The tandem pore, or 2P, family of K channels is a collection of ion channels whose importance in physiologic mechanisms is still emerging (1). They appear to be an ancient class of channels; 2P K channels are found widely, from single-cell yeast to plants to higher mammals (2,3). More than 50 genes with the predicted 2P K channel structure have been identified in Caenorhabditis elegans, as well as 15 human genes (4,5). Figure 1 shows a dendrogram depicting the evolutionary relationships within the human 2P K channel family.
One unique aspect of 2P K channel function is that the activity of some 2P K channels can be enhanced by volatile anesthetics. These drugs have been in clinical use for more than 150 yr, yet their exact mechanism of action remains unknown. The anesthetic state can be conveniently divided into 1) unconsciousness and amnesia, mediated through effects on higher cerebral function, and 2) immobility and muscle relaxation, mediated through spinal cord or peripheral effects (6,7). The term MAC, for minimum alveolar anesthetic concentration, quantitates the ability of volatile anesthetics to produce surgical immobility and is now recognized as a purely spinal cord-mediated function (8). Over the past 20 yr, many ion channels resident in the spinal cord have been studied for their role in MAC; most, except for glycine receptors, N-methyl-d-aspartic acid receptors, and K channels, have now been excluded by various studies (8).
Enhancement of background K channel activity remains a highly plausible mechanism by which volatile anesthetics depress the central nervous system (CNS). Volatile anesthetic potentiation of 2P K channel activity may be conserved across organisms For example, the yeast 2P K channel TOK-1 can be enhanced, as can certain human family members (9). To this point, volatile anesthetic enhancement has been demonstrated for 6 of the 15 mammalian 2P K channels (10,11). However, the concentrations at which in vitro enhancement of these ion channels occurs overlap only at the high end of “clinical” concentrations of these drugs.
Recently a new 2P K channel sequence has been reported (12). Because its expression was found by reverse transcription coupled to the polymerase chain reaction (RT-PCR) only in spinal cord, it has been termed TRESK (for TWIK [tandem pore domain weak inward rectifying channel]-related spinal cord K channel)-related spinal cord K channel). This newest family member has the least sequence similarity to other family members but has similar electrophysiologic properties. When expressed in heterologous systems, it passes outwardly rectifying K currents. Like other 2P K channels, TRESK is inhibited by large concentrations of Ba2+, quinine, quinidine, and the local anesthetic lidocaine. We hypothesized that because of its spinal cord expression pattern, TRESK currents may display volatile anesthetic activation. We have isolated our own clone of human TRESK and report here that TRESK is highly sensitive to activation by volatile anesthetics; this strongly implies a role for this channel in anesthetic mechanisms.
A strategy using RT-PCR was devised for cloning human TRESK. The first step involved reverse-transcribing human spinal cord or brain poly(A)+ RNA (Clontech, Palo Alto, CA) as template RNA by using a Superscript One-Step RT-PCR kit (Invitrogen, Carlsbad, CA). The RT reaction was primed with a DNA oligonucleotide complementary to the 3′ end of the published TRESK sequence. The first round of PCR was performed with oligonucleotides complementary to the 5′ and 3′ ends of the published TRESK sequence: 5′-ATGGAGGTCTCGGG-3′ as the 5′ primer and 5′-TCACTTTTTAACAAGGTGGTAA-3′ as the 3′ primer. A second round of PCR was performed with the aim of increasing the specificity of the PCR products by using two new sets of primers, with the PCR products from the first reaction as a template. This set of primers consisted of 5′-ATGGAGGTCTCGGG-3′ as the 5′ primer and 5′-AAGTTTGGGGCCTGGAAGCTCT-3′ as the 3′ primer. The other set of primers used 5′-AGAGCTTCCAGGCCCCAAACTT-3′ as the 5′ primer and 5′-TCACTTTTTAACAAGGTGGTAA-3′ as the 3′ primer. The PCR products generated from these two reactions produced overlapping fragments that spanned the entire TRESK coding sequence. These products were separated by agarose gel electrophoresis, and the DNA bands were isolated from the gel by using a Qiagen fragment-purification kit. A final round of PCR was performed to assemble the full-length TRESK gene by using 5′-CACCATGGAGGTCTCGGG-3′ as the 5′ primer, 5′-TCACTTTTTAACAAGGTGGTAA-3′ as the 3′ primer, and gel-purified PCR products as a template. This overlapping extension technique produced a single band that was gel-purified and subcloned into a pcDNA3.1-TOPO vector (Invitrogen). Identity of the final cloned sequence with the published human TRESK sequence was confirmed by sequencing the full-length fragment generated.
