Actions of volatile anesthetics on ligand-gated ion channels, such as gamma-aminobutyric acid type A receptors, have been studied extensively. However, actions on other types of channels, such as K+ channels, are poorly understood. The authors previously showed that a Ca2+-activated K+ channel, IK, is sensitive to halothane, whereas SK1, another Ca2+-activated K+ channel, is insensitive. To explore how halothane acts on Ca2+-activated K+ channels, chimeras between IK and SK1 were constructed, and halothane sensitivity was analyzed.
IK, SK1, and chimera channels were expressed in Xenopus laevis oocytes. Currents of expressed channels were measured in the presence of 10 microm Ca2+ by excised patch clamp analysis. Time constants of inhibition by halothane were compared between inside-out and outside-out patch configurations.
Currents from chimera channels possessing the pore domain derived from IK were inhibited by halothane, whereas those possessing the SK1 pore domain were insensitive. Time constants of inhibition by halothane were significantly smaller in the outside-out patches than in the inside-out patches of both wild-type IK and a chimera with pore domain of IK.
It is suggested that halothane interacts with the extracellular part of the ionic pore of IK. Whether this type of interaction is involved in the mechanism of anesthetic actions on ligand-gated ion channels warrants further investigation.
VOLATILE anesthetics affect functions of a variety of ion channels. Recent studies have shown that they have significant effects on neurotransmitter receptor channels, such as γ-aminobutyric acid receptor type A and nicotinic acetylcholine receptor. 1Moreover, the amino acid residues of these channels important in anesthetic action have been identified, and involvement of the second transmembrane region, which is also the pore-forming domain, has been suggested. 2–4However, the molecular mechanism of action of anesthetics on ion channels other than neurotransmitter receptor channels has yet to be explored.
Ca2+-activated K+channels (KCa) are gated by intracellular Ca2+ions, and their activity contributes to repolarization and afterhyperpolarization of the cell membrane. KCaare classified into three subtypes of channels, BK, IK, and SK, according to their unitary conductance. 5BK has a large unitary conductance of 100–220 pS and is gated by combined actions of membrane potential and Ca2+. IK and SK are gated solely by Ca2+and have unitary conductances of 20–85 pS and 2–20 pS, respectively. KCaconsists of four homomeric subunits similar to voltage-gated K+channels. 6Three different SK-channel isotypes have been identified by cDNA cloning and were named SK1, SK2, and SK3. 7Determination of amino acid sequences by cDNA cloning predicted that SK1–3 and IK are similar in structure and have six transmembrane domains (S1–6) and one pore domain located between S5 and S6 (fig. 1A). 7,8
Channels of the KCafamily respond differentially to volatile anesthetics. BK is inhibited in a Ca2+-dependent manner. 9IK is also inhibited but in a Ca2+-independent manner, and SK1–3 are not sensitive. 10Among them, IK is implicated in lymphocyte proliferation, 11chemotaxis and phagocytosis of human granulocytes, 12Ca2+signaling in platelets, 13shape regulation of erythrocytes, 14and action of vascular endothelial cells on smooth muscle cells. 15,16Correspondingly, IK inhibition may have a role in the actions of volatile anesthetics on these cells, 10such as the inhibition of T-cell proliferation, 17phagocytosis, 18platelet aggregation, 19and endothelial derived hyperpolarizing factor production. 20However, the molecular mechanisms of the action of anesthetics on this channel are not known.
To identify the regions that determine the differential halothane sensitivity of KCa, we constructed chimeras between IK and SK1 and analyzed their halothane sensitivity by electrophysiologic methods. We show that chimeras that contain the pore domain of IK are inhibited by halothane, whereas chimeras with that of SK1 are not sensitive. We also show that halothane inhibits the IK currents faster when applied from the extracellular side than from the intracellular side. These results suggest that halothane inhibits IK by interacting with the extracellular part of its ionic pore.
Materials and Methods
Animals were cared for and handled with approval from the Animal Care Committee of Kyoto University Graduate School of Medicine in Kyoto, Japan.
