of Pennsylvania, Philadelphia, PA 19104-6085
Edited by Ramón Latorre, Centro de Estudios Científicos, Valdivia, Chile, and approved September 25, 2007 (received for review June 22, 2007)
 |
Abstract |
|---|
 Top Abstract ResultsDiscussionMethodsAcknowledgementsReferences |
|---|
electrostatic interaction | interdomain interaction | magnesium | Ca2+-activated K+ channel | metal ion regulation
|
Results |
|---|
|
TopAbstractResultsDiscussionMethodsAcknowledgementsReferences |
|---|
31 Å) roughly coincides with that between the C terminus of S4 and the central axis of the Kv1.2 channel (Fig. 1a, yellow circle), suggesting that the Mg2+ site and VSD of the BK channel may lie close together. However, the detailed structure of the BK channel is unknown. Moreover, the BK channel is a complex allosteric protein whose cytosolic domain is thought to undergo conformational changes associated with ligand binding and channel opening (11). Therefore, it is not clear whether Mg2+ and the VSD interact directly or indirectly, through what physical mechanism, or in which conformational state(s). One possibility, depicted in Fig. 1b, is that the binding of the Mg2+ may alter the local electric field that is sensed by charged residues in the VSD, thereby activating the channel. This hypothesis is consistent with the fact that Mg2+ shifts the voltage dependence of channel opening to more negative voltages and that the neutralization of R213 abolishes Mg2+ sensitivity (18). Another line of evidence supporting this hypothesis is that Mg2+-dependent activation is sensitive to the ionic strength of intracellular solution. In our normal intracellular solution with intracellular K+ concentration ([K+]i) = 142 mM (see Methods), 10 mM Mg2+ activates the channel by shifting the voltage dependence of conductance (G-V relationship) to more negative voltages 
V1/2 = –60 ± 3.8 mV, where V1/2 is the voltage at half-maximum conductance (Fig. 1 c and d). The amount of G-V shift caused by 10 mM Mg2+ increases in lower [K+]i and decreases in higher [K+]i (Fig. 1e), as expected if Mg2+ activates the channel by an electrostatic interaction. To the contrary, ionic strength does not significantly affect Ca2+-dependent activation (Fig. 1e), suggesting that Ca2+ does not activate the channel by an electrostatic interaction and the changes in ionic strength do not affect Ca2+ binding. Similarly, the strong response of Mg2+-dependent activation to ionic strength may not reflect changes in Mg2+ binding.
|
Consistent with the above prediction, mutations of Q397 to positively charged amino acids (Q397K, Q397R) activate the channel by shifting the G-V relation to more negative voltages in the absence of intracellular Mg2+ (Fig. 2 a and b). Also consistent with an electrostatic interaction, mutations of Q397 to negatively charged amino acids (Q397E, Q397D) shift the G-V relation to more positive voltages, whereas Q397C has insignificant effects on the G-V relation (P > 0.05). Similar to Mg2+-dependent activation, the G-V shift caused by mutation Q397K is also affected by ionic strength (Fig. 2c). The fractional changes of V1/2 caused by Mg2+ and the mutation superimpose at low [K+]i and only differ at high [K+]i. Because the mutation adds a charge permanently to the channel, this result further suggests that changing ionic strength alters Mg2+-dependent activation primarily by affecting an electrostatic interaction, with only a small effect on Mg2+ binding. BK channels traverse multiple states during activation (11). Interestingly, the change in electric field produced by Mg2+ binding is not felt uniformly by all these transitions. Unlike the G-V relation, the voltage dependence of activation kinetics (
-V relation) is not shifted by increasing intracellular Mg2+ concentration ([Mg2+]i) from 0 to 10 mM (27–29).¶ Similarly, the mutations Q397R and Q397K do not shift the -V relation (Fig. 2d), further supporting that positive charges at position 397 affect the same set of gating transitions as Mg2+.
|
State-Dependent Interaction Between Mg2+ and VSD. The movement of the voltage sensor in response to a change in voltage generates a transient gating current (Ig). If Mg2+ activates the channel through electrostatic interactions with the VSD, it should alter Ig. BK channel voltage sensors can move between the resting and activated state and generate gating current no matter whether channels are closed or open (16, 31) (Fig. 3). However, voltage-sensor activation occurs at less depolarized voltages in the open conformation, thereby promoting voltage-dependent opening (31). The results in Fig. 3 show that, consistent with our hypothesis, Mg2+ enhances voltage-sensor activation, but interestingly, Mg2+ acts primarily when channels are open and has little effect when channels are closed.