A different strategy was adopted to demonstrate the presence of the TRESK transcript in other tissues. A single round of RT-PCR with primers specific for human TRESK, TASK-2 (TWIK-related acid sensitive K channel), and the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase amplified fragments of the correct size (657, 905, and 587 base pairs, respectively) from RNA isolated from human spinal cord, brain, and heart under the following conditions: melting, 94°C for 15 s; annealing, 52°C for 30 s; and extension, 68°C for 60 s (35 cycles for TRESK; 25 cycles for TASK-2 and glyceraldehyde-3-phosphate dehydrogenase). The products of these reactions were separated by agarose gel electrophoresis and visualized with ethidium bromide staining.
All animal experiments were performed according to protocols approved by the University of California San Francisco Institutional Animal Care and Use Committee. Chromosomal RNA (cRNA) transcripts were synthesized from a linearized complementary DNA template of TRESK or TASK-3 by using T3 RNA polymerases (mMessage mMachine; Ambion, Austin, TX). Defolliculated Xenopus laevis oocytes were injected with 1–15 ng of cRNA by using standard methods for oocyte preparation and maintenance (13). One to four days after injection, two-electrode voltage-clamp recordings were performed at room temperature (GeneClamp 500B; Axon Instruments, Foster City, CA). Most voltage pulse protocols were applied from a holding potential of −60 mV by using 1-s voltage pulse steps ranging from −120 to +60 mV in 20-mV increments, with 1.5-s interpulse intervals. In some experiments, single pulses were performed for 2 s at a higher sampling rate to produce higher-resolution recordings. Except where noted, all two-electrode voltage clamp experiments were performed with frog Ringer’s solution (composition [mM]: 115 NaCl, 5 KCl, 1.8 CaCl2, and 10 HEPES, pH 7.6) or high-K frog Ringer’s solution (composition [mM]: 5 NaCl, 115 KCl, 1.8 CaCl2, and 10 HEPES, pH 7.6) as perfusate. An intermediate K solution contained 15 mM KCl and 105 mM NaCl. Recordings were obtained in a 25-μL recording chamber at flow rates of 1–4 mL/min. Signals were filtered with a low-pass filter set at a 50- to 100-Hz cutoff before sampling at 100–1000 Hz. Water-injected oocytes were used as controls and underwent the same treatments as transcript-injected oocytes.
Racemic isoflurane (Baxter Healthcare, New Providence, NJ), halothane with 0.01% thymol (wt/wt) (Ayerst Laboratories, Inc., Philadelphia, PA, and Halocarbon Laboratories, Augusta, SC), sevoflurane (Abbott Laboratories, North Chicago, IL), and desflurane (Ohmeda, Liberty Corner, NJ) were used for experiments. The nonimmobilizer 1,2-dichlorohexafluorocyclobutane (Lancaster Synthesis Inc., Windham, NH) was also studied (14). Methods for volatile anesthetic delivery and analysis have been previously described (9,15). Briefly, solutions and their dilutions to the experimental concentrations were prepared immediately before use. Saturated stock solutions of isoflurane, halothane, desflurane, sevoflurane, and chloroform were prepared by adding approximately 15 mL of volatile anesthetic to an airtight glass bottle with 200 mL of frog Ringer’s solution or high-K frog Ringer’s solution. An equilibration period of 24 h at room temperature was allowed. Aliquots were taken from the bottle through a needle with its tip in a fixed position close to the phase of volatile anesthetic to ensure a constant concentration. Volatile anesthetics were applied by using an airtight perfusion system including glass syringes and Teflon tubing. Their concentrations in the recording chamber were determined by gas chromatography (GowMac Series 750; GowMac, Bethlehem, PA) from samples taken at the outlet of the perfusion system. Nitrous oxide was applied to oocytes by bubbling the gas through the perfusate for a sufficient period of time to reach saturation (15 min) and was then perfused onto voltage-clamped oocytes. Previous experiments have established losses of ∼30% during exposure.