Female Xenopus laevis were obtained from Hamamatsu Seibutsu Kyouzai (Hamamatsu, Japan). Halothane was purchased from Takeda Pharmaceutical (Osaka, Japan). Although this product contains thymol as a preservative, it did not have any significant effects on IK currents up to a concentration corresponding to 10 mm halothane. cDNAs of human IK 8and human SK1 7were kindly provided by John P. Adelman, Ph.D. (The Vollum Institute, Portland, OR).
Chimeric channels between human IK and human SK1 were constructed by the overlap extension method, 21using Pfx DNA polymerase (Invitrogen, Carlsbad, CA). All polymerase chain reaction products were verified by sequencing (BigDye Terminator Cycle Sequencing; Applied Biosciences, Foster City, CA). Chimera constructs were as follows: chimera A, 1–275 of SK1 and 180–427 of IK; chimera B, 1–379 of SK1 and 282–427 of IK; chimera C, 1–281 of IK and 380–561 of SK1; chimera D, 1–275 of SK1, 180–281 of IK and 380–561 of SK1; chimera E, 1–275 of SK1, 180–228 of IK and 325–561 of SK1; chimera F, 1–312 of SK1, 217–226 of IK and 323–561 of SK1; chimera G, 1–322 of SK1, 227–234 of IK and 333–561 of SK1; chimera H, 1–312 of SK1, 217–281 of IK and 380–561 of SK1; and chimera I, 1–339 of SK1, 242–281 of IK and 380–561 of SK1.
Channel Expression in Xenopus Oocytes
cDNAs of human IK, human SK1, and chimeric channels were cloned into an oocyte expression vector, pBF. They were linearized by Pvu I and were in vitro transcribed with SP6 RNA polymerase (Promega, Madison, WI) in the presence of 3 mm m7G(5′)ppp(5′)G, a cap analog (Ambion, Austin, TX). The transcribed cRNA was dissolved in sterile water at approximately 1 μg/μl. A segment of Xenopus ovary was treated with 2% collagenase (Nitta Gelatin, Osaka, Japan) in modified Barth's medium (88 mm NaCl, 1 mm KCl, 2.4 mm NaHCO3, 0.3 mm CaNO3, 0.41 mm CaCl2, 0.82 mm MgSO4, and 15 mm HEPES, pH 7.6) for 1 h at 18°C, and then mature (stages V and VI) oocytes were manually defolliculated and incubated in the medium overnight at 18°C before RNA injection 10; 50 nl of the RNA solution was injected into an oocyte using a microinjector (Nanoject; Drummond, Broomall, PA). Oocytes were incubated for 2–5 days at 18°C in the modified Barth's medium supplemented with 0.1 mg/ml gentamicin until the experiments.
Voltage Clamp Analysis
Before the experiments, the vitelline membrane of the injected oocyte was removed with fine forceps after brief exposure to 200 mm mannitol in Barth's medium. Electrodes were pulled from thin-walled filamented glass capillaries (Narishige, Tokyo, Japan) by using a programmable pipette puller (P-97; Sutter Instrument, Novato, CA) and filled with a perfusion solution containing 116 mm potassium gluconate, 4 mm KCl, and 10 mm HEPES, pH 7.2, supplemented with CaCl2to give a Ca2+concentration of 10 μm. The proportion of the calcium binding to gluconate was calculated using a stability constant for calcium gluconate of 15.9 m−1. Resistance of the electrode was 4–8 MΩ with this solution. Inside-out or outside-out membrane patches of the oocytes were excised into the perfusion solution. The membrane patches were voltage-clamped using Axoclamp2B (Axon Instruments, Foster City, CA). The currents were low-pass filtered at 10 kHz, digitized at a sampling frequency of 40 kHz, and stored through Powerlab 4/s and analyzed using Powerlab software (AD Instruments, Castle Hill, Australia). At the end of each inside-out patch experiment, 0.1 mm EGTA in the intracellular solution was applied to the patches to estimate the leak current, and the value was subtracted from measured currents. The leak currents were typically between 10 and 50 pA at −100 mV. Digitized current data of 100–200 ms from the beginning of inhibition was used to calculate the inhibition time constant. A single exponential function fit (I = Imax· e−t/τ) was performed using a graph software, Kaleidagraph (Synergy Software, Reading, PA).