|
10% increase at +250 mV; data not shown), whereas the time course of IgON is not altered (Fig. 3c). The relationship between gating charge movement and membrane voltage when channels are closed (QC-V relation), determined by integrating the area under the IgON traces at various voltages, exhibits no significant change between 0 and 10 mM [Mg2+]i (Fig. 3b) (10-mV difference in the voltage at the half-maximum of the QC-V relation, Vhc). These results indicate that Mg2+ does not affect the closed channel significantly, which may also explain why K+ ionic current (IK) activation kinetics are insensitive to Mg2+ (Fig. 2d). On the other hand, off Igs (IgOFF), which reflect the return of the voltage sensor from the activated to the resting state, are decreased in amplitude and slowed by 10 mM [Mg2+]i (Fig. 3 c and d). Under our experimental conditions (see Methods) K+ currents through the pore are blocked but channels are still open in response to depolarization (31). Therefore, a fraction of channels are open at the end of a voltage pulse when IgOFF is recorded. To examine whether the change of IgOFF is related to the fraction of open channels, we measured IgOFF after pulses of various duration (Fig. 3c). After a 0.30-ms voltage pulse, few channels are open and IgOFF is similar in the presence or absence of Mg2+. However, as pulse duration increases IgOFF amplitude is greatly reduced in 10 mM [Mg2+]i (red traces, Fig. 3c), indicating that Mg2+ affects voltage sensor movement more significantly when channels are open. Previous studies indicate that voltage-sensor deactivation is slowed when channels are open even in the absence of Mg2+ (31, 32) (Fig. 3d). In our experiment at +250 mV, channels should be almost maximally activated in the presence or absence of Mg2+ (Figs. 1d and 3b). Therefore, the greater effect on IgOFF in 10 mM [Mg2+]i (Fig. 3d) does not merely reflect a change in open probability but rather implies that deactivation of voltage sensors in open channels is slowed by Mg2+.
The charge-voltage relation for open BK channels (QO-V) is difficult to measure directly from gating currents because, in contrast to the case of voltage-sensor activation and channel opening (Fig. 3a), the time course of voltage-sensor deactivation is not much faster than channel closing (31). However, a shift in the QO-V relation to more negative voltages caused by 10 mM [Mg2+]i is evident from the voltage dependence of open probability (Po) at negative voltages (Fig. 3f). BK channels can open even when voltage sensors are in the resting state; thus Po can be measured at voltages more negative than the range of voltage-sensor activation (33). Such openings have weak voltage dependence as indicated by the red line in Fig. 3f. At more positive voltages the Log(Po)-V relation becomes steeper (solid lines in Fig. 3f) because of voltage-sensor activation, and this inflection in slope provides an indication of the position of the QO-V relation along the voltage axis (32). Clearly, Mg2+ does not affect channel opening when the voltage sensor is in the resting state because the Log(Po)-V relations in 0 and 10 mM [Mg2+]i superimpose at the most negative voltages. Mg2+ increases Po only at potentials where voltage sensors can activate and shifts the inflection point in the Log(Po)-V relation to more negative voltages, indicating that Mg2+ enhances voltage-sensor activation when the channel is open. This result is consistent with an earlier report (26). Fitting the Log(Po)-V relations over a wide range of the Po with a well established model that describes the voltage-dependent activation of BK channels (33) (see Methods) reveals that 10 mM [Mg2+]i shifts the Qo-V relation by –50 ± 2 mV (Fig. 3g).
Similar to Mg2+ binding, a positive charge at residue 397 enhances voltage-sensor activation more when channels are open. QC-V relations for the mutant channels Q397R and Q397K are similar to those of the WT (Fig. 3b). Modification of the Q397C channel by MTSET(+) reduces the amplitude and slows the relaxation of IgOFF but leaves IgON unchanged (Fig. 3e). The mutation Q397R, like Mg2+, has no effect on Po at very negative voltages, but shifts voltage-sensor activation for open channels to more negative voltages by –24 ± 2 mV (Fig. 3 f and g), similar to the shift in the G-V relation by Q397R (Fig. 2c). Because the Q397R channel is still sensitive to Mg2+ (21), the charges of the Arg side chain and the bound Mg2+ must be located at different positions. Given this difference, the similar effects on voltage-sensor activation and channel opening produced by the spatially separated Arg at 397 and bound Mg2+ suggest that the interaction between cytosolic and VSDs is mediated by long-range electrostatic interactions.