COS-7 cells were maintained at 37°C in a 95% air/5% CO2 (vol/vol) humidified atmosphere. Cells were grown in Dulbecco’s modified Eagle medium H-21 containing 10% fetal bovine serum, penicillin (100 μg/mL), streptomycin (100 U/mL), and 2 mM glutamine. At 60%–80% confluency, cells were transfected with human pZeo-TRESK (Invitrogen) by using Lipofectamine (GibcoBRL, Gaithersburg, MD). Forty-eight hours after transfection, cells were transferred to and maintained in selection media containing 600 μg/mL Zeocin and then plated on coverslips for patch studies.
Patch electrodes (4–8 MΩ) were pulled from borosilicate glass, coated with Sylgard®, and heat-polished. Seal resistances ranged from 5 to 20 GΩ. Before seal formation, voltage offsets between pipette and bath solution were zeroed. All voltages are referenced to the extracellular side of the membrane. Experiments were performed at room temperature (20°C–23°C). Data were recorded with an Axopatch 200B amplifier (Axon Instruments), filtered at 2 kHz, and digitized at 10 kHz by using an Instrutech ITC-16 analog/digital converter (InstruTECH, Port Washington, NY) and then recorded to disk by using PULSE software (HEKA Elektronik, Lambrecht, Germany). The pipette and bath solutions (symmetrical) contained (mM) 150 potassium aspartate, 3 NaCl, 10 HEPES, 1 EGTA, and 5 MgCl2 (pH 7.4 with KOH). The experimental solutions containing volatile anesthetics were diluted from saturated stocks and directed near the patch under direct visualization by continuous perfusion (20 μL/min). Perfusion with bathing solution alone had no effect on current amplitudes or other measures of channel activity.
Except where noted, data are reported from at least three oocytes, although usually more measurements were collected. Peak response was defined as peak current measured for the −60- to +60-mV pulse during the treatment condition relative to the control condition. Peak currents were defined as the highest current immediately after the capacitance transient, usually within 50 ms of the voltage jump. Mean values are expressed ± se; n values indicate the number of oocytes studied. Concentration-response curves were derived by nonlinear least-square fit performed by the graphing and statistical package Prism 3.0a (Prism, San Diego, CA), which provided 50% effective concentration and nHill (Hill coefficient) values with 95% confidence intervals. For comparison of holding potentials, statistical significance was determined with a Student’s t-test. Statistical significance was defined by P < 0.05.
For single-channel data analysis, all-points current-amplitude histograms were obtained from current records of at least 100 s in duration. Histograms were fit with a multi-Gaussian probability density function by using TAC software (Bruxton Corp., Seattle, WA), and open probability was derived from these fits.
Injection of cRNA transcribed from the cloned human TRESK sequence produced robust baseline currents (1–8 μA of outward current at positive pulses) within 24–48 h of injection into Xenopus oocytes. The resting potentials of TRESK-expressing oocytes were significantly more hyperpolarized than of control water-injected oocytes (−72.8 ± 4.1 mV versus −41.2 ± 4.1 mV; n = 10; P < 0.001), indicating excellent background K channel expression. Likewise, water-injected oocytes showed negligible inward and outward background currents during voltage jumps (<100 nA at the +60-mV pulse). (Figure 2) shows the pattern of currents obtained during a series of step voltage changes from a TRESK cRNA-injected oocyte under two-electrode voltage clamp. Large outward currents are present instantaneously at each voltage step that inactivate slightly over the period of membrane voltage change (1 s). The i-V (current versus voltage) curve is outwardly rectifying, with minimal inward currents at small (5 mM) extracellular K concentrations (Fig. 2B). In large extracellular K concentrations, large inward currents could be observed, and the zero-current potential shifted 57.6 mV (a 10-fold change in K concentration as predicted for a K-selective channel) (Fig. 2C). Currents recorded from oocytes injected with TRESK cRNA were resistant to inhibition by tetraethyl ammonium (TEA+), 4-aminopyridine, and decreases or increases in extracellular pH and were moderately sensitive to quinidine (Table 1). These data confirm that the K channel expressed from our cloned sequence has properties identical to those previously reported (12).