Application of Inhibitors
Halothane was delivered through a vaporizer (Fluotech; Datex-Ohmeda, Helsinki, Finland), which was calibrated with a gas analyzer (anesthetic gas monitor 303; Atom, Tokyo, Japan) and was equilibrated for at least 30 min with the intracellular solution including 10 μm Ca2+at room temperature by bubbling. Concentrations of anesthetics in the oocyte assay chamber were measured by gas chromatography and mass spectrometry (5890A; Hewlett-Packard, Palo Alto, CA) with head space sampler (19395A; Hewlett-Packard) and were 1.0–1.2 mm. This concentration of halothane inhibits IK currents submaximally. 10The halothane solution was applied to the patches by the fast application method with piezo-driven dual-channel theta glass tubing as described. 22,23Open-pipette analysis showed this apparatus changed the solutions with a time constant of 1.1 ± 0.2 ms (mean ± SEM of 5 independent experiments), which is in good agreement with a previous report. 23In some experiments, the perfusion solution was saturated with halothane to achieve the highest concentration possible. An airtight glass syringe was filled with this solution and was connected to the applicator, and the concentration of halothane in the bath was 10–15 mm.
Sodium gluconate solution (116 mm Na gluconate, 4 mm NaCl, and 10 mm HEPES, pH 7.2, including 10 μm Ca2+) was similarly applied to patches.
All values are presented as means ± SEM of patches excised from three or more oocytes of at least two different preparations. To examine the statistical significance, ANOVA followed by Bonferroni correction was performed, and P < 0.05 was regarded to be significant.
Halothane Sensitivity of the Chimera Channels
Human IK and SK1 showed amino acid sequence identity of 44%. To minimize the structural disruption, we constructed chimeric channels by switching the sequences with each other at the S4–5 loop and S6 region, respectively, because amino acid sequences are well conserved in these regions (fig. 1B) (chimeras A–D).
To assess the currents of chimeric channels and wild-type IK and SK1, inside-out membrane patches of the oocytes expressing these channels were examined. Figure 2Ashows the currents of an inside-out patch of oocyte expressing IK. The currents were rapidly and reversibly inhibited by EGTA, indicating that the currents were activated by Ca2+. Consistent with a previous report, 10the currents were rapidly and reversibly inhibited by 1 mm halothane, suggesting that the Ca2+-activated current of IK is halothane-sensitive. Although SK1 currents were also inhibited by EGTA, the currents were not inhibited by halothane, which is in contrast to IK (fig. 2B). Similar to IK and SK1, the currents of chimeras A–D were inhibited by EGTA (figs. 2C and D), but they showed difference in halothane sensitivities. For instance, the current of chimera A was also inhibited by halothane significantly and reversibly (fig. 2C), whereas that of chimera B was not sensitive (fig. 2D). Figure 3Asummarizes the inhibition of IK, SK, and chimera channels by 1 mm halothane. The results indicate that the channels containing the region between S5 and the extracellular half of S6 derived from IK (IK and chimeras A, C, and D) are halothane-sensitive, that chimeras containing the corresponding region of SK1 (SK1 and chimera B) are halothane-insensitive, and that chimeras C and D show diminished sensitivity to 1 mm halothane compared with IK (P < 0.01).
To further investigate the reduced halothane sensitivity of chimeras containing the S5–6 regions of IK, halothane concentration was increased to 10 mm, and the inhibition of IK, SK1, and chimera D was examined (fig. 3). This concentration of halothane inhibited the currents of chimera D comparable with that of IK. This concentration of halothane, however, still failed to inhibit SK1, indicating that the inhibition is specific to the channels containing the pore domain of IK, even at this concentration. These results suggest that the reduced sensitivity of chimera D is the result of an increase in IC50value and not a reduction in maximal inhibition.