R213 in Transmembrane Segment S4 Senses Mg2+. Which residues in the VSD sense the changes in the electric field as a result of Mg2+ binding and are responsible for the functional changes in channel gating? Analysis of the primary sequence of BK channels indicates that 12 charged residues in the VSD (S1–S4 and the S4–S5 linker) may be exposed to cytosol. We mutated each of these residues individually to alter their charge and examined the effects of these mutations on the shift of the G-V relation in response to 10 mM Mg2+ (Fig. 4a). Of all of these mutations, only R213C completely abolishes Mg2+ sensitivity (Fig. 4 a and b), consistent with the previous finding that R213Q abolishes Mg2+ sensitivity (18). Residue 213 is located in the C-terminal half of S4, is accessible to intracellular Cys-modifying reagents when the voltage sensor is in the resting state (18), and contributes to gating charge (14, 15). Thus, this residue is a sensor of both membrane voltage and Mg2+ binding. Fig. 4b shows that when R213 is mutated to Cys the substitution of a positively charged residue at position 397 (the double mutation R213C:Q397R) no longer shifts the G-V relation to negative voltages, supporting that R213 is the main sensor of charges around the Mg2+-binding site. Finally, if Mg2+ activates the channel by an electrostatic interaction with the charge of R213, adding a positive charge back to position 213 should rescue some of the Mg2+ sensitivity that is lost because of the R213C mutation. As expected, modification of R213C channels, but not R213Q, by MTSEA(+) restores Mg2+ sensitivity such that the deactivation of the channel is slowed down and the G-V relation is shifted to more negative voltages by 10 mM Mg2+ (Fig. 4 c and d). The results in Fig. 4 indicate that a positive charge at residue 213 is sufficient and necessary for the VSD to sense Mg2+ binding and bears the primary responsibility for Mg2+-dependent activation.
|
|
Discussion |
|---|
|
TopAbstractResultsDiscussionMethodsAcknowledgementsReferences |
|---|
During voltage-dependent activation, S4 moves in response to voltage changes to open the channel. Our results indicate that R213 in BK channel S4 can be sufficiently close to its tethered cytosolic domain to interact electrostatically with bound Mg2+ and charge at residue 397. Because the energy of the electrostatic interaction is manifested as a shift in the Qo-V relation we can estimate the average distance between the charges of R213 and R397 as follows. Because voltage sensors in each of the four BK channel 
-subunits are thought to activate independently via a two-state process (31, 32), the energy to shift the QO-V is:
 |
where zg is the gating charge per voltage sensor (zg = 0.58e) (15) and V1/2 = –24 mV (Fig. 3 f and g). The energetic effect on voltage-sensor activation produced by electrostatic interaction of Q397R with R213 is:
 |
where 
R and A are the electrostatic potentials imposed by 397R on R213 in the resting and activated state, respectively, and zR213 is the charge of R213 (zR213 = e). Because falls off quickly with distance and in the activated state R213 is likely to be buried in the protein because it is a gating charge (15), we can assume A << R. Thus, combining Eqs. 1 and 2 to solve for G[QO] = G[R213] yields:
 |
This electrostatic potential between the charges of R213 and R397 (R = 13.9 mV) implies that the average distance between them in the resting state is 9.1 Å (see Methods for calculation). This calculation may overestimate the distance if R213 also interacts with 397R when the voltage sensor is in the activated state. Such a short distance between the membrane-spanning and the cytosolic domains in BK channels is consistent with the structure of BK channels, in which a short peptide linker of only 17 aa connects S6 to the cytosolic domain (35). Thus it is reasonable to expect that close interaction between the membrane spanning and the cytosolic domains can occur.
We find that the effect of Mg2+ is state-dependent, enhancing voltage-sensor activation primarily when the channel is open. Although Mg2+ has been proposed to bind with lower affinity to closed than open channels (36), state-dependent binding cannot account for the lack of effect of Mg2+ on QC-V. First, 10 mM [Mg2+]i is higher than the estimated dissociation constant of the closed channel (36) so that more than half of the Mg2+-binding sites should be occupied when channels are closed. Second, the permanent addition of positive charge at residue 397 near the Mg2+-binding site also has no effect on QC-V (Fig. 2b). Taken together, these results indicate that the interaction between the bound Mg2+ and the VSD is state-dependent, presumably involving a conformational change in the cytosolic domain and/or VSD upon channel opening that exposes R213 to the electric field of the bound Mg2+. One possibility is that the distance between R213 and the bound Mg2+ is enlarged in the closed state because of the relative movements between S4 and the cytosolic domain. The crystal structures of MthK channels suggest that the cytosolic gating ring may expand 8 Å at the outer rim in parallel to the membrane from the closed to the open conformation (37). On the other hand, in BK (15) and Shaker (38) K+ channels rearrangements of the VSD have been suggested to occur upon channel opening. Similar changes in BK channels may change the distance between R213 and the bound Mg2+. Another possibility is that the conformational rearrangements in the VSD and the cytosolic domain during channel opening may change the orientation of R213 or the bound Mg2+ such that the side chain of R213 is only exposed to the electric field of the bound Mg2+ in the open state. The detailed mechanism of the state dependence of Mg2+ modulation is not known at this time. Nevertheless, it is clear that VSD activity is regulated not only by Mg2+ bound to the cytosolic domain but also by feedback from the activation gate in the pore domain.