Bath application of volatile anesthetics produced strong enhancement of both outward and inward currents in TRESK-injected oocytes. Figure 3A shows a raw tracing of currents recorded from an oocyte expressing TRESK that was then exposed to 150 μM isoflurane (∼0.7 MAC). During anesthetic application, currents were strongly enhanced and returned to the baseline response with washout. All currently used clinical volatile anesthetics—isoflurane, halothane, sevoflurane, and desflurane—and the archaic volatile anesthetic chloroform produced strong TRESK current enhancement. Concentration-response curves for each of these anesthetics are shown in Figure 3B–E; derived 50% inhibitory concentration and nHill values are reported in Table 2. Within the clinical concentration range of each drug (shaded area, 0.6–1.2 MAC), a doubling (100% increase) of TRESK currents was observed. The degree of enhancement plateaued at a more than 1 mM concentration for all anesthetics. Isoflurane enhanced inward currents (extracellular K concentration, 115 mM) to the same degree as outward currents (see square symbol in Fig. 3B). Nitrous oxide (70%) potentiated TRESK currents by 13.6% ± 4.5% (n = 4). The nonimmobilizer compound 1,2-dichlorohexafluorocyclobutane at concentrations up to 8.8 times larger than the predicted 1.0 MAC concentration (363 μM) had no effect on TRESK currents.
To confirm that strong volatile anesthetic enhancement was a specific effect on TRESK, we also tested the effect of volatile anesthetics on the acid-sensitive 2P K channel TASK-3. With the same oocyte expression system and perfusion method, TASK-3 currents were enhanced only 4.4% ± 1.5% by 769 μM isoflurane and 9.5% ± 3.9% by 948 μM isoflurane.
IV anesthetics and other CNS-active drugs were also tested for their effects on TRESK. These results are summarized in Table 3. Only ethanol enhanced TRESK currents, and at significantly less efficacy than the volatile anesthetics. Other IV drugs produced mild (5%–15%) to moderate (30%–45%) inhibition of TRESK currents at large (100 μM) concentrations.
Because inhibition of K currents has been postulated to enhance the neuraxial block produced by local anesthetics (16), we also tested the effects of local anes-thetics on TRESK currents. Application of amide local anesthetics caused concentration-dependent inhibition of TRESK currents in the rank order of bupivacaine > ropivacaine > mepivacaine/lidocaine (Fig. 4A, Table 4). The ester local anesthetics tetracaine and chloroprocaine produced inhibition in the same concentration range (Fig. 4B). Inhibition by lidocaine was highly dependent on the extracellular pH; at pH 6.0, 100 μM and 1 mM lidocaine produced minimal inhibition, whereas at pH 9.0, the same concentrations inhibited TRESK currents by 25.4% ± 2.1% (n = 4) and 48.2% ± 5.5% (n = 4), respectively. Similarly, the compound QX-314, a permanently charged form of lidocaine, showed no inhibitory effect when superfused onto TRESK-expressing oocytes.
To confirm the effect of volatile anesthetic enhancement of TRESK in mammalian cells, and to better understand the mechanism of enhancement, we recorded the effects of volatile anesthetics on single-channel currents from COS-7 cells stably transfected with a TRESK expression plasmid. After a gigaohm seal was obtained, patches were excised to produce an inside-out configuration. As controls, COS cells transfected with the same expression vector lacking the TRESK coding sequence were also patched. No channels passing outward currents at positive patch potentials could be detected in these cells (n = 10). As shown in Figure 5, patches from COS cells transfected with the TRESK expression plasmid under control conditions displayed outward K currents at positive holding potentials with a low open probability (NPo). The application of 450 μM isoflurane to the patch caused an average 253% increase in channel open probability (control NPo, 0.143 ± 0.01; isoflurane NPo, 0.363 ± 0.04; n = 6 patches), with no change in the single-channel conductance. Halothane (385 μM) produced a 390% increase in open probability (control NPo, 0.076 ± 0.01; halothane NPo, 0.298 ± 2.1; n = 5 patches). These responses show that TRESK expressed in mammalian cells is enhanced by volatile anesthetics to the same degree as when it is expressed in oocytes. The single-channel data also indicate that volatile anesthetics enhance TRESK currents by enhancing channel gating to increase open probability rather than increasing the single-channel conductance or by changing the number of channels in the membrane.