To further narrow the region responsible for halothane sensitivity, several chimeric constructs were made to switch the amino acid sequence at various positions surrounding the pore. These include chimeras containing one of the following IK domains, with the rest of the amino acid sequence derived from SK1: S5 domain (chimera E), S5 to pore domain (chimeras F and G), S5 to S6 domain (chimera H), and pore to S6 domain (chimera I). However, no significant Ca2+-activated currents could be recorded in most of these chimeras, probably because the channel pore structure was disturbed by manipulating the pore-related sequence (data not shown). One construct that has the S5 domain derived from IK (chimera E) showed Ca2+-activated currents; this channel was not sensitive to halothane (fig. 3). This result suggests that halothane sensitivity is determined by the IK domains between the third extracellular loop and extracellular half of S6 (fig. 2B).
Time Constants of Halothane Inhibition
To determine whether halothane acts on the intracellular side or the extracellular side of the ionic pore, we compared the halothane effect on IK when it was applied from the extracellular side and from the intracellular side.
To assess the speed of the test solution to reach the patch, we tested the effect of bath application of sodium gluconate solution on currents in inside-out and outside-out patches. Sodium gluconate solution inhibits the potassium currents by replacing K+with Na+, which is impermeable to IK. When the membrane potential was held at +100 mV in the inside-out configuration, the outward current was inhibited by sodium gluconate applied to the bath (fig. 4A), whereas in the outside-out patch, the inward current measured at −100 mV was inhibited by sodium gluconate applied to the bath (fig. 4B). Thus, IK currents were reversibly inhibited by sodium gluconate in both patch configurations, and the inhibition time constants were not significantly different from each other (19.1 ± 2.4 ms [n = 7] and 20.1 ± 2.8 ms [n = 3] for inside-out and outside-out patches, respectively, P > 0.05). Because Na+and halothane have similar diffusion time constants in aqueous solution (1.3 ×10−9and 1.0 ×10−9m2/s, respectively 24,25), this result suggests that halothane reaches the membrane with a similar time constant from either side of the membrane. Although IK currents were also inhibited by halothane in both patch configurations, inhibition was significantly slower in the inside-out patches than in the outside-out patches (figs. 4C and D, and 5) (37.9 ± 3.2 ms [n = 15] and 6.8 ± 1.8 ms [n = 7] for inside-out and outside-out patches, respectively, P < 0.001). This phenomenon was also observed in chimera D, which contained the shortest region of IK (fig. 5) (50.2 ± 9.4 ms [n = 9] and 7.6 ± 1.0 ms [n = 8] for inside-out and outside-out patches, respectively, P < 0.001).
The aim of this study was to identify regions that are necessary for halothane inhibition of IK. We used the difference in the halothane sensitivity between IK and SK1 to investigate the molecular basis for IK modulation by halothane. Analysis of chimeric channels between IK and SK1 showed that the region around the pore domain is important for determining halothane sensitivity of KCa. This region forms the extracellular side of the ionic pore according to crystal structure analyses of various potassium channels and toxin mapping study of IK. 6,26–28
In the current study, halothane inhibited currents of wild-type IK and chimera D, 30–40 ms more slowly in the inside-out patches than in the outside-out patches. This observation suggests that halothane interacts with IK from the extracellular side. In the study of halothane movement in an artificial phospholipid bilayer, halothane moves rapidly throughout the bilayer and reaches equilibrium within 2–10 ns. 29,30If this is the case with the oocyte plasma membrane, the delay of 30 ms cannot be accounted for by the diffusion time through the plasma membrane. Therefore, we speculate that the delay in inside-out patches may be the result of the time required for the transitions of the anesthetic from bath solution to membrane and/or from membrane to pipette solution.