Aside from elucidating the mechanism of Mg2+-dependent activation in BK channels, our results also demonstrate that a tethered cytosolic protein domain can come close enough to the VSD to regulate its activity through electrostatic interaction. This interaction sets precedence that various cytosolic domains, accessory subunits, or regulatory proteins may, through a similar mechanism, regulate the function of other ion channels or enzymes that contain VSDs. Introduction of charged residues at position 397 in the BK channel alters voltage-sensor activation, suggesting that interactions between the VSD and native charged residues in the cytosolic domain may normally exist and influence VSD function even in the absence of a charged ligand. A similar example may involve the cytosolic tetramerization (T1) domain of Kv channels where mutation of some charged residues facing the transmembrane domain produce large shifts in the voltage dependence of channel activation through unknown mechanisms (39, 40). Importantly, charges in the cytosolic domain of the BK channel do not act merely to bias the voltage detected by the VSD, as in the case of surface-charge effects (41). Rather, the impact of these charges varies as the relative position of cytosolic and VSDs change during channel opening. Similarly, it is possible that conformational changes in the cytosolic domain induced by neutral ligands or ligands distal from the VSD may expose the VSD to native charges to alter the activation of channels and enzymes. Finally, the VSD may be regulated by signaling processes other than ligand binding because posttranslational modifications such as phosphorylation, sulfation, oxidation, protonation, and methylation can increase or decrease the charge of amino acid side chains in cytosolic domains.
|
Methods |
|---|
|
TopAbstractResultsDiscussionMethodsAcknowledgementsReferences |
|---|
Electrophysiology. Ionic currents were recorded with inside-out patches with an Axopatch 200-B patch-clamp amplifier (Axon Instruments, Union City, CA) and Pulse acquisition software (HEKA Electronik, Lambrecht/Pfalz, Germany) (19). The pipette solution contains 140 mM KMeSO3, 20 mM Hepes, 2 mM KCl, and 2 mM MgCl2, pH 7.2. The internal solution contains 140 mM KMeSO3, 20 mM Hepes, 2 mM KCl, and 1 mM N-(2-hydroxyethyl)ethylenediamine-N,N',N'-triacetic acid (HEDTA), pH 7.2. CaCl2 was added to the internal solution to give the appropriate free intracellular Ca2+ concentration ([Ca2+]i), which was measured with a calcium-sensitive electrode. For 0 [Ca2+]i, the same internal solution was used except that HEDTA was substituted by 5 mM EGTA and no CaCl2 was added. The basal internal solution for ionic strength experiments contains 20 mM Hepes, 15 mM KOH and 5 mM EGTA for 0 [Ca2+]i and 20 mM Hepes, 6.5 mM KOH for 200 µM [Ca2+]i, pH 7.2. KCl was added to reach the target [K+]is. A sewer pipe flow system was used to perfuse the internal solution on the cytosolic face of the patch. Po (<0.01–0.1) at negative voltages was measured by single-channel recordings in patches containing hundreds of channels (17, 33).
Gating currents were recorded with inside-out patches (31). The pipette solution contains 127 mM tetraethylammonium hydroxide, 125 mM HMeSO3, 2 mM HCl, 2 mM MgCl2, and 20 mM Hepes, pH 7.2. The internal solution contains 141 mM N-methyl-D-glucamine, 135 mM HMeSO3, 6 mM HCl, 20 mM Hepes, and 5 mM EGTA, pH 7.2. MgCl2 was added to the internal solution to reach 10 mM [Mg2+]i. Voltage commands were filtered at 20 kHz with an eight-pole Bessel filter (Frequency Devices, Haverhill, MA) to prevent the saturation of fast capacitive transients (31). Data were sampled at 100,000 Hz with an 18-bit A/D converter (ITC-18; Instrutech, Mineola, NY) and filtered at 10 kHz with Axopatch's internal filter. Capacitive transients and leak currents were subtracted by using a P/5 protocol with a holding potential of –120 mV.