Although TRESK expression was originally reported only in spinal cord, we tested whether the TRESK transcript could be detected at a higher level of the CNS. With RT-PCR, the TRESK transcript could be detected at slightly higher levels in human whole-brain RNA but was not present in heart RNA (Fig. 6). In contrast, the 2P K channel TASK-2 showed greater expression in spinal cord compared with faint expression in human brain and heart. These results confirm, for both channels, previously reported rodent expression patterns (17,18).
In this study, we demonstrated that currents passed by the most recently isolated 2P K channel, TRESK, when expressed in both mammalian and amphibian cells, are strongly enhanced by volatile anesthetics in the clinically used concentration range. In contrast, TRESK currents were not enhanced significantly by any of the IV anesthetics tested and were inhibited by a variety of local anesthetics. This pattern of response is similar to that found previously for other members of the 2P K channel family. However, no other 2P K channel has shown the degree of sensitivity to volatile anesthetics that we have shown here for TRESK. For example, 1 mM halothane and 2 mM isoflurane enhance TREK-1 (KCNK2) and TASK-1 (KCNK3) in the range of 20%–60% (10); TRESK currents were increased by 150% to more than 300% at smaller concentrations. Furthermore, significant enhancement was found at concentrations less than 100 μM, aqueous concentrations that in clinical practice would be considered light or subsurgical levels. Because TRESK has been identified in spinal cord tissue, we propose that TRESK channels play an important role in causing immobility to surgical stimulation produced by volatile anesthetics. In addition, by identifying TRESK transcript in brain RNA, a role for TRESK in mediating other components of general anesthesia is also feasible.
TRESK differs from other members of the 2P K channel in having a significantly longer segment connecting the second and third predicted transmembrane domains. By aligning all the known human 2P K channels, this region codes for approximately 92 amino acids, which would be predicted to form an elaborated intracellular domain. Only the tandem pore domain halothane-inhibited K channel (THIK)-1 and THIK-2 (KCNK13 and KCNK12) 2P K channels, which have approximately 14 extra amino acids in this region, differ from the other 12 known human 2P K channels. The unique sensitivity of TRESK to volatile anesthetics suggests that this region may be important for mediating the response. Mutagenesis experiments may permit a more detailed definition of the domains of TRESK that confer high volatile anesthetic sensitivity.
Although the primary mechanism of conduction block by local anesthetics is via inhibition of voltage-gated Na channels, K channel blockade may deepen or extend the block (16). The potencies of local anesthetics on TRESK reported here are generally less than those found for other 2P K channels previously studied (19,20), but they do correlate well with the rank order of in vivo potencies of these drugs. The site of action of local anesthetics appears to be at an intracellular or membrane location, because charged local anesthetic molecules, produced by decreasing extracellular pH and converting local anesthetic molecules from uncharged to charged form, exerted little inhibitory effect. In addition, the permanently charged congener QX314, when superfused to the extracellular surface of the oocyte, had no effect on TRESK currents, also supporting the view that the local anesthetic molecule must be in the un-ionized form to cross the cell membrane to produce inhibitory effects from the intracellular side. Our data cannot confirm whether it is the ionized or un-ionized form or both forms that produce channel inhibition. The relative insensitivity of TRESK to local anesthetics indicates that it probably does not contribute strongly to conduction block.
By RT-PCR we were able to demonstrate that the TRESK transcript can be found not only in human spinal cord, but also in brain. This result indicates that TRESK may also play a role in volatile anesthetic modulation of higher cerebral functions, such as consciousness and memory. Localization studies in the spinal cord and brain will be needed to assess more fully the role of TRESK in general anesthetic mechanisms. A recent study of mouse TRESK has confirmed our finding by detecting TRESK transcript in mouse cortex, cerebellum, and brainstem, as well as testis (17). Future in situ hybridization and immunohistochemical localization studies will identify in which cells and at what levels TRESK is expressed in the CNS.
The authors wish to acknowledge the help of Drs. Michael Laster, James Sonner, and Edmond I. Eger II for fruitful discussions and for assisting in the analysis of anesthetic concentrations.