In this study, we showed that the extracellular part of the IK pore is important for the action of halothane, although contribution from other regions cannot be excluded. For example, substantial reduction of inhibition was observed in chimeras C and D (fig. 3) (P < 0.01 vs . IK), suggesting roles of IK regions that are excluded in these chimeras. That inhibition of chimera A was not significantly reduced (fig. 3) (P > 0.05 vs . IK), although that of chimera C was reduced, suggests a role for S6 or the C-terminal tail of IK. However, the role of this region may be marginal, because chimera B showed no sensitivity to halothane. Consistently, chimera D was fully inhibited at higher concentrations of halothane. It is likely that the C-terminal tail affects the structures of pore region so that halothane affinity is altered. Alternatively, it is possible that the C-terminal tail would modify the Hill slope of halothane inhibition of chimera D. This could mean that halothane also binds to the C-terminal tail but that it only facilitates and does not mediate the inhibition. In any case, it seems that the pore domain plays the most critical role in halothane inhibition and that the other regions have only supportive roles at most.
Studies of γ-aminobutyric acid receptor type A, nicotinic acetylcholine receptor, and glutamate receptors also suggested that extracellular regions are important for the action of the volatile anesthetics in these channels. For example, mutations of amino acids in the extracellular part of transmembrane regions of the α2subunit of γ-aminobutyric acid receptor type A attenuate the action of volatile anesthetics. 2,3In the α subunit of nicotinic acetylcholine receptor, a photoreactive general anesthetic, azioctanol, binds at the extracellular half of the pore-forming M2 region. 4In a glutamate receptor subunit, GluR6, halothane action is abolished by a mutation of Gly819, which is located at the extracellular end of a pore-forming domain for glutamate receptors. 31,32These amino acids are proposed to form pockets for anesthetics distinct from that for the ligands. 1,3,33,34It is noteworthy that so many ion channels appear to have such pockets at or close to the extracellular side of pore, although gating mechanisms of these channels are different from each other. Because as many as 10–15% of a randomly selected group of proteins are predicted to have pockets for general anesthetics, 34it is possible that only the proteins that have such pockets at or close to their pores are affected by anesthetics. In any case, it is likely that anesthetic interaction at the extracellular side of a pore-forming domain is one of the shared features of a certain group of ion channels.
Some K+channels other than KCaare also modulated by volatile anesthetics. Voltage-gated K+channels, Kv1.1, Kv2.1, and Kv3, are inhibited by halothane. 35,36Two-pore-domain K+channels, which are determinants of the resting membrane potential, have been suggested to be a candidate of important molecular targets for anesthetics. Some of them (TOK-1, TREK-1, and TASK-1) are activated, and others, such as TWIK-2 and THIK-1, are inhibited by volatile anesthetics. 37–39Another family of potential targets for anesthetics in the central nervous system is G-protein coupled inward rectifying potassium channels, GIRK-1 and GIRK-2. 40Although the mechanisms of these effects are largely unknown, there is a possibility that anesthetics also act on the extracellular side of the pore of these channels, considering that the structure of the pore of potassium channels are well conserved. 41Further investigations are required to define the roles of the pore domain in functional modulation of these channels by anesthetics.
Although the immunostaining experiment shows expression of IK in a subpopulation of neurons, 42IK is found mainly in nonneuronal cells, such as lymphocytes, platelets, and smooth muscle cells. The inhibition of IK would modulate functions of these cells and may be involved in the side effects of inhaled anesthetics, such as immune suppression, 11inhibition of platelet aggregation, 13and alteration in vascular smooth muscle tone. 15,16The current study should contribute to the elucidation of the molecular mechanism of these side effects of volatile anesthetics.
The authors thank John P. Adelman, Ph.D., Senior Scientist, The Vollum Institute, Portland, Oregon, for cDNA clones; R. Adron Harris, Ph.D., Director, and members of his laboratory, Waggoner Center for Alcohol and Addiction Research, Austin, Texas, for critical reading of the manuscript; and Kazuyo Kamiyama, M.S., Researcher, Central Research Laboratories, Maruishi Pharmaceutical Co., Ltd., Osaka, Japan, for assistance with the measurement of halothane concentrations.