Chemical Modification. MTSET ([2-(trimethylammonium)ethyl] methanethiosulfonate bromide) and MTSEA (2-aminoethyl methanethiosulfonate hydrobromide) were purchased from Toronto Research Chemicals (Downsview, Canada). An aliquot of 100 mM MTS reagents stock solution was thawed and diluted 500-fold into the basal internal solution immediately before use. In ionic current recordings, currents were recorded after 2.5 min of MTS reagents treatment and 0.5 min of washing of patches with the inside-out configurations. In gating current recordings, MTSET was added directly into the bath.
Data Analysis.
For macroscopic current, relative conductance was determined by measuring tail current amplitudes at –80 mV. The gating charge movements were determined by integrating the area under the rising phase and single exponential fits to the decaying phase of IgON at various voltages. The conductance-voltage (G-V) relations or the charge-voltage (Qc-V) relations of the WT and mutant channels were fitted with the Boltzmann equation:
 |
 |
In Eq. 4, z is the number of equivalent charges, V1/2 is the voltage for channel in half-activation, e is the elementary charge, k is Boltzmann's constant, and T is the absolute temperature. In Eq. 5, zJ is the gating charge associated with voltage sensor movement, Vhc is the voltage for half of the gating charge movements at the closed conformation of the channel, and the other parameters are as in Eq. 4. The data were obtained from n = 4–16 patches. Error bars represent the SEM in all figures.
HCA Model Fitting.
Po-V relation at negative voltages, G-V and Qc-V relations at 0 [Ca2+]i were fitted to Eq. 6 derived from the HCA model (33),
 |
where Vhc and Vho are the voltages for half of the gating charge movements at the closed and the open conformation, respectively; zJ is the gating charge associated with voltage sensor movement; L0 is the intrinsic equilibrium constant for the activation gate, and zL is the charge associated with channel opening.
Our results indicate that mutation Q397R shifts the Qo-V relation via changing the local voltage sensed by R213. R213 is likely to be located in the crevice among the VSD, pore and cytosolic domains and exposed to the aqueous solution when interacting with R397, which is located at the border between a low dielectric (protein of the cytosolic domain) and a high dielectric (water in the crevice) medium. Thus, the distance between the charges of R213 and R397, r = 9.1 Å, can be calculated from the following equation (42, 43):
 |
where 
R = 0.58*V1/2, V1/2 is the the Qo-V shift (–24 mV; Fig. 3 f and g), e is the elementary charge, 
r is the dielectric constant of water, 0 is the permittivity of free space, and 1/
(8.07 Å) is the Debye length in the aqueous phase and is calculated as in ref. 26.
Structural Model. The structure of Mg2+ binding site of the mSlo1 channel was generated based on the crystal structure of the MthK channel by using the PyMol molecular graphics system (www.pymol.org).
|
Acknowledgements |
|---|
|
TopAbstractResultsDiscussionMethodsAcknowledgementsReferences |
|---|
|
Footnotes |
|---|
Abbreviations: VSD, voltage-sensor domain; Kv, voltage-dependent K+; V1/2, voltage at half-maximum conductance; Vhc, voltage for half of the gating charge movements at the closed conformation of the channel; Ig, gating current; IgON, on Ig; IgOFF, off Ig; IK, K+ ionic current; [Mg2+]i, intracellular Mg2+ concentration; [K+]i, intracellular K+ concentration; Po, open probability; [Ca2+]i, intracellular Ca2+ concentration.
To whom correspondence should be addressed. E-mail: jcui{at}biomed.wustl.edu
Author contributions: H.Y. and L.H. contributed equally to this work; H.Y., L.H., and J.C. designed research; H.Y., L.H., J.S., and K.D. performed research; F.T.H. contributed new reagents/analytic tools; H.Y. and L.H. analyzed data; and H.Y., F.T.H., and J.C. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
¶ Horrigan FT (2005) Biophys J 88:100A (abstr). 
© 2007 by The National Academy of Sciences of the USA
|
References |
|---|
|
TopAbstractResultsDiscussionMethodsAcknowledgementsReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg What's this?
This article has been cited by other articles in HighWire Press-hosted journals:
|
|
 |
|
  Z. Ma, K. Y. Wong, and F. T. Horrigan An Extracellular Cu2+ Binding Site in the Voltage Sensor of BK and Shaker Potassium Channels J. Gen. Physiol., May 1, 2008; 131(5): 483 - 502. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Lingle Mg2+-dependent Regulation of BK Channels: Importance of Electrostatics J. Gen. Physiol., December 31, 2007; 131(1): 5 - 11. [Full Text] [PDF]
|
|
|
|
|
|
F. T. Horrigan and Z. Ma Mg2+ Enhances Voltage Sensor/Gate Coupling in BK Channels J. Gen. Physiol., December 31, 2007; 131(1): 13 - 32. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