1. Goldstein SA, Bockenhauer D, O’Kelly I, Zilberberg N. Potassium leak channels and the KCNK family of two-P-domain subunits. Nat Rev Neurosci 2001;2:175–84.
2. Ketchum KA, Joiner WJ, Sellers AJ, et al. A new family of outwardly rectifying potassium channel proteins with two pore domains in tandem. Nature 1995;376:690–5.
3. Czempinski K, Zimmermann S, Ehrhardt T, Muller-Rober B. New structure and function in plant K+
channels: KCO1, an outward rectifier with a steep Ca2+
dependency. EMBO J 1997;16:2565–75.
4. Bargmann CI. Neurobiology of the Caenorhabditis elegans
genome. Science 1998;282:2028–33.
5. Yost CS. Update on tandem pore (2P) domain K+
channels. Curr Drug Targets 2003;4:347–51.
6. Antognini JF, Schwartz K. Exaggerated anesthetic requirements in the preferentially anesthetized brain. Anesthesiology 1993;79:1244–9.
7. Rampil IJ. Anesthetic potency is not altered after hypothermic spinal cord transection in rats. Anesthesiology 1994;80:606–10.
8. Sonner JM, Antognini JF, Dutton RC, et al. Inhaled anesthetics and immobility: mechanisms, mysteries, and minimum alveolar anesthetic concentration. Anesth Analg 2003;97:718–40.
9. Gray AT, Leonoudakis DL, Winegar BD, et al. TOK-1 is an anesthetic activated K+
channel. Anesthesiology 1998;88:1076–84.
10. Patel AJ, Honore E, Lesage F, et al. Inhalational anesthetics activate two-pore-domain background K+
channels. Nat Neurosci 1999;2:422–6.
11. Yost CS, Kindler CH. Tandem pore (2P) domain K+
channels as anesthetic targets. In: Urban BW, Barann M, eds. Molecular and basic mechanisms of anesthesia. Berlin: Pabst Science Publishers, 2002:203–10.
12. Sano Y, Inamura K, Miyake A, et al. A novel two-pore domain K+
channel, TRESK, is localized in the spinal cord. J Biol Chem 2003;278:27406–12.
13. Leonoudakis D, Gray AT, Winegar BD, et al. An open rectifier potassium channel with two pore domains in tandem cloned from rat cerebellum. J Neurosci 1998;18:868–77.
14. Koblin DD, Chortkoff BS, Laster MJ, et al. Polyhalogenated and perfluorinated compounds that disobey the Meyer-Overton hypothesis. Anesth Analg 1994;79:1043–8.
15. Paul M, Fokt RM, Kindler CH, et al. Characterization of the interactions between volatile anesthetics and neuromuscular blockers at the muscle nicotinic acetylcholine receptor. Anesth Analg 2002;95:362–7.
16. Drachman D, Strichartz G. Potassium channel blockers potentiate impulse inhibition by local anesthetics. Anesthesiology 1991;75:1051–61.
17. Czirjak G, Toth ZE, Enyedi P. The two-pore domain K+
channel, TRESK, is activated by the cytoplasmic calcium signal through calcineurin. J Biol Chem 2004;279:18550–8.
18. Reyes R, Duprat F, Lesage F, et al. Cloning and expression of a novel pH-sensitive two pore domain K+
channel from human kidney. J Biol Chem 1998;273:30863–9.
19. Kindler CH, Yost CS, Gray AT. Local anesthetic inhibition of baseline potassium channels with two pore domains in tandem. Anesthesiology 1999;90:1092–102.
20. Kindler CH, Paul M, Zou H, et al. Amide local anesthetics potently inhibit the human tandem pore domain background K+
channel TASK-2 (KCNK5). J Pharmacol Exp Ther 2003;306:84–92.
21. Franks NP, Lieb WR. Molecular and cellular mechanisms of general anaesthesia. Nature 1994;367:607–14.
22. Jenkins A, Franks NP, Lieb WR. Effects of temperature and volatile anesthetics on GABA(A) receptors. Anesthesiology 1999;90:484–91.
© 2004 International Anesthesia Research Society
23. Friederich P, Benzenberg D, Trellakis S, et al. Interaction of volatile anesthetics with human Kv channels in relation to clinical concentrations. Anesthesiology 2001;95:954–8.